Validation of the HTC VIVE Ultimate Trackers Compared with the Vicon Motion Capture System at Slow, Moderate and Fast Gait Speeds

Validation of the HTC VIVE Ultimate Trackers Compared with the Vicon Motion Capture System at Slow, Moderate and Fast Gait Speeds | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Validation of the HTC VIVE Ultimate Trackers Compared with the Vicon Motion Capture System at Slow, Moderate and Fast Gait Speeds Yixuan He, Matthew A. Brodie, Juno Kim, Peter Humburg, Stephen R. Lord, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6989733/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study investigated the performance of the HTC VIVE Ultimate Tracker against the Vicon motion capture system during human walking. Ten healthy participants (aged 24–44 years) walked on a treadmill at four speeds (0.5, 1.0, 1.5, and 2.0 m/s) while tracking both feet and the pelvis. Agreement between the two systems was evaluated using a linear mixed model, Bland-Altman plots, and concordance correlation coefficients (CCC). Absolute errors presented the accuracy of Ultimate Trackers ranging from millimetre- to centimetre-level. Linear mixed model [speed (0.5-2.0 m/s) × tracker location (sacrum, left foot, and right foot) × movement direction (vertical, medio-lateral, and anterior-posterior)] indicated the absolute errors increased with higher gait speed. Foot trackers exhibited larger errors than the sacrum location, with greater errors in the medio-lateral and anterior-posterior directions compared to the vertical direction ( p < 0.001). Bland-Altman analyses revealed widening limits of agreement at different speeds (e.g., left foot, AP direction: -20.29 to 20.20 at 0.5 m/s, -29.97 to 27.80 mm at 2.0 m/s). Ultimate Trackers demonstrated almost perfect agreement (CCC: >0.99) for sacrum tracking across all speeds and directions, and excellent agreement for foot trackers (CCC: >0.98). These findings highlight the Ultimate Trackers’ potential as cost-effective alternatives for the analysis of human movement, demonstrating research- and clinical-grade performance for sacrum and foot tracking during normal gait speeds (≤ 1.5 m/s). However, the finding that accuracy declined at the 2.0 m/s speed, particularly for foot trajectories and in the anterior-posterior direction, indicates the need for further technical refinements for higher speed movements. motion capture VIVE Ultimate Tracker Vicon virtual reality Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. INTRODUCTION Motion capture systems enable the precise recording and analysis of human movement within defined spatial environments (Rybnikár et al., 2023). These technologies have proved valuable across a wide range of fields, including healthcare, physical rehabilitation, sports biomechanics, education, ergonomics, robotics, filmmaking, and virtual reality (VR), driving interdisciplinary innovation through accurate quantification of human motion (Gu et al., 2023; Menolotto et al., 2020; Ranjan et al., 2025). Among these, optical motion capture systems—particularly marker-based systems such as Vicon (Vicon Motion Systems Ltd., Oxford, UK), Qualisys (Qualisys, Gothenburg, Sweden), and OptiTrack (OptiTrack, Oregon, USA) —are widely regarded as the “gold standard” in motion analysis due to their excellent positional accuracy, often within sub-centimetre ranges (~ 1 mm error) (Gu et al., 2023; Topley & Richards, 2020; Zhou & Hu, 2008). However, these systems come with notable limitations. Their dependence on fixed arrays of infrared cameras, specialized mounting infrastructure, high-performance computing resources, and controlled laboratory environments imposes significant spatial and financial constraints. These requirements limit their accessibility and scalability, particularly in low-resource settings and for personal or home-based applications (Gu et al., 2023). With advancements in graphics technology, VR now enables the creation of diverse, immersive environments (Slater & Sanchez-Vives, 2016), enhancing our ability to study and train human performance in simulated, real-world scenarios (Lee et al., 2024; Sousa et al., 2023). This capability has been increasingly applied in rehabilitation (Demeco et al., 2023), sports training (Putranto et al., 2023), and exercise interventions (Lee et al., 2024), offering more task-specific and contextually relevant applications. Several studies have integrated VR head-mounted displays (HMDs) with gait training to explore innovative strategies for improving performance in simulated daily environments (He et al., 2025; Okubo et al., 2025; Rhiel et al., 2024). Despite these promising developments, translating research findings into real-world applications remain challenging—largely due to the limitations of current motion capture technologies. While some studies have employed low-cost motion capture systems such as OpenCap (Stanford University, USA), Kinect (Microsoft, Washington, USA), VIVE Trackers (HTC, Taiwan, China), and Xsens (Xsens, Enschede, Netherlands), each has its limitations. OpenCap needs post-capture data processing at USA, has lower sample rate (60 Hz), and reduced kinematic accuracy (Lima et al., 2024; Svetek et al., 2025). Kinect’s low sample rate (30–37 Hz) and single camera-setup hinder 360° motion capture (Galna et al., 2014; Kim et al., 2020; Pfister et al., 2014). VIVE Trackers (2.0/3.0) have required multiple infrared base stations, limiting use outside controlled in non-laboratory or home-based settings (Merker et al., 2023; Spitzley & Karduna, 2019). Xsens sensors also show reduced accuracy in measuring flexion and movements in transversal and frontal planes (Kim et al., 2020; Nijmeijer et al., 2023). Emerging consumer-grade technologies offer promising solutions to the limitations of traditional motion capture systems. One such advancement is the use of Simultaneous Localization and Mapping (SLAM) algorithms, which enable real-time estimation of a device’s position while concurrently mapping the surrounding environment using onboard sensors (Al-Tawil et al., 2024). In VR, SLAM forms the foundation of inside-out tracking systems, eliminating the need for external cameras (Sheng et al., 2024). The HTC VIVE Ultimate Tracker leverages SLAM technology with dual inside-out cameras, and advanced computer vision algorithms to achieve real-time spatial awareness, and synchronization between physical movements and virtual environments (HTC VIVE, 2025). With a simplified setup requiring only a computer and tracker units, the Ultimate Tracker presents good potential for accessible and portable motion capture (Filippidis et al., 2024; Kulozik & Jarrassé, 2024). For instance, Kulozik et al. reported the Ultimate Tracker compared well against the OptiTrack gold standard for robotic arm tracking under unusually slow, controlled conditions (0.05–0.40 m/s) (Kulozik & Jarrassé, 2024). However, it remains unclear how the Ultimate Tracker performs in capturing human movement, particularly under common dynamic conditions involving higher peak velocities, such as walking. To address these knowledge gaps, this study aimed to validate the Ultimate Tracker system against the Vicon gold-standard motion capture system during human walking. Participants walked at four speeds—slow, normal, fast, and very fast—within a virtual suburban environment (He et al., 2025). Ultimate Tracker sensors were positioned at the sacrum, right foot, and left foot to capture key gait parameters. A comprehensive analytical approach was used to evaluate agreement between systems, including mixed-effects modelling to account for within-subject variability, as well as assessments of systematic bias and concordance. This multifaceted evaluation provides robust insights into the technical accuracy and consistency of the Ultimate Tracker and its suitability for use in human movement research within both real-world and virtual environments. 2. METHODS 2.1. Study design and ethics This study was conducted at Neuroscience Research Australia (NeuRA) in accordance with the Declaration of Helsinki and approved by the University of New South Wales (UNSW) Human Research Ethics Committee (iRECS6818). Written informed consent was obtained from all participants before participation. 2.2. Participants and sample size Ten participants (men/women: 5/5) were recruited at NeuRA. Participants had a mean age of 32.2 ± 6.5 years (range: 24–44), with an average weight of 66.5 ± 8.0 kg, height of 1.69 ± 0.009 m, and body mass index (BMI) of 23.2 ± 1.8 kg/m². Inclusion criteria were: (i) able to communicate in the English language; (ii) aged 20–70 years. Exclusion criteria were: (i) diagnosis of a neurological condition (e.g., Parkinson’s disease, multiple sclerosis, dementia); (ii) inability to walk for 20 minutes without a mobility aid or rest; (iii) history of lower limb, pelvic or vertebral fractures or lower limb joint replacements in the past 6 months; (iv) conditions that prevent exercising (e.g., severe pain, fatigue, exercise intolerance, heel ulcers); (v) advice from a medical practitioner to avoid exercise; (vi) history of dizziness, vertigo and vestibular disorders (e.g., Meniere's disease). 2.3. Experimental setup and procedure Participants attended NeuRA for a 1.5-hour laboratory session. This session commenced with assessments of body weight and height. Participants were then fitted with an HMD for VR (VIVE XR Elite, HTC, Taiwan, China) and a safety harness. Three Ultimate Trackers were attached to the sacrum using a waist belt and to the instep of the left and right feet over the participant’s sports shoes using gaffer tape. Vicon reflective markers were placed at the diagonal intersection points on the surface of the Ultimate Trackers (Fig. 1 ). SteamVR (Steam, Valve, USA) running on a desktop PC served as the platform for running the virtual environment and wirelessly streamed content to the HMD via Wi-Fi 6e. The virtual environment included concrete footpaths and roadways (Fig. 1 ). Participants undertook a 30-second familiarisation walk on a treadmill (M-Gait, Motek®, Netherlands) with a belt speed of 0.5 m/s in the VR environment. Participants then completed four consecutive 1-minute walking tasks at treadmill speeds of 0.5, 1.0, 1.5, and 2.0 m/s. The tasks were performed continuously, with participants notified of each upcoming speed change via a countdown prompt. An 8-camera Vicon motion capture system (Bonita, Vicon Motion Systems Ltd., Oxford, UK) and Vicon Nexus 2.16.0 x64 software were used to capture three-dimensional (3D) coordinates of the reflective markers at a sampling frequency of 100 Hz. An open-source Python script (Kulozik, 2024) and Visual Studio 2022 (Microsoft, USA) were utilised to collect 3D coordinates from the Ultimate Trackers during experiments at 100 Hz. The setting and data collection of Ultimate Trackers were based on the SteamVR platform. To aid Ultimate Tracker’s spatial recognition, black and white checkerboards were printed in A4 and posted on each wall. 2.4. Data processing The Ultimate Trackers pose matrix was provided by SteamVR to transform the local coordinate system of Ultimate Tracker into the global coordinate system and map it to the Vicon position. Temporal synchronisation was performed using the same directional data recorded by the Ultimate Tracker system. After synchronising the time axes of the data from the two systems using a customised script in MATLAB R2022a (The MathWorks, Inc., USA), it was observed that the Ultimate Tracker experienced a sampling loss of a median of 12 Hz, resulting in an actual sampling rate quartile of 87–88 Hz (Appendix A). To compensate for missing frames, piecewise cubic Hermite interpolating polynomial (PCHIP) interpolation was applied to the Ultimate Tracker data to match the data length of the Vicon system for each second (Gupta & Semwal, 2022). PCHIP was selected for its ability to preserve monotonicity and ensure continuity of first-order derivatives at nodes, thereby reducing overshoot artifacts commonly observed with cubic spline interpolation in gait data (Barker & McDougall, 2020). The Ultimate Trackers in the SteamVR coordinate system and Vicon data in the Vicon coordinate system were transformed into the same coordinate system prior to further data analysis (Fig. 2 ). Thus, the Kabsch algorithm and singular value decomposition (SVD) (Kabsch, 1976; Subramanian & Sarkar, 2019) were utilised to compute transformations between the SteamVR and Vicon coordinate systems to find the translation and the rotation between the two measurements. During the overall transformation of the Ultimate Trackers from the SteamVR coordinate system to the Vicon coordinate system, it was observed that the Ultimate Tracker positions did not align precisely with the corresponding Vicon marker locations. Considering the Ultimate Trackers were not placed at exactly the same positions during each calibration session, this discrepancy suggests the presence of potential calibration errors inherent in the SteamVR-based calibration process (Fig. 3 ). To address this, the Kabsch algorithm and SVD were applied to individually align the Ultimate Trackers to the corresponding Vicon markers, thereby correcting for the SteamVR-based calibration-induced spatial misalignment. Due to manual operation delays in both Vicon Nexus and the open-source scripts during data collection, a temporal offset emerged between the two datasets. To address this, the first and last 5 seconds were removed from both the Vicon and Ultimate Tracker recordings of each participant at each speed, retaining the central 50 seconds of data. Given the Vicon system’s sampling rate of 100 Hz, this 50-second segment contained 5000 frames. Using MATLAB, the experimental data were integrated and categorized based on participant ID (1–10), gait speed (0.5, 1.0, 1.5 and 2.0 m/s), tracker location (sacrum, right foot, and left foot), and direction (medio-lateral [ML], anterior-posterior [AP], and vertical [VT] axes). Additionally, the differences and absolute errors between the Ultimate Trackers and the Vicon system were calculated for each corresponding frame. 