智能假腿的CMAC控制与实例仿真

CMAC Control and Case Simulation for Intelligent Prosthetic Leg

针对智能假腿系统模型的非线性与参数的不确定性等系统特性,提出了一种基于小脑模型神经网络控制器(CMAC)的假腿实时智能控制方法。该方法首先根据一种自制的假肢膝关节自适应结构,建立了智能假腿摆动相动力学数学模型,以描述智能假腿膝关节阻尼器控制参数与摆动运动参数之间的直接耦合关系。以此动力学模型为控制对象,设计了一种基于PD-CMAC的假腿系统智能控制器,并进行了实例仿真。仿真结果表明,假腿膝关节可以很快(约在0.5s时间内)跟踪好目标曲线,具有良好的实时性与精度;此外,膝关节阻尼器针阀开口位置与相应的假腿膝关节的角速度变化具有明显的负相关性;可以通过对假腿阻尼器针阀开口位置的调节,达到假腿跟踪健康腿摆动步态的目的。

An intelligent controller based on cerebellar model articulation controller （CMAC） is proposed for the intelligent prosthetic leg（IPL） system feathured with model nonlinearity and parameter uncertainty.The mathematical dynamics model for the swing phase of IPL system,which describes the relationship between the damper controlling parameters and the kinemics of IPL knee,is established for a self-made IPL with adaptive control structure.An intelligent PD-CMAC controller for the IPL system is designed with the controlling objective of the established dynamics model.The case simulation shows that an arbitrary trajectory such as a desired walking pattern can be tracked in less than 0.5 s,which proves that the contrller has high real-time and precision..In addition,the negative correlation exists between the opening of needle valve in the hydraulic knee damper and the rotation speed of the prosthetic knee.By adjusting the opening of needle valve in the damper,the gait of sound leg can be follwed by the prosthetic leg.