From theoretical differentiation methods to low-cost digital implementation

In this paper we review different differentiation methods and we investigate their digital implementation. The motivation is the design of a numerical differentiator based on a low-cost microcontroller device. Standard digital differentiation algorithms are presented and an alternative approach based on weighted H_∞ filtering is introduced. In this approach, the differentiator design is based on signal and noise spectral density modelisation for the derivative estimation accuracy improvement. Experimental results are presented in order to prove the effectiveness of such approach.