摘要： 为导览机器人实现服务目标的引导规划，提出一种以机器人操作系统为实验平台，结合马尔科夫决策过程和微重启技术的任务规划方法。该方法在全面考虑服务对象身份需求信息及服务过程的总代价后，利用马尔科夫决策模型确立最优执行方案。采用基于机器人分布式操作系统建立的微重启自我修复机制解决功能失效问题。仿真结果验证了该规划方案在执行导览任务中的有效性，同时表明微重启技术在处理功能失效问题时相对于传统处理方法具有优越性，其在随机附加障碍的情况下可获得91.03%的规划成功率。
关键词:
导览机器人,
操作系统,
马尔科夫决策过程,
任务规划,
微重启,
马尔科夫决策模型
Abstract: Towards the planning problem of robot for guiding service objects, Robot Operating System(ROS) as experimental platform, combined with Markov Decision Process(MDP) and micro-reboot technology, this paper presents a robot task planning scheme suitable for guide service. After considering comprehensive service object identity information and total cost of service process, the scheme establishes optimal execution plan using MDP model. And based on the ROS distributed system, the scheme uses micro-reboot self-repairing mechanism to solve the functional failure problem. The simulation results show that the proposal is effective in the implementation of navigation mission. Compared with the traditional processing method, it shows the advantages of micro-reboot technology in dealing with functional failures. The planning scheme gets 91.03% success rate in the case of additional barriers randomly.
Key words:
tour guide robot,
operating system,
Markov Decision Process(MDP),
task planning,
micro-reboot,
Markov decision model
中图分类号:
李月光，尹东，张荣. 一种导览机器人的任务规划方法研究[J]. 计算机工程.
LI Yue-guang, YIN Dong, ZHANG Rong. Study on a Task Planning Method for Tour Guide Robot[J]. Computer Engineering.