2.5. Statistical analysis To examine the impact of gait speed, tracker location and direction on the absolute errors between the Ultimate Trackers and the Vicon system, a linear mixed-effects model (LMM) was utilised (Gałecki et al., 2013). Four gait speeds, three tracker locations, and three directions were set as fixed interaction effects in the LMM, including main effects (speed, tracker location, and direction), two-way interaction effects (speed × tracker location, speed × direction, and tracker location × direction), and three-way interaction effects (speed × tracker location × direction), while participant IDs were treated as random effects. Given the time-ordered nature of the collected data, where consecutive observations within each participant are likely to be autocorrelated, a first-order autoregressive (AR(1)) covariance structure was introduced for repeated measures within participants to account for frame-to-frame dependence (0.98). The speed 0.5 m/s, sacrum and vertical axis showing smallest absolute errors were set as the baseline level in the LMM. Based on the VT direction data of left and right foot trackers, participants’ gait cycles were segmented into four distinct phases: foot strike, stance, foot off, and swing (Agostini et al., 2014). Absolute error was calculated for three Ultimate Trackers for each phase to capture variations in accuracy throughout locomotion. One-way ANOVA was utilised to compare the differences between the four gait phases. Since the sample size of gait cycles was estimated to be 5000, the statistical test of the normal distribution could be omitted (Blanca et al., 2017). Levene's test was conducted to assess the homogeneity of variances. When the assumption of homogeneity was met, a standard ANOVA was used; otherwise, Welch’s ANOVA was applied in the presence of heterogeneity of variances. The limits of agreement (LoA) between the two systems, proportional bias and systematic bias were examined using Bland-Altman plots (Giavarina, 2015). Bland-Altman plots were generated for the four gait speeds, with data further subdivided by tracker location and three directions, resulting in 36 combinations. The X-axis represented the gold standard Vicon system measurements, while the Y-axis showed the differences between the Vicon and Ultimate Tracker systems. Since the walking gait trajectory had strong autocorrelation and each combination included 50,000 data points from all participants, plotting all the data points in the plot would reduce readability. Bland-Altman plots were constructed by randomly sampling from the gait trajectory every 100 frames (n = 500). The LoAs are displayed as \(\:Mean\pm\:1.96\times\:standard\:deviation\) . The concordance correlation coefficient (CCC) was used to quantify the extent to which paired observations deviated from the Vicon gold standard (Akoglu, 2018; Lin, 1989). Unlike correlation coefficients (Benesty et al., 2009), the CCC offers a robust measure of agreement by accounting for both correlation and bias (Lin, 1989). The CCC analysis were grouped by four speeds, with each group further subdivided based on direction and tracker location. The CCC > 0.99 are considered to be almost perfect, > 0.95 to 0.90 to < 0.95 as moderate, and < 0.90 as poor (McBride, 2005). The fundamental calculation of the CCC is shown in Eq. (1). $$\:\begin{array}{c}CCC=\:\frac{2{S}_{HVUT\_Vicon}}{{S}_{HVUT}^{2}+{S}_{Vicon}^{2}+{\left({\mu\:}_{HVUT}-{\mu\:}_{Vicon}\right)}^{2}}\#\left(1\right)\end{array}$$ where: \(\:{S}_{HVUT\_Vicon}\) is the covariance between the two sets of Ultimate Tracker and Vicon, \(\:{S}_{HVUT}^{2}\) ​ and \(\:{S}_{Vicon}^{2}\) are the variances of the respective Ultimate Tracker and Vicon sets, \(\:{\mu\:}_{HVUT}\) ​ and \(\:{\mu\:}_{Vicon}\) ​ are the means of the Ultimate Tracker and Vicon sets. Statistical analyses were conducted using MATLAB R2022a and R (version 4.5.0; R Core Team, Vienna, Austria). p < 0.05 was considered statistically significant. 3. RESULTS 3.1. Absolute error Absolute errors (median [interquartile range]) are presented in Fig. 4 . Absolute errors of sacrum location (all directions) at 0.5, 1.0 and 1.5 and 2.0 m/s were 1.04 (0.41, 2.58), 0.86 (0.37, 1.76), 0.92 (0.31, 1.96), and 1.24 (0.43, 2.49) mm, respectively (Fig. 4 a). Absolute errors of right foot location at 0.5, 1.0, 1.5 and 2.0 m/s were 1.85 (0.81, 4.05), 2.43 (1.06, 5.13), 2.46 (0.81, 5.87), and 2.81 (0.93, 6.50) mm, respectively. Absolute errors of 1.0 m/s in the ML, AP and VT directions (all locations) were 1.32 (0.59, 2.52), 3.86 (1.51, 8.20), 1.09 (0.45, 2.47) mm, respectively (Fig. 4 b). Absolute errors of sacrum, right and left foot locations at the AP direction (all speeds) were 1.92 (0.76, 3.90), 6.13 (2.52, 12.05), and 5.80 (2.38, 11.20) mm, respectively (Fig. 4 c). See Fig. 4 for other conditions. 3.2. Linear mixed-effects model Table 1 summarises the total effect for each main effect and interaction from the LMM. There are statistically significant main effects of gait speed, direction, tracker location, and their interactions on the absolute error between Vicon and VIVE systems. Appendix B gives the parameter estimates. At the reference condition—0.5 m/s at the sacrum in the vertical (VT) direction—the estimated mean absolute error was 0.51 mm ( p = 0.271). Raising gait speed to 1.0 m/s, 1.5 m/s and 2.0 m/s increased error by 0.11 mm, 0.43 mm and 1.02 mm, respectively (all p < 0.001). Compared with the sacrum, the right- and left-foot trackers incurred additional errors of 1.07 mm and 1.00 mm, respectively (both p < 0.001). Directional main effects showed that errors in the ML direction exceeded VT by 1.15 mm, and errors in the AP direction exceeded VT by 3.15 mm (both p < 0.001). Table 1 Summary of fixed effects from the linear mixed model assessing the influence of gait speed, tracker location, direction and their interactions on absolute errors between Vicon and VIVE systems. Effects χ² Df p Gait speed 16057.5 3 < .001 Direction 156662.6 2 < .001 Tracker location 506210.1 2 < .001 Gait speed × Direction 10773.6 6 < .001 Gait speed × Tracker location 6688.2 6 < .001 Direction × Tracker location 101034.6 4 < .001 Gait speed × Direction × Tracker location 10677.7 12 < .001 χ² : Chi-square, Df : degrees of freedom. The larger the χ², the more strongly the effect contributes to explaining the variance in absolute error between Vicon and VIVE. Appendix B gives the parameter estimates. Two-way interactions revealed that the speed-related error increase differed by location and direction. Although AP motion carried the largest standalone error, its additional effect diminished at higher speeds (gait speed × AP, β = − 1.04 mm at 1.0 m/s, − 2.09 mm at 2.0 m/s, all p < 0.001). Similarly, gait speed × ML interactions were negative (e.g., − 1.31 mm at 2.0 m/s, p < 0.001), showing a relatively smaller ML error at fast speeds. Foot-specific interactions (gait speed × tracker location) remained positive, though reduced at 2.0 m/s (right foot β = 0.17 mm, left foot β = 0.20 mm, both p < 0.001). 3.3. Gait phases Due to heterogeneity of variances, Welch’s ANOVA was applied, and the results (Table 2 ) showed that the absolute errors of Ultimate Tracker on the sacrum were not significantly different among the four gait phases ( p = 0.176, ω 2 = 0.0001). The Ultimate Tracker errors on the right and left foot were significantly higher during the swing phase compared to foot strike, stance, and foot off phases ( p < 0.001). There were no significant differences between foot strike, stance, and foot off for either the right or left foot. Table 2 Welch’s ANOVA results, effect sizes, descriptive statistics and post hoc tests of root mean absolute error of gait phases for different locations. Tracker location Significance Effect size Sacrum F 3, 32222 = 1.65, p = 0.176 ω 2 = 0.0001, Small effect Right foot F 3, 34070 = 42.90, p < 0.001 ω 2 = 0.0001, Small effect Left foot F 3, 34032 = 22.29, p < 0.001 ω 2 = 0.0001, Small effect Phase of gait cycle Median (IQR) (mm) Comparison p (post hoc) Sacrum Foot strike 0.92 (0.37, 2.06) Foot strike-Stance 0.176 Stance 1.11 (0.60, 2.10) Foot strike-Foot off 0.996 Foot off 1.02 (0.49, 2.02) Foot strike-Swing 0.901 Swing 1.07 (0.59, 1.95) Stance-Foot off 0.269 Stance-Swing 0.536 Foot off-Swing 0.966 Right foot Foot strike 2.85 (1.38, 5.23) Foot strike-Stance 0.900 Stance 2.45 (1.33, 4.70) Foot strike-Foot off 0.993 Foot off 2.37 (1.14, 4.86) Foot strike-Swing < 0.001 Swing 3.18 (1.82, 5.85) Stance-Foot off 0.976 Stance-Swing < 0.001 Foot off-Swing < 0.001 Left foot Foot strike 2.76 (1.37, 5.26) Foot strike-Stance 0.396 Stance 2.32 (1.30, 4.79) Foot strike-Foot off 0.534 Foot off 2.21 (1.02, 4.70) Foot strike-Swing < 0.001 Swing 3.01 (1.70, 5.58) Stance-Foot off 0.996 Stance-Swing < 0.001 Foot off-Swing < 0.001 IQR: Interquartile Range. 3.4. Bland-Altman plots The Bland-Altman plots (Fig. 5 and Appendix C) provide a visual and quantitative assessment of agreement between the Vicon and Ultimate Tracker systems. The significant slopes ( p < 0.001) suggest that the discrepancy between the methods changes with the magnitude of the measurement. At a gait speed of 0.5 m/s, there were no significant slopes in the VT direction for all tracker locations. At gait speeds of 1.5 m/s and 2.0 m/s, there were no significant slopes between Vicon and Ultimate Tracker in the VT direction for the sacrum location. 3.5. Concordance correlation coefficient Figure 6 presents the CCC comparison between the Vicon and Ultimate Tracker measurements, alongside the mean and SD of the absolute errors. The heatmaps highlight CCC variations across conditions with values approaching 1 displayed in yellow. According to McBride’s strength-of-agreement criteria (McBride, 2005), all measurements achieved almost perfect (> 0.99) correlation at a speed of 0.5 m/s. At 1.0 m/s, 1.5 m/s, and 2.0 m/s, only the VT direction measurements of the right and left foot reached excellent correlation (0.95–0.99); the other measurements all reached almost perfect correlation (> 0.99). 4. DISCUSSION This study systematically evaluated the performance and concordance between the Ultimate Tracker and the Vicon motion capture system at different gait speeds, tracker locations, and movement directions. The findings indicate the Ultimate Tracker is sufficiently accurate for use in clinical- and research-settings as an alternative to gold standard systems for assessing gait at common speeds. 4.1. Factors Affecting the Accuracy of Ultimate Trackers Our findings indicate that the magnitude of error is influenced by gait speed, movement direction, and tracker location. The estimated absolute error of the Ultimate Tracker was lowest (0.51 mm) at the reference condition defined by the LMM: a gait speed of 0.5 m/s, sacrum tracker location, and VT movement direction (Appendix B). As gait speed increased to 1.0, 1.5, and 2.0 m/s, the estimated absolute error rose slightly by 0.11 mm, 0.43 mm, and 1.02 mm, respectively (Appendix B). This trend aligns with findings by Kulozik and Jarrasse, who reported an increase in Ultimate Tracker absolute error from 1.49 ± 0.50 mm to 3.07 ± 1.66 mm as the speed of a robotic arm increased from 0.05 m/s to 0.4 m/s (Kulozik & Jarrassé, 2024). Compared to the sacrum-mounted tracker, the estimated absolute errors for trackers placed on the right and left feet increased slightly by 1.07 mm and 1.00 mm, respectively (Appendix B). This increase is likely due to the rapid accelerations experienced by the feet during the gait cycle, as well as the greater kinematic complexity associated with multi-joint coordination in the lower limbs (Czajkowska et al., 2024; Raffegeau et al., 2022). Additionally, because the SLAM method relies on visual input from environmental landmarks, the camera views from foot-mounted trackers may be partially obstructed by the user’s body during walking, potentially reducing tracking accuracy. Directional asymmetries were also observed, with the largest deviations occurring in the AP direction (3.15 mm), which is characterized by the greatest movement amplitude during the foot swing phase. This finding is consistent with results showing higher absolute errors during the swing phase compared to the foot strike, stance, or foot-off phases (Table 2 ). Together, these results suggest that high-amplitude and high-speed movements consistently reduce the accuracy of Ultimate Trackers. The absolute error findings (Fig. 4 ) demonstrate that Ultimate Tracker maintains accuracy ranging from the millimetre to centimetre scale. These results contrast with robotic arm validation studies (Kulozik & Jarrassé, 2024), where slow, controlled movements (0.05–0.4 m/s) yielded smaller errors—for instance, an absolute error of 1.49 ± 0.50 mm at 0.05 m/s—compared to the errors observed at typical human gait speeds (1.0–1.5 m/s). Our slow gait condition (0.5 m/s) with the tracker on the sacrum produced a comparable error of 1.04 [0.41, 2.58] mm, emphasizing the limitations of applying rigid, predictable motion models to complex human biomechanics. Other robotic arm studies using earlier versions of the VIVE Tracker (Borges et al., 2018; Sitole et al., 2020; Weber et al., 2023) reported different outcomes. For example, gold-standard comparisons showed higher errors for the VIVE Tracker (Sitole et al., 2020), prompting some researchers to develop accuracy improvement algorithms (Borges et al.). However, since these earlier systems relied on infrared base stations—unlike the current Ultimate Tracker setup—their findings serve only as general references and are not directly comparable. The observed discrepancies across different speeds were consistent with previously reported Ultimate Tracker results. For instance, at 0.4 m/s, an absolute error of 3.07 ± 1.66 mm was reported (Kulozik & Jarrassé, 2024), while our findings at 0.5 m/s with the tracker on the left foot showed a comparable absolute error of 1.91 [0.83, 4.16] mm (Fig. 4 a). These results represent an improvement over earlier VIVE trackers, such as the VIVE Tracker 3.0, which demonstrated significantly higher errors (e.g., at 80 rpm with a motion radius of 133.53 ± 19.24 mm, the mean absolute error was 11.33 ± 5.14 mm (Merker et al., 2023)). At higher speeds, such as 2.0 m/s, we observed an error of 2.81 [0.93, 6.50] mm at the right foot location (Fig. 4 a). However, the AP direction at this speed exhibited a notably higher error of 4.24 [1.29, 11.01] mm (Fig. 4 b), exceeding those in the ML and VT directions. This suggests that consumer-grade systems like Ultimate Tracker still exhibit centimetre-level errors in high-velocity applications. Systematic bias and agreement The measurement discrepancy between the Vicon and Ultimate Tracker systems appears to increase with distance from the calibration origin, as indicated by the significant slopes in the Bland–Altman plots (Fig. 5 and Appendix C). Similar trends were reported in a previous study (Kulozik & Jarrassé, 2024), suggesting that the Ultimate Tracker may systematically overestimate or underestimate positional data as the distance from the calibration centre grows. This pattern is particularly evident in the AP direction, which involves the largest range of motion during gait. This directional bias may help explain the greater deviations observed in the AP direction (Table 1 , Appendix B and Fig. 4 ). Regarding error magnitude, the LoAs for the sacrum-mounted tracker were all below 10 mm across gait speeds, indicating millimetre-level accuracy. Similarly, for foot-mounted trackers, the LoAs in the ML and VT directions remained under 10 mm. However, in the AP direction, LoAs exceeded 10 mm, reflecting a transition from millimetre- to centimetre-level accuracy. The Bland–Altman analyses and CCC further define the operational boundaries of the Ultimate Tracker (Fig. 5 , Fig. 6 , and Appendix C). The sacrum tracker demonstrated near-perfect agreement across all gait speeds and directions (CCC > 0.999), suggesting that proximal body segments—with their lower kinematic complexity and minimal occlusion—are well-suited for Ultimate Tracker-based tracking. In contrast, foot-mounted tracker performance, particularly in the VT direction, declined slightly at higher gait speeds (CCC range: right foot 0.9929–0.9897, left foot 0.9932–0.9875), though still reflecting excellent agreement (CCC: 0.95–0.99 or above) (McBride, 2005). At higher gait speeds (2.0 m/s), the AP direction exhibited wide LoAs for foot-mounted trackers (right foot: − 26.58 to 24.12 mm; left foot: − 29.97 to 27.80 mm), indicating both systematic bias and substantial random error. These discrepancies are likely due to cumulative inaccuracies in the SLAM algorithm or hardware limitations during the swing phase—when foot motion in the AP direction is most pronounced (Agostini et al., 2014; Kulozik & Jarrassé, 2024). Potential sources of errors Both biomechanical and system-level factors may contribute to the observed errors in Ultimate Tracker measurements, particularly at higher gait speeds. First, spatial inaccuracies are likely to accumulate during rapid swing phases, where transient occlusions—such as the foot tracker facing downward—can disrupt positional synchronisation. Second, velocity-dependent error amplification may stem from limitations in the Trackers’ SLAM algorithm and hardware capabilities. However, the proprietary nature of HTC’s tracking methods limits detailed interpretation. Third, the presence of 12–13 missing frames at a nominal 100 Hz sampling rate suggests a reduction in effective sampling frequency, likely compromising the system’s ability to capture fast movements accurately. In contrast, high-end systems like Vicon mitigate these issues by increasing the sampling rate (up to 250 Hz) and optimising imaging parameters—including camera gain, strobe intensity, and aperture size—to better accommodate rapid motion. 