Embedded intelligence for fast vertical handover decision and network selection in hetnets

Abstract

The hybrid artificial neural networks (hybrid ANNs) combined between the learning vector quantization (LVQ) and radial basis function (RBF) for the vertical handover decision algorithm. The development and hardware implementation of this algorithm are presented in the fast vertical handover decision process owing to keep the always best connected (ABC) in heterogeneous networks (HetNets). The LVQ is based on unsupervised learning and also the RBF is suitable for the non-linear data that is considered the Gaussian distribution. The received signal strength indicator (RSSI), bandwidth requirement (BW), mobile speed (MS) and monetary cost (MC) coefficient of service metrics are introduced the inputs of hybrid ANNs using field programmable gate arrays (FPGAs) logical architecture design. In addition, VLSI hardware description language (VHDL) is described the hybrid architecture. The heterogeneous wireless networks are cooperated the WCDMA, Advanced LTE and WLAN, respectively. The experimental results, the proposed algorithm that is illustrated the high correlation by simulating with MATLAB program, outperforms compared with other approaches as the learning vector quantization and radial basis function. Also, the FPGA can increase the computation time compared to the standard personal computer (PC) so that the FPGA is proper the real time and non-real time applications for the future wireless communication system. © Int. J. of GEOMATE.