Design and optimization of energy-efficient wireless sensor networks for industrial automation

Design and optimization of energy-efficient wireless sensor networks for industrial automation | 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 Design and optimization of energy-efficient wireless sensor networks for industrial automation Maha Abbas Hutaihit, Samir I. Badrawi, Haider Makki Hameed, Riyadh Khlf Ahmed, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5731209/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Jul, 2025 Read the published version in Journal of Intelligent Systems and Internet of Things → Version 1 posted You are reading this latest preprint version Abstract In order to improve the overall performance of edge-integrated edge IoT networks, this research introduces a combined technique based on profound learning for booking assets. If an IoT network wants to finish a task quickly and effectively, it has to get the greatest resources from the edge layer. Thorough asset booking is crucial to the identification and transfer of optimal assets. The integration of edge networks with IoT applications and the reduction of data transmission latency were previously addressed using profound learning algorithms. If we want to make an Internet of Things application more feasible and provide better service overall, we should think about other metrics like reaction time, waiting time, and bandwidth needs. Combining a convolutional neural network with a gated repeating unit in a certain manner achieves this enhanced performance. The suggested asset booking model considers the features and requirements of the assets in order to select the most suitable ones from the pool and allocate them to the IoT networks. Here, we give a comprehensive analysis of the method-data combination. Design and Optimization For Industrial Automation cognitive industrial internet of things (CIIOT) electric-field measurement system (EFMS) Radio-Access Network-As-A-Service (RANAAS) Multi-InputMulti-Output (MIMO) Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1. Introduction As of late, there has been a ton of interest in wireless sensor networks (WSNs) from both industry and academia. Industrial field control, smart homes, smart factories, environmental observing, and industrial field checking are some of the many communicated observation and control areas that make extensive use of WSNs[1].WSNs are often made up of several sensor hubs that perform a variety of tasks, including data handling, transmission, and gathering. A WSN clearly offers advantages over conventional wired systems with regards to cost, adaptability, and ease[2]. WSNs are often fueled by batteries in industrial settings, and these batteries' constrained capacity and dreary replacement prerequisites have turned into the primary barriers to WSN adoption [3]. In WSN applications, lifetime power utilization management is crucial [4]. The goal of creating energy-assortment advances is to increase the valuable lifetime of WSNs by using various procedures for energy harvesting [5]. Showcase a vibration energy-controlled WSN demonstration testbed. Vibration energy is planned to be harvested by an electromagnetic harvester [6]. Berkeley researchers offer a design for a WSN that runs ceaselessly on renewable energy from the climate [7].Two-stage storage systems made contained a rechargeable solar lithium battery and a supercapacitor are utilized in this method [8]. A gadget for harvesting human energy is built and examined, using a power hardware module to harness the energy created by a walking human [9–12]. Look into a physically autonomous sensor that uses mechanical nano energy derived from human beings to take advantage of movement location and physiological signal monitoring [13]. A wireless energy harvesting cognitive industrial internet of things (CIIOT) is proposed, which simultaneously performs range detecting and transmissions and harvests wireless energy from a primary client [14–18]. In any case, these research' energy-harvesting gadgets produce less energy and need movement, such mechanical vibrations in the system [19–22]. They cannot be broadly applied to industrial settings and have stayed in the laboratory design and verification stages. Data gathering, transmission, and handling are a WSN's primary energy-consuming operations, and these activities are also the vital targets of energy usage improvement [23]. Artificial neural networks are utilized to manage data sampling and preserve hub energy [24]. To increase the WSN's useful life, an adaptive sampling strategy that considers the temporal and spatial correlation of the sensor data is proposed. created a ZigBee and Geiger Muller tube system that reasonably estimates radiation and temperature monitoring [25]. improved uptime during power outages by settling on the ideal portable hub antenna configuration and the most efficient asset management strategy for relay hub selection. suggested an approach to adaptive data collection that would reduce energy consumption and advance data transmission [26–30]. These methods, which are more commonly used in labs, can precisely control energy, reduce energy consumption, and lengthen the lifespan of systems. However, these methods have on occasion been employed in contexts including difficult outside conditions, such as present HVDC transmission line electromagnetic measurement systems. China currently has a large number of HVDC transmission lines operational. The electromagnetic environment must be considered throughout the whole transmission line lifecycle, from design to installation and operation [31]. When studying the electromagnetic field, it is crucial to take into account the electric field beneath the(HVDC) transmission lines. With the help of a wireless sensor network (WSN) electric-field measurement system (EFMS), the electric field under the HVDC transmission cables can be observed. One important part of the EFMS and a popular area of study is the electric-field sensor. Optical, field factory, and MEMS sensors are just a few of the several possible techniques [32]. The charge enrollment guideline states that MEMS sensors convert the DC electric field into measurable electrical values using micromechanical resonators. One-way optical sensors detect electric fields is by using the electro-optic impact concept, which involves watching how a field changes the refractive index of a crystal [33]. The widely recognized charging standard is employed by the field factory sensors. At regular intervals, the engine opens and protects the electric field, keeping the rotor spinning at a constant speed. By cycling between charging and discharging charges, the enrollment cathode can provide an AC signal that is proportionate to the applied DC electric field. Power consumption of the EFMS is significantly increased due to the constant operation of the electric-field factory sensor's motor [34]. Since electric field sensors do not require an engine, they are able to utilize power-efficient MEMS and optical sensors. Unfortunately, most current electric-field measurement systems rely on electric-field factories as their electric-field sensors, mostly because of affordability and stability concerns [35]. In order to study the electric-field circulation under the transmission lines, a range of sensors can continuously measure the electric-field. It is common practice to run EFMSs on batteries due to the lack of a reliable power source at many testing sites [36]. Because of the system's usual data advancement requirements and the battery capacity, extending the usable life of a battery-powered EFMS is necessary yet problematic. New features, like data transfer, are incorporated into the EFMS as a consequence of real-time checking and distributed sensor networks; nevertheless, this considerably raises the system's energy consumption. The major objective of this work is to provide an energy-efficient booking technique that may be used to extend the lifespan of an EFMS. 1.1 Industrial Automation and the Need for WSNs WSNs have played an important role in the development of industrial automation, especially within the context of Industry 4.0, which emphasizes the integration of cyber-physical systems, Internet of Things (IoT), and data-driven decision-making, so that in WSNs, through real-time monitoring and control of industrial processes, improvements in efficiency and productivity and safety can be facilitated [37]. One of the important applications includes environmental monitoring, condition monitoring of equipment, predictive maintenance, and process optimization where sensor nodes collect and transmit critical data to enhance operational intelligence [38]. Nevertheless, although WSNs can be deployed in industrial automation, challenges faced include energy consumption, network scalability, and data reliability in harsh environments. These challenges bring opportunities to innovation toward energy-efficient protocols, advanced sensor technologies, and robust communication strategies; they are crucial for optimizing the performance and sustainability of industrial systems in the industry 4.0 era. 1.2 Energy Efficiency in Wireless Sensor Networks Energy efficiency is very sensitive to WSNs, particularly for industrial applications in large-scale remote environments or in harsh environments where the nodes are set up [39]. The energy usage of WSNs directly influences the life of the battery-powered nodes, the reliability of networks, and the operational costs in general. In industrial environments, where the requirement is for high continuity monitoring of data, the challenge is how to balance these data transmission frequencies with the limited energy sources available on the sensor nodes [40]. This can be solved by using several strategies-including energy-efficient routing protocols, adaptive techniques of data transmission, and duty cycling. Techniques that involve data aggregation, energy harvesting, and the use of low-power communication protocols are also underway in reducing energy consumption but without sacrificing performance and reliability in industrial automation systems for WSNs [41]. 