Adaptive derivative estimation for delay-constrained acceleration measurement

The well-known noise problem in digital differentiation is addressed by mea ns of using adaptive digital filtering for signal preprocessing. Rapidly re sponding differentiators with low-noise output can be constructed by using the adaptive filter in a predictor configuration. As the prefilter is adapt ive, the approximation is more flexible than polynomial fitting. The recurs ive least-squares adaptive algorithm is used for prediction.