A digital phase-locked loop with variable loop gains derived from RLS method

We propose a DPLL (digital phase-locked loop) with time-varying loop gains which are derived from the RLS method. This method, in contrast to the one based on the standard Kalman filtering algorithm, is appropriate for the situation where no statistics are known. If the forgetting factor in the RLS method is constant, the acquisition performance only improves in the initial stage. However, with the variable forgetting factor adaptive to the measurements, the DPLL shows excellent acquisition and tracking performance under a non-stationary environment which are comparable to that based on the adaptive Kalman filter with lock detection algorithm.