RPCA-Based UAV Trajectory Restoration in Strong-Interference Environment

Addressing unmanned swarm threats to low-altitude security requires use of a UAV monitoring system to obtain trajectory data for intent inference and threat assessment. However, the high mobility of unmanned swarms, susceptibility to deceptive interference, and the issue of data scale compromise the quality and reliability of trajectory data. Thus, trajectory restoration is required to enhance the quality and reliability of the compromised trajectory data. We consider this problem in this paper and make the following contributions. Firstly, we introduce RPCA to remove outliers based on the linearity of the trajectory data, mitigating the impact of highly maneuverable targets with varying motion models. Secondly, the parallel and decentralized characteristics of Symmetric Alternating Direction Method of Multipliers (S-ADMM) make it suitable for processing large-scale data. So we employ S-ADMM for processing large-scale data to enhance computational efficiency. Thirdly, the numerical results demonstrate the proposed method can improve the quality and reliability of trajectory data.