融合风险势场的离散优化局部路径规划方法研究

Research on discrete optimization local path planning algorithm integrating risk potential field

针对结构化道路环境下智能汽车在紧急避障过程中的局部路径规划问题,提出一种融合风险势场的离散优化局部路径规划方法。采用离散优化的路径规划框架,基于路径采样与局部Frenet坐标系生成一组有限的候选路径;综合考虑路径的安全性、偏移量和连续性,构建代价函数对候选路径进行评价,根据评价结果选出最优路径,引导车辆完成对障碍物的规避。针对路径的安全性,基于势场理论分别建立道路和障碍物的风险势场,结合离散高斯方法设计路径安全代价函数。构建多种复杂交通场景并进行仿真实验验证,结果表明:所提出的局部路径规划方法能够实时为车辆规划出一条安全、可行的路径,有效实现对静止和移动障碍物的规避。

To deal with the problem for intelligent vehicles driving with avoidance of obstacles under structured road environments, a local path planning algorithm integrating risk potential field and discrete optimization is proposed.Firstly, adopting the discrete optimization path planning framework, a limited set of candidate paths is generated based on the path sampling and the local Frenet coordinate system.Secondly, considering vehicle safety, offset and path consecutiveness, the candidate paths are evaluated by the cost function.Finally, according to the evaluation results, the optimal path is obtained to guide the vehicle to avoid the obstacles.For the safety of paths, the risk potential field of road and obstacle are established based on potential field theory, where the safety cost function is established combined with the Gaussian convolution.A variety of complex traffic scenarios are constructed to operation simulation experiments, the results show that the proposed algorithm can generate a safe and feasible path in real time, and the vehicle can effectively avoid static and moving obstacles.