基于遗传算法的欠驱动机器人运动规划  

Motion Planning of Underactuated Robots Using GA Method

作　　者：刘庆波[1] 余跃庆[1] 王浩[1] 

机构地区：[1]北京工业大学机电学院,北京100022

出　　处：《控制工程》2008年第1期64-67,共4页Control Engineering of China

基　　金：国家自然科学基金资助项目(50575002;60705036);北京市基金资助项目(3062004);北京市教委人才强教计划拔尖人才基金资助项目(PHR(IHLB));北京市教委科技发展计划基金资助项目(KM200610005003)

摘　　要：研究了平面欠驱动机器人的非完整运动规划问题,建立欠驱动机器人系统动力学模型,提出了一种利用遗传算法(GA)解决欠驱动机器人运动规划的方法。引入部分稳定控制器的思想,提出基于能量最优的评价函数,利用遗传算法对评价函数进行离线优化,得到有关部分稳定控制器的切换规则。此方法可以推广到任意平面欠驱动机器人的规划问题上,以平面3R欠驱动机器人为研究对象进行了数值仿真,仿真结果验证了该方法的有效性。The nonbolonomic motion planning problem of underactuated planar robots is discussed. The dynamic model of the underactuated robot system is built. A method for motion planning of underactuated robots based on the genetic algorithm is proposed, The thoughts of partly stable controllers is introduced, and the energy based on cost function is optimized by GA method off-line, The switching rule of partly stable controllers are carried out. This method can be extended to the arbitrary motion planning problem of underactuated planar robots. The numerical simulation on the underactuated 3R planar robot shows the validity of the proposed method.

关 键 词：运动规划 欠驱动 非完整系统 遗传算法 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]