Performance of multiuser MMSE decision feedback receiver with convolutional codes and trellis coded modulation

This paper examines the performance of a coded CDMA system with a minimum mean square error decision feedback (MMSE DF) receiver over a Gaussian channel. In the absence of error propagation the MMSE DF perfectly cancels interference from k-1 already detected users. If the information bits are decoded at the receiver and re-encoded before being processed by a feedback filter, the number of erroneous decisions passed to other users can be decreased. This improves interference estimation, hence lowering the bit error rate (BER) and increasing the capacity. Convolutional, trellis and turbo trellis codes are employed to reduce error propagation. The performance of various coded systems is compared for the same bandwidth and maximal spectral efficiency.