Design And Implementation Of Learning Based Handover Method And System In Satellite-terrestrial Integrated Network

With the advantages of wide coverage and less environmental constraints,satellite-terrestrial integration network can meet the communication needs of remote areas and make up for the shortcomings of ground network,which has gradually become the current research hotspot.At present,the design of satellite constellation in the space-time fusion network has the characteristics of low orbit height and large number of satellites.The coverage of each low orbit satellite will overlap to some extent.In addition,the satellites move around the earth at high speed,resulting in the high dynamic characteristics of the space-time fusion network.In order to ensure continuous communication,the terminal must frequently handover satellites to ensure the continuity of service.However,satellite handover will lead to many problems,such as the increase of delay,transmission loss and signaling overhead.The design of satellite handover algorithm directly affects the quality of service of users.Therefore,designing a reasonable and effective handover strategy according to the high dynamic characteristics of satellite network is of great significance to the future development of heaven earth fusion network.This paper summarizes the research status and main methods of satellite handover in space-time fusion network.According to the business continuity requirements in space-time fusion network,a satellite handover method based on two-level fuzzy learning system is proposed,which solves the problem that end users choose the best satellite among multiple visual satellites to continue to provide services.In the multidimensional,multi-layer and heterogeneous heaven and earth integration network,according to the different types of terminal services,it is necessary to realize flexible handover between different levels of networks.Due to the high mobility of satellites and the difference of satellite orbit,the network topology changes dynamically,and the optional access points and link information also change constantly.The dynamic change of network topology causes the terminal to face the problem of difficult and complex access point selection,and the handover complexity increases sharply with the increase of the number of optional access points or handover requests.Therefore,this paper will use multi-attribute dynamic graph to model the heaven earth fusion network,comprehensively consider the elevation,coverage time,idle channel condition and policy weight,and introduce dynamic graph to represent various possible conversion schemes,so as to convert the handover problem model into the best routing problem from the reachable path.Aiming at the high complexity of reasoning rules in fuzzy learning system,a handover algorithm based on two-level fuzzy learning is proposed.The output decision value is generated by using fuzzy learning system and rule adaptive matching.Finally,the best handover satellite is selected according to the decision value of each candidate satellite.This paper designs and develops the heaven earth integration network handover management system,realizes the module functions of terminal management,handover control and satellite management,establishes the satellite network simulation environment,and tests and verifies the function and performance of the proposed method.