REDUCED-COMPLEXITY EQUALIZATION TECHNIQUES FOR BROAD-BAND WIRELESS CHANNELS

This paper presents reduced complexity equalization techniques for bro adband wireless communications, both outdoors (fixed or mobile wireles s asynchronous transfer mode (ATM) networks) and indoors [high-speed l ocal-area networks (LAN's)]. The two basic equalization techniques inv estigated are decision-feedback equalization (DFE) and delayed decisio n-feedback sequence estimation (DDFSE). We consider the use of these t echniques in highly dispersive channels, where the impulse response ca n last up to 100 symbol periods, The challenge is in minimizing the co mplexity as well as providing fast equalizer start-up for transmission s of short packets, We propose two techniques which, taken together, p rovide an answer to this challenge, One is an open-loop timing recover y approach (for both DFE and DDFSE) which can be executed prior to equ alization; the other is a modified DFE structure for precanceling post cursors without requiring training of the feedback filter, Simulation results are presented to demonstrate the feasibility of the proposed t echniques for both indoor and outdoor multipath channel models, The pr oposed open-loop timing recovery technique plays a crucial role in max imizing the performance of DFE and DDFSE with short feedforward spans (the feedforward section of DDFSE is a Viterbi sequence estimator), A feedforward span of only five is quite sufficient for channels with sy mbol rate-delay spread products approaching 100, The modified DPE stru cture speeds up the training process for these channels by 10-20 times , compared to the conventional structure without postcursor precancell ation. The proposed techniques offer the possibility of practical equa lization for broadband wireless systems.