Research On Multi-layer Satellite Network Access And Handoff Technology For Space-air-ground Integration

Space-air-ground integration is being discussed and implemented.The space-air-ground integrated networks is constructed of satellite backbone network, high-level aircraft network and ground networks. But for now though, the satellite systems are independent of each other, there is no uniform communication interface and standard, resulting in no information sharing, repeated construction of communication links, waste of communication resources and so on. In order to break this situation, it is necessary to establish a backbone network composed of geostationary satellites, medium orbit satellites and low orbit satellites as access nodes of independent special satellites, aircraft and terrestrial terminals to enhance the comprehensive utility of space communication resources.This paper presents a space-air-ground integrated backbone network architecture under the above-mentioned idea where GEO and MEO mainly act as the backbone of space-air-ground network and the LEO as the acess points.Specifically, the GEO satellites are used as space aircraft access points and network traffic forwarding nodes(NTFN) of MEO, the MEO satellites as network distributed management nodes and network traffic forwarding nodes of LEO, and the LEO satellites as ground access points. Three orbit satellites work in coordination to form an independent and complete space backbone network.To solve the problem of multi-layer satellite network access, this paper presents an optimized multi-attribute sorting algorithm, which introduces a closeness index in place of tranditional normalization method to solve the problem of unreliable sorting; presents using Euclidean distance to measure the distance between attribute values and ideal solutions to solve the rank reversal problem.On the basis of this algorithm, new rules of selection and reselection,named O rules, are presented. Compared with the traditional access selection and reselection rules, the O rules are designed for the satellite environment. The algorithm is more stable, which can reduce new call blocking rate and forced outage rate, and increase average elevation angle.In order to solve the problems existing in multi-layer satellite network handoff,like large handoff delay, unnecessary handoffs, which affecting network load balancing. This paper presents a parallel fuzzy neural network algorithm which takes parallel input module,including the RSRP module, satellite load modlue and other modules.The algorithm takes the T-S fuzzy neural network and adaptively adjusts the membership function and the decision rule, resulting in smooth handoff.This algorithm has good adaptability and learning ability. Compared with the traditional handover algorithm, it can reduce the system handoff delay and handoff times, improve handoff success rate and system utilization rate, resulting in achieving fast and smooth handoff. It can also bring traffic balance, reduce network congestion and other additional benefits.