智能空间中的服务机器人路径规划  

Path planning for service robots in an intelligent space

作　　者：薛英花[1,2] 田国会[2] 吴皓[1] 吉艳青[1] 

机构地区：[1]山东大学控制科学与工程学院,山东济南250061 [2]山东财政学院计算机信息工程学院,山东济南250014

出　　处：《智能系统学报》2010年第3期260-265,共6页CAAI Transactions on Intelligent Systems

基　　金：国家"863"计划重点资助项目(2006AA040206);国家"863"计划资助项目(2009AA04Z220)

摘　　要：为了加深服务机器人对环境的理解,实现安全高效的智能空间导航,建立了一种信息更为丰富的环境模型——危险度地图;并针对智能空间环境部分未知的特点,设计了分层的路径规划方法.静态规划层根据已知环境信息,采用改进的粒子群优化算法规划初始最优路径,动态规划层利用基于动态危险度地图的改进A*算法进行避障.该方法克服了常规算法只追求路径最短的缺点,增加了对路径危险度的评价,规划出的路径既安全又较短;且该方法实现简单,实时性好.仿真结果验证了该方案的可行性.A service robot must be capable of deeply understand its environment,so that it may safely navigate in-telligent spaces with high performance.In order for this to be possible,a new environmental model,or danger de-gree map （ DDM）,was created to provide enriched environmental information for service robots.As the intelligent space is partly unknown,the layered path planning method was used.A modified particle swarm optimization （ PSO） algorithm based on a known environment was introduced to get a static optimized path.The dynamic layer used a modified A algorithm for avoidance of dynamic obstacles on the basis of the dynamic DDM.The proposed method adds evaluation of the degree of danger in a path to overcome the disadvantages of conventional methods,which follow the shortest path and ignore safety.The new method is simple and meets the real-time requirements of robot navigation.The resulting dynamic path is not only safe enough but also comparatively short.Simulation re-sults demonstrated the feasibility of the method.

关 键 词：智能空间 服务机器人 路径规划 危险度地图 粒子群优化算法 A＊算法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]