Core Concepts
A novel approach utilizing Graph Convolutional Networks (GCN) to predict jamming areas and enable multi-UAV swarms to efficiently navigate around them and reach target destinations.
Abstract
This paper addresses the challenge of jamming technology disrupting the operations of UAV (Unmanned Aerial Vehicle) swarms. It proposes a novel approach that leverages the collective intelligence of UAV swarms to predict the location and intensity of jamming signals, and then employs a multi-agent control algorithm to disperse the swarm, avoid the jamming areas, and regroup upon reaching the target destination.
The key highlights of the approach are:
Prediction of jamming areas: The paper utilizes GCN to gather information from each UAV, including their current position, velocity, and the probability and rate of change of jamming signal impact. This data is used to accurately predict the location and intensity of the jamming area.
Anti-jamming path planning: Based on the predicted jamming information, the multi-agent control algorithm generates paths for the UAVs to disperse, avoid the jamming areas, and regroup at the target destination. This leverages the physical movement capabilities of the UAVs to mitigate the effects of jamming.
Robustness and scalability: The approach is designed to be effective even when the jammer's location is unknown, making it widely applicable. The use of GCN and multi-agent control also enables scalability to handle larger UAV swarms.
Computational efficiency: The distributed nature of the algorithm, with each UAV contributing to the prediction and path planning, reduces the computational burden compared to centralized approaches.
Through simulations, the paper demonstrates the effectiveness of the proposed method in accurately predicting jamming areas and successfully guiding the UAV swarm to the target destination while avoiding the jamming zones.
Stats
The probability P of a UAV experiencing communication disruption due to the jammer is inversely related to the distance r between the center of the jamming area and the UAV, as described by the equation: P = kA/r, where k and A are constants that vary depending on the signal strength of the jammer.
The distance r_tau between the jammer and the UAV when P reaches the threshold value P_tau is given by: r_tau = log_A(P_tau/k).
Quotes
"Even in situations where the information about the jammer's location and intensity is unknown in advance, this method can predict them based on the given information, making it robust and widely applicable."
"By leveraging the algorithm structure of GCN and utilizing collective intelligence from the information each UAV possesses, this framework achieves accuracy in prediction."
"By utilizing the physical movement of UAVs to evade jamming, effective anti-jamming is achieved."