Smart Anti-jamming Mobile Communication for Cloud and Edge-Aided UAV Network

2020 KSII Transactions on Internet and Information Systems  
The Unmanned Aerial Vehicles (UAV) networks consisting of low-cost UAVs are very vulnerable to smart jammers that can choose their jamming policies based on the ongoing communication policies accordingly. In this article, we propose a novel cloud and edge-aided mobile communication scheme for low-cost UAV network against smart jamming. The challenge of this problem is to design a communication scheme that not only meets the requirements of defending against smart jamming attack, but also can be
more » ... deployed on low-cost UAV platforms. In addition, related studies neglect the problem of decision-making algorithm failure caused by intermittent ground-to-air communication. In this scheme, we use the policy network deployed on the cloud and edge servers to generate an emergency policy tables, and regularly update the generated policy table to the UAVs to solve the decision-making problem when communications are interrupted. In the operation of this communication scheme, UAVs need to offload massive computing tasks to the cloud or the edge servers. In order to prevent these computing tasks from being offloaded to a single computing resource, we deployed a lightweight game algorithm to ensure that the three types of computing resources, namely local, edge and cloud, can maximize their effectiveness. The simulation results show that our communication scheme has only a small decrease in the SINR of UAVs network in the case of momentary communication interruption, and the SINR performance of our algorithm is higher than that of the original Q-learning algorithm. 4683 gradually become the focus of researchers. In [8], the authors used a multi-agent Q-learning algorithm called femto-based distributed and sub-carrier-based distributed power controls using Q-learning (FBDPC-Q) to deal with the aggregate macrocell and femtocell capacities. Considering the impact of the channel estimation error, a Q-learning based power control algorithm using non-cooperative game theory was proposed to suppress the joint smart jamming attack [9] . In [10], the authors used UAVs to relay the messages and improve the communication performance of VANETs with a Q-learning based scheme. The application areas of Q-learning algorithms continue to expand. In [11] , the authors solve the anti-jamming problem of UAV radar network based on the double greedy Q-learning algorithm. The Q-learning algorithm associated with an onedevice federated jamming detection mechanism [12] played an important role in the defense against smart jamming attack in FANET. However, the shortcomings of the table-driven Q-learning algorithm have gradually emerged. In the multi-agent system anti-intelligence interference problem, the state space of the problem to be solved is very large. Consider a typical communication system composed of multiple UAVs, where each UAV's receiver has multiple channels and the UAVs can perform autonomous maneuvering. The high dimensionality of the state space of the system can easily cause dimensional disasters, making the Q-learning algorithm unable to work. The combination of deep learning technology and reinforcement learning technology can perfectly solve this problem. The authors in [13] proposed a hotbooting deep Q-network based 2-D mobile communication scheme to address the smart jamming problem. The algorithm is deployed in a robot that can move on the ground, not in a UAV platform. In [14] , a deep Q-learning-based UAV power allocation strategy combines Q-learning and deep learning was proposed, but this research only studied the scenario of one UAV rather than UAV network. Communication model As shown in Fig. 2 , the communication model consists of three parts. The first part is UAV network, and its network topology is hierarchical mesh network. The backbone mesh network consists of backbone UAVs. The backbone UAVs have strong communication capabilities and a long distance for sending and receiving information. Each backbone UAV establishes a homeland area around it, and there are several small and low-cost mission UAVs in the area. Cross-domain communication of mission UAVs requires relay services from backbone UAVs. GPS is responsible for providing timing and positioning services for UAVs. Jamming UAVs can approach backbone UAVs and mission UAVs and launch smart jamming attack based on cognitive radio technology. The second part is the ground station network. The ground stations maneuver within the communication range of the backbone UAV, and use data link technology to maintain communication with the backbone UAV. Ground stations maintain communication through microwave relays. Since the ground station is far from the mission area, it is assumed that the communication link between the ground stations cannot be jammed. The working area of mission UAVs is supported by base stations. The third part is the cloud computing center. The UAV network communicates with the cloud computing center through the base station with 4G or 5G cellular communication technology. The base station is connected to the cloud computing center through an optical fiber link. The base station is in an urban environment with a complex electromagnetic environment, and it is difficult to effectively control jamming sources in the city. Therefore, it is assumed that the air interface links between the base stations and the UAVs will be interfered with and intermittently interrupted. The links between the base stations are fiber-optic and are not subject to interference.
doi:10.3837/tiis.2020.12.004 fatcat:osdb3en4freebgfxzzghzevxma