Skip to Main Content
Adaptive signal processing algorithms are often used in order to "track" an unknown time-varying parameter vector. This work develops an upper bound on the mean of the norm-squared error between the unknown parameter vector being tracked and the value obtained by the algorithm. The results require very mild covariance decay rate conditions on the training data and a bounded algorithm. The upper bound illustrates the relationship between the algorithm step size and the maximum rate of variation in the parameter vector being tracked.
Acoustics, Speech and Signal Processing, IEEE Transactions on (Volume:29 , Issue: 3 )
Date of Publication: Jun 1981