Time-varying formation tracking for noisy multi-agent systems with unknown leader trajectory

• Published in: Proceedings of the Institution of Mechanical Engineers. Part I, Journal of systems and control engineering Journal of systems and control engineering, 2022-08, Vol.236 (7), p.1386-1399
• Main Authors: Zaeri Amirani, Mojtaba, Bigdeli, Noshin, Haeri, Mohammad
• Format: Article
• Language: English
• Subjects: Control theory; Data integration; Feedback control; Kalman filters; Mechanical engineering; Multiagent systems; Obstacle avoidance; Reference signals; State feedback; State observers; Tracking problem; Unmanned aerial vehicles
• ISSN: 0959-6518; 2041-3041
• DOI: 10.1177/09596518221083717

In this article, time-varying formation tracking problem for multi-agent systems is investigated. It is assumed the number of agents is changing and obstacles should be avoided along the trajectory. In addition, the tracking formation should be reached while the leader trajectory is unknown and communication data are noisy. To reach the desired formation tracking, first, a virtual leader is constructed using the local neighbouring data. Second, the reference signal is created based on the virtual leader and the leader data (which is available for some agents). The data fusion at this stage is done based on Kalman filtering to reduce the effect of communication noise. Third, an integral state feedback controller as a new consensus formation controller based on the relative position and velocity vectors of the agents and the estimated (virtual) leader are proposed to achieve the time-varying formation tracking. Next, closed-loop stability of the system under the proposed control law is investigated using the Lyapunov stability theorem. In order to implement the proposed control strategy, an extended state observer is designed to estimate the required velocity and accelerations. Finally, an obstacle avoidance strategy would be embedded to the controller via a repelling strategy. That is, once the leader identifies the obstacle and enters a repulsion region, its reference trajectory is modified to keep a safe distance from the obstacle. The agents modify their reference trajectory according to the new leader trajectory while trying to keep their formation. As an illustrative example, various simulations are performed for formation tracking of a unmanned aerial vehicle multi-agent system to evaluate its performance.