基于DQN的VDES异构星座兼容策略

Compatible strategy of VDES heterogeneous constellation based on DQN

异构VDES(VHF data exchange system)星座采用相同的通信频率和时分多址通信机制,使得异构星座重复覆盖区域内存在大量由时隙冲突造成的同频干扰,严重影响通信质量。针对此问题,提出一种基于深度Q网络(DQN)的星座间兼容策略。基于VDES通信流程,设置船站作为资源信息中转节点,赋予卫星对通信环境的感知能力。在此基础上,将异构星座场景下的资源分配问题建模为强化学习问题,提出一种基于DQN的时隙资源分配算法。通过重构历史资源信息和当前资源信息,规划最优时隙资源分配方案,并根据结果对算法迭代优化。仿真结果表明,所提出的策略可以有效提高通信性能。

As heterogeneous VHF data exchange system(VDES)constellations use the same communication frequency and time division multiple access(TDMA)communication mechanism,a large number of co-channel interference is caused by slot conflicts in the overlapping areas of constellations,reducing the communication quality.To tackle this problem,an inter-constellation compatibility strategy for deep Q-network(DQN)is proposed.Based on the VDES communication process,the ship station is set as the resource information transfer node,which enables the satellite to perceive the communication environment.On this basis,the resource allocation problem in the heterogeneous constellation scenario is modeled as a reinforcement learning(RL)problem,and a DQN-based slot resource allocation algorithm is proposed.By reconstructing the historical and current resource information,the optimal slot resource allocation scheme is planned and the algorithm is iteratively optimized according to the results.Simulation results show that the proposed strategy can effectively enhance communication performance.