4.2. Clinical and Research Implications The demonstrated accuracy of the Ultimate Trackers at typical gait speeds appears for a range of applications, including VR experiences (Caserman et al., 2019), rehabilitation and gait training (Okubo et al., 2025), ergonomic risk assessment (Vox et al., 2021), and robot inverse kinematics (Weber et al., 2023). These trackers may be particularly valuable in home and clinical settings, where usability and affordability often take precedence over laboratory-grade precision (Zhou et al., 2024). Their integration with VR platforms enables quantification of human-environment interactions, and real-time feedback on gross motor movements such as foot-obstacle collisions (He et al., 2025). Furthermore, Ultimate Trackers can function independently of a VR headset, enhancing their accessibility for clinical motion analysis. The near-perfect concordance correlation coefficient (CCC > 0.999) for sacrum tracking across all gait speeds and directions supports its use in assessing trunk control during balance tasks, with potential applications in monitoring trunk stability (Ordoñez Nuñez et al., 2024) or estimating the dynamic postural stability (Hof et al., 2005). However, a slight reduction in concordance observed in foot trackers at higher gait speeds may limit their reliability in precision-demanding applications, such as sports biomechanics and detailed stride length analysis (Merker et al., 2023; Vox et al., 2021). 4.3. Study Limitations This study has several limitations. First, the study was conducted in a highly controlled laboratory environment. Since Ultimate Tracker performance is sensitive to lighting conditions (Kulozik & Jarrassé, 2024), the positional errors observed here may underestimate those encountered in less-controlled clinical or home settings. Second, the 12–13 frame loss at a nominal 100 Hz sampling rate (Kulozik & Jarrassé, 2024; Merker et al., 2023; Sitole et al., 2020; Vox et al., 2021) was addressed through PCHIP interpolation, and SteamVR calibration errors (Fig. 3 ) were corrected using the Kabsch algorithm. These steps highlight the need for practical procedures to routinely correct calibration errors in applied settings. Third, we analysed only repetitive gait patterns with continuous Ultimate Tracker visibility. Consequently, our findings may not generalise to complex movements involving occlusions, where intermittent tracking could introduce additional errors. CONCLUSIONS The findings of this study suggest that the Ultimate Tracker is a promising low-cost motion capture system, capable of achieving tracking accuracy ranging from millimetre- to centimetre-level, depending on gait speed, tracker location, and movement direction. While tracking performance remains robust overall, accuracy slightly declines at higher gait speeds (e.g., 2.0 m/s), particularly at the feet during the swing phase, where movement velocities are greatest. Despite this limitation, the Ultimate Tracker demonstrates sufficient precision for clinical gait assessment. Furthermore, its integration with immersive VR platforms offers valuable opportunities for quantifying human–environment interactions in dynamic, real-world contexts. With continued technical refinement and targeted validation, the Ultimate Tracker has strong potential to bridge the gap between laboratory-grade motion capture systems and more accessible, real-world applications—supporting broader adoption across clinical, research, and industrial domains. Declarations ACKNOWLEDGEMENTS We would like to thank Mr George Mitri for his contribution during the technical development of the VR system and Ms Carly Chaplin for her contribution to data collection and processing. We thank all study participants who volunteered to participate in this study. CONTRIBUTORSHIP Y.H. – Software, methodology, formal analysis, data curation, visualization, writing – original draft. J.K. – Conceptualisation, software, supervision, writing – review & editing. M.A.B. – Conceptualisation, supervision, writing – review & editing. P.H. – Formal analysis, writing – review & editing. S.R.L. – Conceptualisation, supervision, writing – review & editing. Y.O. – Conceptualisation, methodology, validation, investigation, resources, supervision, project administration, funding acquisition, writing – review & editing. Competing interests The authors declare no competing interests. Ethical approval This study was approved by the University of New South Wales (UNSW) Human Research Ethics Committee (No: iRECS6818) and was performed according to the Declaration of Helsinki. Informed consent All participants provided written informed consent prior to the experiments. FUNDING This work was supported by the UNSW University International Postgraduate Award (UIPA). References Agostini, V., Balestra, G., & Knaflitz, M. (2014). Segmentation and Classification of Gait Cycles. IEEE transactions on neural systems and rehabilitation engineering , 22 (5), 946-952. https://doi.org/10.1109/TNSRE.2013.2291907 Akoglu, H. (2018). User's guide to correlation coefficients. Turkish journal of emergency medicine , 18 (3), 91-93. https://doi.org/10.1016/j.tjem.2018.08.001 Al-Tawil, B., Hempel, T., Abdelrahman, A., & Al-Hamadi, A. (2024). A review of visual SLAM for robotics: evolution, properties, and future applications. Frontiers in robotics and AI , 11 , 1347985-1347985. https://doi.org/10.3389/frobt.2024.1347985 Barker, P. M., & McDougall, T. J. (2020). Two Interpolation Methods Using Multiply-Rotated Piecewise Cubic Hermite Interpolating Polynomials. Journal of atmospheric and oceanic technology , 37 (4), 605-619. https://doi.org/10.1175/JTECH-D-19-0211.1 Benesty, J., Chen, J., Huang, Y., & Cohen, I. (2009). Pearson Correlation Coefficient. In (pp. 1-4). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-00296-0_5 Blanca, M. J., Alarcon, R., Arnau, J., Bono, R., & Bendayan, R. (2017). Non-normal data: Is ANOVA still a valid option? Psicothema , 29 (4), 552. https://doi.org/10.7334/psicothema2016.383 Borges, M., Symington, A., Coltin, B., Smith, T., & Ventura, R. (2018, 2018). HTC Vive: Analysis and Accuracy Improvement. Caserman, P., Garcia-Agundez, A., Konrad, R., Göbel, S., & Steinmetz, R. (2019). Real-time body tracking in virtual reality using a Vive tracker. Virtual Reality : the Journal of the Virtual Reality Society , 23 (2), 155-168. https://doi.org/10.1007/s10055-018-0374-z Czajkowska, U., Żuk, M., Pezowicz, C., Popek, M., Łopusiewicz, M., & Bulińska, K. (2024). Low-Cost Motion Tracking Systems in the Kinematics Analysis of VR Game Users–Preliminary Study. Scientific Conference" Medical and Sport Technologies”, Demeco, A., Zola, L., Frizziero, A., Martini, C., Palumbo, A., Foresti, R., Buccino, G., & Costantino, C. (2023). Immersive Virtual Reality in Post-Stroke Rehabilitation: A Systematic Review. Sensors (Basel) , 23 (3). https://doi.org/10.3390/s23031712 Filippidis, A., Marmaras, N., Maravgakis, M., Plexousaki, A., Kamarianakis, M., & Papagiannakis, G. (2024). VR Isle Academy: A VR Digital Twin Approach for Robotic Surgical Skill Development. arXiv preprint arXiv:2406.00002 . Gałecki, A., Burzykowski, T., Gałecki, A., & Burzykowski, T. (2013). Linear mixed-effects model . Springer. Galna, B., Barry, G., Jackson, D., Mhiripiri, D., Olivier, P., & Rochester, L. (2014). Accuracy of the Microsoft Kinect sensor for measuring movement in people with Parkinson's disease. Gait & Posture , 39 (4), 1062-1068. https://doi.org/10.1016/j.gaitpost.2014.01.008 Giavarina, D. (2015). Understanding Bland Altman analysis. Biochemia medica , 25 (2), 141-151. https://doi.org/10.11613/BM.2015.015 Gu, C., Lin, W., He, X., Zhang, L., & Zhang, M. (2023). IMU-based motion capture system for rehabilitation applications: A systematic review. Biomimetic intelligence and robotics , 3 (2), 100097. https://doi.org/10.1016/j.birob.2023.100097 Gupta, A., & Semwal, V. B. (2022). Occluded Gait reconstruction in multi person Gait environment using different numerical methods. Multimedia tools and applications , 81 (16), 23421-23448. https://doi.org/10.1007/s11042-022-12218-2 He, Y., Lee, J., Kim, J., Brodie, M. A., Mitri, G., van Schooten, K. S., Lovell, N. H., Lord, S. R., & Okubo, Y. (2025). Virtual Obstacle-Avoidance Training Using Daily-Life Obstacles with Physical Feedback in Older People: A Cross-Over Trial. Available at SSRN 5104022 . https://doi.org/10.2139/ssrn.5104022 Hof, A. L., Gazendam, M. G. J., & Sinke, W. E. (2005). The condition for dynamic stability. Journal of Biomechanics , 38 (1), 1-8. https://doi.org/10.1016/j.jbiomech.2004.03.025 HTC VIVE. (2025). https://www.vive.com/au/accessory/vive-ultimate-tracker/ . Kabsch, W. (1976). A solution for the best rotation to relate two sets of vectors. Acta Crystallographica Section A , 32 (5), 922-923. https://doi.org/10.1107/S0567739476001873 Kim, W., Huang, C., Yun, D., Saakes, D., Xiong, S., Murata, A., Goonetilleke, R. S., Xiong, S., Goossens, R. H. M., & Karwowski, W. (2020). Comparison of Joint Angle Measurements from Three Types of Motion Capture Systems for Ergonomic Postural Assessment. In (Vol. 1215, pp. 3-11). Springer International Publishing AG. https://doi.org/10.1007/978-3-030-51549-2_1 Kulozik, J. (2024). Ultimate tracker python https://github.com/jkulozik . gitHub repository. Retrieved 30/Nov/2024 from Kulozik, J., & Jarrassé, N. (2024). Evaluating the precision of the HTC VIVE Ultimate Tracker with robotic and human movements under varied environmental conditions. arXiv preprint arXiv:2409.01947 . Lee, J., Phu, S., Lord, S. R., & Okubo, Y. (2024). Effects of immersive virtual reality training on balance, gait and mobility in older adults: A systematic review and meta-analysis. Gait & Posture , 110 , 129-137. https://doi.org/10.1016/j.gaitpost.2024.03.009 Lima, Y. L., Collings, T., Hall, M., Bourne, M. N., & Diamond, L. E. (2024). Validity and reliability of trunk and lower-limb kinematics during squatting, hopping, jumping and side-stepping using OpenCap markerless motion capture application. Journal of sports sciences , 42 (19), 1847-1858. https://doi.org/10.1080/02640414.2024.2415233 Lin, L. I. (1989). A Concordance Correlation Coefficient to Evaluate Reproducibility. Biometrics , 45 (1), 255-268. https://doi.org/10.2307/2532051 McBride, G. (2005). A proposal for strength-of-agreement criteria for Lin’s concordance correlation coefficient. NIWA client report: HAM2005-062 , 45 , 307-310. Menolotto, M., Komaris, D.-S., Tedesco, S., O'Flynn, B., & Walsh, M. (2020). Motion Capture Technology in Industrial Applications: A Systematic Review. Sensors (Basel, Switzerland) , 20 (19), 5687. https://doi.org/10.3390/s20195687 Merker, S., Pastel, S., Bürger, D., Schwadtke, A., & Witte, K. (2023). Measurement Accuracy of the HTC VIVE Tracker 3.0 Compared to Vicon System for Generating Valid Positional Feedback in Virtual Reality. Sensors (Basel, Switzerland) , 23 (17), 7371. https://doi.org/10.3390/s23177371 Nijmeijer, E. M., Heuvelmans, P., Bolt, R., Gokeler, A., Otten, E., & Benjaminse, A. (2023). Concurrent validation of the Xsens IMU system of lower-body kinematics in jump-landing and change-of-direction tasks. Journal of Biomechanics , 154 , 111637-111637. https://doi.org/10.1016/j.jbiomech.2023.111637 Okubo, Y., He, Y., Brodie, M., Hicks, C., Schooten, K. v., Lovell, N. H., Anstey, K., J., Lord, S. R., & Kim, J. (2025). Virtual reality obstacle avoidance training can be enhanced by physical feedback via perturbations: a proof-of-concept study. Applied Ergonomics . https://doi.org/10.1016/j.apergo.2024.104442 Ordoñez Nuñez, T., Garcia Ramirez, A. R., & Becherán Marón, L. (2024). Analysis of waist and wrist positioning wearable machine learning models to detect falls. Electronics letters , 60 (2), n/a. https://doi.org/10.1049/ell2.13086 Pfister, A., West, A. M., Bronner, S., & Noah, J. A. (2014). Comparative abilities of Microsoft Kinect and Vicon 3D motion capture for gait analysis. Journal of medical engineering & technology , 38 (5), 274-280. https://doi.org/10.3109/03091902.2014.909540 Putranto, J. S., Heriyanto, J., Kenny, Achmad, S., & Kurniawan, A. (2023). Implementation of virtual reality technology for sports education and training: Systematic literature review. Procedia computer science , 216 , 293-300. https://doi.org/10.1016/j.procs.2022.12.139 Raffegeau, T. E., Brinkerhoff, S. A., Kellaher, G. K., Baudendistiel, S., Terza, M. J., Roper, J. A., & Hass, C. J. (2022). Changes to margins of stability from walking to obstacle crossing in older adults while walking fast and during a dual-task. Experimental Gerontology , 161 , 111710-111710. https://doi.org/10.1016/j.exger.2022.111710 Ranjan, R., Ahmedt-Aristizabal, D., Armin, M. A., & Kim, J. (2025). Computer vision for clinical gait analysis: A gait abnormality video dataset. IEEE Access , 13 , 45321-45339. https://doi.org/10.1109/ACCESS.2025.3545787 Rhiel, S., Kläy, A., Keller, U., van Hedel, H. J. A., & Ammann-Reiffer, C. (2024). Comparing Walking-Related Everyday Life Tasks of Children with Gait Disorders in a Virtual Reality Setup With a Physical Setup: Cross-Sectional Noninferiority Study. JMIR serious games , 12 , e49550-e49550. https://doi.org/10.2196/49550 Rybnikár, F., Kačerová, I., Hořejší, P., & Šimon, M. (2023). Ergonomics Evaluation Using Motion Capture Technology—Literature Review. Applied sciences , 13 (1), 162. https://doi.org/10.3390/app13010162 Sheng, X., Mao, S., Yan, Y., & Yang, X. (2024). Review on SLAM algorithms for Augmented Reality. Displays , 84 , 102806. https://doi.org/10.1016/j.displa.2024.102806 Sitole, S. P., LaPre, A. K., & Sup, F. C. (2020). Application and Evaluation of Lighthouse Technology for Precision Motion Capture. IEEE sensors journal , 20 (15), 8576-8585. https://doi.org/10.1109/JSEN.2020.2983933 Slater, M., & Sanchez-Vives, M. V. (2016). Enhancing our lives with immersive virtual reality. Frontiers in robotics and AI , 3 , 74. Sousa, C. V., Lee, K., Alon, D., Sternad, D., & Lu, A. S. (2023). A Systematic Review and Meta-analysis of the Effect of Active Video Games on Postural Balance. Arch Phys Med Rehabil , 104 (4), 631-644. https://doi.org/10.1016/j.apmr.2023.01.002 Spitzley, K. A., & Karduna, A. R. (2019). Feasibility of using a fully immersive virtual reality system for kinematic data collection. Journal of Biomechanics , 87 , 172-176. https://doi.org/10.1016/j.jbiomech.2019.02.015 Subramanian, R., & Sarkar, S. (2019). Evaluation of Algorithms for Orientation Invariant Inertial Gait Matching. IEEE transactions on information forensics and security , 14 (2), 304-318. https://doi.org/10.1109/TIFS.2018.2850032 