Delayless recursive differentiator with efficient noise attenuation for control instrumentation

Acceleration feedback can improve the performance of motion control in servo drives. Closed-loop acceleration control is, however, seldom implemented in practical drive systems due to dissatisfying performance of most existing acceleration measurement methods. In this article, we introduce a feedback extension scheme for finite impulse response (FIR) polynomial differentiators making it possible to differentiate the velocity signal computationally efficiently while providing considerable noise attenuation. The proposed recursive differentiator does not delay the primary acceleration signal but, instead, the forward prediction step can be adjusted to compensate for various delays in the physical implementation. Predictive properties are obtained by taking advantage of the deterministic nature of polynomial signals. The utilization of the polynomial model, on the other hand, is justified because within narrow time windows, low-degree polynomials can be fitted accurately into the smoothly changing velocity and acceleration signals. The performance of the novel differentiation scheme is evaluated by applying it to an acceleration measurement application. (C) 1998 Elsevier Science B.V. All rights reserved.