25 January 2021 in IEEE Transactions on Automatic Control
This article studies the distributed optimal output agreement problem for multiagent systems described by uncertain nonlinear models. By using the partial information of an objective function, the design aims to steer the outputs of the agents to an agreement on the optimal solution to the objective function. To solve this problem, this article introduces distributed coordinators to calculate the desired outputs, and designs reference-tracking controllers for the agents to follow the desired outputs. To deal with the nonlinear uncertain dynamics, the closed-loop multiagent system is considered as a dynamical network, and Sontag's input-to-state stability is employed to characterize the interconnections. It is shown that output agreement in multiagent nonlinear systems is achievable by means of distributed optimal controllers via a small-gain approach. The proposed design features a three-layer architecture, and the reference-tracking controllers can be implemented as successive nonlinear proportional-integral loops. A numerical example is employed to show the effectiveness of the design.
23 July 2024 in IEEE Transactions on Industrial Electronics
This article presents a novel design algorithm for the cooperative source-seeking problem using a group of quadrotors equipped with sensors for source scalar field measurement with limited communications. The main objective is to drive the average position state of quadrotors to the optimal point of a given scalar field function while maintaining the desired formation, based on the field value measured by each quadrotor. To achieve this objective, the formation-based source-seeking algorithm is proposed with a cascade structure and the consensus filter, which is composed of a formation controller and a gradient-free optimization algorithm: the formation controller implements the desired formation and provides optimization direction information for the optimization algorithm, while the optimization algorithm generates a movement direction to guide the average position state of quadrotors collectively moving toward the source without estimating gradient, and the consensus filters estimate the centralized quantities. The Lyapunov stability analysis and invariancelike theorem are applied to establish the convergence and robustness of quadrotors’ positions and velocities to the formation and the average position to the optimal point. Finally, the effectiveness of the proposed method is illustrated by a physical experiment and a numerical simulation.
10 January 2022 in IEEE Transactions on Control of Network Systems
This article presents a composite fuzzy distributed control law to solve a cooperative control problem of fuzzy multiagent systems with a nonlinear leader system based on the distributed observer. The contributions of this article are three-fold. First, the small-gain theorem is extended to fuzzy systems, and input-to-output stability (IOS) and output-Lagrange input-to-output stability (OL-IOS) fuzzy small-gain theorem are proposed for fuzzy systems. Second, the OL-IOS small-gain theorem of fuzzy systems with multiple loop is given. Third, based on fuzzy small-gain theorem, the fuzzy distributed control law is composed of a purely fuzzy distributed control law and a distributed observer for the leader system. Finally, the effectiveness of the proposed method is illustrated by using a numerical example and the leader-following consensus problem of the quadrotor.
4 January 2022 in IEEE Transactions on Circuits and Systems II: Express Briefs
Based on the framework of distributed optimization algorithms, this brief studies distributed optimal formation behaviors of multi-agent systems modeled as a group of uncertain Euler-Lagrange systems with collision avoidance. Assume that each agent can only obtain measured position-based gradient value of a local objective function. By integrating a modified distributed optimization algorithm and an adaptive control law, the proposed distributed optimal formation controllers can lead the positions of agents asymptotically converging to the optimal solution to the total objective function with collision avoidance. The proposed design is verified by a numerical example of mobile relays for communication.
21 September 2022 in IEEE Transactions on Cybernetics
The distributed optimization problem (DOP) of Takagi–Sugeno (T–S) fuzzy cyber–physical systems is studied under the framework of weight-balanced graphs and quasistrongly connected characteristics. The objective is to drive the outputs of all agents to the optimal solution of a given global objective function regarded as the desired output, based on the partial information of the local objective functions. To this end, distributed optimal coordinators (DOCs) are used to generate optimal solutions of local objective functions that converge to the desired output, and fuzzy reference-tracking controllers are designed to ensure that all agents can track the optimal solutions. As novel technical results, two Lyapunov-based fuzzy input-to-state stability (ISS) small-gain theorems are proposed for the T–S fuzzy interconnected system. Thus, the overall closed-loop system is an interconnected system involving the modules of optimal coordinators and fuzzy tracking controllers with T–S fuzzy subsystems. The fuzzy ISS cyclic-small-gain theorem is applied to analyze the system stability. The DOP of T–S fuzzy cyber–physical systems is solved using the DOCs and fuzzy reference-tracking controllers through the fuzzy small-gain approach. A numerical example is presented to demonstrate the effectiveness and superiority of the proposed method.
14 December 2023 in IEEE Transactions on Automatic Control
This paper studies the distributed feedback optimization problem for nonlinear uncertain multi-agent systems subject to inequality constraints. A new class of distributed optimization algorithms is proposed by extending the standard primal-dual dynamics and introducing two new inputs to deal with the couplings arising from feedback optimization. With each controlled agent satisfying a mild dissipation assumption, the proposed distributed feedback optimization algorithms, using only the output-dependent gradient value of each agent's corresponding local objective function and the information from its neighboring agents, can steer the outputs of the agents to a common set-point which minimizes the total objective function while satisfying the inequality constraints. A composite Lyapunov function is constructed to prove global asymptotic stability of the closed-loop system at the equilibrium corresponding to the optimal point.