Svetek, A., Morgan, K., Burland, J., & Glaviano, N. R. (2025). Validation of OpenCap on lower extremity kinematics during functional tasks. Journal of Biomechanics , 183 , 112602. https://doi.org/10.1016/j.jbiomech.2025.112602 Topley, M., & Richards, J. G. (2020). A comparison of currently available optoelectronic motion capture systems. Journal of Biomechanics , 106 , 109820-109820. https://doi.org/10.1016/j.jbiomech.2020.109820 Vox, J. P., Weber, A., Wolf, K. I., Izdebski, K., Schüler, T., König, P., Wallhoff, F., & Friemert, D. (2021). An Evaluation of Motion Trackers with Virtual Reality Sensor Technology in Comparison to a Marker-Based Motion Capture System Based on Joint Angles for Ergonomic Risk Assessment. Sensors (Basel, Switzerland) , 21 (9), 3145. https://doi.org/10.3390/s21093145 Weber, M., Hartl, R., Zäh, M. F., & Lee, J. (2023). Dynamic Pose Tracking Accuracy Improvement via Fusing HTC Vive Trackers and Inertia Measurement Units. International journal of precision engineering and manufacturing , 24 (9), 1661-1674. https://doi.org/10.1007/s12541-023-00891-8 Zhou, C., Qian, Y., & Kaner, J. (2024). A study on smart home use intention of elderly consumers based on technology acceptance models. PLoS One , 19 (3), e0300574-e0300574. https://doi.org/10.1371/journal.pone.0300574 Zhou, H., & Hu, H. (2008). Human motion tracking for rehabilitation—A survey. Biomedical signal processing and control , 3 (1), 1-18. https://doi.org/10.1016/j.bspc.2007.09.001 Additional Declarations No competing interests reported. Supplementary Files HoloWalk3AppendixAFramespersecond.pptx HoloWalk3AppendixBLinearMixedModeltable.xlsx HoloWalk3AppendixCBlandAltmanplotsV4.pptx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6989733","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":518743415,"identity":"e1165929-a8a4-4137-b914-e59c14e1682f","order_by":0,"name":"Yixuan He","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9UlEQVRIiWNgGAWjYDACZgiVwMDeABNKIFYLzwFitcCVSSQg2HiBwXEeM4mfOxjy+CXfHnzMu8eOgZ89x4DhZxseLYfZ0iR7zzAUS87OSzbmeZbMINnzxoCxF68W5mMSvG0MiRtu55hJ8xxgZjC4AbSFF68WxjbJv0At+2+eMf/Nc6CewR6ohfEvAVukwbZI8Jgx8xw4zGAgkWPAjM8WycNsydaybRKJM87kGEvOOXCcR+LMs4LDMudwa+E7f8bw5ts2m8T+9jOGH94cqJbjb0/e+PBNGW4tCgfAlARcgAdEHMCtgYFBvgGf7CgYBaNgFIwCEAAA5uNMSGgL5xMAAAAASUVORK5CYII=","orcid":"","institution":"University of New South Wales","correspondingAuthor":true,"prefix":"","firstName":"Yixuan","middleName":"","lastName":"He","suffix":""},{"id":518743416,"identity":"015e8f59-2001-46a2-93cf-2de13e6a6ff1","order_by":1,"name":"Matthew A. Brodie","email":"","orcid":"","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"A.","lastName":"Brodie","suffix":""},{"id":518743417,"identity":"845412cf-7888-4fc7-8d73-25cc4a2ff9b6","order_by":2,"name":"Juno Kim","email":"","orcid":"","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Juno","middleName":"","lastName":"Kim","suffix":""},{"id":518743418,"identity":"a9e13be8-bb4b-410e-a29f-d2fe24729b25","order_by":3,"name":"Peter Humburg","email":"","orcid":"","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Humburg","suffix":""},{"id":518743419,"identity":"797649b8-4318-4a78-9fcb-7ac2c7ccde07","order_by":4,"name":"Stephen R. Lord","email":"","orcid":"","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"R.","lastName":"Lord","suffix":""},{"id":518743420,"identity":"68c21225-4197-4f80-8664-43ce337a6086","order_by":5,"name":"Yoshiro Okubo","email":"","orcid":"","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Yoshiro","middleName":"","lastName":"Okubo","suffix":""}],"badges":[],"createdAt":"2025-06-27 08:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6989733/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6989733/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92615588,"identity":"e5b9c7e5-edd1-46ea-9af1-2674d069d131","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3238615,"visible":true,"origin":"","legend":"","description":"","filename":"TextHoloWalk3PaperV102024.06.22.docx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/59deb64c01f2f38f3e37a642.docx"},{"id":92615583,"identity":"b829e5a3-4613-4cdf-a05d-ca384d8ed4a3","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7793,"visible":true,"origin":"","legend":"","description":"","filename":"b2ebfb4ef3484a399a1078abddcde700.json","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/4a253daf4d971f9cc688d6a9.json"},{"id":92615585,"identity":"6378273c-fd6f-490e-82e7-2f04e653db38","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"pptx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":61245,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixAFramespersecond.pptx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/32270d42ba90019bf10caf5f.pptx"},{"id":92615066,"identity":"8c35e0f5-2c20-47e4-8a87-d466cd54f4d2","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12027,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixBLinearMixedModeltable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/d482f22b1c8c7ef65916ca72.xlsx"},{"id":92615589,"identity":"bce63e38-b52e-4c0e-80b3-234cd7a8790d","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"pptx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1084634,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixCBlandAltmanplotsV4.pptx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/c75e8563cd8106a7eb8b8625.pptx"},{"id":92615070,"identity":"4d3f0abd-82e5-4f33-bff6-294abcd8cbac","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"xml","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106641,"visible":true,"origin":"","legend":"","description":"","filename":"b2ebfb4ef3484a399a1078abddcde7001enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/d5950ea88a03dbedc905bf7f.xml"},{"id":92615077,"identity":"62656fc6-5375-47b0-9cc8-d6689af178ce","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1801088,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/32e1151ad1ce60a55a80305b.png"},{"id":92615072,"identity":"42da7f8e-8e41-45f1-b7e0-4e7021f8b4f6","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":549903,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/70d3bab125d9e6a203697f3e.png"},{"id":92615068,"identity":"23d5df80-d06b-4918-9174-fa7c6beddfcb","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":73030,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/2e191a0546df306edbd6e3f6.png"},{"id":92615593,"identity":"4ad3e10e-20b1-4099-a388-64a28aea8fe4","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":152850,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/dc9beb5e5917dac928d03bf3.png"},{"id":92615082,"identity":"20358ed3-7675-4b32-a6de-5da316406f8b","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":329423,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/0d73381c3745d5ea645f0078.png"},{"id":92615087,"identity":"b63a9d2f-b753-4ce5-927c-24d517a6d815","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":161228,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/8d98e344c36966be6a0bf707.png"},{"id":92615596,"identity":"79e114d6-a458-41e8-bdb1-8914543b239e","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":225956,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/7e71fbaab53f0d7afb287a26.png"},{"id":92615590,"identity":"8c05f34a-6bbf-408a-b387-a1ee1e35eebb","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":127529,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/d7593b2eea9663ac412d9ad0.png"},{"id":92615074,"identity":"baeaeb4a-d9a5-4a86-8c99-6036a6181998","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20323,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/6595761d36948a118159dca2.png"},{"id":92615085,"identity":"f398692b-529b-4338-b3ba-7667677f0ae2","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35716,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/1150e1175b75a716bd6d4b6e.png"},{"id":92615592,"identity":"a217fece-24c4-44c6-ab0a-fa68aadf7721","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":71276,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/8f42e56e594b275b992f0969.png"},{"id":92615079,"identity":"3a4abd81-7d9f-4c0c-94f5-b7397a89425b","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48946,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/3f8c4b6f3a91881363f740b5.png"},{"id":92617506,"identity":"269bea3c-87a0-4566-ba56-264975ae92ff","added_by":"auto","created_at":"2025-10-01 17:52:41","extension":"xml","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103666,"visible":true,"origin":"","legend":"","description":"","filename":"b2ebfb4ef3484a399a1078abddcde7001structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/0318f2b534204973d36f0661.xml"},{"id":92617181,"identity":"d205db9c-4ab7-4a7c-b1f6-01010801d907","added_by":"auto","created_at":"2025-10-01 17:44:41","extension":"html","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":116429,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/b6cc7d99e25bf823740db332.html"},{"id":92617178,"identity":"051a08d0-9294-4b0a-9b28-48432e3730e9","added_by":"auto","created_at":"2025-10-01 17:44:41","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1404738,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e A participant walking on the treadmill. \u003cstrong\u003eb\u003c/strong\u003e VIVE Ultimate tracker and reflective maker on the back sacrum. \u003cstrong\u003ec\u003c/strong\u003e VIVE Ultimate trackers and reflective markers on the right and left foot. \u003cstrong\u003ed\u003c/strong\u003e The participant’s view while in the head-mounted display with apples in the distance\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/42ab013337fc736674040198.png"},{"id":92615064,"identity":"16c29935-367d-42d6-a5d2-ed5ab430d999","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":549903,"visible":true,"origin":"","legend":"\u003cp\u003eVicon and HTC VIVE Ultimate Tracker (HVUT) data of a participant’s 1.5 m/s speed after transformation and synchronisation. ML: medio-lateral, AP: anterior-posterior, VT: vertical\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/7dacaf4568e35365f9380d30.png"},{"id":92617505,"identity":"b983e5bb-2de7-40e0-bf66-1b3f3ee602d3","added_by":"auto","created_at":"2025-10-01 17:52:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":73030,"visible":true,"origin":"","legend":"\u003cp\u003eThe SteamVR platform calibration errors when drawing the number 8. ML: medio-lateral, AP: anterior-posterior, VT: vertical.These errors were corrected using Kabsch algorithm prior to the subsequent analyses\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/122196219be9b0689162b454.png"},{"id":92615587,"identity":"d47b2d54-2574-4951-8b69-b7c618c26719","added_by":"auto","created_at":"2025-10-01 17:36:41","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":177086,"visible":true,"origin":"","legend":"\u003cp\u003eThe visualisation of absolute errors between Vicon and Ultimate Trackers based on (\u003cstrong\u003ea\u003c/strong\u003e) the gait speed and tracker location interaction, (\u003cstrong\u003eb\u003c/strong\u003e) the direction and gait speed interaction, and (\u003cstrong\u003ec\u003c/strong\u003e) the tracker location and direction interaction. ML: medio-lateral, AP: anterior-posterior, VT: vertical. Error bar: standard error of the mean\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/914597c13fe6c6c07ef6890a.png"},{"id":92615065,"identity":"e853f99a-a0f6-41b0-8c28-11a9869c2f01","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":421137,"visible":true,"origin":"","legend":"\u003cp\u003eThe Bland-Altman plots of trackers on the sacrum and right foot during 0.5 m/s speed (sample size n = 500). mean diff: mean difference, SD: standard deviation, 95% LoA: limits of agreement, ML: medio-lateral, AP: anterior-posterior, VT: vertical\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/d6af03d90c35e3bf803729a5.png"},{"id":92615081,"identity":"4e9e2b6a-51b2-4f19-be33-4bc722022338","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":161228,"visible":true,"origin":"","legend":"\u003cp\u003eThe concordance correlation coefficient for all speeds, tracker locations, directions, and related 95% confidence interval. ML: medio-lateral, AP: anterior-posterior, VT: vertical\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/1b2541bfcddf52a3a65f8ca7.png"},{"id":96083038,"identity":"f33d7993-b46c-4729-ad6c-38f1b6e76298","added_by":"auto","created_at":"2025-11-17 11:54:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3864184,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/ddc04195-5413-46a8-a762-13e0f69f1a10.pdf"},{"id":92615059,"identity":"4bc7f23f-a617-4f83-be3b-9d07ad3d2901","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"pptx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":61245,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixAFramespersecond.pptx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/5a1cee756ecbd4a5e2f3c2d6.pptx"},{"id":92615062,"identity":"f3fd1e79-370b-4570-b3c5-ab78a0b0a549","added_by":"auto","created_at":"2025-10-01 17:28:41","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":12027,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixBLinearMixedModeltable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/d15e603caad6baa224ae8f7f.xlsx"},{"id":92617180,"identity":"4a79ad9f-2bd6-4c4f-826e-266cca5caea2","added_by":"auto","created_at":"2025-10-01 17:44:41","extension":"pptx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1084634,"visible":true,"origin":"","legend":"","description":"","filename":"HoloWalk3AppendixCBlandAltmanplotsV4.pptx","url":"https://assets-eu.researchsquare.com/files/rs-6989733/v1/c6ada6918f887c7d8be8b236.pptx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Validation of the HTC VIVE Ultimate Trackers Compared with the Vicon Motion Capture System at Slow, Moderate and Fast Gait Speeds","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eMotion capture systems enable the precise recording and analysis of human movement within defined spatial environments (Rybnik\u0026aacute;r et al., 2023). These technologies have proved valuable across a wide range of fields, including healthcare, physical rehabilitation, sports biomechanics, education, ergonomics, robotics, filmmaking, and virtual reality (VR), driving interdisciplinary innovation through accurate quantification of human motion (Gu et al., 2023; Menolotto et al., 2020; Ranjan et al., 2025). Among these, optical motion capture systems\u0026mdash;particularly marker-based systems such as Vicon (Vicon Motion Systems Ltd., Oxford, UK), Qualisys (Qualisys, Gothenburg, Sweden), and OptiTrack (OptiTrack, Oregon, USA) \u0026mdash;are widely regarded as the \u0026ldquo;gold standard\u0026rdquo; in motion analysis due to their excellent positional accuracy, often within sub-centimetre ranges (~\u0026thinsp;1 mm error) (Gu et al., 2023; Topley \u0026amp; Richards, 2020; Zhou \u0026amp; Hu, 2008). However, these systems come with notable limitations. Their dependence on fixed arrays of infrared cameras, specialized mounting infrastructure, high-performance computing resources, and controlled laboratory environments imposes significant spatial and financial constraints. These requirements limit their accessibility and scalability, particularly in low-resource settings and for personal or home-based applications (Gu et al., 2023).