基于遗传算法和模糊神经网络的康复机器人上肢阻抗控制  

Upper Limb Impedance Control of Rehabilitation Robot Based on Genetic Algorithm and Fuzzy Neural Network

作　　者：穆海芳[1] 郭凯[2,3] 胡波 Mu Haifang;Guo Kai;Hu Bo(Suzhou University,Suzhou 234000,China;Anhui Provincial Engineering Laboratory on Information Fusion and Control of Intelligent Robot,Wuhu 241002,China)

机构地区：[1]宿州学院机械与电子工程学院,安徽宿州234000 [2]宿州学院,安徽宿州234000 [3]安徽省智能机器人信息融合与控制工程实验室,安徽芜湖241002

出　　处：《廊坊师范学院学报（自然科学版）》2021年第4期56-59,64,共5页Journal of Langfang Normal University(Natural Science Edition)

基　　金：宿州学院科研平台开放课题“康复机器人上肢运动控制策略研究”(2020ykf10);安徽省智能机器人信息融合与控制工程实验室开放课题(IFCIR2020005)。

摘　　要：康复机器人用于患者上肢的辅助康复训练时,需要根据患肢病情的变化及时调整控制策略,保证系统运动的稳定。在传统阻抗控制方法基础上,提出一种基于模糊神经网络的新阻抗控制方法,通过在线辨识得到机械阻抗参数,根据目标阻抗力模型建立模糊神经网络阻抗控制器,运用控制器调整目标阻抗参数。具体调整采用离线优化和在线微调相结合的方式。仿真结果表明,该方法控制的上肢康复机器人能更好地适应患者病情的变化,系统具有较好的稳定性和鲁棒性。When rehabilitation robot for patients with upper limb is applied in rehabilitation training,in order to ensure the stability of system movement,it must adjust control strategy according to the change of limb condition.Traditional impedance control method is an effective force control method,on this basis,this paper presents a new impedance control method based on fuzzy neural network.The method obtains mechanical impedance parameters by online identification,establishes fuzzy neural network impedance controller based on target impedance force model,adjusting target impedance parameters by controller.Specific adjustment is a combination of offline optimization and online fine tuning.Simulation results show that,compared with traditional impedance control methods,the upper limb rehabilitation robot controlled by this method can better adapt to the changes of the condition,the system has better stability and robustness.

关 键 词：康复机器人 阻抗控制 模糊神经网络 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]