一种肌电假手的自适应模糊控制方法  

Adaptive fuzzy control method for EMG prosthetic hand

作　　者：吴常铖[1] 宋爱国[1] 章华涛[1] 

机构地区：[1]东南大学仪器科学与工程学院,南京210096

出　　处：《仪器仪表学报》2013年第6期1339-1345,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(61272379);江苏省自然科学基金重点项目(BK2010063);江苏省产学研前瞻性项目(BY2012201)资助

摘　　要：针对肌电假手的力控制问题,提出了一种基于肌电信号自适应学习的动作识别方法,同时结合模糊神经网络PID控制算法实现肌电信号对假手的控制。设计的肌电信号自适应学习动作识别方法在时域内进行,减小运算复杂度和计算量的同时保证了动作识别精度。采用模糊神经网络PID算法设计了假手握力控制器,在没有位置传感器的情况下保证了假手握力的控制精度。进行了肌电信号动作识别跟踪实验、假手握力跟踪实验、肌电信号控制假手抓取实验,实验结果证明了肌电信号自适应学习动作识别方法和模糊神经网络PID握力控制方法在肌电假手控制中的有效性。Aiming at the force control issue of EMG prosthetic hand, an action identification method based on adaptive learning of EMG signal is proposed and combined with fuzzy neural network PID control algorithm, which implements the prosthetic hand control with EMG signal. The designed action identification method based on adaptive learning of EMG signal is performed in time domain, which ensures the action identification accuracy, while reducing the compu- tational complexity and computational task. A prosthetic hand grip force controller is designed with fuzzy neural net- work PID algorithm, which ensures the grip force control accuracy without using a position sensor. The EMG signal action identification and tracking experiment, prosthetic hand grip force tracking experiment and EMG signal con- trolled prosthetic hand grip experiment were carried out. The experiment results prove the effectiveness of the de- signed EMG signal adaptive learning action identification method in EMG prosthetic hand control. method and fuzzy neural network PID grip force control

关 键 词：肌电假手 自适应学习 模糊神经网络PID 

分 类 号：TP911.7[自动化与计算机技术] TP241