| To support multimedia services, wideband direct sequence code division multiple access (W-CDMA) is emerging as an important wireless technology for future wireless multimedia application. While wider bandwidth allocation results in more resolvable multipath available at the receiver, the multiple paths are spread over a larger unknown time interval in comparison to narrowband CDMA. In principle, an optimal RAKE receiver could combine all the multipath components by maximal ratio combining (MRC) through brute-force search. Due to the prohibitive complexity of the optimal RAKE receiver, in this thesis, we propose a sub-optimal low complexity variable-size RAKE receiver with a small performance degradation.; We develop a novel PN code acquisition algorithm to detect uplink short-code CDMA multipath signals. A tree-structured parallel adaptive network of RAKE fingers is created for all possible delay shifts and users. Employing sequential detection, only the strongest paths are used to generate the output decision statistics. The performance of the proposed code acquisition scheme in multipath channels is analytically determined. We quantify the complexity savings in multipath channels. A lower bound of the signal combining loss is calculated and verified by Monte Carlo simulations. Both the simulation and numerical calculation results demonstrate that the proposed algorithm can, at the expense of slight signal combining loss, significantly reduce the computation complexity and latency of the receiver. For example, in a CDMA system with 10 users and 4 paths per user, the overall benefit of our proposed sequential detection/tree search approach when compared to the alternative full-search one is demonstrated by at least one order of magnitude complexity savings (over 90%) with only 1.8 dB loss over optimal signal combining performance at an SNR of 5 dB. Incorporating channel estimation errors and test design mismatch in the study, we have found that the performance degradation is only slightly higher. |
