摘要： 以实现月球探测器的避障路径规划为目标，采用立体视觉技术感知车前环境，设计一套基于特征匹配的车前障碍物检测算法。构造极坐标直方图描述环境空间，提出一种基于视觉感知的直方图Bug路径规划算法。通过朝目标移动和沿障碍物边界滑动2种行为的切换来保证算法全局收敛。实景重构与仿真实验验证了算法的有效性。

关键词: 月球探测器, 立体视觉, 路径规划, 直方图Bug

Abstract: To realize obstacle avoidance path planning for lunar rover, stereo vision technology is applied to precept front environments, and a feature-based obstacle detection method is designed. Histogram-Bug algorithm is presented to realize obstacle avoidance path planning. The world model is described using a polar coordinates histogram. Global convergence is ensured by motion-to-goal behavior and boundary following behavior working together. 3-D reconstruction technologies and simulations confirm the effectiveness of the method.

Key words: lunar rover, stereo vision, path planning, Histogram-Bug

中图分类号: 

• TP18