← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9191 · 生成模型:GPT-6

RRT循环先提案再验可连接性

理解等级:中级

先按目标偏置或空间采样生成q_rand,找最近树节点,按步长与运动约束得到q_new,生成局部路径并验证后才入树。书中伪码返回的是树,未找到目标时不能把返回值当成功路径;应另外检查目标连接和恢复结果。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Let 𝒢⊂𝒳 be a d−dimensional sphere in the state space that is considered to be the goal, max_dist the longest permissible edge length, t the allowed time, k the maximum number of vertices to allow in the tree, and goal_bias the percentage of the time the algorithm should try to connect to a goal state. An RRT planner would follow this pseudo-code:

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第13章《路径规划》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_13.xhtml) - 结论依据:摘录中关于RRT循环先提案再验可连接性的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Let 𝒢⊂𝒳 be a d−dimensional sphere in the state space that is considered to be the goal, max_dist the longest permissible edge length, t the allowed time, k the maximum number of vertices to allow in the tree, and goal_bias the percentage of the time the algorithm should try to connect to a goal state. An RRT planner would follow this pseudo-code: