Low complexity detectors for bit-interleaved coded modulation over flat-fading multiple-input multiple-output wireless channels

A well-know transceiver structure that provides high spectral efficiency and good performance is that of bit-interleaved coded modulation over multiple-input multiple-output fading channels. Although the interleaver is very beneficial in alleviating deep fades of the channel and introducing independence among the coded bit, it renders au optimal receiver impossible to realize. For this reason, a two-stage suboptimal receiver is usually assumed. The first stage is called the detector and the second is called the decoder. They can operate iteratively to provide very high performance. While the decoder block is quite standard in most cases, the detector on the other hand, involves a great deal of a trade-off between complexity and performance. For this goal, two complexity-reduction techniques are considered in this thesis.;The first technique focuses on the 2 x 2 channel. The complexity reduction is done by using a multi-strata space time code. The idea is that the upper left block of the H matrix is always a scaled identity for any channel realization. The minimum determinant of the new code is 2.3874 with a QPSK constellation. A way to generate bit log-likelihood ratios for iterative decoding is also shown.;In the second technique, a generalization of the depth-first branch-and-bound Schnorr-Euchner sphere decoders is used to build two families of iterative soft detectors. The Schnorr-Euchner concept can be generalized to objective functions more general than squared Euclidean norms. This generalization is exploited in iterative detectors to incorporate the a priori probabilities of the coded bits. The first family in this technique is called the most-contributive-2 N generalized Schnorr-Euchner detectors. It searches for the most contributive N lattice points in the two log-sum-exponents of the bit log-likelihood ratios, for N ≤ 2. The case of N = 1 is the max - log approximation. The case of N = 2 provides a better approximation.;The second family is called the pragmatic iterative generalized Sclmorr-Euchner detector. It looks for a subset of lattice points to use as a representative of the whole lattice. The new detector is based on a depth-first search and shrinks the sphere radius deliberately during the search in order to end it soon and yet, collect lattice points that are likely to impact the log-likelihood ratio values. The search radius is controlled via a so-called radius backlash update strategy which confines the search to within a certain distance from the best found poin. The main advantages of the two families of iterative detectors are that they (i) take into consideration the input a priori probabilities fed back from the decoder and (ii) do not require initial radius estimation and their search step is void of any square root or division operations. The case of non-perfect channel estimates is also considered. For this, an expression of the a posteriori probability is first derived under the assumption of uncorrelated channel estimation errors. The performance gain with dense constellations is very significant.