Complex amplitudes tracking loop for multipath slow fading channel estimation in OFDM systems
Résumé
This Gipsa-lab research report deals with channel estimation for Orthogonal Frequency Division Multiplexing systems over slowly-varying multi-path fading channels. Most of the conventional methods exploit the frequency domain correlation in estimating the channel at pilot subcarriers position, and then interpolating it over the entire frequency grid. More advanced algorithms exploit the time-domain correlation as well, by employing Kalman filter based on the approximation of the time-varying (assumed Rayleigh with Jakes'spectrum) channel. Adopting a parametric approach and assuming a primary acquisition of delay related information, channel estimator has to track the Complex Amplitudes (CAs) of the channel paths. In this perspective, we propose a reduced complexity algorithm compared to Kalman methods, based on a second-order CA Tracking Loop. Inspired by Phase-Locked Loops, an error signal is built from the pilot-aided Least-Squares (LS) estimate of the CAs, and is integrated by the loop to carry out the final CAs estimate. We derive closed-form expressions of the Mean Square Error (MSE) of the algorithm, and of the corresponding optimal loop coefficients. Simulation results show that our algorithm outperforms the conventional (frequency-domain interpolation) method and has almost the same performance as a second-order Kalman based algorithm. Moreover, efficiency of a second-order versus first-order approach is demonstrated for the slow-fading case, with a MSE closer to the Bayesian-Cramer-Rao-Bound.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...