\u003c/p\u003e\u003cp\u003eWith advancements in graphics technology, VR now enables the creation of diverse, immersive environments (Slater \u0026amp; Sanchez-Vives, 2016), enhancing our ability to study and train human performance in simulated, real-world scenarios (Lee et al., 2024; Sousa et al., 2023). This capability has been increasingly applied in rehabilitation (Demeco et al., 2023), sports training (Putranto et al., 2023), and exercise interventions (Lee et al., 2024), offering more task-specific and contextually relevant applications. Several studies have integrated VR head-mounted displays (HMDs) with gait training to explore innovative strategies for improving performance in simulated daily environments (He et al., 2025; Okubo et al., 2025; Rhiel et al., 2024). Despite these promising developments, translating research findings into real-world applications remain challenging\u0026mdash;largely due to the limitations of current motion capture technologies.\u003c/p\u003e\u003cp\u003eWhile some studies have employed low-cost motion capture systems such as OpenCap (Stanford University, USA), Kinect (Microsoft, Washington, USA), VIVE Trackers (HTC, Taiwan, China), and Xsens (Xsens, Enschede, Netherlands), each has its limitations. OpenCap needs post-capture data processing at USA, has lower sample rate (60 Hz), and reduced kinematic accuracy (Lima et al., 2024; Svetek et al., 2025). Kinect\u0026rsquo;s low sample rate (30\u0026ndash;37 Hz) and single camera-setup hinder 360\u0026deg; motion capture (Galna et al., 2014; Kim et al., 2020; Pfister et al., 2014). VIVE Trackers (2.0/3.0) have required multiple infrared base stations, limiting use outside controlled in non-laboratory or home-based settings (Merker et al., 2023; Spitzley \u0026amp; Karduna, 2019). Xsens sensors also show reduced accuracy in measuring flexion and movements in transversal and frontal planes (Kim et al., 2020; Nijmeijer et al., 2023).\u003c/p\u003e\u003cp\u003eEmerging consumer-grade technologies offer promising solutions to the limitations of traditional motion capture systems. One such advancement is the use of Simultaneous Localization and Mapping (SLAM) algorithms, which enable real-time estimation of a device\u0026rsquo;s position while concurrently mapping the surrounding environment using onboard sensors (Al-Tawil et al., 2024). In VR, SLAM forms the foundation of inside-out tracking systems, eliminating the need for external cameras (Sheng et al., 2024). The HTC VIVE Ultimate Tracker leverages SLAM technology with dual inside-out cameras, and advanced computer vision algorithms to achieve real-time spatial awareness, and synchronization between physical movements and virtual environments (HTC VIVE, 2025). With a simplified setup requiring only a computer and tracker units, the Ultimate Tracker presents good potential for accessible and portable motion capture (Filippidis et al., 2024; Kulozik \u0026amp; Jarrass\u0026eacute;, 2024). For instance, Kulozik et al. reported the Ultimate Tracker compared well against the OptiTrack gold standard for robotic arm tracking under unusually slow, controlled conditions (0.05\u0026ndash;0.40 m/s) (Kulozik \u0026amp; Jarrass\u0026eacute;, 2024). However, it remains unclear how the Ultimate Tracker performs in capturing human movement, particularly under common dynamic conditions involving higher peak velocities, such as walking.\u003c/p\u003e\u003cp\u003eTo address these knowledge gaps, this study aimed to validate the Ultimate Tracker system against the Vicon gold-standard motion capture system during human walking. Participants walked at four speeds\u0026mdash;slow, normal, fast, and very fast\u0026mdash;within a virtual suburban environment (He et al., 2025). Ultimate Tracker sensors were positioned at the sacrum, right foot, and left foot to capture key gait parameters. A comprehensive analytical approach was used to evaluate agreement between systems, including mixed-effects modelling to account for within-subject variability, as well as assessments of systematic bias and concordance. This multifaceted evaluation provides robust insights into the technical accuracy and consistency of the Ultimate Tracker and its suitability for use in human movement research within both real-world and virtual environments.\u003c/p\u003e"},{"header":"2. METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Study design and ethics\u003c/h2\u003e\u003cp\u003e This study was conducted at Neuroscience Research Australia (NeuRA) in accordance with the Declaration of Helsinki and approved by the University of New South Wales (UNSW) Human Research Ethics Committee (iRECS6818). Written informed consent was obtained from all participants before participation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Participants and sample size\u003c/h2\u003e\u003cp\u003eTen participants (men/women: 5/5) were recruited at NeuRA. Participants had a mean age of 32.2\u0026thinsp;\u0026plusmn;\u0026thinsp;6.5 years (range: 24\u0026ndash;44), with an average weight of 66.5\u0026thinsp;\u0026plusmn;\u0026thinsp;8.0 kg, height of 1.69\u0026thinsp;\u0026plusmn;\u0026thinsp;0.009 m, and body mass index (BMI) of 23.2\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8 kg/m\u0026sup2;. Inclusion criteria were: (i) able to communicate in the English language; (ii) aged 20\u0026ndash;70 years. Exclusion criteria were: (i) diagnosis of a neurological condition (e.g., Parkinson\u0026rsquo;s disease, multiple sclerosis, dementia); (ii) inability to walk for 20 minutes without a mobility aid or rest; (iii) history of lower limb, pelvic or vertebral fractures or lower limb joint replacements in the past 6 months; (iv) conditions that prevent exercising (e.g., severe pain, fatigue, exercise intolerance, heel ulcers); (v) advice from a medical practitioner to avoid exercise; (vi) history of dizziness, vertigo and vestibular disorders (e.g., Meniere's disease).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Experimental setup and procedure\u003c/h2\u003e\u003cp\u003eParticipants attended NeuRA for a 1.5-hour laboratory session. This session commenced with assessments of body weight and height. Participants were then fitted with an HMD for VR (VIVE XR Elite, HTC, Taiwan, China) and a safety harness. Three Ultimate Trackers were attached to the sacrum using a waist belt and to the instep of the left and right feet over the participant\u0026rsquo;s sports shoes using gaffer tape. Vicon reflective markers were placed at the diagonal intersection points on the surface of the Ultimate Trackers (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). SteamVR (Steam, Valve, USA) running on a desktop PC served as the platform for running the virtual environment and wirelessly streamed content to the HMD via Wi-Fi 6e. The virtual environment included concrete footpaths and roadways (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Participants undertook a 30-second familiarisation walk on a treadmill (M-Gait, Motek\u0026reg;, Netherlands) with a belt speed of 0.5 m/s in the VR environment. Participants then completed four consecutive 1-minute walking tasks at treadmill speeds of 0.5, 1.0, 1.5, and 2.0 m/s. The tasks were performed continuously, with participants notified of each upcoming speed change via a countdown prompt.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAn 8-camera Vicon motion capture system (Bonita, Vicon Motion Systems Ltd., Oxford, UK) and Vicon Nexus 2.16.0 x64 software were used to capture three-dimensional (3D) coordinates of the reflective markers at a sampling frequency of 100 Hz. An open-source Python script (Kulozik, 2024) and Visual Studio 2022 (Microsoft, USA) were utilised to collect 3D coordinates from the Ultimate Trackers during experiments at 100 Hz. The setting and data collection of Ultimate Trackers were based on the SteamVR platform. To aid Ultimate Tracker\u0026rsquo;s spatial recognition, black and white checkerboards were printed in A4 and posted on each wall.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4. Data processing\u003c/h2\u003e\u003cp\u003eThe Ultimate Trackers pose matrix was provided by SteamVR to transform the local coordinate system of Ultimate Tracker into the global coordinate system and map it to the Vicon position. Temporal synchronisation was performed using the same directional data recorded by the Ultimate Tracker system. After synchronising the time axes of the data from the two systems using a customised script in MATLAB R2022a (The MathWorks, Inc., USA), it was observed that the Ultimate Tracker experienced a sampling loss of a median of 12 Hz, resulting in an actual sampling rate quartile of 87\u0026ndash;88 Hz (Appendix A). To compensate for missing frames, piecewise cubic Hermite interpolating polynomial (PCHIP) interpolation was applied to the Ultimate Tracker data to match the data length of the Vicon system for each second (Gupta \u0026amp; Semwal, 2022). PCHIP was selected for its ability to preserve monotonicity and ensure continuity of first-order derivatives at nodes, thereby reducing overshoot artifacts commonly observed with cubic spline interpolation in gait data (Barker \u0026amp; McDougall, 2020).\u003c/p\u003e\u003cp\u003eThe Ultimate Trackers in the SteamVR coordinate system and Vicon data in the Vicon coordinate system were transformed into the same coordinate system prior to further data analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Thus, the Kabsch algorithm and singular value decomposition (SVD) (Kabsch, 1976; Subramanian \u0026amp; Sarkar, 2019) were utilised to compute transformations between the SteamVR and Vicon coordinate systems to find the translation and the rotation between the two measurements. During the overall transformation of the Ultimate Trackers from the SteamVR coordinate system to the Vicon coordinate system, it was observed that the Ultimate Tracker positions did not align precisely with the corresponding Vicon marker locations. Considering the Ultimate Trackers were not placed at exactly the same positions during each calibration session, this discrepancy suggests the presence of potential calibration errors inherent in the SteamVR-based calibration process (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). To address this, the Kabsch algorithm and SVD were applied to individually align the Ultimate Trackers to the corresponding Vicon markers, thereby correcting for the SteamVR-based calibration-induced spatial misalignment.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eDue to manual operation delays in both Vicon Nexus and the open-source scripts during data collection, a temporal offset emerged between the two datasets. To address this, the first and last 5 seconds were removed from both the Vicon and Ultimate Tracker recordings of each participant at each speed, retaining the central 50 seconds of data. Given the Vicon system\u0026rsquo;s sampling rate of 100 Hz, this 50-second segment contained 5000 frames. Using MATLAB, the experimental data were integrated and categorized based on participant ID (1\u0026ndash;10), gait speed (0.5, 1.0, 1.5 and 2.0 m/s), tracker location (sacrum, right foot, and left foot), and direction (medio-lateral [ML], anterior-posterior [AP], and vertical [VT] axes). Additionally, the differences and absolute errors between the Ultimate Trackers and the Vicon system were calculated for each corresponding frame.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5. Statistical analysis\u003c/h2\u003e\u003cp\u003eTo examine the impact of gait speed, tracker location and direction on the absolute errors between the Ultimate Trackers and the Vicon system, a linear mixed-effects model (LMM) was utilised (Gałecki et al., 2013). Four gait speeds, three tracker locations, and three directions were set as fixed interaction effects in the LMM, including main effects (speed, tracker location, and direction), two-way interaction effects (speed \u0026times; tracker location, speed \u0026times; direction, and tracker location \u0026times; direction), and three-way interaction effects (speed \u0026times; tracker location \u0026times; direction), while participant IDs were treated as random effects. Given the time-ordered nature of the collected data, where consecutive observations within each participant are likely to be autocorrelated, a first-order autoregressive (AR(1)) covariance structure was introduced for repeated measures within participants to account for frame-to-frame dependence (0.98). The speed 0.5 m/s, sacrum and vertical axis showing smallest absolute errors were set as the baseline level in the LMM.\u003c/p\u003e\u003cp\u003eBased on the VT direction data of left and right foot trackers, participants\u0026rsquo; gait cycles were segmented into four distinct phases: foot strike, stance, foot off, and swing (Agostini et al., 2014). Absolute error was calculated for three Ultimate Trackers for each phase to capture variations in accuracy throughout locomotion. One-way ANOVA was utilised to compare the differences between the four gait phases. Since the sample size of gait cycles was estimated to be 5000, the statistical test of the normal distribution could be omitted (Blanca et al., 2017). Levene's test was conducted to assess the homogeneity of variances. When the assumption of homogeneity was met, a standard ANOVA was used; otherwise, Welch\u0026rsquo;s ANOVA was applied in the presence of heterogeneity of variances.\u003c/p\u003e\u003cp\u003eThe limits of agreement (LoA) between the two systems, proportional bias and systematic bias were examined using Bland-Altman plots (Giavarina, 2015). Bland-Altman plots were generated for the four gait speeds, with data further subdivided by tracker location and three directions, resulting in 36 combinations. The X-axis represented the gold standard Vicon system measurements, while the Y-axis showed the differences between the Vicon and Ultimate Tracker systems. Since the walking gait trajectory had strong autocorrelation and each combination included 50,000 data points from all participants, plotting all the data points in the plot would reduce readability. Bland-Altman plots were constructed by randomly sampling from the gait trajectory every 100 frames (n\u0026thinsp;=\u0026thinsp;500). The LoAs are displayed as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Mean\\pm\\:1.96\\times\\:standard\\:deviation\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\u003cp\u003eThe concordance correlation coefficient (CCC) was used to quantify the extent to which paired observations deviated from the Vicon gold standard (Akoglu, 2018; Lin, 1989). Unlike correlation coefficients (Benesty et al., 2009), the CCC offers a robust measure of agreement by accounting for both correlation and bias (Lin, 1989). The CCC analysis were grouped by four speeds, with each group further subdivided based on direction and tracker location. The CCC\u0026thinsp;\u0026gt;\u0026thinsp;0.99 are considered to be almost perfect, \u0026gt;\u0026thinsp;0.95 to \u0026lt;\u0026thinsp;0.99 as excellent, \u0026gt;\u0026thinsp;0.90 to \u0026lt;\u0026thinsp;0.95 as moderate, and \u0026lt;\u0026thinsp;0.90 as poor (McBride, 2005). The fundamental calculation of the CCC is shown in Eq.\u0026nbsp;(1).