Design and Simulation of Multi-User Detection in DS-CDMA Communication Systems

Abstract

Wireless communications rely on Multiple Access Techniques to share limited radio spectrum resources. There are three main types of Multiple Access Techniques i.e. FDMA, TDMA and CDMA. These days CDMA is gaining popularity due to its high capacity and performance. In CDMA different users occupy the same bandwidth (unlike FDMA) at the same time (unlike TDMA). The user data is separated via the use of signature waveforms, sequences or codes assigned to each user. In order to eliminate interference among different user’s data these signature waveforms should be mutually orthogonal (Cross-correlation = 0). DS-CDMA has been the most popular and widely used technique among all other techniques for CDMA. In DS-CDMA each user is assigned a PN Code which is multiplied by the user data before transmission. At the receiver detection is done on the basis of a filter matched to the PN Code of the user. This detector is known as Matched filter detector. Initially this conventional matched filter detector used only orthogonal sequences so there was no issue regarding Multiple Access Interference (MAI), but in order to increase capacity the condition of orthogonality was removed so MAI comes in to the scene. So in order to remove this MAI different techniques were proposed and Multi-user detection (MUD) is one of them. MUD offers high capacity and better performance as compared to Matched Filter Detector. We implemented two MUD Algorithms in Matlab and Simulink, in order to remove and mitigate the effects of MAI. The first one is the Decorrelating Detector, which consists of a bank of matched filters followed by a linear transformation that multiplies the output of the matched filter bank with the inverse of the correlation matrix. The second one is the Successive Interference Cancellation, which involves successive cancellation of MAI from the outputs of the matched filters. First the users are ranked according to increasing received signal strength. Then, the strongest signal is cancelled out first followed by the second strongest signal and so on.