2. Research Methodology In any situation where the Internet of Things is put to substantial use, no matter how large or little. Every single one of the domains makes use of the IoT's feature advantages, which greatly boost the performance of real-time applications. There is a pressing need for efficient data management due to the explosion of Internet of Things (IoT) devices, which has enabled smart cities and smart agriculture. It is possible that distributed computing can efficiently manage the IoT network in terms of storage and figure limitations. Internet of Things devices with low resources gather data and upload it to the cloud for processing. Due to the heterogeneity of the network, data transmissions from IoT devices to the cloud, whether incoming or departing, experience high latency and bandwidth needs. Edge computing is suggested as a means to lessen latency in IoT networks; it involves moving processing power from the cloud to the client end. As an add-on, edge registering is useful for cloud and Internet of Things networks. Distributed computing's figure load is drastically reduced in real-time applications by using edge registering. With their three-layer heterogeneous design, cloud-edge cloud IoT networks require proper asset planning to further increase productivity and service quality. The diverse types of data collected by the Internet of Things network necessitate different approaches to data processing. All data should be handled on the edge or in the cloud when asset demands are booked. The asset requirements can be represented by tasks, and the edge network can obtain these demands in turn. After that, it selects the best cloud assets and arranges for IoT networks to use them for additional processing. To avoid SLA violations, cloud providers should think about things like load balancing, energy consumption, and bandwidth congestion, regardless of how many resources they offer. In order to reduce the real-time real data handling latency in the dispersed computing environment, IoT networks use edge registration. The edge enhances the performance of the network and reduces compute, blockage, and data transmission delay by supplying the appropriate cloud assets to IoT networks. Distributed computing assets should be planned for edge figuring into Internet of Things networks as part of an effective asset planning strategy. Planning must take asset elasticity and scalability into account. When assets in the cloud are shared via edge networks, their scalability and flexibility change since most of these assets are virtual or physical. Because not all applications have the same asset demands, the edge should be aware of these needs when booking assets for IoT networks. This is because different applications require different registering assets. The resource scheduling algorithms that have emerged recently are either machine learning- or statistically-based scheduling procedures. Based on the resource requirements, the best resources are chosen from the resource pool. Deep learning techniques have recently supplanted statistical and machine learning-based scheduling models in order to improve scheduling performance. When it comes to scheduling resources at the edge of a network, deep learning techniques like RL, Q-learning, and deep neural networks are heavily utilized. Cutting down on wait times and scheduling delays is critical for optimizing efficiency, even when deep learning methods work as expected. Asset planning in edge-integrated IoT networks is facilitated by this combined deep learning method. Integrating IoT Services in the Cloud Many applications on the Internet of Things have embraced distributed computing because of its capacity to process, store, and analyze massive amounts of data. The distributed computing environment is another option for the Internet of Things (IoT) that relies on cloud computing to enable device connection. According to their requirements, customers can access cloud services whenever and wherever they like. This internet of things (IoT) approach that integrated the cloud was utilized by numerous smart city, transportation, agricultural, and healthcare applications. However, the data processing experiences latency due to the long-distance data transit between IoT devices to the cloud. Various applications of IoT are applied for actual real-time time operations which should deliver rapid replies by evaluating the data. Delays in answers will harm the quality of services in IoT applications. Meanwhile, it is vital to maintain a steady connection between devices and the cloud49 which is also a challenge while integrating the cloud with IoT. Thus, it is vital to consider few insights as provided below when designing cloud- based integrated IoT apps. An exceptionally fast reaction time and as little delay as possible from beginning to end are essential for improving service quality. The cloud and IoT application hubs are dependent on a reliable and steady network. Significant processing complexity will result from adding more networking conventions. For this reason, limiting computational difficulties requires careful consideration while selecting conventions. 3.1. Edge IoT Integration An edge-integrated m module can help with the limitations of a cloud-integrated IoT ecosystem. Not only does edge registration reduce handling complexity and latency, but it also brings cloud resources directly to the user's device. Some examples of applications of edge figuring include cloudlets, flexible edge processing, and mist registering. Each of these methods reduces the amount of time data must be processed before IoT applications get their responses. Reduced latency and data transfer times are the results of edge processing, which places assets in close proximity to the Internet of Things. The IoT app can take advantage of edge processing to improve both its local and transmitted data handling capabilities. Edge enhances system resilience and fault tolerance by reducing bandwidth requirements and providing Internet of Things clients with great adaptability. The interplay between edge clouds and Internet of Things devices is illustrated in Fig. 3.1. 3.2. Proposed Concatenated Deep Learning Algorithm An in-depth mathematical model of the concatenated deep learning technique that has been suggested is presented in this section. A convolutional neural network and a gated recurrent unit are utilized for initial feature extraction in the most basic form of the proposed model, as illustrated in Fig. 5.1. The asset's category, sub-class, class, duration, and other relevant criteria are taken into account when determining the demand for the asset. Asset needs can be parsed into local and regional details by use of a one-dimensional convolutional neural network. Similar to how features are recovered in the latter stage of the process using a gated recurrent unit, the best asset for planning in edge processing is chosen in the first stage. For this project, we've settled on a gated recurrent unit (GRU) because of its ease of use and high performance. When compared to conventional long short-term memory (LSTM), GRU performs better due to its faster input processing and less parameter usage. By fixing the vanishing gradient issue in RNN, GRU could make current booking algorithms better. Finally, after the combined traits are categorized, the best resources for the jobs are scheduled. Gated Recurrent Unit As a GRU model, a gated recurrent neural network is recommended. With only two gates, GRU stands in stark contrast to LSTM's trio. The GRU’s update and reset gates not only improve union rates but also reduce the number of parameters needed compared to an LSTM. Using its memory cell, the GRU model may retrieve crucial data and identify situations in the input asset requirements. The GRU's reset gate forgets or erases the redundant data. The GRU model often takes period series data as input, even if the asset demand input is typically period grouping data with a single time step. The activation is successful, and the GRU model's outputs are obtained. By feeding the principal layer's output into the next layer and repeating the process, we may extract the important features from the input to the resultant layer. Mathematical descriptions of the GRU model are as: $$\:{\mathcal{G}}_{\varvec{u}}=\varvec{\sigma\:}\left({\mathcal{w}}_{\mathcal{u}}\left({\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}-1\right)},{\varvec{x}}^{\left(\varvec{t}\right)}\right)+{\mathcal{b}}_{\mathcal{u}}\right)$$ 1 $$\:{\mathcal{G}}_{\mathcal{r}}=\varvec{\sigma\:}\left({\mathcal{w}}_{\mathcal{r}}\left({\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}-1\right)},{\varvec{x}}^{\left(\varvec{t}\right)}\right)+{\mathcal{b}}_{\mathcal{r}}\right)$$ 2 Here, Gu handles the update gate while Gr handles the reset gate. Update gates differ from reset gates in that their range is [0,1] rather than [-1,1]. Wr denotes the capability to reset the gate weight and Wu stands for the capability to update the gate weight. Just as how Br takes care of the update gate's bias vector, as does as for the reset gate. The candidate activation capability for the recurrent unit is created using the following formula, which is based on the gate works. $$\:{\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}\right)}=\varvec{t}\varvec{a}\varvec{n}\varvec{h}\left[{\mathcal{w}}_{\mathcal{u}}\left({{\mathcal{G}}_{\varvec{r}}\times\:\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}-1\right)},{\varvec{x}}^{\left(\varvec{t}\right)}\right)+{\mathcal{b}}_{\mathcal{u}}\right]$$ 3 To handle the bias vector, the activation capacity weight factors—Wu for the update gate—are applied when the input training data is labeled as X(t). The final step is to transmit the GRU model's output as: $$\:{\mathcal{v}}^{\left(\varvec{t}\right)}=\left(\left(1-{\mathcal{G}}_{\mathcal{u}}\right)\times\:{\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}-1\right)}\right)+\left({{\mathcal{G}}_{\mathcal{u}}\times\:\stackrel{\sim}{\mathcal{v}}}^{\left(\varvec{t}\right)}\right)$$ 4 where d is a function of the output of the prior unit and v(t-1) is the input of the present unit. Merging the output characteristics of the CNN and GRU models results in extra processing steps that determine which assets are most suitable for booking. Convolutional Neural Network The input is sorted into subclasses according to the specifications by the Convolutional Neural Network Model employed in the proposed work before the Convolution layer takes over. Two max-pooling layers and two convolution layers are employed by the suggested design to glean useful details from the asset needs. In terms of data handling and maintaining local interactions, CNN outperforms traditional neural network models. Compared to earlier models, this neural network method conveys attributes more clearly while jellying the input data's spatial localization. Through autonomous training, the network is able to absorb various data qualities. The CNN module is based on the correlated cycle principle of the convolution interaction. To construct the convolution interaction correctly, one needs to think about the loads of the one-dimensional aspects component, which are represented by the words {W1,W2,...,Wn}, where n is the kernel length. $$\:{\mathcal{y}}_{\varvec{t}}=\varvec{f}\left({\sum\:}_{\varvec{i}=1}^{\varvec{n}}{\varvec{w}}_{\varvec{i}}\varvec{*}{\varvec{x}}_{\varvec{t}-\varvec{i}+1}\right)$$ 5 We say that the data created at time t is Yt and that the input sample is Xt. The proposed model makes use of the Redressed Linear Unit (RELU) as its activation capability. The activation capability can be mathematically expressed as $$\:\varvec{R}\varvec{e}\varvec{L}\varvec{U}\left(\varvec{x}\right)=\left\{\begin{array}{c}x,\\\:0,\end{array}\right.\genfrac{}{}{0pt}{}{\varvec{x}}{\varvec{x}}\genfrac{}{}{0pt}{}{>}{\le\:}\genfrac{}{}{0pt}{}{0}{0}$$ 6 The suggested architecture employs max pooling to cap the feature size subsequent to the convolution layer. The outputs of the convolution layer are down-sampled to reduce unpredictability. Maximum pooling operator mathematically expresses the forwarding of the maximum value as $$\:{\mathcal{P}}_{\mathcal{j}.\mathcal{m}}=\mathbf{max}{{\mathcal{h}}_{\mathcal{j}},}_{\left(\mathcal{m}-1\right)}\mathcal{n}+\mathcal{r}$$ 7 where n is the permitted area-to-area pooling shift, m is the maximum pooled band, and j are the channels. With most convolution bands, the pooling layer reduces their dimensionality. Batch normalization follows pooling capabilities and improves training results by standardizing the features. The mathematical description of batch normalization characteristics is up next. $$\:\mathcal{u}=\frac{1}{{\mathcal{n}}_{\varvec{b}\varvec{a}\varvec{t}}}{\sum\:}_{\mathcal{n}=1}^{{\mathcal{n}}_{\varvec{b}\varvec{a}\varvec{t}}}{\varvec{x}}_{\mathcal{n}}$$ $$\:{\varvec{\sigma\:}}^{2}=\frac{1}{{\mathcal{n}}_{\varvec{b}\varvec{a}\varvec{t}}}{\sum\:}_{\mathcal{n}=1}^{{\mathcal{n}}_{\varvec{b}\varvec{a}\varvec{t}}}{\left({\mathcal{x}}_{\mathcal{n}}-\varvec{\mu\:}\right)}^{2}$$ $$\:\widehat{\mathcal{x}}=\frac{{\mathcal{x}}_{\mathcal{n}}-\varvec{\mu\:}}{\sqrt{{\varvec{\sigma\:}}^{2}+\varvec{\epsilon\:}}}$$ $$\:{\mathcal{y}}_{\mathcal{n}}=\varvec{\gamma\:}{\widehat{\mathcal{x}}}_{\mathcal{n}}+\varvec{\beta\:}$$ 8 where X n is the input data and N bat is the batch size. While σ2 addresses the batch variance, µ indicates the mean. In order to avoid zero gradients, the normalized data is associated with a constant ε, denoted as X̂. Dand K is the graphic depiction of the learning vector parameters. The features that are represented by the output are ʆ and β. Yn stands for the feature that is produced. Afterwards, the CNN and GRU models' properties are integrated. To avoid overfitting the data, apply a dropout layer following concatenation. Finally, the collected features are classified using the fully linked network layer and SoftMax algorithms in order to assign the correct resources to a job. The SoftMax capability can be stated numerically as $$\:\widehat{\mathcal{y}}=\varvec{s}\varvec{o}\varvec{f}\varvec{t}\varvec{m}\varvec{a}\varvec{x}\left(\mathcal{Q}\right)$$ 9 the result of the dropout layer is denoted by Q. Last but not least, the mistake capacity of the suggested model is checked using a cross-entropy capability. Mathematically, it's expressed as $$\:\mathcal{l}=-\frac{1}{\mathcal{b}}{\sum\:}_{\varvec{i}=1}^{\varvec{n}}{\mathcal{y}}_{\varvec{i}}\mathbf{log}{\mathcal{y}}_{\varvec{i}}^{\varvec{{\prime\:}}}$$ 10 As an example, yi' represents the expected component and yi addresses the actual component, whereas b, n, and yi are the sizes of the batch and training samples, respectively. 