Flock is a desired state of unmanned aerial vehicle (UAV) systems, where multiple UAVs collaborate with each other in energy-efficient ways to complete mission-critical tasks while avoiding mutual collision. However, existing flocking control models still have defects such as potential node collision between followers, lack of capacity analysis, and difficulty in implementation. In this backdrop, this paper presents a model, called the diffusion model, to characterize the flock generation process of a UAV system; moreover, the system capacity of the diffusion model is analyzed. Then, a flocking control algorithm based on the diffusion model is detailed. Finally, a series of experiments based on both the OMNeT++ platform and a heterogeneous UAV-based prototype system are conducted to verify the effectiveness of our proposal. The experimental results show that our flocking control algorithm based on the diffusion model can make the UAV system behave as a flock while avoiding collision under diverse parameter settings.

Although dense interconnection datacenter networks (DCNs) (e.g., FatTree) provide multiple paths and high bisection bandwidth for each server pair, the wid

Traditional Internet protocol networks cannot provide the service of selecting a secure path to transmit various types of data with specific security constraints. First, to solve the “secure path transmission” problem, this paper first proposes a search-transmit model for secure network path transmission, i.e., to find a multi-attribute optimized path that meets the specific security requirements in the search phase, and then transmit the packets along the optimized path in the transmission phase. Second, we propose a solution to the search-transmit model. The idea of the solution is to use the particle swarm optimization algorithm to search for a secure path that meets the multi-attribute requirements and then set the source route on the router to control the packet transmission. Finally, a prototype system based on network function virtualization is developed to evaluate the feasibility and performance of the proposed solution. Experimental results show that the proposed solution outperforms existing algorithms in terms of performance.

It is necessary to understand the characteristics of network applications for network planning, network management and network security. This paper proposed a method to classify network applications by using BP neural network to parse network trace files. Firstly, six characteristics of network application flow are proposed by analyzing PCAP files. Secondly, based on the BP neural network model, its connection weights and thresholds are obtained through supervised training, and the architecture classification of network applications can be completed by inputting the PCAP datasets. We tested the proposed method on the PCAP datasets obtained from CERNET, and analyzed the influence of the network application flow characteristics on the BP neural network model. The experimental results show that the proposed method is a simple and practical method for network application classification.

Harsh or hostile environments may lead to node failure and connectivity degradation of a UAV swarm system. In order to maintain or restore the connectivity of a network in case of node failure, this paper proposes a mechanism to adjust the network topology to resist the impact caused by node failure. Firstly, a network model of a UAV swarm network based on k-hop constrained reachability is proposed. Secondly, a k-hop constrained reachability based proactive connectivity maintaining mechanism of UAV swarm network is presented. In this mechanism, each node identifies the network abnormality distributed according to k-hop reachability, and reports the observed abnormality to the master node; then, a virtual edge-based topology reconstruction algorithm is put forward for the master node to derive a topology reconstruction solution in a centralized way; afterwards, the solution is delivered to the slave nodes to reconfigure the network topology in parallel. Thirdly, a quantitative method is introduced to optimize the total travel distance of nodes, and a spanning tree-based method is designed to maintain the connectivity during the topology transformation process. Both theoretical analysis and simulation results have shown that: on the one hand, the proposed mechanism are effective in maintaining a UAV swarm’s connectivity in case of node failure; on the other hand, the proposed mechanism outperforms existing mechanisms in terms of fault tolerance, connectivity, and total travel distance, and it’s less affected by the failure rate.