基于城市车道撒点策略的RT-RRT^*智能车局部路径规划

Intelligent vehicle local path planning based on RT-RRT^* with the strategy of scattering points in urban lanes

针对实时RRT^*算法(real-time RRT^*,RT-RRT^*)探索空间时随机性强、时效性不稳定以及不满足交通规则等问题,提出了一种基于城市车道撒点策略改进的RT-RRT^*算法(scatter points RT-RRT^*,SP-RT-RRT^*)。该算法采用规则的撒点采样思路,取代了原算法的随机位置采样,保证了算法在动态避障过程中探索城市道路空间时的高效、稳定、可靠。同时算法对采样点的连接进行了限制,保证了车辆运动时不会横向行驶、压线行驶。之后,算法针对代理车辆和障碍车辆之间不同的相对速度预留了相应的安全距离,从而在动态障碍物环境下生成安全、无碰的车辆行驶轨迹。将该算法与RT-RRT^*算法在仿真实验中进行比较,并在仿真中使用三次B样条曲线拟合法对两种算法规划结果进行拟合比较,结果显示该算法在搜索稳定性、实时性方面有较大提升,且在满足交通规则的层面上有较高优越性。

Aiming at the problems such as strong randomness,unstable timeliness and violation of traffic rules of realtime RRT^*(RT-RRT^*)algorithm in exploring space,an improved RT-RRT*algorithm based on the strategy of scattering points in urban lanes is proposed,namely(scatter points RT-RRT^*,SP-RT-RRT^*).The algorithm adopts the regular scatter sampling idea,which replaces the random position sampling of the original algorithm,and ensures the efficiency,stability and reliability of the algorithm in exploring urban road space in the process of dynamic obstacle avoidance.At the same time,the algorithm restricts the connection of sampling points to ensure that the vehicle will not drive laterally or drive on the lane line when moving.Then,the algorithm reserved the corresponding safety distance for the different relative speeds between the agent vehicle and the obstacle vehicle,so as to generate a safe and non-collision vehicle trajectory in the dynamic obstacle environment.Compare SP-RT-RRT^*with RT-RRT^*in the simulation experiment,and use cubic b-spline curve fitting method in the simulation to fit and compare the programming results of the two algorithms,the results show that the algorithm has a great improvement in search stability and real-time performance,Moreover,it has higher superiority in satisfying traffic rules.