Improved Artificial Potential Field for Unknown Narrow Environments

摘要

Path planning in unknown environments is very different from that in known environments, because the robot doesn’t know the whole map. So it can only navigate according to the local information get by sensors, and explore the environment step by step. When the robot searches an unknown environment, it need plan its path in order to search all the space rapidly and completely. So it must avoid the obstacles with intentions. We adopt APF (artificial potential field) to achieve it. But APF has many shortcomings, especially in unknown environments. In this paper, we develop a geometry method to judge the situation of obstacles. We also develop a strategy that can make the robot plan the path as human’s behavior. And combining the geometry method with APF, the path is optimized and suitable to applications. We also develop some new ways to overcome other disadvantages of APF.