Polynomial constrained LMS adaptive algorithm for measurement signal processing

The LMS adaptive algorithm is modified in such a way that unbiased polynomial estimation or prediction is guaranteed continuously. This is accomplished by introducing the necessary linear constraints and modifying the coefficient update equations accordingly. The adaptive properties in polynomial prediction allow optimum noise attenuation even when the characteristics of noise and interference are dynamically changing. Also, adaptive differentiation with the constrained LMS algorithm is proposed. Examples of signal processing in a velocity measurement application are shown.