一种改进A^(*)算法在智能车中的应用研究

Application of an Improved A^(*)Algorithm in Intelligent Vehicle

在智能车的自主导航中,要求规划模块在满足一定限制条件下,生成符合智能车运动特性的路径。而传统A^(*)算法存在着路径点不平滑,路径点紧挨障碍物和起始时刻路径不合理的问题。为了解决传统A^(*)算法所存在的问题,首先建立了车辆运动学模型并得到约束条件,同时将方向代价和自适应障碍物惩罚代价加入评价函数中;然后用车辆约束条件优化启发函数和路径优化模块;最后通过自由边界三次插值算法拟合转折点,使A^(*)算法规划的路径能够更好地被跟踪模块跟踪。通过实验分析可知:相比于传统A^(*)算法,改进A^(*)算法规划的路径更适用于实际车辆的运动控制。

It is required that the planning module needs to generate a path that conforms to the characteristics of the intelligent vehicle in the autonomous navigation.However,the traditional A^(*)algorithm has some problems:①the path point is not smooth;②the path is close to obstacles;③the path is unreasonable at the beginning.In order to solve the problems of the A^(*)algorithm,this paper establishes the vehicle kinematics model and obtains the constraint conditions.At the same time,these constraint conditions are added the direction cost and the adaptive obstacle penalty cost.Then,the heuristic function is optimized by the vehicle constraints.By combining with path optimization,the turning point is finally fitted by a free boundary cubic interpolation algorithm,so that the path planned by the improved A^(*)algorithm can be better tracked by the tracking module.Through experimental analysis,it can be seen that compared with the traditional A^(*)algorithm,the improved A^(*)algorithm is more suitable for the actual vehicle motion control.