Research On Key Technologies Of Cooperative Transmission In Satellite-Terrestrial Networks

To meet the global demand for seamless access and connectivity across all devices and locations,the development of a global seamless integrated satellite-terrestrial network(ISTN)has become a critical goal for wireless communication networks.By integrating satellite networks with traditional terrestrial networks,ISTN can provide ubiquitous,efficient,continuous,and intelligent information security for various network applications in multi-dimensional space.However,compared with terrestrial networks and tranditional small-scale satellite networks,ISTN has undergone fundamental changes in network structure(from flat to large-scale spatial span),node quantity(tens of thousands of satellites),and node capability(transmission,mobility,and routing).In addition,heterogeneous network nodes also have different channel characteristics,motion characteristics,coverage characteristics,and resource characteristics.Therefore,traditional single network transmission mechanisms are difficult to be directly applied in ISTN.It is urgently needed to further study efficient satellite-terrestrial cooperative transmission technology to meet the complex and diverse task requirements of users while maximizing the utility of integrated network resources.Focusing on the ISTN with large-scale low Earth orbit(LEO)satellite constellations,this paper aims to optimize network performance by deeply exploring key technologies for satellite-terrestrial cooperative transmission from the perspectives of differential network characteristics,user association,satellite-terrestrial cooperative task offloading,and satellite-terrestrial cooperative content distribution,which can provide theoretical basis and technical support for the optimization design of integrated network systems,cooperative transmission strategies,and resource allocation methods.The main work of this paper includes:1.Characteristics Analysis of Low Earth Orbit Satellite Constellation in Integrated Satellite-Terrestrial NetworksIn view of the heterogeneity of the LEO satellite network and terrestrial wireless network in ISTN,this paper first proposes an integrated architecture based on a large-scale LEO constellation and terrestrial wireless nodes,and discusses and simulates the cooperative transmission paradigm and propagation delay based on the functional characteristics of the satellite payload.Secondly,focusing on the high-speed orbital characteristics of LEO satellites in ISTN,this paper discusses their orbit characteristics,constellation configuration,and coverage characteristics in detail,and analyzes the impact of satellite coverage area,terminal visibility,and terminal mobility on network transmission through simulation.Finally,based on the working frequency band,the transmission characteristics of the LEO satellite channel are modeled and analyzed to clarify the channel parameters that affect the communication link transmission performance.Simulation results show that compared with ground wireless nodes,LEO satellite nodes have highly dynamic and time-varying topological connections and propagation channels.In addition,differential satellite payload functions can lead to different selections of backhaul links and end-to-end delays.Meanwhile,there are significant differences in the operational direction of satellites,the number of visible satellites at terminals,coverage time,and construction methods of inter-satellite link in different constellation configurations.2.Deep Reinforcement Learning Based User Association in Integrated Satellite-Terrestrial NetworksNowadays,the increasing scale of satellite constellations,high dynamic topology of the network,and diversified user demands bring about fierce competition in user association,leading to large end-to-end delays and jitters.To address these problems,this paper proposes a user association method for ISTN based on deep reinforcement learning.Firstly,the satellite-ground and inter-satellite link models are constructed for the Walker Delta-type LEO constellation.An end-to-end propagation delay model is derived based on the minimum hop theory.Secondly,a multi-objective user association problem is formulated by maximizing the user access success rate and satellite service duration while ensuring load balancing.Considering the dynamic movement of satellites,time-varying user requests,and large decision space,the problem is transformed into a multi-agent Markov decision problem.Then,a user association algorithm based on D3 QN is proposed to efficiently solve the optimization problem.Finally,simulation experiments are conducted to evaluate the convergence performance and application performance of the proposed method.The simulation results show that the proposed method is applicable to meeting user requests with different Qo S while ensuring the reliability of end-to-end network.It also has excellent performance under different user concurrency and network load environments compared to traditional user association methods.3.Two-Timescale Learning-Based Task Offloading in Integrated Satellite-Terrestrial NetworksAchieving diversified wireless access and backhaul access modes will become an important feature and development trend of the future ISTN.However,the differentiated channel characteristics,node resource characteristics,and control time granularity of the two networks pose significant challenges for the development of uplink cooperative transmission strategies in ISTN.To address the above challenges,in this paper,we propose an integrated satellite-terrestrial network architecture to support delay-sensitive task offloading,in which satellite networks serve as a complement to terrestrial networks by providing additional communication resources,backhaul capacities,and seamless coverage.Under this architecture,we investigate how to jointly make offloading link selection and bandwidth allocation decisions for BSs and remote users.Considering the differentiated decision-making time granularities,we formulate a two-timescale stochastic optimization problem to minimize the overall task offloading delay.To accommodate the two-timescale network dynamics and characterize state-action relations,we establish an H-MDP framework with two separate agents tackling two-timescale network management decisions,and two evolved MDP-based subproblems are formulated accordingly.To efficiently solve the subproblems,we further develop an H-PPO-based algorithm.Specifically,a hybrid actor-critic architecture is designed to deal with the mixed discrete and continuous actions.In addition,an action mask layer and an action shaping function are designed to sample feasible task offloading decisions from the time-variant action set.Extensive simulation results have validated the superiority of the proposed ISTN architecture and the H-PPO-based algorithm,especially in scenarios with scarce spectrum resources and heavy traffic loads.4.Joint Cache Placement and Cooperative Multicast Beamforming in Integrated Satellite-Terrestrial NetworksIn ISTN,the emerging content-centric services,such as streaming video and content sharing,have become the main downstream demand for wireless networks.Edge caching and multicast transmission can effectively reduce the repetitive transmission of popular content.However,the the significant difference in operation time granularity and coupling between cache placement and content distribution strategies severely restrict the network transmission efficiency.To address this issue,this paper studies joint cache placement and cooperative multicast beamforming to provide content-centric data services for mobile users in ISTN.Specifically,in ISTN,users requesting the same content are arranged into a multicast group and served by the cache-enabled base stations(BSs)and LEO satellite via cooperative beamforming.To maximize the network utility that takes network throughput and backhaul traffic into consideration,the cache placement,LEO satellite and BS clustering,and multicast beamforming are jointly designed and formulated as a two-timescale optimization problem.However,the original problem is anti-causal since the cache placement strategy and content delivery policy are coupled in different timescales.By utilizing historical information,we propose a two-step scheme to decompose the problem into a short-term content delivery subproblem and a long-term cache placement subproblem.As the former subproblem is nonconvex with mixed-integer variables and coupling constraints,we transform it into an equivalent problem and propose a penalty concave-convex procedure based algorithm to solve it.To address the latter subproblem,a centralized iterative algorithm and a distributed alternating algorithm with low complexity are developed,respectively.Simulation results validate that the proposed schemes can effectively enhance the network throughput and reduce the backhaul traffic compared with benchmark schemes.