\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}CCC=\\:\\frac{2{S}_{HVUT\\_Vicon}}{{S}_{HVUT}^{2}+{S}_{Vicon}^{2}+{\\left({\\mu\\:}_{HVUT}-{\\mu\\:}_{Vicon}\\right)}^{2}}\\#\\left(1\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{S}_{HVUT\\_Vicon}\\)\u003c/span\u003e\u003c/span\u003e is the covariance between the two sets of Ultimate Tracker and Vicon, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{S}_{HVUT}^{2}\\)\u003c/span\u003e\u003c/span\u003e​ and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{S}_{Vicon}^{2}\\)\u003c/span\u003e\u003c/span\u003e are the variances of the respective Ultimate Tracker and Vicon sets, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\mu\\:}_{HVUT}\\)\u003c/span\u003e\u003c/span\u003e​ and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\mu\\:}_{Vicon}\\)\u003c/span\u003e\u003c/span\u003e​ are the means of the Ultimate Tracker and Vicon sets. Statistical analyses were conducted using MATLAB R2022a and R (version 4.5.0; R Core Team, Vienna, Austria). \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. RESULTS","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.1. Absolute error\u003c/h2\u003e\u003cp\u003eAbsolute errors (median [interquartile range]) are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Absolute errors of sacrum location (all directions) at 0.5, 1.0 and 1.5 and 2.0 m/s were 1.04 (0.41, 2.58), 0.86 (0.37, 1.76), 0.92 (0.31, 1.96), and 1.24 (0.43, 2.49) mm, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Absolute errors of right foot location at 0.5, 1.0, 1.5 and 2.0 m/s were 1.85 (0.81, 4.05), 2.43 (1.06, 5.13), 2.46 (0.81, 5.87), and 2.81 (0.93, 6.50) mm, respectively.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAbsolute errors of 1.0 m/s in the ML, AP and VT directions (all locations) were 1.32 (0.59, 2.52), 3.86 (1.51, 8.20), 1.09 (0.45, 2.47) mm, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb).\u003c/p\u003e\u003cp\u003eAbsolute errors of sacrum, right and left foot locations at the AP direction (all speeds) were 1.92 (0.76, 3.90), 6.13 (2.52, 12.05), and 5.80 (2.38, 11.20) mm, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). See Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e for other conditions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.2. Linear mixed-effects model\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e summarises the total effect for each main effect and interaction from the LMM. There are statistically significant main effects of gait speed, direction, tracker location, and their interactions on the absolute error between Vicon and VIVE systems. Appendix B gives the parameter estimates. At the reference condition\u0026mdash;0.5 m/s at the sacrum in the vertical (VT) direction\u0026mdash;the estimated mean absolute error was 0.51 mm (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.271). Raising gait speed to 1.0 m/s, 1.5 m/s and 2.0 m/s increased error by 0.11 mm, 0.43 mm and 1.02 mm, respectively (all \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Compared with the sacrum, the right- and left-foot trackers incurred additional errors of 1.07 mm and 1.00 mm, respectively (both \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Directional main effects showed that errors in the ML direction exceeded VT by 1.15 mm, and errors in the AP direction exceeded VT by 3.15 mm (both \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of fixed effects from the linear mixed model assessing the influence of gait speed, tracker location, direction and their interactions on absolute errors between Vicon and VIVE systems.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cem\u003eEffects\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cem\u003eχ\u0026sup2;\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eDf\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGait speed\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e16057.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDirection\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e156662.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTracker location\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e506210.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGait speed \u0026times; Direction\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10773.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGait speed \u0026times; Tracker location\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e6688.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDirection \u0026times; Tracker location\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e101034.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGait speed \u0026times; Direction \u0026times; Tracker location\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10677.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eχ\u0026sup2;\u003c/em\u003e: Chi-square, \u003cem\u003eDf\u003c/em\u003e: degrees of freedom. The larger the χ\u0026sup2;, the more strongly the effect contributes to explaining the variance in absolute error between Vicon and VIVE. Appendix B gives the parameter estimates.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eTwo-way interactions revealed that the speed-related error increase differed by location and direction. Although AP motion carried the largest standalone error, its additional effect diminished at higher speeds (gait speed \u0026times; AP, \u003cem\u003eβ\u003c/em\u003e = \u0026minus;\u0026thinsp;1.04 mm at 1.0 m/s, \u0026minus;\u0026thinsp;2.09 mm at 2.0 m/s, all \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Similarly, gait speed \u0026times; ML interactions were negative (e.g., \u0026minus;\u0026thinsp;1.31 mm at 2.0 m/s, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), showing a relatively smaller ML error at fast speeds. Foot-specific interactions (gait speed \u0026times; tracker location) remained positive, though reduced at 2.0 m/s (right foot \u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.17 mm, left foot \u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.20 mm, both \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.3. Gait phases\u003c/h2\u003e\u003cp\u003eDue to heterogeneity of variances, Welch\u0026rsquo;s ANOVA was applied, and the results (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) showed that the absolute errors of Ultimate Tracker on the sacrum were not significantly different among the four gait phases (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.176, \u003cem\u003eω\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u0026thinsp;\u003cem\u003e=\u003c/em\u003e\u0026thinsp;0.0001). The Ultimate Tracker errors on the right and left foot were significantly higher during the swing phase compared to foot strike, stance, and foot off phases (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). There were no significant differences between foot strike, stance, and foot off for either the right or left foot.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eWelch\u0026rsquo;s ANOVA results, effect sizes, descriptive statistics and post hoc tests of root mean absolute error of gait phases for different locations.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eTracker location\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003eSignificance\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003eEffect size\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eSacrum\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003e\u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003e3, 32222\u003c/em\u003e\u003c/sub\u003e = 1.65, \u003cem\u003ep\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.176\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u003cem\u003eω\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u0026thinsp;\u003cem\u003e=\u003c/em\u003e\u0026thinsp;0.0001, Small effect\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eRight foot\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003e\u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003e3, 34070\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e=\u003c/em\u003e 42.90, \u003cem\u003ep\u0026thinsp;\u0026lt;\u003c/em\u003e\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u003cem\u003eω\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e \u0026thinsp;=\u0026thinsp;0.0001, Small effect\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eLeft foot\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003e\u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003e3, 34032\u003c/em\u003e\u003c/sub\u003e = 22.29, \u003cem\u003ep\u0026thinsp;\u0026lt;\u003c/em\u003e\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u003cem\u003eω\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e \u0026thinsp;=\u0026thinsp;0.0001, Small effect\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e\u003cb\u003ePhase of gait cycle\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003eMedian (IQR) (mm)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003e\u003cb\u003eComparison\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003ep\u003c/b\u003e \u003cb\u003e(post hoc)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSacrum\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot strike\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.92 (0.37, 2.06)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Stance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.176\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eStance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.11 (0.60, 2.10)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.996\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.02 (0.49, 2.02)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.901\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eSwing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.07 (0.59, 1.95)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.269\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.536\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot off-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.966\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRight foot\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot strike\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.85 (1.38, 5.23)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Stance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.900\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eStance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.45 (1.33, 4.70)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.993\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.37 (1.14, 4.86)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eSwing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.18 (1.82, 5.85)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.976\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot off-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLeft foot\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot strike\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.76 (1.37, 5.26)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Stance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.396\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eStance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.32 (1.30, 4.79)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.534\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eFoot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.21 (1.02, 4.70)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot strike-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eSwing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.01 (1.70, 5.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Foot off\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.996\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eStance-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eFoot off-Swing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003eIQR: Interquartile Range.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.4. Bland-Altman plots\u003c/h2\u003e\u003cp\u003eThe Bland-Altman plots (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and Appendix C) provide a visual and quantitative assessment of agreement between the Vicon and Ultimate Tracker systems. The significant slopes (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) suggest that the discrepancy between the methods changes with the magnitude of the measurement. At a gait speed of 0.5 m/s, there were no significant slopes in the VT direction for all tracker locations. At gait speeds of 1.5 m/s and 2.0 m/s, there were no significant slopes between Vicon and Ultimate Tracker in the VT direction for the sacrum location.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e3.5. Concordance correlation coefficient\u003c/h2\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e presents the CCC comparison between the Vicon and Ultimate Tracker measurements, alongside the mean and SD of the absolute errors. The heatmaps highlight CCC variations across conditions with values approaching 1 displayed in yellow. According to McBride\u0026rsquo;s strength-of-agreement criteria (McBride, 2005), all measurements achieved almost perfect (\u0026gt;\u0026thinsp;0.99) correlation at a speed of 0.5 m/s. At 1.0 m/s, 1.5 m/s, and 2.0 m/s, only the VT direction measurements of the right and left foot reached excellent correlation (0.95\u0026ndash;0.99); the other measurements all reached almost perfect correlation (\u0026gt;\u0026thinsp;0.99).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"4. DISCUSSION","content":"\u003cp\u003eThis study systematically evaluated the performance and concordance between the Ultimate Tracker and the Vicon motion capture system at different gait speeds, tracker locations, and movement directions. The findings indicate the Ultimate Tracker is sufficiently accurate for use in clinical- and research-settings as an alternative to gold standard systems for assessing gait at common speeds.\u003c/p\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.1. Factors Affecting the Accuracy of Ultimate Trackers\u003c/h2\u003e\u003cp\u003eOur findings indicate that the magnitude of error is influenced by gait speed, movement direction, and tracker location. The estimated absolute error of the Ultimate Tracker was lowest (0.51 mm) at the reference condition defined by the LMM: a gait speed of 0.5 m/s, sacrum tracker location, and VT movement direction (Appendix B). As gait speed increased to 1.0, 1.5, and 2.0 m/s, the estimated absolute error rose slightly by 0.11 mm, 0.43 mm, and 1.02 mm, respectively (Appendix B). This trend aligns with findings by Kulozik and Jarrasse, who reported an increase in Ultimate Tracker absolute error from 1.49 ± 0.50 mm to 3.07 ± 1.66 mm as the speed of a robotic arm increased from 0.05 m/s to 0.4 m/s (Kulozik \u0026amp; Jarrassé, 2024).\u003c/p\u003e\u003cp\u003eCompared to the sacrum-mounted tracker, the estimated absolute errors for trackers placed on the right and left feet increased slightly by 1.07 mm and 1.00 mm, respectively (Appendix B). This increase is likely due to the rapid accelerations experienced by the feet during the gait cycle, as well as the greater kinematic complexity associated with multi-joint coordination in the lower limbs (Czajkowska et al., 2024; Raffegeau et al., 2022). Additionally, because the SLAM method relies on visual input from environmental landmarks, the camera views from foot-mounted trackers may be partially obstructed by the user’s body during walking, potentially reducing tracking accuracy.