Three Dimensional Path Planning Based on Modified Potential Model for UAV Collision Avoidance

摘要

As the missions of Unmanned Aerial Vehicle (UAV) increase in the various areas, UAV should accomplishgiven tasks under the wide variety of environments. Considering the situations of penetrating into the enemy territory, UAV has to make a detour around radars and ground-to-air missile sites to accomplish the mission. In this study, potential field method based on the information of terrains and obstacles is proposed to avoid dynamic obstacles and to follow the terrain. Improved potential field method defines a new attractive potential function with regard to goal, altitude, relative height from the terrain and a new repulsive potential function with regard to relative position, velocity between the UAV and the obstacles, and relative height deal with the terrain. In three dimensional environments, the classical potential function tends to have many local minima. In order to address local minima problem, direction of attractive potential field for the goal has suggested with the direction of the gradient of the terrain toward the goal at the current position of the UAV. Numerical simulation demonstrates the effectiveness of the improved potential field method.