A Cost-based Vertical Handoff with Combination Prediction of Sinr in Heterogeneous Wireless Networks

It is well known that to provide seamless mobility and transmission in heterogeneous wireless networks, a vertical handoff technique is required to guarantee an Always Best Connected. This paper thus presents a cost-based vertical handoff algorithm with combination prediction of SINR (CPSVH) in heterogeneous wireless networks to make handoff decision. Our approach involves two steps, first SINR is predicted by combining GM (1,1) and BP neural network for accurate timing to trigger handoff, and then a handoff decision on the optimal network is made by way of a cost function. The cost function, on basis of multiattribute QoS consideration, is composed of SINR, user preference, user traffic cost and available bandwidth from accessible networks, with the weight of each attribute in the cost function calculated by a fuzzy judgment matrix constructed for this purpose. Meanwhile, the stability period (defined as the waiting time before handoff) is also taken into regard to reduce unnecessary handoffs. The simulation results in this study reveal that the proposed handoff scheme outperforms other approaches in terms of system throughput, dropping probability, and vertical handoff numbers.