4. Performance Evaluation We empirically validate the performance of the recommended deep learning model by incorporating the package and works as capabilities in a Python simulation study. Using these tools, hyperparameters are automatically generated and adjusted to enhance performance even more. The benchmark data used in the investigation comes from Intel's Berkeley research laboratories and contains 96. For the purpose of this simulation investigation, the hyperparameters used are detailed in Table 1 . Validation is conducted by examining and comparing current strategies, including the hereditary algorithm, the Improved Particle Swarm Optimization (IPSO) algorithm, Long Short-Term Memory (LSTM), and the Bidirectional Recurrent Neural Network (BRNN). Table 1 Hyperparameters of the suggested DL technique S. No Parameters Value 1 Conv filters 1 32 2 Conv filters 2 128 3 GRU Units 64 4 Dropout 0.0 5 Epochs 25 6 Batch Size 64 The proposed model's accuracy and misfortune curves are shown in Fig. 4 . Measuring performance is based on the standard method of testing and training. Each dataset is partitioned for testing, validation, and training at 70:20:10. The results have not altered after more than 25 generations of measurement. According to the results, the suggested model is the most accurate, and it has been verified. The suggested model's asset use is compared to the ongoing models in Fig. 5 . The results show that the suggested model is able to make the most efficient use of assets thanks to the optimal choice of assets. Positions are allocated optimal assets, which expedites data processing and frees up these assets for other uses. The overall asset consumption of the suggested approach is thus larger than that of current strategies. The suggested model and the existing BRNN models perform similarly, however other models display large discrepancies in the values of asset utilization. The response times of the ongoing asset planning methods and the suggested concatenated deep learning strategy are contrasted in Fig. 6 . Reaction time is the amount of time it takes for the planning algorithm to assess and plan for asset demands. The average time is calculated for each strategy that requests a certain asset from edge figuring. The suggested paradigm for asset demands demonstrates a minimal reaction time of 1.25 seconds. However, when other methods are employed, the average rises. By completing asset demands in 1.66s, 1.98s, and 2.20s, respectively, the LSTM model, the BRNN model, GA, and IPSO all outperform the suggested model in terms of reaction time. The suggested model's total execution seasons are compared to those of the state-of-the-art models in Fig. 7 . Execution time includes the time required to evaluate asset demand, choose the best asset from the pool, and schedule that asset. According to the results, the suggested model has the fastest execution time when compared to alternative strategies for planning. Compared to the BRNN model, the LSTM-based booking, the GA, and the IPSO model, the recommended model's execution season of 10.25s is 5s quicker, 8s faster, 11s faster, and 16s faster, respectively. We prioritize any time savings that may be achieved in the asset booking procedure for edge registration. The suggested model's performance is validated by considering the average delay given by both the present and new models. This is done through research into different asset demands. The results, shown in Fig. 8 , show that compared to the state-of-the-art methods, the suggested model minimizes delays more effectively on average. By a margin of five seconds, the IPSO model outperforms the suggested model in terms of latency. When compared to GA-based booking, the suggested process is 4.3 seconds slower. Models trained with LSTM and BRNN outperform those using GA and IPSO by a small margin. However, this model is far from perfect. The LSTM-based booking model is 2 seconds off and the BRNN model is 1.5 seconds off; the recommended model has a minimum delay of 1.15 seconds. Figure 9 showcases an effective comparison of booking algorithms. By comparing the proposed and current models' execution times, reaction times, and delay factors, we may get a sense of how efficient they are overall. The proposed model outperforms the competition across the board, improving the efficiency of both the IoT networks and the edge processing platform. Compared to the current booking methods, the suggested model has a substantially higher maximum efficacy of 99.48%. Table 2 Analysing Performance in Comparison Methods Resource Utilization Response Time Execution Time Average Delay Efficiency IPSO 95.8217% 2.2123s 26.0035s 5.4136s 94.4532% GA 95.7245% 1.9742s 20.6379s 4.2948s 96.1037% LSTM 98.5871% 1.6734s 17.4659s 3.1368s 97.9082% BRNN 99.0213% 1.5529s 14.8164s 2.8547s 98.8045% Proposed 99.5234% 1.2531s 10.2537s 1.1539s 99.4862% Table 2 provides an overview of the overall performance metrics used to compare the proposed model to the current models. The results demonstrate that the proposed approach achieves a higher level of asset utilization and productivity when compared to other existing booking systems. The proposed model also has the quickest execution and reaction times, making it ideal for real-time applications that need to record data produced or received efficiently through asset allocation. 5. Conclusion Here, we provide a mixed-methods deep learning approach to asset planning in IoT networks with embedded edges. In order to prioritize characteristics based on asset demands during the asset planning phase, the suggested work utilizes a gated recurrent unit and a one-dimensional convolutional neural network. The optimal planning assets are identified through the application of deep learning models that quickly assess and combine time-series requests, followed by classification. With the use of simulation analysis, we can see how the suggested model stacks up against other methods in terms of effectiveness, average delay, reaction time, execution time, and asset utilization, as well as against Genetic Algorithm (GA), LSTM, BRNN, and Genetic Algorithm (GA). The overall efficiency of the cycle is supported by the suggested model's best asset utilization, base execution, and reaction times when compared to current strategies. Declarations Author Contribution Authors share their knowledge and efforts to do this work Acknowledgements The research work presented in this paper, Design and Optimization of Energy-Efficient Wireless Sensor Networks for Industrial Automation , has been made possible through the continuous support and guidance of my mentor, (________) Her extensive knowledge in the field of research and her vast experience have been a constant source of inspiration, enabling the development of this paper in the form of a comprehensive research article. I extend my heartfelt gratitude to (_______). for her invaluable contributions and unwavering encouragement throughout this project. References A. Chehri, R. Saadane, N. Hakem, and H. Chaibi, "Enhancing energy efficiency of wireless sensor network for mining industry applications," Procedia Computer Science , vol. 176, pp. 261-270, 2020. Ahmed S, Chowdhury MZ, and Jang YM 2020 Energy-efficient UAV-to-user scheduling to maximize throughput in wireless networks. IEEE Access 8: 21215-21225. Alamu O, Gbenga-Ilori A, Adelabu M, Imoize A, and Ladipo O 2020 Energy efficiency techniques in ultra-dense wireless heterogeneous networks: An overview and outlook. Eng. Sci. Technol. Int. J. 23(6): 1308-1326. Alghamdi TA 2020 A hybrid algorithm for energy-efficient cluster head selection in wireless sensor networks. IEEE Access 8: 189211-189222. Arora VK, Sharma V, and Sachdeva M 2016 A survey on LEACH and other routing protocols in wireless sensor networks. Optik 127(16): 6590–6600. Bagwari, A., Logeshwaran, J., Usha, K., Kannadasan, R., Alsharif, M. H., Uthansakul, P., &Uthansakul, M. (2023). An Enhanced Energy Optimization Model for Industrial Wireless Sensor Networks Using Machine Learning. IEEE Access . Bhola J, Soni S, and Cheema GK 2020 Genetic algorithm based optimized LEACH protocol for energy efficient wireless sensor networks. J. Ambient Intell. Humaniz. Comput. 11: 1281-1288. C. S. Abella et al., "Autonomous energy-efficient wireless sensor network platform for home/office automation," IEEE Sensors Journal , vol. 19, no. 9, pp. 3501-3512, 2019. Chowdhury SM, and Hossain A 2020 Different energy saving schemes in wireless sensor networks: A survey. Wirel. Pers. Commun. 114(3): 2043-2062. D. D. Rao, A. Jain, S. Sharma, S. V. Pandit, and R. Pandey, "Effectual energy optimization stratagems for wireless sensor network collections through fuzzy-based inadequate clustering," SN Computer Science , vol. 5, no. 8, pp. 1-10, 2024. D. Lin, Q. Wang, W. Min, J. Xu, and Z. Zhang, "A survey on energy-efficient strategies in static wireless sensor networks," ACM Transactions on Sensor Networks (TOSN) , vol. 17, no. 1, pp. 1-48, 2020. Dario S, Luca P, Giuseppe F, and Pietro M 2014 Radio-Access Network-As-A-Service (RANAAS): Architecture for energy efficiency in 5G networks. IEEE Commun. Mag. 52(2): 114-122. F. H. El-Fouly and R. A. Ramadan, "Real-time energy-efficient reliable traffic aware routing for industrial wireless sensor networks," IEEE Access , vol. 8, pp. 58130-58145, 2020. H. Kathiriya, A. Pandya, V. Dubay, and A. Bavarva, "State of art: energy efficient protocols for self-powered wireless sensor network in IIoT to support industry 4.0," in Proc. 2020 8th Int. Conf. on Reliability, Infocom Technologies and Optimization (Trends and Future Directions) (ICRITO) , 2020, pp. 1311-1314, IEEE. J. Blanco, A. García, and J. D. L. Morenas, "Design and implementation of a wireless sensor and actuator network to support the intelligent control of efficient energy usage," Sensors , vol. 18, no. 6, p. 1892, 2018. J. Li, J. Lv, P. Zhao, Y. Sun, H. Yuan, and H. Xu, "Research and application of energy-efficient management approach for wireless sensor networks," Sensors , vol. 23, no. 3, p. 1567, 2023. Jiang B, Feng D, and Chen X 2020 Trust-Based Energy-Efficient Data Collection with Unmanned Aerial Vehicle (TEEDC-UAV) in large-scale IoT systems. IEEE Internet Things J. 7(8): 7214-7225. Jin R, He X, and Dai H 2022 Communication efficient federated learning with energy awareness over wireless networks. IEEE Trans. Wireless Commun. 