Efficient Non-Pilot-Aided Channel Length Estimation for Digital Broadcasting Receivers

Abstract

Channel estimation and equalization techniques are crucial for the ubiquitous broadcasting systems. Conventional receivers for most broadcasting or wireless standards preset the channel length to the maximal expected duration of the channel impulse response for the adopted channel estimation and equalization algorithms. The excessive channel length often significantly increases the implementational complexity of the wireless receivers and leads to the redundant information which would induce the additional estimation errors. Moreover, such a scheme does not allow the dynamic memory allocation for variable channel lengths. This could further increase the power consumption and reduce the battery life of a mobile device. The knowledge of the actual channel length would, in principle, help the system designers decrease the complexity of the channel estimators using maximum likelihood (ML) and minimum-mean-square-error (MMSE) algorithms. In this paper, we address this important channel length estimation problem and propose a novel autocorrelation-based algorithm to estimate the channel length without the need of pilots or training sequence. The associated threshold for the channel length estimation depends on the sample size, the signal-to-noise ratio and the leading/last channel coefficients. In addition, we provide the mean-square analysis on the effectiveness of the proposed non-pilot-aided channel length estimator through Monte Carlo simulations.