Mean Bit Error Rate Evaluation of MC-CDMA Cellular Systems Employing Multiuser-Maximum-Likelihood Detector

In this paper, we evaluate the uplink bit error rate (BER) of multicarrier-code-division-multiple-access (MC-CDMA) cellular systems with multiuser-maximum-likelihood detector (MU-MLD). The scenario considers perfect power control, own-cell interference (or multiple access interference—MAI), interference from co-cells (or co-cell interference—CCI), additive white Gaussian noise, exponential path-loss and slow frequency-selective Rayleigh fading. Upper bound expressions to evaluate the mean BER are obtained, which are functions of the signal-to-noise ratio (SNR), the number of users, the spreading factor, and the channel reuse factor. For BPSK and 4-QAM (quadrature amplitude modulation), closed-form expressions involving the analytical calculation of pairwise error probabilities are obtained. For 16-QAM and 64-QAM, we show that the number of pairwise error probabilities needed to obtain the BER can be simplified through the constellations symmetry. The procedure for obtaining this simplification is detailed. Also, by considering BER asymptotes, expressions for the SNR penalty due to MAI are derived. Monte Carlo simulations verify the analytical expressions accuracy. In the simulations, the MU-MLD is implemented through the sphere decoder algorithm with MMSE QR factorization. Results show that MC-CDMA does not lose diversity due to MAI. Moreover, CCI significantly affects the system performance. As a consequence, it produces floors in the BER curves.