Reduced-complexity equalization techniques for broadband wireless channels

Abstract

This paper presents reduced-complexity equalization techniques for broadband wireless communications. The two basic techniques investigated are decision-feedback equalization (DFE) and delayed decision-feedback sequence estimation (DDFSE). We consider the use of these techniques in highly dispersive channels, where the impulse response can last up to 100 symbol periods. The challenge is in minimizing the complexity as well as providing fast equalizer start-up for transmissions of short packets. We propose two techniques which, taken together, provide an answer to this challenge. One is an open-loop timing recovery approach (for both DFE and DDFSE) which can be executed prior to equalization; the other is a modified DFE structure for pre-cancelling postcursors without requiring training of the feedback filter. Simulation results are presented to demonstrate the feasibility of the proposed techniques for both indoor and outdoor multipath channel models. The proposed open-loop timing recovery technique plays a crucial role in maximizing the performance of DFE and DDFSE with short feedforward spans. A feedforward span of only 5 is quite sufficient for channels with symbol rate-delay spread products approaching 100. The modified DFE structure speeds up the training process for these channels by 10 to 20 times, compared to the conventional structure without postcursor pre-cancellation. The proposed techniques offer the possibility of practical equalization for broadband wireless systems.