Development of A* Real-time Path Planning Algorithm of UAV Considering Collision Avoidance

Abstract

Unmanned aerial vehicle (UAV) has developed to perform military missions including patrol, surveillance, and reconnaissance. Recently, the applications of UAV are expanding to the private markets of UAV agriculture, aerial photography and etc. For these various applications, collision avoidance of UAV is an essential problem, because collision to obstacles may cause not only mission failure but also destruction of UAV and fatal accident including loss of human life. In this study, modified A* path planning algorithm is proposed for UAV systems. In various robot applications, standard A* algorithm is widely used, which does not apply dynamics of UAV. Therefore, the results from the path planning algorithm should be post-processed to reflect the dynamic constraint of an UAV such as limited turning angle. To deal with this problem, search direction achievable by the UAV is considered in the proposed A* algorithm. The proposed algorithm can be implemented on the on-board system of the UAV in real-time, and does not need post processing to follow the result of path planning. Nonlinear guidance algorithm and PID controller to follow the path are also designed. The performance of the proposed algorithm is demonstrated using numerical simulations. Six degree-of-freedom simulation model was obtained by system identification flight test. Finally, the integrated algorithm is verified by the flight test.