Semiblind EM-Based Iterative Receivers for Space–Time-Coded Modulation and Quasi-Static Frequency-Selective Fading Channels

This paper introduces two different iterative receivers for space-time-coded data transmitted through frequency-selective fading channels. Both receivers perform maximum a posteriori data detection and decoding, and operate semiblindly by making use of the expectation maximization (EM) algorithm to estimate the unknown channel. The first receiver utilizes a supertrellis to jointly describe the channel and the code, while in the second one, this description is accomplished with two separated trellises. The two approaches are compared in terms of trellis complexity, achievable diversity level, and bit error rate (BER). The supertrellis approach is less computationally demanding and has a faster convergence rate, while the separate approach can achieve a higher diversity level and thus a smaller BER. This tradeoff is explored with examples and computer simulations, which show that, in many cases, the supertrellis approach can be a better solution than the separate approach