Power and performance trade-off in DS-CDMA receivers based on adaptive LMS-MMSE multi-user detector.
Third generation cellular communication systems based on CDMA techniques have shown great scope for improvement in system capacity. Over the last decade, there has been significant interest in DS-CDMA detectors. The conventional detector, the optimal detector and a number of sub-optimal multi-user detectors (MUD) have been extensively analyzed in the literature. Recently, the reduction of power consumption in DS-CDMA systems has also become another important consideration in both system design and in implementation. In order to support wireless multimedia services, all CDMA-based systems for third generation systems have a large bandwidth and a high data rate, therefore the power consumed by the digital signal processor (DSP) is high. This thesis focuses on power consumption in the adaptive Minimum Mean Square Error (MMSE) detector which is based on the Least Mean Square (LMS) algorithm. This thesis presents a literature survey on MUD and adaptive filter algorithms. A system model of the quantized LMS-MMSE MUD is proposed and its performance is analyzed. The quantization effects in the finite precision LMS-MMSE adaptive MUD including the steady-state weight covariance, mean square error (MSE) and bit error rate (BER) versus wordlength of data and coefficient are investigated when both the data and filter coefficients are quantized. The effects of wordlength size on power consumption are investigated and the tradeoff between the power consumption and performance degradation and the optimal allocation of bits to data and to LMS coefficients under power constraint is presented.