\u003c/p\u003e\u003cp\u003eDirectional asymmetries were also observed, with the largest deviations occurring in the AP direction (3.15 mm), which is characterized by the greatest movement amplitude during the foot swing phase. This finding is consistent with results showing higher absolute errors during the swing phase compared to the foot strike, stance, or foot-off phases (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Together, these results suggest that high-amplitude and high-speed movements consistently reduce the accuracy of Ultimate Trackers.\u003c/p\u003e\u003cp\u003eThe absolute error findings (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) demonstrate that Ultimate Tracker maintains accuracy ranging from the millimetre to centimetre scale. These results contrast with robotic arm validation studies (Kulozik \u0026amp; Jarrassé, 2024), where slow, controlled movements (0.05–0.4 m/s) yielded smaller errors—for instance, an absolute error of 1.49 ± 0.50 mm at 0.05 m/s—compared to the errors observed at typical human gait speeds (1.0–1.5 m/s). Our slow gait condition (0.5 m/s) with the tracker on the sacrum produced a comparable error of 1.04 [0.41, 2.58] mm, emphasizing the limitations of applying rigid, predictable motion models to complex human biomechanics. Other robotic arm studies using earlier versions of the VIVE Tracker (Borges et al., 2018; Sitole et al., 2020; Weber et al., 2023) reported different outcomes. For example, gold-standard comparisons showed higher errors for the VIVE Tracker (Sitole et al., 2020), prompting some researchers to develop accuracy improvement algorithms (Borges et al.). However, since these earlier systems relied on infrared base stations—unlike the current Ultimate Tracker setup—their findings serve only as general references and are not directly comparable.\u003c/p\u003e\u003cp\u003eThe observed discrepancies across different speeds were consistent with previously reported Ultimate Tracker results. For instance, at 0.4 m/s, an absolute error of 3.07 ± 1.66 mm was reported (Kulozik \u0026amp; Jarrassé, 2024), while our findings at 0.5 m/s with the tracker on the left foot showed a comparable absolute error of 1.91 [0.83, 4.16] mm (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). These results represent an improvement over earlier VIVE trackers, such as the VIVE Tracker 3.0, which demonstrated significantly higher errors (e.g., at 80 rpm with a motion radius of 133.53 ± 19.24 mm, the mean absolute error was 11.33 ± 5.14 mm (Merker et al., 2023)). At higher speeds, such as 2.0 m/s, we observed an error of 2.81 [0.93, 6.50] mm at the right foot location (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). However, the AP direction at this speed exhibited a notably higher error of 4.24 [1.29, 11.01] mm (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), exceeding those in the ML and VT directions. This suggests that consumer-grade systems like Ultimate Tracker still exhibit centimetre-level errors in high-velocity applications.\u003c/p\u003e\u003cp\u003eSystematic bias and agreement\u003c/p\u003e\u003cp\u003eThe measurement discrepancy between the Vicon and Ultimate Tracker systems appears to increase with distance from the calibration origin, as indicated by the significant slopes in the Bland–Altman plots (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and Appendix C). Similar trends were reported in a previous study (Kulozik \u0026amp; Jarrassé, 2024), suggesting that the Ultimate Tracker may systematically overestimate or underestimate positional data as the distance from the calibration centre grows. This pattern is particularly evident in the AP direction, which involves the largest range of motion during gait. This directional bias may help explain the greater deviations observed in the AP direction (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Appendix B and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eRegarding error magnitude, the LoAs for the sacrum-mounted tracker were all below 10 mm across gait speeds, indicating millimetre-level accuracy. Similarly, for foot-mounted trackers, the LoAs in the ML and VT directions remained under 10 mm. However, in the AP direction, LoAs exceeded 10 mm, reflecting a transition from millimetre- to centimetre-level accuracy.\u003c/p\u003e\u003cp\u003eThe Bland–Altman analyses and CCC further define the operational boundaries of the Ultimate Tracker (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, and Appendix C). The sacrum tracker demonstrated near-perfect agreement across all gait speeds and directions (CCC \u0026gt; 0.999), suggesting that proximal body segments—with their lower kinematic complexity and minimal occlusion—are well-suited for Ultimate Tracker-based tracking. In contrast, foot-mounted tracker performance, particularly in the VT direction, declined slightly at higher gait speeds (CCC range: right foot 0.9929–0.9897, left foot 0.9932–0.9875), though still reflecting excellent agreement (CCC: 0.95–0.99 or above) (McBride, 2005).\u003c/p\u003e\u003cp\u003eAt higher gait speeds (2.0 m/s), the AP direction exhibited wide LoAs for foot-mounted trackers (right foot: − 26.58 to 24.12 mm; left foot: − 29.97 to 27.80 mm), indicating both systematic bias and substantial random error. These discrepancies are likely due to cumulative inaccuracies in the SLAM algorithm or hardware limitations during the swing phase—when foot motion in the AP direction is most pronounced (Agostini et al., 2014; Kulozik \u0026amp; Jarrassé, 2024).\u003c/p\u003e\u003cp\u003ePotential sources of errors\u003c/p\u003e\u003cp\u003eBoth biomechanical and system-level factors may contribute to the observed errors in Ultimate Tracker measurements, particularly at higher gait speeds. First, spatial inaccuracies are likely to accumulate during rapid swing phases, where transient occlusions—such as the foot tracker facing downward—can disrupt positional synchronisation. Second, velocity-dependent error amplification may stem from limitations in the Trackers’ SLAM algorithm and hardware capabilities. However, the proprietary nature of HTC’s tracking methods limits detailed interpretation. Third, the presence of 12–13 missing frames at a nominal 100 Hz sampling rate suggests a reduction in effective sampling frequency, likely compromising the system’s ability to capture fast movements accurately. In contrast, high-end systems like Vicon mitigate these issues by increasing the sampling rate (up to 250 Hz) and optimising imaging parameters—including camera gain, strobe intensity, and aperture size—to better accommodate rapid motion.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e4.2. Clinical and Research Implications\u003c/h2\u003e\u003cp\u003eThe demonstrated accuracy of the Ultimate Trackers at typical gait speeds appears for a range of applications, including VR experiences (Caserman et al., 2019), rehabilitation and gait training (Okubo et al., 2025), ergonomic risk assessment (Vox et al., 2021), and robot inverse kinematics (Weber et al., 2023). These trackers may be particularly valuable in home and clinical settings, where usability and affordability often take precedence over laboratory-grade precision (Zhou et al., 2024). Their integration with VR platforms enables quantification of human-environment interactions, and real-time feedback on gross motor movements such as foot-obstacle collisions (He et al., 2025). Furthermore, Ultimate Trackers can function independently of a VR headset, enhancing their accessibility for clinical motion analysis.\u003c/p\u003e\u003cp\u003eThe near-perfect concordance correlation coefficient (CCC \u0026gt; 0.999) for sacrum tracking across all gait speeds and directions supports its use in assessing trunk control during balance tasks, with potential applications in monitoring trunk stability (Ordoñez Nuñez et al., 2024) or estimating the dynamic postural stability (Hof et al., 2005). However, a slight reduction in concordance observed in foot trackers at higher gait speeds may limit their reliability in precision-demanding applications, such as sports biomechanics and detailed stride length analysis (Merker et al., 2023; Vox et al., 2021).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e4.3. Study Limitations\u003c/h2\u003e\u003cp\u003eThis study has several limitations. First, the study was conducted in a highly controlled laboratory environment. Since Ultimate Tracker performance is sensitive to lighting conditions (Kulozik \u0026amp; Jarrassé, 2024), the positional errors observed here may underestimate those encountered in less-controlled clinical or home settings. Second, the 12–13 frame loss at a nominal 100 Hz sampling rate (Kulozik \u0026amp; Jarrassé, 2024; Merker et al., 2023; Sitole et al., 2020; Vox et al., 2021) was addressed through PCHIP interpolation, and SteamVR calibration errors (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) were corrected using the Kabsch algorithm. These steps highlight the need for practical procedures to routinely correct calibration errors in applied settings. Third, we analysed only repetitive gait patterns with continuous Ultimate Tracker visibility. Consequently, our findings may not generalise to complex movements involving occlusions, where intermittent tracking could introduce additional errors.\u003c/p\u003e\u003c/div\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eThe findings of this study suggest that the Ultimate Tracker is a promising low-cost motion capture system, capable of achieving tracking accuracy ranging from millimetre- to centimetre-level, depending on gait speed, tracker location, and movement direction. While tracking performance remains robust overall, accuracy slightly declines at higher gait speeds (e.g., 2.0 m/s), particularly at the feet during the swing phase, where movement velocities are greatest. Despite this limitation, the Ultimate Tracker demonstrates sufficient precision for clinical gait assessment. Furthermore, its integration with immersive VR platforms offers valuable opportunities for quantifying human–environment interactions in dynamic, real-world contexts. With continued technical refinement and targeted validation, the Ultimate Tracker has strong potential to bridge the gap between laboratory-grade motion capture systems and more accessible, real-world applications—supporting broader adoption across clinical, research, and industrial domains.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eACKNOWLEDGEMENTS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Mr George Mitri for his contribution during the technical development of the VR system and Ms Carly Chaplin for her contribution to data collection and processing. We thank all study participants who volunteered to participate in this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCONTRIBUTORSHIP\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eY.H. \u0026ndash; Software, methodology, formal analysis, data curation, visualization, writing \u0026ndash; original draft.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eJ.K. \u0026ndash; Conceptualisation, software, supervision, writing \u0026ndash; review \u0026amp; editing.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eM.A.B. \u0026ndash; Conceptualisation, supervision, writing \u0026ndash; review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003eP.H. \u0026ndash; Formal analysis, writing \u0026ndash; review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003eS.R.L. \u0026ndash; Conceptualisation, supervision, writing \u0026ndash; review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003eY.O. \u0026ndash; Conceptualisation, methodology, validation, investigation, resources, supervision, project administration, funding acquisition, writing \u0026ndash; review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003eThe authors declare no competing interests.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u0026nbsp;\u003c/strong\u003eThis study was approved by the University of New South Wales (UNSW) Human Research Ethics Committee (No: iRECS6818) and was performed according to the Declaration of Helsinki.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed consent\u003c/strong\u003e All participants provided written informed consent prior to the experiments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFUNDING\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the UNSW University International Postgraduate Award (UIPA).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAgostini, V., Balestra, G., \u0026amp; Knaflitz, M. (2014). Segmentation and Classification of Gait Cycles. \u003cem\u003eIEEE transactions on neural systems and rehabilitation engineering\u003c/em\u003e,\u003cem\u003e 22\u003c/em\u003e(5), 946-952. https://doi.org/10.1109/TNSRE.2013.2291907 \u003c/li\u003e\n\u003cli\u003eAkoglu, H. (2018). User\u0026apos;s guide to correlation coefficients. \u003cem\u003eTurkish journal of emergency medicine\u003c/em\u003e,\u003cem\u003e 18\u003c/em\u003e(3), 91-93. https://doi.org/10.1016/j.tjem.2018.08.001 \u003c/li\u003e\n\u003cli\u003eAl-Tawil, B., Hempel, T., Abdelrahman, A., \u0026amp; Al-Hamadi, A. (2024). A review of visual SLAM for robotics: evolution, properties, and future applications. \u003cem\u003eFrontiers in robotics and AI\u003c/em\u003e,\u003cem\u003e 11\u003c/em\u003e, 1347985-1347985. https://doi.org/10.3389/frobt.2024.1347985 \u003c/li\u003e\n\u003cli\u003eBarker, P. M., \u0026amp; McDougall, T. J. (2020). Two Interpolation Methods Using Multiply-Rotated Piecewise Cubic Hermite Interpolating Polynomials. \u003cem\u003eJournal of atmospheric and oceanic technology\u003c/em\u003e,\u003cem\u003e 37\u003c/em\u003e(4), 605-619. https://doi.org/10.1175/JTECH-D-19-0211.1 \u003c/li\u003e\n\u003cli\u003eBenesty, J., Chen, J., Huang, Y., \u0026amp; Cohen, I. (2009). Pearson Correlation Coefficient. In (pp. 1-4). