Cross-Layer Design for WiMax System Based on Adaptive HARQ and LS-SVM Channel Prediction

In this paper, a cross-layer design for WiMax system is proposed to enhance the average spectral efficiency of the system by combining two adaptive transmission techniques, i.e. adaptive modulation and coding (AMC) scheme at physical (PHY) layer and hybrid automatic repeat request (HARQ) protocol at medium access control (MAC) layer. We deduce the relationship between cross-layer coupling parameters and the system performance, and propose an adaptive HARQ scheme to maximize the system average spectrum efficiency, which selects the optimum maximum number of HARQ and way of AMC under specified quality-of-service (QOS) requirements. Besides, a channel prediction scheme for WiMax system based on least squares support vector machines (LS-SVM) is presented and applied in order to decrease the impact on AMC performance of rapidly changing channel conditions caused by high-speed movement of receivers. Compared with traditional channel prediction algorithm based on auto regression (AR) model, LS-SVM scheme can achieve higher system performance under low SNR conditions.