Search for a command to run...
With the rapid development of the autonomous driving industry, obstacle detection and avoidance technologies for unmanned vehicles have made significant progress. This paper provides an overview of the current research status of obstacle detection and avoidance systems for unmanned vehicles, analyzes the advantages and limitations of single-sensor systems and multi-sensor systems, and proposes a 3D obstacle avoidance system for unmanned vehicles based on multi-sensor fusion. First, the overall design of the 3D obstacle avoidance system is presented. The system integrates three types of sensors: ultrasonic sensors, cameras, and LiDAR. Ultrasonic sensors are used to establish a low-position 3D obstacle avoidance space, while cameras and LiDAR are used to construct a high-position 3D obstacle avoidance space. By incorporating YOLOv5 object detection and intelligent control, the system enables intelligent vehicle control and human target safety alerts. Next, the 3D obstacle avoidance system was deployed on an unmanned vehicle. Ultrasonic sensor arrays were installed around the vehicle body, while a 360° panoramic camera and LiDAR were mounted centrally on the top rear of the vehicle, forming a 3D obstacle avoidance space with a radius of 3 meters and a height of 40 meters. Finally, the modified unmanned vehicle underwent testing for the 3D obstacle avoidance system. The test results showed that the system could achieve 100% automatic braking when obstacles were within 1 meter of the vehicle. For obstacles in the range of 1 to 3 meters, the system successfully cooperated with the vehicle control system to achieve speed reduction.