一种最优异步自学习控制方法

On the Asynchronous Optimal Learning Control Method

本文在文[1]的基础上,通过引入L_2(-∞,∞)空间范数指标,而将PID型的异步自学习控制发展为具有加权矩阵的最优异步自学习控制,文中给出了相应学习控制规律收敛性的严格证明,提供了计算加权矩阵的BFGS校正公式。本文最后将其应用于由于摩擦非线性引起的三轴运动模拟台正弦振动失真的补偿中。仿真结果表明,利用此法可使相应正弦振动精度得到进一步提高。

Using the norm criterion in L_2(-∞, ∞) space, the asynchronous optimal learning control method with a weighting matrix, on the basis of an asynchronous PID-type learning control presented by ref. [1], is proposed. In this paper the proof on the convergence of learning control law is strictly given, and the BFGS formula for calculating weighting matrix is obtained. Finally, it is applied to correct the distortion error of the sinusoidal vibration of Three-Axis Motion Simulator (TAMS). The simulation result shows that by means of this method the sinusoidal accuracy is improved greatly.