Turbo Space–Time Equalization of TCM for Broadband Wireless Channels

Abstract

This paper presents a space-time turbo (iterative) equalization method for trellis-coded modulation (TCM) signals over broadband wireless channels. For fixed wireless systems operating at high data rates, the multipath delay spread becomes large, making it impossible to apply trellis-based equalization methods. The equalizer proposed here consists of a broadband beamformer which processes antenna array measurements to shorten the observed channel impulse response, followed by a conventional scalar turbo equalizer. Since the applicability of trellis-based equalizers is limited to additive white noise channels, the beamformer is required to preserve the whiteness of the noise at its output. This constraint is equivalent to requiring that the finite-impulse response (FIR) beamforming filters must have a power complementarity property. The power complementarity property imposes nonnegative definite quadratic constraints on the beamforming filters, so the beamformer design is expressed as a constrained quadratic optimization problem. The composite channel impulse response at the beamformer output is shortened significantly, making it possible to use a turbo equalizer for the joint equalization and decoding of trellis modulated signals. The proposed receiver structure is simulated for two-dimensional TCM signals such as 8-PSK and 16-QAM and the results indicate that the use of antenna arrays with only two or three elements allows a large decrease in the channel signal-to-noise ratio needed to achieve a 10 4 bit-error rate.Index Terms-Broadband beamforming, power complementary filter banks, soft-output decoding, turbo equalization.

Turbo Equalization: An Overview

“…Extensions to trellis-coded modulation [63] and BICM [64] as well as frequency domain-based formulations using single-carrier transmission with and without cyclic prefix [65]- [69] as well as MIMO OFDM were also developed [70], [71]. The increase in problem dimensionality lends itself naturally to complexity reduction methods through rank reduction [72] and a variety of other MAP approximation methods.…”