Regularised Equalisation for OFDM Systems With BEM-Based Channel Estimation

Over the last few decades, wireless communication systems have extraordinary development. Great research efforts have been committed, and a huge amount of incredible outcomes have been made. One of the most challenging issues is how to improve the efficiency of transmission when channel is not stationary and varies with time. This thesis studies the channel estimation and equalisation of orthogonal frequency-division multiplexing (OFDM) systems over doubly selective channels. We consider practical channel estimation schemes based on the basis expansion model (BEM) and clustered pilots. We investigate the influence of the parameters of the channel estimation scheme, including the dimensionality of the BEM basis and bandwidth chosen for the frequency-domain (FD) channel matrix. We address the performance degradation of linear and non-linear equalisation due to imperfect channel estimation. We study the usage of regularisation techniques to enhance the performance of both linear and non-linear equalisation schemes. We also consider reduced-complexity segment-by-segment implementations. The numerical results show that regularisation can effectively enhance the detection performance.