论文部分内容阅读
RRT*(快速搜索随机树)算法在以往研究中存在收敛速度慢、结果不稳定的缺点。针对此问题,文章在现有RRT*基础之上提出一种新型改进算法。该改进算法结合环境约束、车辆自身约束和运动学约束,舍弃原算法贪心思想并引入启发式采样节点插入算法,提高路径规划的速度和质量;接着对改进算法进行理论分析,证明算法具有概率完整性、渐近最优性,从理论上保证算法能快速收敛到最优路径。通过各种仿真环境的测试,验证改进算法的有效性、稳定性和正确性,也验证理论分析的正确性。
The RRT * (fast search random tree) algorithm has the disadvantage of slow convergence and unstable results in previous studies. In response to this problem, the article proposes a new improved algorithm based on the existing RRT *. The improved algorithm combined with environmental constraints, vehicle constraints and kinematic constraints, abandoned the greedy thought of the original algorithm and introduced heuristic sampling node insertion algorithm to improve the speed and quality of the path planning; then theoretical analysis of the improved algorithm to prove that the algorithm has the probability of complete Asymptotically optimality, theoretically ensure that the algorithm converges to the optimal path quickly. Through the testing of various simulation environments, the validity, stability and correctness of the improved algorithm are verified, and the correctness of the theoretical analysis is verified.