State Key Laboratory of Integrated Service Network Theory and Key Technology

The resource scheduling problem of UAV communication networks has always been a hot research issue in the current environment. However, most of the current research focuses on meeting the user’s time cost requirements and rarely considers the user’s security requirements in the scheduling process. To solve the problems of low efficiency, more resource fragments, low network reliability, and unbalanced load, this paper improves the virtual network mapping model into a three-layer structure consisting of a security service request layer, virtual network provider layer, and security resource layer, so that it can be applied to the special scenario of security resource scheduling in an unmanned aerial communication network. An intelligent UAV communication network security resource scheduling algorithm based on optimized subgraph isomorphism and link segmentation is proposed. Intelligent UAVs can predict all possible access nodes and intelligently provide network services. The algorithm first improves and enriches the node constraint conditions and evaluation indicators, integrates the node resource attributes and topology attributes, optimizes the sorting process, then calculates the mapping domain in advance, considers the regional comprehensive resource capacity to balance the load, and finally combines the subgraph isomorphism and link segmentation technology to carry out a one-stage mapping scheduling to achieve the security service request in the network and the efficient allocation of network resources required to provide security services. The simulation results of the OMNeT++ show that the proposed algorithm improves the mapping success rate and the long-term average cost-benefit ratio of the UAVs.

Based on the characteristics of game theory and wireless Ad Hoc networks,a novel concept of Incompletely Cooperative Game Theory(ICGT) is proposed to improve the performance of MAC protocols in wireless Ad Hoc networks.In ICGT,every node estimats the Game state by detecting the channel,and adjusts its equilibrium strategy.By repeating the game finitely,an optimal performance is achieved.As an example,the number of competing nodes is estimated by virtual DCF as the game state,and each node adjusts its equilibrium strategy by tuning the minimum contention window.Simulation results prove that ICGT is valid,the optimized saturated system throughput increasesby 7%~25%,while system delay decreasesby 20%~30%.