An Application of ICA to DS-CDMA Detection

This paper presents the application of the theory and algorithms of independent component analysis (ICA) to solve the symbol estimation problem of the multi users in a direct-sequence code division multiple access (DS-CDMA) communication system. The main focus is on blind separation of convolved CDMA mixture and the improvement of the downlink symbol estimation. A combined scheme of ICA-SUD (single user detection) detector has been simulated. The simulation demonstrated that ICA-SUD gave the lower error rate comparing to the conventional SUD receiver and pure ICA detector with training sequences.