基于多曲率拟合模型的无人车路径规划研究

摘要

针对无人驾驶车辆的局部路径规划问题,提出一种基于多次曲率拟合模型的路径规划算法.鉴于无人车辆需要生成无碰路径的特点,建立了由三次曲率多项式产生的候选路径集,采用四重参数循环法解决了三次曲率多项式参数计算的问题,进而能够根据评价函数从候选路径集中选择出当前环境下的最优路径.所提出的无人车局部路径规划算法经现场多个路口通过性实验证明了有效性.%To solve the problem of local path planning of unmanned vehicles,this paper proposes a path planning algorithm based on the multi-curvature fitting-model. We set up the candidate path set produced by the three times curvature polynomials after considering the actual characteristics of unmanned vehicles,and use the four times repetitive cycle parameters method to solve the problem of calculation of the three times curvature polynomials' parameters. We can choose the optimal path in the current environment from the candidate path set according to the evaluation function. With the path planning algorithm put forward by this paper,unmanned vehicles passed a few intersections successfully,which proved that this algorithm is effective.