21(7): 5204-5219. K. Karunanithy and B. Velusamy, "Cluster-tree based energy efficient data gathering protocol for industrial automation using WSNs and IoT," Journal of Industrial Information Integration , vol. 19, p. 100156, 2020. Hammadi, Y.I. and Mahmood, O.A., 2022. Transmission performance analysis of 67.5 Tbps MC-WDM system incorporating optical wireless communication. Optoelectronics and Advanced Materials-Rapid Communications , 16 (March-April 2022), pp.130-136.. M. Faheem and V. C. Gungor, "Energy efficient and QoS-aware routing protocol for wireless sensor network-based smart grid applications in the context of industry 4.0," Applied Soft Computing , vol. 68, pp. 910-922, 2018. Mahmuddin M, Alabadleh WA, and Kamarudin LM 2019 A comparative study on hoping mechanism of LEACH protocol in Wireless Sensor Networks: A survey. IOP Conf. Ser. Mater. Sci. Eng. 551: 012056. Mohanty SN, and Patra MR 2020 Deep Learning-Based Distributed Data Mining (DDM) model for energy-efficient wireless sensor networks. IEEE Access 8: 67890-67901. Mohanty SN, Lydia EL, Elhoseny M, Al Otaibi MMG, and Shankar K 2020 Deep learning with LSTM based distributed data mining model for energy efficient wireless sensor networks. Phys. Commun. 40: 101097. Nakas C, Kandris D, and Visvardis G 2020 Energy efficient routing in wireless sensor networks: A comprehensive survey. Algorithms 13(3): 72. Ogbebor JO, Imoize AL, and Atayero AAA 2020 Energy efficient design techniques in next-generation wireless communication networks: emerging trends and future directions. Wirel. Commun. Mobile Comput. 2020: 1-19. P. Sreedevi and S. Venkateswarlu, "Comparative analysis of energy efficient routing protocols with optimization in WSN," International Journal on Interactive Design and Manufacturing (IJIDeM) , pp. 1-16, 2022. Hammadi, Yousif I., Riyadh Khlf Ahmed, Omar Abdulkareem Mahmood, and Ammar Muthanna. "New Filtering Method to Reduce PAPR and OOBE of UFMC in 5G Communication System." In International Conference on Distributed Computer and Communication Networks , pp. 3-16. Cham: Springer Nature Switzerland, 2022.. R. E. Mohamed, A. I. Saleh, M. Abdelrazzak, and A. S. Samra, "Survey on wireless sensor network applications and energy efficient routing protocols," Wireless Personal Communications , vol. 101, pp. 1019-1055, 2018. R. Nagarajan and R. Dhanasekaran, "Energy efficient data transmission approaches for wireless industrial automation," Current Signal Transduction Therapy , vol. 13, no. 1, pp. 37-43, 2018. R. Tanash, M. AlQudah, and S. Al-Agtash, "Enhancing energy efficiency of IEEE 802.15.4-based industrial wireless sensor networks," Journal of Industrial Information Integration , vol. 33, p. 100460, 2023. S. Sivakumar, J. Logeshwaran, R. Kannadasan, M. Faheem, and D. Ravikumar, "A novel energy optimization framework to enhance the performance of sensor nodes in Industry 4.0," Energy Science & Engineering , vol. 12, no. 3, pp. 835-859, 2024. Sahoo BM, Amgoth T, and Pandey HM 2020 Particle swarm optimization based energy efficient clustering and sink mobility in heterogeneous wireless sensor networks. Ad Hoc Netw. 106: 102237. Hammadi YI. Fiber Bragg grating-based monitoring system for fiber to the home (FTTH) passive optical network. Journal of Optical Communications. 2022 Oct 26;43(4):573-83. Tinatin M, Guibin Q, and Jinghui Z 2012 Evolution of wireless communication network and its impact on human life. J. Comput. Sci. Technol. 27(5): 857-868. Deheyab AO, Alwan MH, Rezzaqe IK, Mahmood OA, Hammadi YI, Kareem AN, Ibrahim M. An overview of challenges in medical image processing. InProceedings of the 6th International Conference on Future Networks & Distributed Systems 2022 Dec 15 (pp. 511-516). Wang T, Li X, Liu Y, and Zhang Y 2017 Fog-based data collection for wireless sensor networks with mobile sinks. IEEE Trans. Ind. Informatics 13(2): 850-859. Hussain, Abdul Hussain Ali, Montadar Abas Taher, Omar Abdulkareem Mahmood, Yousif I. Hammadi, Reem Alkanhel, Ammar Muthanna, and Andrey Koucheryavy. "Urban traffic flow estimation system based on gated recurrent unit deep learning methodology for Internet of Vehicles." IEEE Access 11 (2023): 58516-58531. Yang Z, Chen M, Saad W, Xu W, Shikh-Bahaei M, Poor HV, and Cui S 2021 Energy-efficient wireless communications with distributed reconfigurable intelligent surfaces. IEEE Trans. Wireless Commun. 21(1): 665-679. Mahmood, Omar Abdulkareem, Abdukodir Khakimov, Ammar Muthanna, and Alexander Paramonov. "Effect of heterogeneous traffic on quality of service in 5G network." In Distributed Computer and Communication Networks: 22nd International Conference, DCCN 2019, Moscow, Russia, September 23–27, 2019, Revised Selected Papers 22 , pp. 469-478. Springer International Publishing, 2019. Zhuo X, Liu M, Wei Y, Yu G, Qu F, and Sun R 2020 AUV-aided energy-efficient data collection in underwater acoustic sensor networks. IEEE Internet Things J. 7(10): 10010-10022. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 11 Jul, 2025 Read the published version in Journal of Intelligent Systems and Internet of Things → 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-5731209","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":400966997,"identity":"dc39fab9-f2ff-4afd-95cf-3749b3fc94e7","order_by":0,"name":"Maha Abbas Hutaihit","email":"","orcid":"","institution":"University of Diyala","correspondingAuthor":false,"prefix":"","firstName":"Maha","middleName":"Abbas","lastName":"Hutaihit","suffix":""},{"id":400966998,"identity":"655b9901-2243-4098-a8a2-c5c56eb18966","order_by":1,"name":"Samir I. Badrawi","email":"","orcid":"","institution":"Al Mustaqbal University","correspondingAuthor":false,"prefix":"","firstName":"Samir","middleName":"I.","lastName":"Badrawi","suffix":""},{"id":400966999,"identity":"20372a15-2b26-4ecb-8f50-8880c345ef44","order_by":2,"name":"Haider Makki Hameed","email":"","orcid":"","institution":"University of Diyala","correspondingAuthor":false,"prefix":"","firstName":"Haider","middleName":"Makki","lastName":"Hameed","suffix":""},{"id":400967000,"identity":"6be86f7c-2a59-4ccf-834b-b796543d6e76","order_by":3,"name":"Riyadh Khlf Ahmed","email":"","orcid":"","institution":"University of Diyala","correspondingAuthor":false,"prefix":"","firstName":"Riyadh","middleName":"Khlf","lastName":"Ahmed","suffix":""},{"id":400967001,"identity":"5c653fc1-10d7-40a8-ab52-609e90b2c638","order_by":4,"name":"Marwa Flaah Hasan","email":"","orcid":"","institution":"University of Diyala","correspondingAuthor":false,"prefix":"","firstName":"Marwa","middleName":"Flaah","lastName":"Hasan","suffix":""},{"id":400967002,"identity":"b636d8a6-0bd6-4a35-94a8-3837aa22b9c9","order_by":5,"name":"Omar Abdulkareem Mahmood","email":"data:image/png;base64,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","orcid":"","institution":"University of Diyala","correspondingAuthor":true,"prefix":"","firstName":"Omar","middleName":"Abdulkareem","lastName":"Mahmood","suffix":""}],"badges":[],"createdAt":"2024-12-29 18:08:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5731209/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5731209/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.54216/JISIoT.180212","type":"published","date":"2025-07-12T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":73651698,"identity":"4a5027fd-4ece-41ff-949c-3703717696ec","added_by":"auto","created_at":"2025-01-13 09:42:23","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":42585,"visible":true,"origin":"","legend":"\u003cp\u003eIoT edge networks\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/7a4979b61a10e3b62236da72.png"},{"id":73651699,"identity":"8d1a1e2c-d275-4361-bf9a-8c1ad2f89986","added_by":"auto","created_at":"2025-01-13 09:42:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":106910,"visible":true,"origin":"","legend":"\u003cp\u003eConnection between IoT devices and the cloud and edge\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/83e37c8e296fd8e7abc85ef1.png"},{"id":73651998,"identity":"a56a6118-9ab9-4eaf-8b65-9616f6c29432","added_by":"auto","created_at":"2025-01-13 09:50:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":66227,"visible":true,"origin":"","legend":"\u003cp\u003eAn extended deep learning model that has been proposed\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/a6c81239115a3afe339a3958.png"},{"id":73653741,"identity":"81ac53b7-bf98-41d1-b47e-fcf499b27bdf","added_by":"auto","created_at":"2025-01-13 09:58:23","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":55252,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy \u0026amp; Loss\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/b95ee81f323ee5611975161d.png"},{"id":73653742,"identity":"59ceaeb9-2515-41be-8136-0a98ef98ea7c","added_by":"auto","created_at":"2025-01-13 09:58:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":98497,"visible":true,"origin":"","legend":"\u003cp\u003eResource utilization (%)\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/e8eb7fbcc2bd8ff1a716a03b.png"},{"id":73653743,"identity":"b516dc55-e8cd-4805-8bf0-7969efaa7a09","added_by":"auto","created_at":"2025-01-13 09:58:23","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":106218,"visible":true,"origin":"","legend":"\u003cp\u003eResponse time (s)\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/121b034b2d5f658bf03481fb.png"},{"id":73652004,"identity":"cdd123eb-55d4-4585-a2f7-bb8ad602091d","added_by":"auto","created_at":"2025-01-13 09:50:23","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":92006,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal Analysis of Execution\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/e15addc4226c3f8f29c452ae.png"},{"id":73651721,"identity":"c88a9231-8b74-4a43-b0e5-f9958f740008","added_by":"auto","created_at":"2025-01-13 09:42:23","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":89656,"visible":true,"origin":"","legend":"\u003cp\u003eAnalysis of Average Delays\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/729a23cf06d70aeab42fec18.png"},{"id":73651704,"identity":"e7d28fc3-3d16-47a8-bafc-cceeb5cfceb6","added_by":"auto","created_at":"2025-01-13 09:42:23","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":108536,"visible":true,"origin":"","legend":"\u003cp\u003eEfficiency Analysis\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/c4297d0fe83b28c1b4ec05da.png"},{"id":92526783,"identity":"0db23385-3b6a-499b-b406-1024251c7a54","added_by":"auto","created_at":"2025-09-30 15:54:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1396589,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5731209/v1/27e06334-4e56-4e12-a552-e334edb94f22.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Design and optimization of energy-efficient wireless sensor networks for industrial automation","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eAs of late, there has been a ton of interest in wireless sensor networks (WSNs) from both industry and academia. Industrial field control, smart homes, smart factories, environmental observing, and industrial field checking are some of the many communicated observation and control areas that make extensive use of WSNs[1].WSNs are often made up of several sensor hubs that perform a variety of tasks, including data handling, transmission, and gathering. A WSN clearly offers advantages over conventional wired systems with regards to cost, adaptability, and ease[2].\u003c/p\u003e \u003cp\u003eWSNs are often fueled by batteries in industrial settings, and these batteries' constrained capacity and dreary replacement prerequisites have turned into the primary barriers to WSN adoption [3]. In WSN applications, lifetime power utilization management is crucial [4].\u003c/p\u003e \u003cp\u003eThe goal of creating energy-assortment advances is to increase the valuable lifetime of WSNs by using various procedures for energy harvesting [5]. Showcase a vibration energy-controlled WSN demonstration testbed. Vibration energy is planned to be harvested by an electromagnetic harvester [6]. Berkeley researchers offer a design for a WSN that runs ceaselessly on renewable energy from the climate [7].Two-stage storage systems made contained a rechargeable solar lithium battery and a supercapacitor are utilized in this method [8]. A gadget for harvesting human energy is built and examined, using a power hardware module to harness the energy created by a walking human [9\u0026ndash;12]. Look into a physically autonomous sensor that uses mechanical nano energy derived from human beings to take advantage of movement location and physiological signal monitoring [13]. A wireless energy harvesting cognitive industrial internet of things (CIIOT) is proposed, which simultaneously performs range detecting and transmissions and harvests wireless energy from a primary client [14\u0026ndash;18]. In any case, these research' energy-harvesting gadgets produce less energy and need movement, such mechanical vibrations in the system [19\u0026ndash;22]. They cannot be broadly applied to industrial settings and have stayed in the laboratory design and verification stages.