An adaptive Kalman filter for the enhancement of noisy AR signals

Abstract

In this paper, we describe a new adaptive system for the enhancement of autoregressive (AR) signals which are disturbed by additive broadband noise. The system is comprised of an adaptive Kalman filter operating as a fixed-lag smoother and a subsystem for AR parameter estimation. As opposed to the conventional approach of employing an extended Kalman filter, we estimate the Kalman filter parameters using the enhanced signal and thus establishing a feedback between the Kalman filter output and the estimated parameters. Our system is capable of tracking short-time stationary signals. It is computationally efficient and can easily be implemented on today's integrated digital signal processors.