Cooperative Output Regulation of MASs With a Novel Parameter Convergence Condition and Prescribed-Time Event-Triggered Observers and Its Application to Robot Formation

An enhanced control scheme for multiagent cooperative output regulation (COR) is presented in this paper. The proposed scheme incorporates prescribed-time observers and controllers with concurrent learning adaptive laws. By introducing prescribed-time event-triggered (ET) observers, the estimation of exosystem information can be achieved within a prescribed time, eliminating continuous communication between followers and avoiding reliance on global information. Additionally, Zeno behavior is excluded. The convergence of controller parameters is ensured by the established concurrent learning adaptive laws. Specifically, the challenge of designing correlation filter signals caused by the lack of a matching condition between the exosystem matrix and the follower system matrix, essential for concurrent learning, is addressed by employing the regulator equation. The desired relationship between two sets of developed filtered signals is ensured by the developed prescribed-time observers after a prescribed time. This establishes the basis for a concurrent learning strategy. Moreover, a cooperative controller with cooperative adaptive laws is developed, effectively addressing the difficulty arising from the mismatch between the matrix estimation of the exosystem and the regulator equation. Compared with existing studies, the developed controller can accurately identify controller parameters, resulting in better system performance. The successful application of the cooperative finite-time excitation (CFTE) condition relaxes the widely used persistent excitation (PE) and cooperative persistent excitation (CPE) conditions in parameter identification. Finally, a multi-robot formation simulation verifies the effectiveness of the developed method.
Note to Practitioners
—The COR problem explored in this paper can be applied to address the actual robot formation issue, utilizing coordinate transformation. Unlike existing studies, this paper employs easily satisfied cooperative persistent excitation conditions, making them readily applicable to real-world systems. The developed approach enables the convergence of the designed controller parameters to their true values, achieving superior control effects. Moreover, this paper overcomes some current research difficulties. It successfully integrates the most advanced event-triggered mechanism (ETM) into the prescribed-time observer, realizing the prescribed-time observation of the exosystem. Uniquely, the method presented here does not depend on global information, making it suitable for complex large-scale multiagent connections and the Internet of Things. It holds significant guidance for practical multiagent networked systems. The effectiveness of the proposed method is demonstrated through a nonholonomic dynamic robot formation problem. It is anticipated to be scalable to multiagent systems (MASs) with broader practical constraints, finding applications in more realistic engineering scenarios.