Design of an Improved Autonomous Tracked Vehicle Based on SLAM Algorithm

Abstract

In the design of autonomous vehicles, relying solely on the SLAM algorithm for obstacle avoidance poses a certain response delay issue, which may affect the real-time safe driving of unmanned vehicles. To address this issue, a solution that identifies simulated obstacles by running YOLOv5 on the Jetson module is proposed to achieve optimized obstacle avoidance processing. At the same time, a tracked vehicle was designed using the ROS robot development system to achieve the specific effects of this solution. In the experiment, when the tracked vehicle encounters simulated obstacles during driving, it can automatically turn and deviate from the original driving trajectory, proving that this visual obstacle avoidance scheme is effective.