Multi-objective Intelligent Handover in Satellite-Terrestrial Integrated Networks

The satellite-terrestrial integrated networks (STIN-s), which consists of satellite and terrestrial nodes, aims to realize wide-area coverage, universal multi-access. However, the fast periodic motion of low earth orbit (LEO) satellites results in the dynamic topology and the intermittent connectivity of STIN. In this paper, we address the handover problem caused by the frequent changes in the connectivity of satellite-terrestrial links, and proposed a multi-objective intelligent handover (MIHO) scheme to increase throughput and balance load under the constraint of handover delay via executing the optimal network selection. Firstly, a STIN model is constructed with LEO satellites, base stations (BSs), and user terminals. Continuously, the achievable rate and handover delay are derived after analyzing the coverage and association relationship among LEO satellites, BSs and users. After that, the handover in STIN is modeled as a multi-objective optimization problem, and the MIHO scheme is proposed to make the optimal network selection. In MIHO, a handover algorithm based on improved discrete binary particle swarm optimization (IBPSO-HO) is designed via jointly considering the achievable rate and load balance. Simulation results show that the proposed IBPSO-HO algorithm can outperform the existing handover algorithms in terms of throughput, load balance, as well as handover delay.