Maximum likelihood sequence estimation in the presence of constant envelope interference [cellular radio networks]

Gaussian models of signal disturbance are widely used in conventional designs of mobile receivers. However, such models do not accurately reflect the statistics of co-channel interference in narrow-band cellular networks. In this paper, we investigate the potential of suppressing co-channel interference using a simple non-Gaussian disturbance model in equalization. In particular, the model exploits the constant-envelope nature of GMSK-modulated interference, which commonly appears in GSM/EDGE networks. We show that the metric of the corresponding MLSE equalizer is simply the sum of the conventional Euclidean-distance metric and a correction term that depends on the estimated amplitude of the interference. The new metric sagaciously biases the decision regions in accordance with the non-Gaussian statistics. Using the GSM/EDGE air interface, we demonstrate that a large gain (up to 18 dB) in C/I can be achieved by the new metric with little increase in receiver complexity.