Fixed-Point Performance of Space-Time Turbo Trellis Coded Modulation

To achieve both diversity gain and coding gain, the design of space-time turbo trellis coded modulation (ST turbo TCM), which is based on the turbo structure, has been proposed in the literature. The aim of this paper is to study the effect of finite-word length representation on the performance of ST turbo TCM. We present the investigation of bit-width optimization for an 4-state QPSK ST turbo TCM decoder based on the symbol-by-symbol (SBS) logarithm-maximum a posteriori (log-MAP) algorithm on a quasistatic fading channel. The decoder structure utilizes a sliding window algorithm for reducing memory requirements and decoding latency. Simulation results show that ST turbo TCM is feasible for finite word length representation without significant degradation in the frame error rate (FER) performance. The optimum word length configurations are defined for all quantities external and internal to the SBS log-MAP decoder