基于业务特征的高级在轨系统虚拟信道复用方法研究

Aiming at the real-time and efficient transmission demand of self-similar and multi-type service in space information network, based on AOS data multiplexing mechanism, an AOS virtual channel multiplexing method based on service characteristics is proposed. Firstly, an AOS virtual channel multiplexing optimization model is established, which can provide technical architecture for multiplexing method. Secondly, a AOS queue management method for self-similar traffic is researched, which includes the combination prediction model of space information network traffic based on optimization Extreme Learning Machine and ARMA, Hurst- priority Random Early Detection algorithm for self-similar traffic, therefor, the real-time transmission problem of self-similar service can be solved. Thirdly, an AOS virtual channel scheduling method for multi-type service is researched, which includes emergency, synchronous and asynchronous time slot movable virtual channel scheduling model, asynchronous virtual channel dynamic scheduling algorithm based on entropy estimation reinforcement genetic algorithm, therefor, the efficient transmission problem of multi-type service can be solved. Finally, simulation system is built, which can realize the performance testing for key technologies. The breakthrough of key technologies will greatly improve the high-speed transmission capacity for large capacity data in space information network, and provide technical reserve for the research of space information network protocol.

针对空间信息网络中自相似、多类型业务的实时高效传输需求，根据高级在轨系统(AOS, Advanced Orbiting System)数据复用机制，开展基于业务特征的高级在轨系统虚拟信道复用方法研究。首先，建立AOS虚拟信道复用优化模型，为复用方法的研究提供技术架构；其次，研究面向自相似业务的AOS队列管理方法，包括基于优化极限学习机与ARMA的空间信息网络流量组合预测模型、适应自相似业务的Hurst-优先级随机早检测算法，以解决自相似业务的实时传输问题；再次，研究面向多类型业务的AOS虚拟信道调度方法，包括适应多类型业务的紧急/同步/异步时隙可移动虚拟信道调度模型、基于熵估计强化遗传算法的异步虚拟信道动态调度算法，以解决多类型数据的高效传输问题；最后，搭建仿真系统，完成关键技术性能测试。关键技术的突破将提高空间信息网络中大容量数据的高速传输能力，为空间信息网络协议的研究提供技术储备。

针对空间信息网络中自相似、多类型业务的实时高效传输需求，基于高级在轨系统(AOS, Advanced Orbiting System)数据复用机制，开展了基于业务特征的高级在轨系统虚拟信道复用方法研究。首先，建立了AOS虚拟信道复用优化模型，为复用方法的研究提供技术架构；其次，研究了面向自相似业务的AOS队列管理方法，包括基于分解果蝇优化极限学习机的空间信息网络流量组合预测模型、基于流量预测的Hurst加权队列管理算法，以解决自相似业务的实时传输问题；再次，研究了面向多类型业务的AOS虚拟信道调度方法，包括适应多类型业务的紧急/同步/异步时隙可移动虚拟信道调度模型、基于深度Q网络的异步虚拟信道动态调度算法，以解决多类型数据的高效传输问题；最后，搭建了仿真系统，完成了关键技术性能测试。关键技术的突破可提高空间信息网络中大容量数据的高速传输能力，为空间信息网络协议的研究提供技术储备。