一种无机械平衡调节器自行车机器人的惯性参数辨识  

Inertial parameters identification of a bicycle robot without mechanical regulators

作　　者：黄用华[1,2] 黄渭[1] 庄未[1] 黄美发[1] 钟永全[1] 

机构地区：[1]桂林电子科技大学机电工程学院,桂林541004 [2]北京邮电大学自动化学院,北京100876

出　　处：《东南大学学报（自然科学版）》2013年第A01期131-137,共7页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(61105103);广西制造系统与先进制造技术重点实验室基金资助项目(PF12133X;PF12131X);高等学校博士科研启动基金资助项目(UF12002Y)

摘　　要：针对一种前轮驱动自行车机器人样机的惯性参数难以直接获得、控制器参数整定效率低和平衡控制效果差的问题,基于欠驱动力学特性对系统进行惯性参数辨识,并对辨识结果进行平衡控制试验验证.采用拉格朗日方法分析系统的力学机理,结果发现,其中包含有欠驱动子系统,若将该子系统进行变量分离,可以推导出系统的惯性参数辨识模型;根据惯性单元等传感器的实测信息,引入遗传算法对模型进行求解,辨识出了系统的惯性参数.样机试验结果表明,根据辨识出的惯性参数可以快速整定出控制器参数并实现样机在水平面上的稳定平衡行走,验证了辨识方法的有效性.To solve the difficulties of obtaining the inertial parameters and the lower effectiveness of turning the controller parameters and the poor effect of keeping balance of a front-w heel drive bicycle robot,an approach to identify its inertial parameters w as proposed,and the identification results w ere validated by balancing experiments.The robot’s dynamics w as analyzed by Lagrange method.The dynamics show s the robot system consists of an under-actuated subsystem.Through variables separation for the under-actuated subsystem,an inertial parameters identification model w as developed.The model w as solved by the Genetic algorithm according to the measurement data from the inertial measurement unit and other sensors on the robot,and then the inertial parameters of the system are identified.The results of balanced motion experiments for the physical prototype show that,according to the identified inertial parameters,the controller parameters can be quickly tuned and the robot can stably run on the horizontal plane,indicating the effectiveness of the identification method.

关 键 词：自行车机器人 机械调节器 欠驱动 力学模型 惯性参数辨识 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]