The Research On Dynamic Resource Management Technques For Satellite Communication Systems

As a major infrastructure of the space information network(SIN),the satellite communication system plays an irreplaceable role in protecting the national economy and the people's livelihood with its wide coverage,good communication quality,and low operation and maintenance costs.With the increasing demand for diversity and high quality of service(QoS)provided by satellite communication systems,satellite communication systems are constantly evolving in terms of network architecture and key technologies to meet the growing demands of users.As a typical resource-limited system,the spectrum,power,storage and other available resources are extremely scarce and valuable.The dynamic resource management(DRM)technology aims to improve the utilization of system resources through dynamic,flexible and efficient resource management methods,which is of great significance to improve the performance of satellite communication systems.This dissertation focuses on DRM technologies in satellite communication systems.The main innovative results are as follows:First of all,aiming at the problem that the current dynamic channel allocation algorithms in satellite communication systems do not consider the temporal correlation and result in a high blocking probability,a deep reinforcement learning based dynamic channel allocation(DRL-DCA)algorithm is proposed.This algorithm models the dynamic channel allocation problem as a Markov Decision Process(MDP)and adopts the convolutional neural network(CNN)to extract the spatial correlation of services between the beams.Simulation results show that the proposed DRL-DCA algorithm can effectively decrease the system blocking probability and improve the utilization of spectrum resources in satellite communication systems.Secondly,aiming at the problem that the existing satellite-terrestrial spectrum sharing algorithms do not consider the mobility of the satellite terminal and cause the terminal to receive excessive interference,an adaptive interference mitigation based spectrum sharing(AIM-SS)algorithm is proposed.The algorithm models and analyzes the antenna radiation pattern and satellite terminal mobility,and obtains the required precoding update period so that the system can support satellite terminals at different motion speeds meanwhile ensuring that the co-channel interference(CCI)does not exceed a certain threshold.The results show that the proposed algorithm solves the problem that the terminal suffers from the excessive interference,and also effectively decreases the required separation distance between the components of the satellite and the terrestrial,and improves the spectrum efficiency.Thirdly,aiming at the problem that the satellite communication system is unable to meet the demands for massive and low-latency multimedia services due to the limited spectrum,a dynamic cache allocation based on Exchange Stable Matching(ESM)algorithm is proposed based on the research thinking of caching exchange spectrum.This algorithm models the allocation of cache content in distributed satellite nodes as a many-to-many matching model.Based on the mutual influence of cache contents between satellite nodes,the preference is designed rationally and exchange stable matching is obtained through bidirectional multiple iterations.The results show that the proposed algorithm decreases the content access delay(CAD)of the user terminals and the required on-board storage overheads.Finally,aiming at the problem that current satellite communication system simulation platforms lack the performance evaluation for multi-scenario and multi-resource DRM algorithms,a dynamic resource management platform(DRMP)for satellite communication system is designed and implemented.The platform can support evaluation on different resources(ex.channel,power,and caching)in single GEO satellite scenarios or Walker constellations composed of multiple LEO satellites.In the DRMP,multiple spatial coordinate system conversions and flexible function calls among various functional modules are described.Various performance indicators such as the system blocking probability and content access delay are supported.The results show that the platform has high reliability in scenario and service modeling,performance evaluation,which can provide a reliable evaluation platform for DRM algorithms for satellite communication systems.