Enhancing Unmanned Rotorcraft Guidance With Lidar And Ads-B Integrated PgfLOW Algorithm

Abstract

This paper details the development of a Detect and Avoid (DAA) system for UAVs using potential flow theory to avoid static and dynamic obstacles. The research features the integration of an existing Potential Guidance Flow algorithm into a ROS2-based DAA system with SLAM-based environment mapping and PX4 Autopilot. Utilizing LiDAR and ADS-B data, the system detects collision risks and generates avoidance commands, applicable to various UAV platforms and supporting real-world experiments. The system’s effectiveness is validated through Software-in-the-Loop simulations in Gazebo, which realistically models sensor noise, measurement uncertainties, and vehicle dynamics. The evaluation includes diverse scenarios, from navigation around convex obstacles to navigation in complex urban environments. Experiments test system responsiveness, obstacle detection accuracy, and maneuver reliability at different speeds. Results show that the PGFlow algorithm, combined with SLAM-based mapping, allows multicopter UAVs to effectively avoid obstacles in unknown environments. The mapping function accurately represents static obstacles and identifies collision risks in time to initiate avoidance maneuvers. The ROS2 framework efficiently manages computational demands, ensuring low latency in command execution.