A novel obstacle avoidance strategy of nonholonomic mobile robot based on virtual simulation platform

Abstract

The obstacle avoidance is the key technology of nonholonomic mobile robot, where obstacle avoidance algorithm is the core. Artificial potential field (APF) algorithm has the advantage of sample mathematical model, which is understood and applied into practice easily. However, there are some problems in APF algorithm, for examples, local minimum and GNRON problem, which limits the effect of APF algorithm. In order to solve these problems, we propose a modified APF algorithm in this paper. This modified algorithm utilizes the improved potential field formulas for GNRON problem. Otherwise, this algorithm uses a method of boundary detection and setting a secondary target for optimal path and local minimum. Finally, we use MRPT virtual simulation platform to verify the modified algorithm. The simulation result and field test show that the modified is feasible.