基于改进CMAC模型的电动伺服加载算法研究

Research on Electric Loading Simulator Algorithms Based on Improved CMAC

在电动伺服加载系统中,多余力矩以及摩擦等非线性因素会严重降低系统的跟踪精度,传统的基于结构不变性原理的方法,在消除多余力矩和抑制非线性环节时表现欠佳。首先,建立了电动伺服加载系统的数学模型,分析了多余力矩的产生原理。其次,提出了一种改进的CMAC算法,采用基于Sigmoid函数的变学习率替代传统的固定学习率,解决了常规CMAC的收敛速度和稳态性能之间的矛盾。然后,采用改进的CMAC算法,设计了一种CMAC/PID的复合控制器,应用于电动伺服加载系统中。仿真结果表明,所设计的复合控制器,可以有效地抑制了系统的多余力矩,克服摩擦等非线性因素的干扰,改善加载系统的动态性能,提高了跟踪精度,增强了稳定性。

In the electric loading simulator,surplus torque and friction,as well as other nonlinear factors,can seriously reduce tracking accuracy of the system. The traditional approach based on structure invariance principle could not get good results at eliminating surplus torque and suppressing nonlinear aspects. Firstly,the mathematical model of electric loading simulator was established,the causes of surplus torque were analysed. Secondly,proposed an improved CMAC（ Cerebellar Model Articulation Controller） algorithm using variable learning rate based on Sigmoid function instead of the traditional fixed learning rate,which can solve the contradiction between convergence speed and steady state performance. Then,designed a compound controller of CMAC and PID with the improved CMAC algorithm,and applied it to the electric loading simulator. The results of simulation prove that the compound controller can not only effectively eliminate the surplus torque,suppress the interference of friction and other nonlinear factors,improve the dynamic performance of loading simulator,but also increase the tracking precision and enhance the stability.