Resilient Time-Varying Formation Tracking for Mobile Robot Networks under Deception Attacks on Positioning

TL;DR

This paper develops resilient control and localization algorithms for mobile robot networks to maintain formation tracking despite deception attacks on positioning, combining stability analysis, attack detection, and recovery strategies validated through simulations and experiments.

Contribution

It introduces a novel decentralized localization method and a gain-tuning resilient control framework to enhance robustness against deception attacks in mobile robot formations.

Findings

Proposed algorithms maintain formation under deception attacks.

Lyapunov analysis confirms system stability and bounded errors.

Experimental results validate effectiveness with quadrotor teams.

Abstract

This paper investigates the resilient control, analysis, recovery, and operation of mobile robot networks in time-varying formation tracking under deception attacks on global positioning. Local and global tracking control algorithms are presented to ensure redundancy of the mobile robot network and to retain the desired functionality for better resilience. Lyapunov stability analysis is utilized to show the boundedness of the formation tracking error and the stability of the network under various attack modes. A performance index is designed to compare the efficiency of the proposed formation tracking algorithms in situations with or without positioning attacks. Subsequently, a communication-free decentralized cooperative localization approach based on extended information filters is presented for positioning estimate recovery where the identification of the positioning attacks is based on Kullback-Leibler divergence. A gain-tuning resilient operation is proposed to strategically synthesize the formation control and cooperative localization for accurate and rapid system recovery from positioning attacks. The proposed methods are tested using both numerical simulation and experimental validation with a team of quadrotors.