Control-oriented fast numerical approaches of fractional-order models
面向控制的分数阶微分模型的快速数值计算

In the computation of fractional order derivatives, the crucial point is to balance the computation speed and the computation accuracy. The existing short memory principle or variable memory principle helps little in relaxing the contradiction. To deal with this problem, we proposed an equal-weight memory principle, in which an equal-weight is applied to all past data in history, and the result is reserved instead of being discarded. In each subsequent sampling period, only one new data is collected for consideration with the historical data. Therefore, the computation accuracy is improved and the computation complexity is reduced, thus, the contradiction is effectively relaxed. Results in numerical examples demonstrate the feasibility and superiority in applying the proposed principle to the design of fractional-order control systems.