Online adaptive parameter estimator design and tuning

Focuses on the derivation of a methodology for the tuning of discrete online adaptive parameter estimation. A normalized least-mean-squares algorithm is implemented. The primary analysis is done for linear time invariant discrete parameter estimation. The proposed scheme is designed using the Lyapunov direct method. Stability conditions are checked for both noise-free and noisy measurements. Adaptation gain tuning methods are proposed based on stability and optimal estimator performance considerations. The criteria are partly based on uncertain data taking into account the potential deviations of the predictions from the actual plant output.