Relative Position-based Collision Avoidance System for Swarming UAV Applications

Abstract

This paper presents the development of quadrotor unmanned aerial vehicle (UAV) for close proximity swarming applications that is capable of quad-directional collision avoidance with obstacles through the implementation of relative position-based cascaded PID position and velocity controllers. A collision avoidance algorithm that decides evasive maneuvers in two dimensional flight by the means of net error calculation is developed. Sensor fusion of ultrasonic (US) and infrared (IR) sensors is performed to obtain a reliable relative position data of obstacles which then fed into collision avoidance controller (CAC) for generating necessary response in terms of attitude commands. Flight tests performed in both indoor and outdoor environments proved the capability of UAV to avoid collisions with the obstacles and dummy non-flying UAVs that existed at a closer distance in its four primary directions of detections during flight successfully.