【计算机科学】【2009】自主式机器人路径规划与避障算法

本文为美国北卡大学夏洛特分校（作者：Sharayu Yogesh Ghangrekar）的硕士论文，共103页。

机器人路径规划涉及到机器人导航设计的具体逻辑。考虑到路径规划对机器人功能的深刻影响，路径规划仍然有很长的路要走。各种路径规划技术都已经进行过研究与尝试，包括概率、积分和遗传方法。大多数这些算法的实现细节都是针对特定要求的。本文针对全地形车辆（All-Terrain Vehicle，ATV）无碰撞协调导航的自定义策略需求，开展了这项研究。作为本研究的一部分内容，已经开发并模拟了算法实现的可视化效果。本文所提出的算法是不断进化的，能够在完全已知和新发现障碍物存在条件下对ATV进行路径规划。该算法有助于ATV以特定模式在开放场景中机动，并规避沿路障碍物。作为研究的一部分，使用C和WINAPI实现了算法并进行仿真。因此，给定已知障碍物和现场环境的数据，ATV可以通过避开路径上的所有障碍物以系统的最优的方式朝目标机动。该算法还可以利用激光雷达和GPS的实时数据部署在ATV上，算法的逻辑可扩展到完全动态环境中的路径规划。

Path planning in robotics is concerned with developing the logic fornavigation of a robot. Path planning still has a long way to go considering itsdeep impact on any robot’s functionality. Various path planning techniques havebeen tried and tested earlier, including probabilistic, integral and geneticapproaches. The implementation details of most of these algorithms areproprietary to specific organizations. The requirement of a customized strategyfor collision free and concerted navigation of an All-Terrain Vehicle (ATV) ledto the activities of this research. As a part of this research an algorithm hasbeen developed and simulated to give a visual effect. The algorithm presentedis evolutionary and capable of path planning for ATVs in the presence ofcompletely known and newly-discovered obstacles. This algorithm helps the ATVto maneuver in an open field in a specific pattern and avoid the obstacles, ifany, along its path. As part of the research the actual algorithm isimplemented and simulated using C and WINAPI. As a result, given the data ofknown obstacles and the field, the ATV can maneuver in a systematic and optimummanner towards its goal by avoiding all the obstacles in its path. Thisalgorithm can also be deployed on an ATV using real time data from LIDAR and GPS.The logic of the algorithm can be extended for path planning in a completely dynamicenvironment.

1 引言
2 路径规划回顾
3 算法设计
4 逻辑与假设在路径规划中的作用
5 算法仿真
6 探索性研究
7 结论与未来工作展望

下载英文原文地址：

http://page5.dfpan.com/fs/8lcj8221129166905a8/

更多精彩文章请关注微信号：