Signal Equalization and User Detection Scheme for Multi-Carrier Code Division Multiple Access Systems based on the Support Vector Machines

Shariff Abdul Rahmab (0151208)


In order to obtain a reliable communication in the time varying multipath radio channels, efficient equalization techniques are required. Linear equalizers do not perform well in mobile radio channels where the Doppler spreads are common. In that case, nonlinear equalizers are required to compensate for the signal distortions caused by channels.

In this work, a novel nonlinear signal equalization and user detection scheme for the Multi-carrier Code Division Multiple Access(MC-CDMA) systems applying the Support Vector Machines(SVM) classification is proposed . SVM is an emerging machine learning technique initially developed for the classification applications. It is gaining popularity due to its many promising empirical performance.

In the propsed scheme, the SVM classifier is directly trained with the frequency domain samples of MC-CDMA signals in order to effectively utilize the frequency diversity. The performance of the proposed scheme is evaluated via computer simulations under the time varying multipath radio channel. Results obtained confirmed that the proposed scheme is capable of both channel equalization and user detection. The results confirmed that it outperforms linear equlizers with perfect channel estimates.