\u003c/p\u003e \u003cp\u003eData gathering, transmission, and handling are a WSN's primary energy-consuming operations, and these activities are also the vital targets of energy usage improvement [23]. Artificial neural networks are utilized to manage data sampling and preserve hub energy [24]. To increase the WSN's useful life, an adaptive sampling strategy that considers the temporal and spatial correlation of the sensor data is proposed. created a ZigBee and Geiger Muller tube system that reasonably estimates radiation and temperature monitoring [25]. improved uptime during power outages by settling on the ideal portable hub antenna configuration and the most efficient asset management strategy for relay hub selection. suggested an approach to adaptive data collection that would reduce energy consumption and advance data transmission [26\u0026ndash;30]. These methods, which are more commonly used in labs, can precisely control energy, reduce energy consumption, and lengthen the lifespan of systems. However, these methods have on occasion been employed in contexts including difficult outside conditions, such as present HVDC transmission line electromagnetic measurement systems.\u003c/p\u003e \u003cp\u003eChina currently has a large number of HVDC transmission lines operational. The electromagnetic environment must be considered throughout the whole transmission line lifecycle, from design to installation and operation [31]. When studying the electromagnetic field, it is crucial to take into account the electric field beneath the(HVDC) transmission lines. With the help of a wireless sensor network (WSN) electric-field measurement system (EFMS), the electric field under the HVDC transmission cables can be observed. One important part of the EFMS and a popular area of study is the electric-field sensor. Optical, field factory, and MEMS sensors are just a few of the several possible techniques [32]. The charge enrollment guideline states that MEMS sensors convert the DC electric field into measurable electrical values using micromechanical resonators. One-way optical sensors detect electric fields is by using the electro-optic impact concept, which involves watching how a field changes the refractive index of a crystal [33]. The widely recognized charging standard is employed by the field factory sensors. At regular intervals, the engine opens and protects the electric field, keeping the rotor spinning at a constant speed. By cycling between charging and discharging charges, the enrollment cathode can provide an AC signal that is proportionate to the applied DC electric field. Power consumption of the EFMS is significantly increased due to the constant operation of the electric-field factory sensor's motor [34]. Since electric field sensors do not require an engine, they are able to utilize power-efficient MEMS and optical sensors. Unfortunately, most current electric-field measurement systems rely on electric-field factories as their electric-field sensors, mostly because of affordability and stability concerns [35].\u003c/p\u003e \u003cp\u003eIn order to study the electric-field circulation under the transmission lines, a range of sensors can continuously measure the electric-field. It is common practice to run EFMSs on batteries due to the lack of a reliable power source at many testing sites [36]. Because of the system's usual data advancement requirements and the battery capacity, extending the usable life of a battery-powered EFMS is necessary yet problematic. New features, like data transfer, are incorporated into the EFMS as a consequence of real-time checking and distributed sensor networks; nevertheless, this considerably raises the system's energy consumption. The major objective of this work is to provide an energy-efficient booking technique that may be used to extend the lifespan of an EFMS.\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003e1.1 Industrial Automation and the Need for WSNs\u003c/h2\u003e \u003cp\u003eWSNs have played an important role in the development of industrial automation, especially within the context of Industry 4.0, which emphasizes the integration of cyber-physical systems, Internet of Things (IoT), and data-driven decision-making, so that in WSNs, through real-time monitoring and control of industrial processes, improvements in efficiency and productivity and safety can be facilitated [37]. One of the important applications includes environmental monitoring, condition monitoring of equipment, predictive maintenance, and process optimization where sensor nodes collect and transmit critical data to enhance operational intelligence [38]. Nevertheless, although WSNs can be deployed in industrial automation, challenges faced include energy consumption, network scalability, and data reliability in harsh environments. These challenges bring opportunities to innovation toward energy-efficient protocols, advanced sensor technologies, and robust communication strategies; they are crucial for optimizing the performance and sustainability of industrial systems in the industry 4.0 era.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e1.2 Energy Efficiency in Wireless Sensor Networks\u003c/h2\u003e \u003cp\u003eEnergy efficiency is very sensitive to WSNs, particularly for industrial applications in large-scale remote environments or in harsh environments where the nodes are set up [39]. The energy usage of WSNs directly influences the life of the battery-powered nodes, the reliability of networks, and the operational costs in general. In industrial environments, where the requirement is for high continuity monitoring of data, the challenge is how to balance these data transmission frequencies with the limited energy sources available on the sensor nodes [40]. This can be solved by using several strategies-including energy-efficient routing protocols, adaptive techniques of data transmission, and duty cycling. Techniques that involve data aggregation, energy harvesting, and the use of low-power communication protocols are also underway in reducing energy consumption but without sacrificing performance and reliability in industrial automation systems for WSNs [41].\u003c/p\u003e \u003c/div\u003e"},{"header":"2. Research Methodology","content":"\u003cp\u003eIn any situation where the Internet of Things is put to substantial use, no matter how large or little. Every single one of the domains makes use of the IoT's feature advantages, which greatly boost the performance of real-time applications. There is a pressing need for efficient data management due to the explosion of Internet of Things (IoT) devices, which has enabled smart cities and smart agriculture. It is possible that distributed computing can efficiently manage the IoT network in terms of storage and figure limitations. Internet of Things devices with low resources gather data and upload it to the cloud for processing. Due to the heterogeneity of the network, data transmissions from IoT devices to the cloud, whether incoming or departing, experience high latency and bandwidth needs. Edge computing is suggested as a means to lessen latency in IoT networks; it involves moving processing power from the cloud to the client end. As an add-on, edge registering is useful for cloud and Internet of Things networks. Distributed computing's figure load is drastically reduced in real-time applications by using edge registering.\u003c/p\u003e \u003cp\u003eWith their three-layer heterogeneous design, cloud-edge cloud IoT networks require proper asset planning to further increase productivity and service quality. The diverse types of data collected by the Internet of Things network necessitate different approaches to data processing. All data should be handled on the edge or in the cloud when asset demands are booked. The asset requirements can be represented by tasks, and the edge network can obtain these demands in turn. After that, it selects the best cloud assets and arranges for IoT networks to use them for additional processing. To avoid SLA violations, cloud providers should think about things like load balancing, energy consumption, and bandwidth congestion, regardless of how many resources they offer.\u003c/p\u003e \u003cp\u003eIn order to reduce the real-time real data handling latency in the dispersed computing environment, IoT networks use edge registration. The edge enhances the performance of the network and reduces compute, blockage, and data transmission delay by supplying the appropriate cloud assets to IoT networks. Distributed computing assets should be planned for edge figuring into Internet of Things networks as part of an effective asset planning strategy. Planning must take asset elasticity and scalability into account. When assets in the cloud are shared via edge networks, their scalability and flexibility change since most of these assets are virtual or physical. Because not all applications have the same asset demands, the edge should be aware of these needs when booking assets for IoT networks. This is because different applications require different registering assets.\u003c/p\u003e \u003cp\u003eThe resource scheduling algorithms that have emerged recently are either machine learning- or statistically-based scheduling procedures. Based on the resource requirements, the best resources are chosen from the resource pool. Deep learning techniques have recently supplanted statistical and machine learning-based scheduling models in order to improve scheduling performance. When it comes to scheduling resources at the edge of a network, deep learning techniques like RL, Q-learning, and deep neural networks are heavily utilized. Cutting down on wait times and scheduling delays is critical for optimizing efficiency, even when deep learning methods work as expected. Asset planning in edge-integrated IoT networks is facilitated by this combined deep learning method.\u003c/p\u003e \u003cp\u003eIntegrating IoT Services in the Cloud Many applications on the Internet of Things have embraced distributed computing because of its capacity to process, store, and analyze massive amounts of data. The distributed computing environment is another option for the Internet of Things (IoT) that relies on cloud computing to enable device connection. According to their requirements, customers can access cloud services whenever and wherever they like. This internet of things (IoT) approach that integrated the cloud was utilized by numerous smart city, transportation, agricultural, and healthcare applications. However, the data processing experiences latency due to the long-distance data transit between IoT devices to the cloud. Various applications of IoT are applied for actual real-time time operations which should deliver rapid replies by evaluating the data. Delays in answers will harm the quality of services in IoT applications. Meanwhile, it is vital to maintain a steady connection between devices and the cloud49 which is also a challenge while integrating the cloud with IoT. Thus, it is vital to consider few insights as provided below when designing cloud- based integrated IoT apps.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eAn exceptionally fast reaction time and as little delay as possible from beginning to end are essential for improving service quality.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe cloud and IoT application hubs are dependent on a reliable and steady network.