Obstacle Detection and Avoidance System for a UAV using LIDAR

This presentation talks about the research on autonomous collision avoidance system for a octocopter unmanned aerial vehicle (UAV) using LIDAR. Lack of obstacle avoidance capabilities have limited the use of UAVs for many applications such as package delivery, traffic monitoring, search & rescue, power line inspection, and environmental gas monitoring. An S1000 octocopter is equipped with a 16 channel LIDAR, Pixhawk flight controller, and an Intel NUC onboard computer, which is used to process the LIDAR data and implement the obstacle avoidance algorithm. Data collected by the LIDAR is transferred to a planar grid system for analysis and obstacle detection. Vector Field Histogram Plus (VFH+) method is used for obstacle avoidance. A successful implementation of the algorithm yields a horizontal, bidirectional (left or right) avoidance maneuver by the vehicle. Obstacle avoidance capability helps safely integrate UAVs in the National Airspace System for many applications. Abilities to avoid both static and dynamic obstacles are important for UAVs to possess a human equivalent level of safety and to satisfy federal and local safety regulations. Improvements to the VFH+ method include using a multiplanar grid system to increase its accuracy and effectiveness for the avoidance of static and dynamic obstacles. The presentation will show and discuss simulation and flight test results.