Sinusoidal time delay tracking by a self-tuned LMS filter with interpolation based on system response coefficient ratios

Summary form only given. In this paper we address the estimation of time delay (TD) between two spatially separated noisy sinusoidal signals by system identification modeling with the input and output corrupted by additive white Gaussian noise. Input noise introduces bias in the TD estimates. Normally the solution requires a priori knowledge of the input-output noise variance ratio. We utilize the cascade of a self-tuned filter with the TD estimator, which makes the delay estimates robust to input noise. Furthermore, we propose a new method of TD interpolation based on the FIR filter coefficient ratio. The proposed TD estimation method requires no frequency domain transformations, and is suitable for entirely time-domain applications. Simulation results confirm the superiority of the proposed approach at low signal-to-noise ratios.