ADAPTIVE ANALOG NETWORK FOR REAL-TIME ESTIMATION OF BASIC WAVE-FORMS OF VOLTAGES AND CURRENTS

A new parallel algorithm for estimation of parameters of sinewave, distorted by DC exponential signal and contaminated by noise, is proposed. The method may be seen as an extension and generalisation of the standard least-squares (L2-norm) optimisation approach. The advantage of the developed algorithm is that it is more robust with respect to the modelling error and impulsive (wild) noise than when a standard least-squares criterion is employed. The implementation of the algorithm by a suitable adaptive analogue network is also given. Computer simulation results are presented to confirm the validity and performance of the proposed network. The proposed method seems to be particularly useful for real-time high-speed and low-cost estimation of parameters of sinusoidal signals.