Complexity-reduced iterative MAP receiver for spatial multiplexing systems

A complexity-reduced iterative algorithm for joint maximum a posteriori (MAP) detection and channel estimation for spatial multiplexing systems is presented. To reduce the complexity of the MAP detection, which increases exponentially with the number of transmit antennas and the modulation order, a modified sphere decoding (SD) algorithm is proposed. It just enumerates the symbol sequences with high a priori probabilities, which fall inside a sphere centred at the received signal point. Extensive simulations demonstrate that the proposed algorithm, combined with convolutional code, reduces the complexity as compared to that of the MAP detection with moderate performance degradation. The effects of design parameters, e.g. sphere radius and the number of symbol sequences with high a priori probabilities, on the performance/complexity trade-offs are also discussed.