Adaptive kernel least square support vector machines applied to recover DS-CDMA signals
Xin Zhao, Anthony Kuh · 2003
This paper discusses an adaptive nonlinear learning algorithm for direct-sequence code division multiple access (DS-CDMA) system. The algorithm is based on the least square support vector machine (LS-SVM), a nonlinear kernel based machine. The LS-SVM detectors have advantages in that they have moderate complexity, can realize nonlinear decision regions, can be implemented adaptively, and require only training sequence data from the desired user. Through simulations, the performance of bit error rate (BER) of the designed LS-SVM receiver is compared to other conventional CDMA receivers and observes that the LS-SVM detector's performance approaches that of the Bayesian receiver. The simulation results also show that the proposed adaptive LS-SVM receiver can track data in time varying environment.