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSignificant processing complexity will result from adding more networking conventions. For this reason, limiting computational difficulties requires careful consideration while selecting conventions.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Edge IoT Integration\u003c/h2\u003e \u003cp\u003eAn edge-integrated m module can help with the limitations of a cloud-integrated IoT ecosystem. Not only does edge registration reduce handling complexity and latency, but it also brings cloud resources directly to the user's device. Some examples of applications of edge figuring include cloudlets, flexible edge processing, and mist registering. Each of these methods reduces the amount of time data must be processed before IoT applications get their responses. Reduced latency and data transfer times are the results of edge processing, which places assets in close proximity to the Internet of Things. The IoT app can take advantage of edge processing to improve both its local and transmitted data handling capabilities. Edge enhances system resilience and fault tolerance by reducing bandwidth requirements and providing Internet of Things clients with great adaptability. The interplay between edge clouds and Internet of Things devices is illustrated in Fig.\u0026nbsp;3.1.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Proposed Concatenated Deep Learning Algorithm\u003c/h2\u003e \u003cp\u003eAn in-depth mathematical model of the concatenated deep learning technique that has been suggested is presented in this section. A convolutional neural network and a gated recurrent unit are utilized for initial feature extraction in the most basic form of the proposed model, as illustrated in Fig.\u0026nbsp;5.1. The asset's category, sub-class, class, duration, and other relevant criteria are taken into account when determining the demand for the asset.\u003c/p\u003e \u003cp\u003eAsset needs can be parsed into local and regional details by use of a one-dimensional convolutional neural network. Similar to how features are recovered in the latter stage of the process using a gated recurrent unit, the best asset for planning in edge processing is chosen in the first stage. For this project, we've settled on a gated recurrent unit (GRU) because of its ease of use and high performance. When compared to conventional long short-term memory (LSTM), GRU performs better due to its faster input processing and less parameter usage. By fixing the vanishing gradient issue in RNN, GRU could make current booking algorithms better. Finally, after the combined traits are categorized, the best resources for the jobs are scheduled.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eGated Recurrent Unit\u003c/b\u003e \u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eAs a GRU model, a gated recurrent neural network is recommended. With only two gates, GRU stands in stark contrast to LSTM's trio. The GRU\u0026rsquo;s update and reset gates not only improve union rates but also reduce the number of parameters needed compared to an LSTM. Using its memory cell, the GRU model may retrieve crucial data and identify situations in the input asset requirements. The GRU's reset gate forgets or erases the redundant data. The GRU model often takes period series data as input, even if the asset demand input is typically period grouping data with a single time step. The activation is successful, and the GRU model's outputs are obtained. By feeding the principal layer's output into the next layer and repeating the process, we may extract the important features from the input to the resultant layer. Mathematical descriptions of the GRU model are as:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{G}}_{\\varvec{u}}=\\varvec{\\sigma\\:}\\left({\\mathcal{w}}_{\\mathcal{u}}\\left({\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}-1\\right)},{\\varvec{x}}^{\\left(\\varvec{t}\\right)}\\right)+{\\mathcal{b}}_{\\mathcal{u}}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{G}}_{\\mathcal{r}}=\\varvec{\\sigma\\:}\\left({\\mathcal{w}}_{\\mathcal{r}}\\left({\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}-1\\right)},{\\varvec{x}}^{\\left(\\varvec{t}\\right)}\\right)+{\\mathcal{b}}_{\\mathcal{r}}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eHere, Gu handles the update gate while Gr handles the reset gate. Update gates differ from reset gates in that their range is [0,1] rather than [-1,1]. Wr denotes the capability to reset the gate weight and Wu stands for the capability to update the gate weight. Just as how Br takes care of the update gate's bias vector, as does as for the reset gate. The candidate activation capability for the recurrent unit is created using the following formula, which is based on the gate works.\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:{\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}\\right)}=\\varvec{t}\\varvec{a}\\varvec{n}\\varvec{h}\\left[{\\mathcal{w}}_{\\mathcal{u}}\\left({{\\mathcal{G}}_{\\varvec{r}}\\times\\:\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}-1\\right)},{\\varvec{x}}^{\\left(\\varvec{t}\\right)}\\right)+{\\mathcal{b}}_{\\mathcal{u}}\\right]$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eTo handle the bias vector, the activation capacity weight factors\u0026mdash;Wu for the update gate\u0026mdash;are applied when the input training data is labeled as X(t). The final step is to transmit the GRU model's output as:\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{v}}^{\\left(\\varvec{t}\\right)}=\\left(\\left(1-{\\mathcal{G}}_{\\mathcal{u}}\\right)\\times\\:{\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}-1\\right)}\\right)+\\left({{\\mathcal{G}}_{\\mathcal{u}}\\times\\:\\stackrel{\\sim}{\\mathcal{v}}}^{\\left(\\varvec{t}\\right)}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere d is a function of the output of the prior unit and v(t-1) is the input of the present unit. Merging the output characteristics of the CNN and GRU models results in extra processing steps that determine which assets are most suitable for booking.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eConvolutional Neural Network\u003c/b\u003e \u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThe input is sorted into subclasses according to the specifications by the Convolutional Neural Network Model employed in the proposed work before the Convolution layer takes over. Two max-pooling layers and two convolution layers are employed by the suggested design to glean useful details from the asset needs. In terms of data handling and maintaining local interactions, CNN outperforms traditional neural network models. Compared to earlier models, this neural network method conveys attributes more clearly while jellying the input data's spatial localization. Through autonomous training, the network is able to absorb various data qualities. The CNN module is based on the correlated cycle principle of the convolution interaction. To construct the convolution interaction correctly, one needs to think about the loads of the one-dimensional aspects component, which are represented by the words {W1,W2,...,Wn}, where n is the kernel length.\u003cdiv id=\"Equ5\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ5\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{y}}_{\\varvec{t}}=\\varvec{f}\\left({\\sum\\:}_{\\varvec{i}=1}^{\\varvec{n}}{\\varvec{w}}_{\\varvec{i}}\\varvec{*}{\\varvec{x}}_{\\varvec{t}-\\varvec{i}+1}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e5\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWe say that the data created at time t is Yt and that the input sample is Xt. The proposed model makes use of the Redressed Linear Unit (RELU) as its activation capability. The activation capability can be mathematically expressed as\u003cdiv id=\"Equ6\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ6\" name=\"EquationSource\"\u003e\n$$\\:\\varvec{R}\\varvec{e}\\varvec{L}\\varvec{U}\\left(\\varvec{x}\\right)=\\left\\{\\begin{array}{c}x,\\\\\\:0,\\end{array}\\right.\\genfrac{}{}{0pt}{}{\\varvec{x}}{\\varvec{x}}\\genfrac{}{}{0pt}{}{\u0026gt;}{\\le\\:}\\genfrac{}{}{0pt}{}{0}{0}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e6\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThe suggested architecture employs max pooling to cap the feature size subsequent to the convolution layer. The outputs of the convolution layer are down-sampled to reduce unpredictability. Maximum pooling operator mathematically expresses the forwarding of the maximum value as\u003cdiv id=\"Equ7\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ7\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{P}}_{\\mathcal{j}.\\mathcal{m}}=\\mathbf{max}{{\\mathcal{h}}_{\\mathcal{j}},}_{\\left(\\mathcal{m}-1\\right)}\\mathcal{n}+\\mathcal{r}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e7\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere n is the permitted area-to-area pooling shift, m is the maximum pooled band, and j are the channels. With most convolution bands, the pooling layer reduces their dimensionality. Batch normalization follows pooling capabilities and improves training results by standardizing the features. The mathematical description of batch normalization characteristics is up next.\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\mathcal{u}=\\frac{1}{{\\mathcal{n}}_{\\varvec{b}\\varvec{a}\\varvec{t}}}{\\sum\\:}_{\\mathcal{n}=1}^{{\\mathcal{n}}_{\\varvec{b}\\varvec{a}\\varvec{t}}}{\\varvec{x}}_{\\mathcal{n}}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:{\\varvec{\\sigma\\:}}^{2}=\\frac{1}{{\\mathcal{n}}_{\\varvec{b}\\varvec{a}\\varvec{t}}}{\\sum\\:}_{\\mathcal{n}=1}^{{\\mathcal{n}}_{\\varvec{b}\\varvec{a}\\varvec{t}}}{\\left({\\mathcal{x}}_{\\mathcal{n}}-\\varvec{\\mu\\:}\\right)}^{2}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\widehat{\\mathcal{x}}=\\frac{{\\mathcal{x}}_{\\mathcal{n}}-\\varvec{\\mu\\:}}{\\sqrt{{\\varvec{\\sigma\\:}}^{2}+\\varvec{\\epsilon\\:}}}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ8\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ8\" name=\"EquationSource\"\u003e\n$$\\:{\\mathcal{y}}_{\\mathcal{n}}=\\varvec{\\gamma\\:}{\\widehat{\\mathcal{x}}}_{\\mathcal{n}}+\\varvec{\\beta\\:}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e8\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere X n is the input data and N bat is the batch size. While σ2 addresses the batch variance, \u0026micro; indicates the mean. In order to avoid zero gradients, the normalized data is associated with a constant ε, denoted as X̂. Dand K is the graphic depiction of the learning vector parameters. The features that are represented by the output are ʆ and β. Yn stands for the feature that is produced.\u003c/p\u003e \u003cp\u003eAfterwards, the CNN and GRU models' properties are integrated. To avoid overfitting the data, apply a dropout layer following concatenation. Finally, the collected features are classified using the fully linked network layer and SoftMax algorithms in order to assign the correct resources to a job. The SoftMax capability can be stated numerically as\u003cdiv id=\"Equ9\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ9\" name=\"EquationSource\"\u003e\n$$\\:\\widehat{\\mathcal{y}}=\\varvec{s}\\varvec{o}\\varvec{f}\\varvec{t}\\varvec{m}\\varvec{a}\\varvec{x}\\left(\\mathcal{Q}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e9\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ethe result of the dropout layer is denoted by Q. Last but not least, the mistake capacity of the suggested model is checked using a cross-entropy capability. Mathematically, it's expressed as\u003cdiv id=\"Equ10\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ10\" name=\"EquationSource\"\u003e\n$$\\:\\mathcal{l}=-\\frac{1}{\\mathcal{b}}{\\sum\\:}_{\\varvec{i}=1}^{\\varvec{n}}{\\mathcal{y}}_{\\varvec{i}}\\mathbf{log}{\\mathcal{y}}_{\\varvec{i}}^{\\varvec{{\\prime\\:}}}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e10\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eAs an example, yi' represents the expected component and yi addresses the actual component, whereas b, n, and yi are the sizes of the batch and training samples, respectively.