Path Planning of an Autonomous Quadcopter Based Delivery System

This paper presents a novel offline path planning algorithm for the quadcopter based delivery system. For any quadcopter based system, time of execution needs to be minimized as it affects the system performance. This proposed algorithm is a combination of sampling based Bidirectional RRT algorithm and improved Artificial potential field (APF) algorithm, so that the former is made more goal oriented. Algorithm rely on the knowledge of the environment from sensors, to plan a path efficiently with reduced execution time. Improved APF limits the search space to a smaller section of the environment in turn reducing the execution time. Validation of the algorithm is done using ROS TUM simulator AR Drone in presence of obstacles. A proportional position and altitude controller are used to track the path which is planned using proposed algorithm.