Modified RLS algorithm with enhanced tracking capability for MIMO channel estimation

A new RLS adaptive filtering algorithm has been developed by enforcing a constraint on the a posteriori error of the conventional exponentially-weighted RLS algorithm and modifying its coefficient update equation. The motivation behind the new algorithm is to attain improved tracking capability when the target system features fast variations in time. The proposed algorithm requires only slightly more computations than the conventional RLS algorithm and is specifically suitable for estimation and tracking of fast-fading MIMO channels. Simulations demonstrate that a MIMO channel estimator utilising the proposed algorithm outperforms those using the conventional RLS algorithm, the set-membership-based BEACON algorithm and the extended RLS algorithm without requiring prior knowledge of the Doppler shift.