Collective Iterative Learning Control: Exploiting Diversity in Multi-Agent Systems for Reference Tracking Tasks

Multi-agent systems (MASs) can autonomously learn to solve previously unknown tasks by means of each agent's individual intelligence as well collaborating and exploiting collective intelligence. This article considers a group autonomous agents learning track the same given reference trajectory in possibly small number trials. We propose novel control method that combines iterative (ILC) with update strategy. derive conditions for desirable convergence properties such systems. show proposed allows combine advantages agents' strategies thereby overcomes trade-offs limitations single-agent ILC. benefit is achieved designing heterogeneous collective, i.e., different law assigned agent. All theoretical results are confirmed simulations experiments two-wheeled-inverted-pendulum robots (TWIPRs) jointly perform desired maneuver.