A Loss Differentiation Method Based on Heterogeneous Ensemble Learning Model for Low Earth Orbit Satellite Networks.

In light of the high bit error rate in satellite network links, the traditional transmission control protocol (TCP) fails to distinguish between congestion and wireless losses, and existing loss differentiation methods lack heterogeneous ensemble learning models, especially feature selection for loss differentiation, individual classifier selection methods, effective ensemble strategies, etc. A loss differentiation method based on heterogeneous ensemble learning (LDM-HEL) for low-Earth-orbit (LEO) satellite networks is proposed. This method utilizes the Relief and mutual information algorithms for selecting loss differentiation features and employs the least-squares support vector machine, decision tree, logistic regression, and K-nearest neighbor as individual learners. An ensemble strategy is designed using the stochastic gradient descent method to optimize the weights of individual learners. Simulation results demonstrate that the proposed LDM-HEL achieves higher accuracy rate, recall rate, and F1-score in the simulation scenario, and significantly improves throughput performance when applied to TCP. Compared with the integrated model LDM-satellite, the above indexes can be improved by 4.37%, 4.55%, 4.87%, and 9.28%, respectively.

Intelligent Hierarchical Admission Control for Low-Earth Orbit Satellites Based on Deep Reinforcement Learning.

Low-Earth orbit (LEO) satellites have limited on-board resources, user terminals are unevenly distributed in the constantly changing coverage area, and the service requirements vary significantly. It is urgent to optimize resource allocation under the constraint of limited satellite spectrum resources and ensure the fairness of service admission control. Therefore, we propose an intelligent hierarchical admission control (IHAC) strategy based on deep reinforcement learning (DRL). This strategy combines the deep deterministic policy gradient (DDPG) and the deep Q network (DQN) intelligent algorithm to construct upper and lower hierarchical resource allocation and admission control frameworks. The upper controller considers the state features of each ground zone and satellite resources from a global perspective, and determines the beam resource allocation ratio of each ground zone. The lower controller formulates the admission control policy based on the decision of the upper controller and the detailed information of the users' services. At the same time, a designed reward and punishment mechanism is used to optimize the decisions of the upper and lower controllers. The fairness of users' services admissions in each ground zone is achieved as far as possible while ensuring the reasonable allocation of beam resources among zones. Finally, online decision-making and offline learning were combined, so that the controller could make full use of a large number of historical data to learn and generate intelligent strategies with stronger adaptive ability while interacting with the network environment in real time. A large number of simulation results show that IHAC has better performance in terms of a successful service admission rate, service drop rate, and fair resource allocation. Among them, the number of accepted services increased by 20.36% on average, the packet loss rate decreased by 17.56% on average, and the resource fairness increased by 17.16% on average.