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-00296-0_5 \u003c/li\u003e\n\u003cli\u003eBlanca, M. J., Alarcon, R., Arnau, J., Bono, R., \u0026amp; Bendayan, R. (2017). Non-normal data: Is ANOVA still a valid option? \u003cem\u003ePsicothema\u003c/em\u003e,\u003cem\u003e 29\u003c/em\u003e(4), 552. https://doi.org/10.7334/psicothema2016.383 \u003c/li\u003e\n\u003cli\u003eBorges, M., Symington, A., Coltin, B., Smith, T., \u0026amp; Ventura, R. (2018, 2018). HTC Vive: Analysis and Accuracy Improvement. \u003c/li\u003e\n\u003cli\u003eCaserman, P., Garcia-Agundez, A., Konrad, R., G\u0026ouml;bel, S., \u0026amp; Steinmetz, R. (2019). Real-time body tracking in virtual reality using a Vive tracker. \u003cem\u003eVirtual Reality : the Journal of the Virtual Reality Society\u003c/em\u003e,\u003cem\u003e 23\u003c/em\u003e(2), 155-168. https://doi.org/10.1007/s10055-018-0374-z \u003c/li\u003e\n\u003cli\u003eCzajkowska, U., Żuk, M., Pezowicz, C., Popek, M., Łopusiewicz, M., \u0026amp; Bulińska, K. (2024). Low-Cost Motion Tracking Systems in the Kinematics Analysis of VR Game Users\u0026ndash;Preliminary Study. Scientific Conference\u0026quot; Medical and Sport Technologies\u0026rdquo;, \u003c/li\u003e\n\u003cli\u003eDemeco, A., Zola, L., Frizziero, A., Martini, C., Palumbo, A., Foresti, R., Buccino, G., \u0026amp; Costantino, C. (2023). Immersive Virtual Reality in Post-Stroke Rehabilitation: A Systematic Review. \u003cem\u003eSensors (Basel)\u003c/em\u003e,\u003cem\u003e 23\u003c/em\u003e(3). https://doi.org/10.3390/s23031712 \u003c/li\u003e\n\u003cli\u003eFilippidis, A., Marmaras, N., Maravgakis, M., Plexousaki, A., Kamarianakis, M., \u0026amp; Papagiannakis, G. (2024). VR Isle Academy: A VR Digital Twin Approach for Robotic Surgical Skill Development. \u003cem\u003earXiv preprint arXiv:2406.00002\u003c/em\u003e. \u003c/li\u003e\n\u003cli\u003eGałecki, A., Burzykowski, T., Gałecki, A., \u0026amp; Burzykowski, T. (2013). \u003cem\u003eLinear mixed-effects model\u003c/em\u003e. Springer. \u003c/li\u003e\n\u003cli\u003eGalna, B., Barry, G., Jackson, D., Mhiripiri, D., Olivier, P., \u0026amp; Rochester, L. (2014). Accuracy of the Microsoft Kinect sensor for measuring movement in people with Parkinson\u0026apos;s disease. \u003cem\u003eGait \u0026amp; Posture\u003c/em\u003e,\u003cem\u003e 39\u003c/em\u003e(4), 1062-1068. https://doi.org/10.1016/j.gaitpost.2014.01.008 \u003c/li\u003e\n\u003cli\u003eGiavarina, D. (2015). Understanding Bland Altman analysis. \u003cem\u003eBiochemia medica\u003c/em\u003e,\u003cem\u003e 25\u003c/em\u003e(2), 141-151. https://doi.org/10.11613/BM.2015.015 \u003c/li\u003e\n\u003cli\u003eGu, C., Lin, W., He, X., Zhang, L., \u0026amp; Zhang, M. (2023). IMU-based motion capture system for rehabilitation applications: A systematic review. \u003cem\u003eBiomimetic intelligence and robotics\u003c/em\u003e,\u003cem\u003e 3\u003c/em\u003e(2), 100097. https://doi.org/10.1016/j.birob.2023.100097 \u003c/li\u003e\n\u003cli\u003eGupta, A., \u0026amp; Semwal, V. B. (2022). Occluded Gait reconstruction in multi person Gait environment using different numerical methods. \u003cem\u003eMultimedia tools and applications\u003c/em\u003e,\u003cem\u003e 81\u003c/em\u003e(16), 23421-23448. https://doi.org/10.1007/s11042-022-12218-2 \u003c/li\u003e\n\u003cli\u003eHe, Y., Lee, J., Kim, J., Brodie, M. A., Mitri, G., van Schooten, K. S., Lovell, N. H., Lord, S. R., \u0026amp; Okubo, Y. (2025). Virtual Obstacle-Avoidance Training Using Daily-Life Obstacles with Physical Feedback in Older People: A Cross-Over Trial. \u003cem\u003eAvailable at SSRN 5104022\u003c/em\u003e. https://doi.org/10.2139/ssrn.5104022 \u003c/li\u003e\n\u003cli\u003eHof, A. L., Gazendam, M. G. J., \u0026amp; Sinke, W. E. (2005). The condition for dynamic stability. \u003cem\u003eJournal of Biomechanics\u003c/em\u003e,\u003cem\u003e 38\u003c/em\u003e(1), 1-8. https://doi.org/10.1016/j.jbiomech.2004.03.025 \u003c/li\u003e\n\u003cli\u003eHTC VIVE. (2025). \u003cem\u003ehttps://www.vive.com/au/accessory/vive-ultimate-tracker/\u003c/em\u003e. \u003c/li\u003e\n\u003cli\u003eKabsch, W. (1976). A solution for the best rotation to relate two sets of vectors. \u003cem\u003eActa Crystallographica Section A\u003c/em\u003e,\u003cem\u003e 32\u003c/em\u003e(5), 922-923. https://doi.org/10.1107/S0567739476001873 \u003c/li\u003e\n\u003cli\u003eKim, W., Huang, C., Yun, D., Saakes, D., Xiong, S., Murata, A., Goonetilleke, R. S., Xiong, S., Goossens, R. H. M., \u0026amp; Karwowski, W. (2020). Comparison of Joint Angle Measurements from Three Types of Motion Capture Systems for Ergonomic Postural Assessment. In (Vol. 1215, pp. 3-11). Springer International Publishing AG. https://doi.org/10.1007/978-3-030-51549-2_1 \u003c/li\u003e\n\u003cli\u003eKulozik, J. (2024). \u003cem\u003eUltimate tracker python \u003c/em\u003e\u003cem\u003ehttps://github.com/jkulozik\u003c/em\u003e. gitHub repository. Retrieved 30/Nov/2024 from \u003c/li\u003e\n\u003cli\u003eKulozik, J., \u0026amp; Jarrass\u0026eacute;, N. (2024). Evaluating the precision of the HTC VIVE Ultimate Tracker with robotic and human movements under varied environmental conditions. \u003cem\u003earXiv preprint arXiv:2409.01947\u003c/em\u003e. \u003c/li\u003e\n\u003cli\u003eLee, J., Phu, S., Lord, S. R., \u0026amp; Okubo, Y. (2024). Effects of immersive virtual reality training on balance, gait and mobility in older adults: A systematic review and meta-analysis. \u003cem\u003eGait \u0026amp; Posture\u003c/em\u003e,\u003cem\u003e 110\u003c/em\u003e, 129-137. https://doi.org/10.1016/j.gaitpost.2024.03.009 \u003c/li\u003e\n\u003cli\u003eLima, Y. L., Collings, T., Hall, M., Bourne, M. N., \u0026amp; Diamond, L. E. (2024). Validity and reliability of trunk and lower-limb kinematics during squatting, hopping, jumping and side-stepping using OpenCap markerless motion capture application. \u003cem\u003eJournal of sports sciences\u003c/em\u003e,\u003cem\u003e 42\u003c/em\u003e(19), 1847-1858. https://doi.org/10.1080/02640414.2024.2415233 \u003c/li\u003e\n\u003cli\u003eLin, L. I. (1989). A Concordance Correlation Coefficient to Evaluate Reproducibility. \u003cem\u003eBiometrics\u003c/em\u003e,\u003cem\u003e 45\u003c/em\u003e(1), 255-268. https://doi.org/10.2307/2532051 \u003c/li\u003e\n\u003cli\u003eMcBride, G. (2005). A proposal for strength-of-agreement criteria for Lin\u0026rsquo;s concordance correlation coefficient. \u003cem\u003eNIWA client report: HAM2005-062\u003c/em\u003e,\u003cem\u003e 45\u003c/em\u003e, 307-310. \u003c/li\u003e\n\u003cli\u003eMenolotto, M., Komaris, D.-S., Tedesco, S., O\u0026apos;Flynn, B., \u0026amp; Walsh, M. (2020). Motion Capture Technology in Industrial Applications: A Systematic Review. \u003cem\u003eSensors (Basel, Switzerland)\u003c/em\u003e,\u003cem\u003e 20\u003c/em\u003e(19), 5687. https://doi.org/10.3390/s20195687 \u003c/li\u003e\n\u003cli\u003eMerker, S., Pastel, S., B\u0026uuml;rger, D., Schwadtke, A., \u0026amp; Witte, K. (2023). Measurement Accuracy of the HTC VIVE Tracker 3.0 Compared to Vicon System for Generating Valid Positional Feedback in Virtual Reality. \u003cem\u003eSensors (Basel, Switzerland)\u003c/em\u003e,\u003cem\u003e 23\u003c/em\u003e(17), 7371. https://doi.org/10.3390/s23177371 \u003c/li\u003e\n\u003cli\u003eNijmeijer, E. M., Heuvelmans, P., Bolt, R., Gokeler, A., Otten, E., \u0026amp; Benjaminse, A. (2023). Concurrent validation of the Xsens IMU system of lower-body kinematics in jump-landing and change-of-direction tasks. \u003cem\u003eJournal of Biomechanics\u003c/em\u003e,\u003cem\u003e 154\u003c/em\u003e, 111637-111637. https://doi.org/10.1016/j.jbiomech.2023.111637 \u003c/li\u003e\n\u003cli\u003eOkubo, Y., He, Y., Brodie, M., Hicks, C., Schooten, K. v., Lovell, N. H., Anstey, K., J., Lord, S. R., \u0026amp; Kim, J. (2025). Virtual reality obstacle avoidance training can be enhanced by physical feedback via perturbations: a proof-of-concept study. \u003cem\u003eApplied Ergonomics\u003c/em\u003e. https://doi.org/10.1016/j.apergo.2024.104442 \u003c/li\u003e\n\u003cli\u003eOrdo\u0026ntilde;ez Nu\u0026ntilde;ez, T., Garcia Ramirez, A. R., \u0026amp; Becher\u0026aacute;n Mar\u0026oacute;n, L. (2024). Analysis of waist and wrist positioning wearable machine learning models to detect falls. \u003cem\u003eElectronics letters\u003c/em\u003e,\u003cem\u003e 60\u003c/em\u003e(2), n/a. https://doi.org/10.1049/ell2.13086 \u003c/li\u003e\n\u003cli\u003ePfister, A., West, A. M., Bronner, S., \u0026amp; Noah, J. A. (2014). Comparative abilities of Microsoft Kinect and Vicon 3D motion capture for gait analysis. \u003cem\u003eJournal of medical engineering \u0026amp; technology\u003c/em\u003e,\u003cem\u003e 38\u003c/em\u003e(5), 274-280. https://doi.org/10.3109/03091902.2014.909540 \u003c/li\u003e\n\u003cli\u003ePutranto, J. S., Heriyanto, J., Kenny, Achmad, S., \u0026amp; Kurniawan, A. (2023). Implementation of virtual reality technology for sports education and training: Systematic literature review. \u003cem\u003eProcedia computer science\u003c/em\u003e,\u003cem\u003e 216\u003c/em\u003e, 293-300. https://doi.org/10.1016/j.procs.2022.12.139 \u003c/li\u003e\n\u003cli\u003eRaffegeau, T. E., Brinkerhoff, S. A., Kellaher, G. K., Baudendistiel, S., Terza, M. J., Roper, J. A., \u0026amp; Hass, C. J. (2022). Changes to margins of stability from walking to obstacle crossing in older adults while walking fast and during a dual-task. \u003cem\u003eExperimental Gerontology\u003c/em\u003e,\u003cem\u003e 161\u003c/em\u003e, 111710-111710. https://doi.org/10.1016/j.exger.2022.111710 \u003c/li\u003e\n\u003cli\u003eRanjan, R., Ahmedt-Aristizabal, D., Armin, M. A., \u0026amp; Kim, J. (2025). Computer vision for clinical gait analysis: A gait abnormality video dataset. \u003cem\u003eIEEE Access\u003c/em\u003e,\u003cem\u003e 13\u003c/em\u003e, 45321-45339. https://doi.org/10.1109/ACCESS.2025.3545787 \u003c/li\u003e\n\u003cli\u003eRhiel, S., Kl\u0026auml;y, A., Keller, U., van Hedel, H. J. A., \u0026amp; Ammann-Reiffer, C. (2024). Comparing Walking-Related Everyday Life Tasks of Children with Gait Disorders in a Virtual Reality Setup With a Physical Setup: Cross-Sectional Noninferiority Study. \u003cem\u003eJMIR serious games\u003c/em\u003e,\u003cem\u003e 12\u003c/em\u003e, e49550-e49550. https://doi.org/10.2196/49550 \u003c/li\u003e\n\u003cli\u003eRybnik\u0026aacute;r, F., Kačerov\u0026aacute;, I., Hořej\u0026scaron;\u0026iacute;, P., \u0026amp; \u0026Scaron;imon, M. (2023). Ergonomics Evaluation Using Motion Capture Technology\u0026mdash;Literature Review. \u003cem\u003eApplied sciences\u003c/em\u003e,\u003cem\u003e 13\u003c/em\u003e(1), 162. https://doi.org/10.3390/app13010162 \u003c/li\u003e\n\u003cli\u003eSheng, X., Mao, S., Yan, Y., \u0026amp; Yang, X. (2024). Review on SLAM algorithms for Augmented Reality. \u003cem\u003eDisplays\u003c/em\u003e,\u003cem\u003e 84\u003c/em\u003e, 102806. https://doi.org/10.1016/j.displa.2024.102806 \u003c/li\u003e\n\u003cli\u003eSitole, S. P., LaPre, A. K., \u0026amp; Sup, F. C. (2020). Application and Evaluation of Lighthouse Technology for Precision Motion Capture. \u003cem\u003eIEEE sensors journal\u003c/em\u003e,\u003cem\u003e 20\u003c/em\u003e(15), 8576-8585. https://doi.org/10.1109/JSEN.2020.2983933 \u003c/li\u003e\n\u003cli\u003eSlater, M., \u0026amp; Sanchez-Vives, M. V. (2016). Enhancing our lives with immersive virtual reality. \u003cem\u003eFrontiers in robotics and AI\u003c/em\u003e,\u003cem\u003e 3\u003c/em\u003e, 74. \u003c/li\u003e\n\u003cli\u003eSousa, C. V., Lee, K., Alon, D., Sternad, D., \u0026amp; Lu, A. S. (2023). A Systematic Review and Meta-analysis of the Effect of Active Video Games on Postural Balance. \u003cem\u003eArch Phys Med Rehabil\u003c/em\u003e,\u003cem\u003e 104\u003c/em\u003e(4), 631-644. https://doi.org/10.1016/j.apmr.2023.01.002 \u003c/li\u003e\n\u003cli\u003eSpitzley, K. A., \u0026amp; Karduna, A. R. (2019). Feasibility of using a fully immersive virtual reality system for kinematic data collection. \u003cem\u003eJournal of Biomechanics\u003c/em\u003e,\u003cem\u003e 87\u003c/em\u003e, 172-176. https://doi.org/10.1016/j.jbiomech.2019.02.015 \u003c/li\u003e\n\u003cli\u003eSubramanian, R., \u0026amp; Sarkar, S. (2019). Evaluation of Algorithms for Orientation Invariant Inertial Gait Matching. \u003cem\u003eIEEE transactions on information forensics and security\u003c/em\u003e,\u003cem\u003e 14\u003c/em\u003e(2), 304-318. https://doi.org/10.1109/TIFS.2018.2850032 \u003c/li\u003e\n\u003cli\u003eSvetek, A., Morgan, K., Burland, J., \u0026amp; Glaviano, N. R. (2025). Validation of OpenCap on lower extremity kinematics during functional tasks. \u003cem\u003eJournal of Biomechanics\u003c/em\u003e,\u003cem\u003e 183\u003c/em\u003e, 112602. https://doi.org/10.1016/j.jbiomech.2025.112602 \u003c/li\u003e\n\u003cli\u003eTopley, M., \u0026amp; Richards, J. G. (2020). A comparison of currently available optoelectronic motion capture systems. \u003cem\u003eJournal of Biomechanics\u003c/em\u003e,\u003cem\u003e 106\u003c/em\u003e, 109820-109820. https://doi.org/10.1016/j.jbiomech.2020.109820 \u003c/li\u003e\n\u003cli\u003eVox, J. P., Weber, A., Wolf, K. I., Izdebski, K., Sch\u0026uuml;ler, T., K\u0026ouml;nig, P., Wallhoff, F., \u0026amp; Friemert, D. (2021). An Evaluation of Motion Trackers with Virtual Reality Sensor Technology in Comparison to a Marker-Based Motion Capture System Based on Joint Angles for Ergonomic Risk Assessment. \u003cem\u003eSensors (Basel, Switzerland)\u003c/em\u003e,\u003cem\u003e 21\u003c/em\u003e(9), 3145. https://doi.org/10.3390/s21093145 \u003c/li\u003e\n\u003cli\u003eWeber, M., Hartl, R., Z\u0026auml;h, M. F., \u0026amp; Lee, J. (2023). Dynamic Pose Tracking Accuracy Improvement via Fusing HTC Vive Trackers and Inertia Measurement Units. \u003cem\u003eInternational journal of precision engineering and manufacturing\u003c/em\u003e,\u003cem\u003e 24\u003c/em\u003e(9), 1661-1674. https://doi.org/10.1007/s12541-023-00891-8 \u003c/li\u003e\n\u003cli\u003eZhou, C., Qian, Y., \u0026amp; Kaner, J. (2024). A study on smart home use intention of elderly consumers based on technology acceptance models. \u003cem\u003ePLoS One\u003c/em\u003e,\u003cem\u003e 19\u003c/em\u003e(3), e0300574-e0300574. https://doi.org/10.1371/journal.pone.0300574 \u003c/li\u003e\n\u003cli\u003eZhou, H., \u0026amp; Hu, H. (2008). Human motion tracking for rehabilitation\u0026mdash;A survey. \u003cem\u003eBiomedical signal processing and control\u003c/em\u003e,\u003cem\u003e 3\u003c/em\u003e(1), 1-18. https://doi.org/10.1016/j.bspc.2007.09.001 \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"motion capture, VIVE Ultimate Tracker, Vicon, virtual reality","lastPublishedDoi":"10.21203/rs.3.rs-6989733/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6989733/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study investigated the performance of the HTC VIVE Ultimate Tracker against the Vicon motion capture system during human walking. Ten healthy participants (aged 24\u0026ndash;44 years) walked on a treadmill at four speeds (0.5, 1.0, 1.5, and 2.0 m/s) while tracking both feet and the pelvis. Agreement between the two systems was evaluated using a linear mixed model, Bland-Altman plots, and concordance correlation coefficients (CCC). Absolute errors presented the accuracy of Ultimate Trackers ranging from millimetre- to centimetre-level. Linear mixed model [speed (0.5-2.0 m/s) \u0026times; tracker location (sacrum, left foot, and right foot) \u0026times; movement direction (vertical, medio-lateral, and anterior-posterior)] indicated the absolute errors increased with higher gait speed. Foot trackers exhibited larger errors than the sacrum location, with greater errors in the medio-lateral and anterior-posterior directions compared to the vertical direction (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Bland-Altman analyses revealed widening limits of agreement at different speeds (e.g., left foot, AP direction: -20.29 to 20.20 at 0.5 m/s, -29.97 to 27.80 mm at 2.0 m/s). Ultimate Trackers demonstrated almost perfect agreement (CCC: \u0026gt;0.99) for sacrum tracking across all speeds and directions, and excellent agreement for foot trackers (CCC: \u0026gt;0.98). These findings highlight the Ultimate Trackers\u0026rsquo; potential as cost-effective alternatives for the analysis of human movement, demonstrating research- and clinical-grade performance for sacrum and foot tracking during normal gait speeds (\u0026le;\u0026thinsp;1.5 m/s). However, the finding that accuracy declined at the 2.0 m/s speed, particularly for foot trajectories and in the anterior-posterior direction, indicates the need for further technical refinements for higher speed movements.\u003c/p\u003e","manuscriptTitle":"Validation of the HTC VIVE Ultimate Trackers Compared with the Vicon Motion Capture System at Slow, Moderate and Fast Gait Speeds","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-01 17:28:36","doi":"10.21203/rs.3.rs-6989733/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"670e9f33-54bc-4aff-8472-987c8bd0bfb4","owner":[],"postedDate":"October 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-17T11:53:47+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-01 17:28:36","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6989733","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6989733","identity":"rs-6989733","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}