假肢控制命令识别的优化研究

摘要

研究假肢的优化控制问题.针对假肢控制动作是通过对脑部发出的控制命令信号中,单一动作特征进行识别,然后控制假肢驱动完成动作执行.如果脑部发出的控制命令信号特征不是单一的,而是包含多个特征的控制命令,会使得提取的命令特征混乱,造成假肢指令识别正确率不高的问题.为了解决上述问题,提出信号归属分类的假肢控制命令信号识别方法,通过采集多特征的控制命令信号,在识别中加入贝叶斯决策理论,运用决策思维判断多个控制命令信号的具体指令归属,完成多特征指令的准确执行.对系统进行仿真,结果表明,多特征假肢控制命令信号归属识别的正确率有了明显提高.%Study prosthetic optimal control problem. Artificial limbs movement control is based on the recognition of single movement characteristics from the brain command signals and then drive the prosthetic to complete the movement. If the brain command signals not only contain a single feature, the confusion of features extracted may occurs, causing low recognition rate of artificial limbs command characteristics. In order to solve this problem, based on the recognition method of artificial limbs control command signal classification and through the acquisition of the various characteristics of control command signals, the recognition combined with Bayesian decision theory, used the decision thinking to judg the classification of the control command signals, and exactly execute the instruction which contains various characteristics. Simulation results show that the recognition of the multi - characteristics of control command signal achieves high accuracy.