Design and Implementation of an Integrated Autonomous Navigation Framework

Autonomous navigation is the capability of a vehicle to plan and execute a trajectory without human intervention by combining perception, decision-making, and control subsystems. While lane/path following, obstacle avoidance, traffic sign recognition, and safety fallback mechanisms have been widely studied in isolation, their integration as a framework for an autonomous vehicle prototype remains underreported. This paper presents the design and implementation of such a unified framework that merges these four functions and is implemented in Python on ROS 2 within the Cognipilot autopilot stack (Airy release). The system was validated both in Gazebo simulation and on a Cognipilot-enabled NXP MR-B3RB robotic buggy equipped with an NXP NavQPlus, a 2D LiDAR, and a monocular camera.