Co-simulation-based evaluation of UAV obstacle avoidance using 3DVFH* and stereo cameras

The paper aims to simulate UAVs in complex network environments to evaluate their performance in obstacle avoidance using stereo sensors. Since there are strict security regulations limiting real-world UAV flight testing, this research addresses the critical need for comprehensive simulation-based approaches.. The system integrates real-time obstacle detection and avoidance algorithms, enabling UAVs to navigate safely in environments with obstacles. Developed on the ROS framework with PX4 SITL and Gazebo, the system supports comprehensive end-to-end testing, from stereo image acquisition to UAV navigation using the 3DVFH* algorithm. It utilizes the MAVLink protocol for control and QGroundControl for monitoring. Simulation results confirm the system’s effectiveness, establishing a solid foundation for advanced autonomous UAV navigation in real-world applications.