基于CMAC的电动负载模拟器的研究

Research on Electric Load Simulator Based on CMAC

为了抑制电动负载模拟器中存在的多余力矩,增强加载系统的非线性抑制能力,提出小脑模型(CMAC)神经网络与经典PID结合的复合控制策略,CMAC网络用于前馈控制,经典PID实现反馈控制,能够快速实时地消除多余力矩对系统的干扰。将存储单元的先前学习次数作为可信度,提出基于可信度分配的权值更新算法,来消除常规CMAC网络的过学习现象。建立了电动负载模拟器的数学模型,并给出了具体的算法流程。经仿真和试验结果表明,该方法具有很强的鲁棒性,有效地抑制了系统的多余力矩,提高了系统的加载精度。

In order to prevent the surplus torque in the electric load simulator（ELS） and improve the ability of prevent the nonlinear, hybrid control based on cerebella model articulation controller（CMAC） neural network and classical PID is proposed, the feedforward control is realized by CMAC while the feedback control is realized by classical PID controller, it is able to eliminate surplus torque. Weight update algorithm based on credibility is proposed, the former learned times of the storage unit is used as credibility in order to eliminate the excess self learning phenomenon. The ELS’s model is established and the detailed control structure is put forward. simulation and experimental results demonstrate that the method has good robustness and can effectively eliminate the surplus torqure, improve the loading precision.