Kalman prediction based VFH of dynamic obstacle avoidance for intelligent vehicles

Abstract

a mixed obstacle avoidance algorithm for intelligent vehicles to avoid dynamic obstacles is presented in uncertain environments. Traditional Vector Field Histogram method is combined with kalman prediction algorithm in this algorithm. The kalman predictor forecasts the optimal position estimation of dynamic obstacles at the next moment, then the intelligent vehicle calls VFH algorithm to avoid obstacles according the current position and the next position predicted by the predictor. This method solves the problem that intelligent vehicle can not choose optimal path for the reason that the intelligent vehicle has no priori knowledge about the local environment. It is more suitable for dynamic obstacles avoidance of the intelligent vehicle. The simulation results show that the method has a good real-time performance, and the intelligent vehicle can avoid the dynamic obstacles accurately and reach the target point.