RRT_star_smart
所属分类:人工智能/神经网络/深度学习
开发工具:matlab
文件大小:412KB
下载次数:2
上传日期:2022-02-21 13:12:02
上 传 者:
sh-1993
说明: RRT_star_smart,机器人路径规划
(RRT_star_smart,robot path planning)
文件列表:
code (0, 2022-02-21)
code\main.m (823, 2022-02-21)
code\rrt_start_smart.m (9041, 2022-02-21)
code\utils (0, 2022-02-21)
code\utils\out_of_range.m (249, 2022-02-21)
code\utils\vaild_path.m (536, 2022-02-21)
code\utils\will_collide.m (936, 2022-02-21)
code\world (0, 2022-02-21)
code\world\general_fixed_obstacles.m (352, 2022-02-21)
code\world\generate_obstacles.m (978, 2022-02-21)
code\world\plot_world.m (1704, 2022-02-21)
pictures (0, 2022-02-21)
pictures\rrt1.jpg (92844, 2022-02-21)
pictures\rrt2.jpg (98318, 2022-02-21)
pictures\rrt3.jpg (90872, 2022-02-21)
pictures\rrt4.jpg (82830, 2022-02-21)
pictures\rrt5.jpg (88344, 2022-02-21)
# RRT_star_smart
robot path planning
RRT*-Smart has made improvements on the basis of RRT*, mainly optimizing the path. The path explored through RRT and
RRT* is often tortuous and slightly wavy (after all, the nodes are randomly generated), but in fact, the best path in
the open space is generally a straight line. RRT*-Smart is completely consistent with RRT * in the previous stage of
operation, but after finding a feasible path from the starting point to the end point, it starts to consider optimizing
the path and turning the curve into straight. This process actually starts from the leaf node and constantly looks for
whether it can be directly connected to the predecessor node without obstacles. If you directly connect one layer forward,
there will be more straight lines and less curves. (in order to increase the calculation speed, we might as well
directly treat the obstacles as rectangular.) in the process of turning the curve, we can find several anchor points,
which are often near the obstacles, and they can't help our descendants optimize directly.
reference:
https://zhuanlan.zhihu.com/p/161829703
https://github.com/adnanmunawar/matlab-rrt-variants
近期下载者:
相关文件:
收藏者: