The major contribution of CDMA towards communication is noise mitigation... The inherent temporal interference minimization of spread spectrum is a unique to CDMA which is termed as processing gain. However event then, there are number of situations where interference becomes so stronger, so that extra interference cancellation is required. Especially in multi-user environment with spread spectrums sharing same bandwidth develops with each other arising from cross correlations of signals. In the detection phase these cross correlation factors arising from multi-user environment are the reasons for the need for additional mitigation capabilities in the receivers. In this thesis Independent Component Analysis (ICA) is used for the noise cancellation. The ICA approach is applied as an advanced pr-processing tool before conventional detection process. The ICA incorporated with PCA (Principal Component Analysis) as a preprocessing and whitening tool is used prior to ICA. CDMA base band BPSK signal after coding stage in bipolar form simulated using Matlab� random bipolar source generation according to the number of users. These generated signals are the spread by using Gold spreading signature sequences. The channel is assumed as AWGN . Hence the received signal is sum of all user’s spread sequences, carrying channel noise. This signal is then detected by using MATCHED filter tuned to spreading sequences separately at each of its fingers. Bits of each user along with cross correlation interference due to other users are separated in this stage. Further out put from MATCHED Filter is processed through ICA. In this phase cross correlation mitigation is carried out. Afterwards comparative study carried out with bit error rate comparison with Independent Component Analysis combined with conventional detection and only using conventional detection system.

BEMTS