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Performance Evaluation","content":"\u003cp\u003eWe empirically validate the performance of the recommended deep learning model by incorporating the package and works as capabilities in a Python simulation study. Using these tools, hyperparameters are automatically generated and adjusted to enhance performance even more. The benchmark data used in the investigation comes from Intel's Berkeley research laboratories and contains 96. For the purpose of this simulation investigation, the hyperparameters used are detailed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Validation is conducted by examining and comparing current strategies, including the hereditary algorithm, the Improved Particle Swarm Optimization (IPSO) algorithm, Long Short-Term Memory (LSTM), and the Bidirectional Recurrent Neural Network (BRNN).\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\u003eHyperparameters of the suggested DL technique\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eS. No\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eParameters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eValue\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eConv filters 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eConv filters 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e128\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGRU Units\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDropout\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEpochs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBatch Size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe proposed model's accuracy and misfortune curves are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Measuring performance is based on the standard method of testing and training. Each dataset is partitioned for testing, validation, and training at 70:20:10. The results have not altered after more than 25 generations of measurement. According to the results, the suggested model is the most accurate, and it has been verified.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe suggested model's asset use is compared to the ongoing models in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. The results show that the suggested model is able to make the most efficient use of assets thanks to the optimal choice of assets. Positions are allocated optimal assets, which expedites data processing and frees up these assets for other uses. The overall asset consumption of the suggested approach is thus larger than that of current strategies. The suggested model and the existing BRNN models perform similarly, however other models display large discrepancies in the values of asset utilization.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe response times of the ongoing asset planning methods and the suggested concatenated deep learning strategy are contrasted in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. Reaction time is the amount of time it takes for the planning algorithm to assess and plan for asset demands. The average time is calculated for each strategy that requests a certain asset from edge figuring. The suggested paradigm for asset demands demonstrates a minimal reaction time of 1.25 seconds. However, when other methods are employed, the average rises. By completing asset demands in 1.66s, 1.98s, and 2.20s, respectively, the LSTM model, the BRNN model, GA, and IPSO all outperform the suggested model in terms of reaction time.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe suggested model's total execution seasons are compared to those of the state-of-the-art models in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. Execution time includes the time required to evaluate asset demand, choose the best asset from the pool, and schedule that asset. According to the results, the suggested model has the fastest execution time when compared to alternative strategies for planning. Compared to the BRNN model, the LSTM-based booking, the GA, and the IPSO model, the recommended model's execution season of 10.25s is 5s quicker, 8s faster, 11s faster, and 16s faster, respectively.\u003c/p\u003e \u003cp\u003eWe prioritize any time savings that may be achieved in the asset booking procedure for edge registration. The suggested model's performance is validated by considering the average delay given by both the present and new models. This is done through research into different asset demands. The results, shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, show that compared to the state-of-the-art methods, the suggested model minimizes delays more effectively on average. By a margin of five seconds, the IPSO model outperforms the suggested model in terms of latency. When compared to GA-based booking, the suggested process is 4.3 seconds slower. Models trained with LSTM and BRNN outperform those using GA and IPSO by a small margin. However, this model is far from perfect. The LSTM-based booking model is 2 seconds off and the BRNN model is 1.5 seconds off; the recommended model has a minimum delay of 1.15 seconds.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e showcases an effective comparison of booking algorithms. By comparing the proposed and current models' execution times, reaction times, and delay factors, we may get a sense of how efficient they are overall. The proposed model outperforms the competition across the board, improving the efficiency of both the IoT networks and the edge processing platform. Compared to the current booking methods, the suggested model has a substantially higher maximum efficacy of 99.48%.\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\u003eAnalysing Performance in Comparison\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\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=\"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=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMethods\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResource Utilization\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eResponse Time\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eExecution Time\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAverage Delay\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEfficiency\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIPSO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e95.8217%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.2123s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26.0035s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5.4136s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e94.4532%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e95.7245%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.9742s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20.6379s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.2948s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e96.1037%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLSTM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e98.5871%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.6734s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17.4659s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.1368s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e97.9082%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBRNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e99.0213%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.5529s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14.8164s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.8547s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e98.8045%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProposed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e99.5234%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.2531s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.2537s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.1539s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e99.4862%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e provides an overview of the overall performance metrics used to compare the proposed model to the current models. The results demonstrate that the proposed approach achieves a higher level of asset utilization and productivity when compared to other existing booking systems. The proposed model also has the quickest execution and reaction times, making it ideal for real-time applications that need to record data produced or received efficiently through asset allocation.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eHere, we provide a mixed-methods deep learning approach to asset planning in IoT networks with embedded edges. In order to prioritize characteristics based on asset demands during the asset planning phase, the suggested work utilizes a gated recurrent unit and a one-dimensional convolutional neural network. The optimal planning assets are identified through the application of deep learning models that quickly assess and combine time-series requests, followed by classification. With the use of simulation analysis, we can see how the suggested model stacks up against other methods in terms of effectiveness, average delay, reaction time, execution time, and asset utilization, as well as against Genetic Algorithm (GA), LSTM, BRNN, and Genetic Algorithm (GA). The overall efficiency of the cycle is supported by the suggested model's best asset utilization, base execution, and reaction times when compared to current strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthors share their knowledge and efforts to do this work\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eThe research work presented in this paper, \u003cem\u003eDesign and Optimization of Energy-Efficient Wireless Sensor Networks for Industrial Automation\u003c/em\u003e, has been made possible through the continuous support and guidance of my mentor, \u003cb\u003e(________)\u003c/b\u003e Her extensive knowledge in the field of research and her vast experience have been a constant source of inspiration, enabling the development of this paper in the form of a comprehensive research article.\u003c/p\u003e \u003cp\u003eI extend my heartfelt gratitude to (_______). for her invaluable contributions and unwavering encouragement throughout this project.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eA. Chehri, R. Saadane, N. Hakem, and H. Chaibi, \u0026quot;Enhancing energy efficiency of wireless sensor network for mining industry applications,\u0026quot; \u003cem\u003eProcedia Computer Science\u003c/em\u003e, vol. 176, pp. 261-270, 2020.\u003c/li\u003e\n\u003cli\u003eAhmed S, Chowdhury MZ, and Jang YM 2020 Energy-efficient UAV-to-user scheduling to maximize throughput in wireless networks. IEEE Access 8: 21215-21225.\u003c/li\u003e\n\u003cli\u003eAlamu O, Gbenga-Ilori A, Adelabu M, Imoize A, and Ladipo O 2020 Energy efficiency techniques in ultra-dense wireless heterogeneous networks: An overview and outlook. Eng. Sci. Technol. Int. J. 23(6): 1308-1326.\u003c/li\u003e\n\u003cli\u003eAlghamdi TA 2020 A hybrid algorithm for energy-efficient cluster head selection in wireless sensor networks. IEEE Access 8: 189211-189222.\u003c/li\u003e\n\u003cli\u003eArora VK, Sharma V, and Sachdeva M 2016 A survey on LEACH and other routing protocols in wireless sensor networks. Optik 127(16): 6590\u0026ndash;6600.\u003c/li\u003e\n\u003cli\u003eBagwari, A., Logeshwaran, J., Usha, K., Kannadasan, R., Alsharif, M. H., Uthansakul, P., \u0026amp;Uthansakul, M. (2023). An Enhanced Energy Optimization Model for Industrial Wireless Sensor Networks Using Machine Learning. \u003cem\u003eIEEE Access\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eBhola J, Soni S, and Cheema GK 2020 Genetic algorithm based optimized LEACH protocol for energy efficient wireless sensor networks. J. Ambient Intell. Humaniz. Comput. 