Formation Control of Uncertain Multi-agent Systems with Non-Identical Actuation Capacities
Atahan Kurttisi, N. Eren Sarioglu, K. Merve Dogan · 2024
Multi-agent systems offer low-cost and flexible solutions for complex tasks thanks to their advantages and efficient algorithms. One of the most important features of multi-agent systems is operating in a formation for example, carrying a slung-load by using several multi-copters or platooning autonomous trucks. However, building control algorithms to provide formation control is not straightforward as real physical systems are affected by non-linear uncertainties and non-identical actuator-related issues. Being inspired by the nature of the real physical systems, in this research, we focus on designing distributed adaptive control algorithms to obtain formation control for multi-agent systems with uncertainties, unknown actuation efficiency, and actuator dynamics. Specifically, we utilize a leader-follower distance-based consensus algorithm together with adaptive terms to cancel out unknown and undesired effects. To do this, we use hedging-based reference models to decouple adaptation from the actuator dynamics. We analyze the stability of the multiagent system by using the Lyapunov Stability Theory, and investigate feasible actuator bandwidth limits with the Linear Matrix Inequalities (LMI) method. Finally, we show the efficacy of the designed control law on a multi-agent system with both on undirected line graph and a fully connected graph. In addition, the results of this paper is compared with the previous results given in [1].