11: 1281-1288.\u003c/li\u003e\n\u003cli\u003eC. S. Abella et al., \u0026quot;Autonomous energy-efficient wireless sensor network platform for home/office automation,\u0026quot; \u003cem\u003eIEEE Sensors Journal\u003c/em\u003e, vol. 19, no. 9, pp. 3501-3512, 2019.\u003c/li\u003e\n\u003cli\u003eChowdhury SM, and Hossain A 2020 Different energy saving schemes in wireless sensor networks: A survey. Wirel. Pers. Commun. 114(3): 2043-2062.\u003c/li\u003e\n\u003cli\u003eD. D. Rao, A. Jain, S. Sharma, S. V. Pandit, and R. Pandey, \u0026quot;Effectual energy optimization stratagems for wireless sensor network collections through fuzzy-based inadequate clustering,\u0026quot; \u003cem\u003eSN Computer Science\u003c/em\u003e, vol. 5, no. 8, pp. 1-10, 2024.\u003c/li\u003e\n\u003cli\u003eD. Lin, Q. Wang, W. Min, J. Xu, and Z. Zhang, \u0026quot;A survey on energy-efficient strategies in static wireless sensor networks,\u0026quot; \u003cem\u003eACM Transactions on Sensor Networks (TOSN)\u003c/em\u003e, vol. 17, no. 1, pp. 1-48, 2020.\u003c/li\u003e\n\u003cli\u003eDario S, Luca P, Giuseppe F, and Pietro M 2014 Radio-Access Network-As-A-Service (RANAAS): Architecture for energy efficiency in 5G networks. IEEE Commun. Mag. 52(2): 114-122.\u003c/li\u003e\n\u003cli\u003eF. H. El-Fouly and R. A. Ramadan, \u0026quot;Real-time energy-efficient reliable traffic aware routing for industrial wireless sensor networks,\u0026quot; \u003cem\u003eIEEE Access\u003c/em\u003e, vol. 8, pp. 58130-58145, 2020.\u003c/li\u003e\n\u003cli\u003eH. Kathiriya, A. Pandya, V. Dubay, and A. Bavarva, \u0026quot;State of art: energy efficient protocols for self-powered wireless sensor network in IIoT to support industry 4.0,\u0026quot; in \u003cem\u003eProc. 2020 8th Int. Conf. on Reliability, Infocom Technologies and Optimization (Trends and Future Directions) (ICRITO)\u003c/em\u003e, 2020, pp. 1311-1314, IEEE.\u003c/li\u003e\n\u003cli\u003eJ. Blanco, A. Garc\u0026iacute;a, and J. D. L. Morenas, \u0026quot;Design and implementation of a wireless sensor and actuator network to support the intelligent control of efficient energy usage,\u0026quot; \u003cem\u003eSensors\u003c/em\u003e, vol. 18, no. 6, p. 1892, 2018.\u003c/li\u003e\n\u003cli\u003eJ. Li, J. Lv, P. Zhao, Y. Sun, H. Yuan, and H. Xu, \u0026quot;Research and application of energy-efficient management approach for wireless sensor networks,\u0026quot; \u003cem\u003eSensors\u003c/em\u003e, vol. 23, no. 3, p. 1567, 2023.\u003c/li\u003e\n\u003cli\u003eJiang B, Feng D, and Chen X 2020 Trust-Based Energy-Efficient Data Collection with Unmanned Aerial Vehicle (TEEDC-UAV) in large-scale IoT systems. IEEE Internet Things J. 7(8): 7214-7225.\u003c/li\u003e\n\u003cli\u003eJin R, He X, and Dai H 2022 Communication efficient federated learning with energy awareness over wireless networks. IEEE Trans. Wireless Commun. 21(7): 5204-5219.\u003c/li\u003e\n\u003cli\u003eK. Karunanithy and B. Velusamy, \u0026quot;Cluster-tree based energy efficient data gathering protocol for industrial automation using WSNs and IoT,\u0026quot; \u003cem\u003eJournal of Industrial Information Integration\u003c/em\u003e, vol. 19, p. 100156, 2020.\u003c/li\u003e\n\u003cli\u003eHammadi, Y.I. and Mahmood, O.A., 2022. Transmission performance analysis of 67.5 Tbps MC-WDM system incorporating optical wireless communication. \u003cem\u003eOptoelectronics and Advanced Materials-Rapid Communications\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(March-April 2022), pp.130-136..\u003c/li\u003e\n\u003cli\u003eM. Faheem and V. C. Gungor, \u0026quot;Energy efficient and QoS-aware routing protocol for wireless sensor network-based smart grid applications in the context of industry 4.0,\u0026quot; \u003cem\u003eApplied Soft Computing\u003c/em\u003e, vol. 68, pp. 910-922, 2018.\u003c/li\u003e\n\u003cli\u003eMahmuddin M, Alabadleh WA, and Kamarudin LM 2019 A comparative study on hoping mechanism of LEACH protocol in Wireless Sensor Networks: A survey. IOP Conf. Ser. Mater. Sci. Eng. 551: 012056.\u003c/li\u003e\n\u003cli\u003eMohanty SN, and Patra MR 2020 Deep Learning-Based Distributed Data Mining (DDM) model for energy-efficient wireless sensor networks. IEEE Access 8: 67890-67901.\u003c/li\u003e\n\u003cli\u003eMohanty SN, Lydia EL, Elhoseny M, Al Otaibi MMG, and Shankar K 2020 Deep learning with LSTM based distributed data mining model for energy efficient wireless sensor networks. Phys. Commun. 40: 101097.\u003c/li\u003e\n\u003cli\u003eNakas C, Kandris D, and Visvardis G 2020 Energy efficient routing in wireless sensor networks: A comprehensive survey. Algorithms 13(3): 72.\u003c/li\u003e\n\u003cli\u003eOgbebor JO, Imoize AL, and Atayero AAA 2020 Energy efficient design techniques in next-generation wireless communication networks: emerging trends and future directions. Wirel. Commun. Mobile Comput. 2020: 1-19.\u003c/li\u003e\n\u003cli\u003eP. Sreedevi and S. Venkateswarlu, \u0026quot;Comparative analysis of energy efficient routing protocols with optimization in WSN,\u0026quot; \u003cem\u003eInternational Journal on Interactive Design and Manufacturing (IJIDeM)\u003c/em\u003e, pp. 1-16, 2022.\u003c/li\u003e\n\u003cli\u003eHammadi, Yousif I., Riyadh Khlf Ahmed, Omar Abdulkareem Mahmood, and Ammar Muthanna. \u0026quot;New Filtering Method to Reduce PAPR and OOBE of UFMC in 5G Communication System.\u0026quot; In \u003cem\u003eInternational Conference on Distributed Computer and Communication Networks\u003c/em\u003e, pp. 3-16. Cham: Springer Nature Switzerland, 2022..\u003c/li\u003e\n\u003cli\u003eR. E. Mohamed, A. I. Saleh, M. Abdelrazzak, and A. S. Samra, \u0026quot;Survey on wireless sensor network applications and energy efficient routing protocols,\u0026quot; \u003cem\u003eWireless Personal Communications\u003c/em\u003e, vol. 101, pp. 1019-1055, 2018.\u003c/li\u003e\n\u003cli\u003eR. Nagarajan and R. Dhanasekaran, \u0026quot;Energy efficient data transmission approaches for wireless industrial automation,\u0026quot; \u003cem\u003eCurrent Signal Transduction Therapy\u003c/em\u003e, vol. 13, no. 1, pp. 37-43, 2018.\u003c/li\u003e\n\u003cli\u003eR. Tanash, M. AlQudah, and S. Al-Agtash, \u0026quot;Enhancing energy efficiency of IEEE 802.15.4-based industrial wireless sensor networks,\u0026quot; \u003cem\u003eJournal of Industrial Information Integration\u003c/em\u003e, vol. 33, p. 100460, 2023.\u003c/li\u003e\n\u003cli\u003eS. Sivakumar, J. Logeshwaran, R. Kannadasan, M. Faheem, and D. Ravikumar, \u0026quot;A novel energy optimization framework to enhance the performance of sensor nodes in Industry 4.0,\u0026quot; \u003cem\u003eEnergy Science \u0026amp; Engineering\u003c/em\u003e, vol. 12, no. 3, pp. 835-859, 2024.\u003c/li\u003e\n\u003cli\u003eSahoo BM, Amgoth T, and Pandey HM 2020 Particle swarm optimization based energy efficient clustering and sink mobility in heterogeneous wireless sensor networks. Ad Hoc Netw. 106: 102237.\u003c/li\u003e\n\u003cli\u003eHammadi YI. Fiber Bragg grating-based monitoring system for fiber to the home (FTTH) passive optical network. Journal of Optical Communications. 2022 Oct 26;43(4):573-83.\u003c/li\u003e\n\u003cli\u003eTinatin M, Guibin Q, and Jinghui Z 2012 Evolution of wireless communication network and its impact on human life. J. Comput. Sci. Technol. 27(5): 857-868.\u003c/li\u003e\n\u003cli\u003eDeheyab AO, Alwan MH, Rezzaqe IK, Mahmood OA, Hammadi YI, Kareem AN, Ibrahim M. An overview of challenges in medical image processing. InProceedings of the 6th International Conference on Future Networks \u0026amp; Distributed Systems 2022 Dec 15 (pp. 511-516).\u003c/li\u003e\n\u003cli\u003eWang T, Li X, Liu Y, and Zhang Y 2017 Fog-based data collection for wireless sensor networks with mobile sinks. IEEE Trans. Ind. Informatics 13(2): 850-859.\u003c/li\u003e\n\u003cli\u003eHussain, Abdul Hussain Ali, Montadar Abas Taher, Omar Abdulkareem Mahmood, Yousif I. Hammadi, Reem Alkanhel, Ammar Muthanna, and Andrey Koucheryavy. \u0026quot;Urban traffic flow estimation system based on gated recurrent unit deep learning methodology for Internet of Vehicles.\u0026quot; \u003cem\u003eIEEE Access\u003c/em\u003e 11 (2023): 58516-58531.\u003c/li\u003e\n\u003cli\u003eYang Z, Chen M, Saad W, Xu W, Shikh-Bahaei M, Poor HV, and Cui S 2021 Energy-efficient wireless communications with distributed reconfigurable intelligent surfaces. IEEE Trans. Wireless Commun. 21(1): 665-679. \u003c/li\u003e\n\u003cli\u003eMahmood, Omar Abdulkareem, Abdukodir Khakimov, Ammar Muthanna, and Alexander Paramonov. \u0026quot;Effect of heterogeneous traffic on quality of service in 5G network.\u0026quot; In \u003cem\u003eDistributed Computer and Communication Networks: 22nd International Conference, DCCN 2019, Moscow, Russia, September 23\u0026ndash;27, 2019, Revised Selected Papers 22\u003c/em\u003e, pp. 469-478. Springer International Publishing, 2019.\u003c/li\u003e\n\u003cli\u003eZhuo X, Liu M, Wei Y, Yu G, Qu F, and Sun R 2020 AUV-aided energy-efficient data collection in underwater acoustic sensor networks. IEEE Internet Things J. 7(10): 10010-10022.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"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":"Design and Optimization, For Industrial Automation, cognitive industrial internet of things (CIIOT), electric-field measurement system (EFMS), Radio-Access Network-As-A-Service (RANAAS), Multi-InputMulti-Output (MIMO)","lastPublishedDoi":"10.21203/rs.3.rs-5731209/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5731209/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn order to improve the overall performance of edge-integrated edge IoT networks, this research introduces a combined technique based on profound learning for booking assets. If an IoT network wants to finish a task quickly and effectively, it has to get the greatest resources from the edge layer. Thorough asset booking is crucial to the identification and transfer of optimal assets. The integration of edge networks with IoT applications and the reduction of data transmission latency were previously addressed using profound learning algorithms. If we want to make an Internet of Things application more feasible and provide better service overall, we should think about other metrics like reaction time, waiting time, and bandwidth needs. Combining a convolutional neural network with a gated repeating unit in a certain manner achieves this enhanced performance. The suggested asset booking model considers the features and requirements of the assets in order to select the most suitable ones from the pool and allocate them to the IoT networks. Here, we give a comprehensive analysis of the method-data combination.\u003c/p\u003e","manuscriptTitle":"Design and optimization of energy-efficient wireless sensor networks for industrial automation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-01-13 09:42:18","doi":"10.21203/rs.3.rs-5731209/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":"d2eeb18a-d802-4bb1-9020-e1a46079c238","owner":[],"postedDate":"January 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-09-30T15:54:24+00:00","versionOfRecord":{"articleIdentity":"rs-5731209","link":"https://doi.org/10.54216/JISIoT.180212","journal":{"identity":"journal-of-intelligent-systems-and-internet-of-things","isVorOnly":true,"title":"Journal of Intelligent Systems and Internet of Things"},"publishedOn":"2025-07-12 00:00:00","publishedOnDateReadable":"July 12th, 2025"},"versionCreatedAt":"2025-01-13 09:42:18","video":"","vorDoi":"10.54216/JISIoT.180212","vorDoiUrl":"https://doi.org/10.54216/JISIoT.180212","workflowStages":[]},"version":"v1","identity":"rs-5731209","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5731209","identity":"rs-5731209","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}