The workshop will be held on July 9th. The tentative schedule for the workshop, along with the talk abstracts, is as follows:
1:00 — 1:10 pm
Speaker: Selahattin Burak Sarsılmaz, Utah State University
Title: Welcome
1:10 — 1:50 pm
Speaker: Jie Huang, The Chinese University of Hong Kong
Title: An Overview of the Cooperative Output Regulation of Multi-agent Systems and its Applications
Abstract: The cooperative output regulation problem is an extension of the classical output regulation problem from a single system to a multi-agent system. The problem aims to design a distributed control law for a multi-agent system composed of multiple follower systems and a leader system so that the output of each follower system asymptotically tracks a class of reference inputs and rejects a class of disturbances generated by the leader system. The problem includes several cooperative control problems such as consensus/synchronization, formation, and containment as special cases. Since 2010, the problem has been studied by the distributed observer approach and the distributed internal model approach, respectively. The distributed observer approach is an extension of the classical feedforward control approach and is able to handle the cooperative output regulation problem for general heterogeneous linear multi-agent systems and some weak nonlinear multi-agent systems over (disconnected) switching networks. However, as this approach relies on the solution of the regulator equations, it is less capable in dealing with the uncertainty of the system by itself. On the other hand, the distributed internal model approach is an extension of the classical internal model approach and is effective in dealing with the strong nonlinearity and uncertainty of the system. Nevertheless, as this approach makes use of the virtual output of the system which requires the follower network to be connected for all time, it cannot handle disconnected switching networks by itself. In this talk, we will first give an overview of the cooperative output regulation problem by both the distributed observer approach and the distributed internal model approach. Then we further present an approach that integrates the distributed observed approach and the distributed internal model approach, and is thus capable of handling the strong nonlinearity and uncertainty of multi-agent systems over switching networks. Applications of the cooperative output regulation problem to a variety of other cooperative control problems such as consensus, formation, flocking, and rendezvous will also be highlighted.
1:50 — 2:00 pm: Break
2:00 — 2:40 pm
Speaker: Changran He, The Chinese University of Hong Kong
Title: Cooperative Output Regulation over Jointly Connected Switching Networks
Abstract: A distributed dynamic compensator is called a distributed observer for the leader system if it can provide each follower with an estimate of the state of the leader. Ascertaining the existence of a distributed observer is a key step in applying the distributed observer-based approach. In this talk, we will first introduce a distributed observer for a linear leader system over jointly connected switching networks. Based on this distributed observer, we will synthesize a distributed control law to solve the cooperative output regulation problem for a continuous-time linear multi-agent system. We will also discuss extensions of the distributed observer from state-based to output-based.
2:40 — 3:20 pm
Speaker: Yamin Yan, Nanyang Technological University
Title: Cooperative Output Regulation of Discrete-time Multi-agent Systems by the Distributed Observer Approach
Abstract: The cooperative output regulation problem, considered one of the foundational challenges in cooperative control, has gained substantial attention within the research community over the past decade. Notably, the distributed observer-based approach has emerged as an effective control strategy for addressing this problem. In this talk, we delve into the cooperative output regulation problem within the context of discrete-time linear multi-agent systems, applying both a distributed observer approach and an adaptive distributed observer approach. Our investigation considers various challenges, including time-delays and unknown leader dynamics. The examination reveals intriguing distinctions between the cooperative output regulation problem in discrete-time systems and its continuous-time counterpart, emphasizing the need for a comprehensive and dedicated study in this domain.
3:20 — 3:30 pm: Break
3:30 — 4:10 pm
Speaker: Selahattin Burak Sarsılmaz, Utah State University
Title: Solvability of the Cooperative Robust Output Regulation with p-copy Internal Model Approach: Local and Global Design Methods
Abstract: The celebrated internal model approach has been a powerful approach to tackling the robust output regulation of a sole system since the late 1970s. The natural follow-up question is whether this can be extended to emerging multi-agent systems. For heterogeneous (in dynamics and dimension) linear multi-agent systems over general directed graphs, we first present the solvability of the cooperative robust output regulation problem with p-copy internal model-based distributed control laws. Though the overall dynamics of the multi-agent system is more complicated than a sole system, the fundamental lemma for the p-copy internal model approach is extended to the heterogeneous multi-agent systems. It is fascinating to see that the only additional assumption in this result is related to graph connectivity. Moreover, the extended lemma is key to the design of any p-copy internal model-based distributed control laws. We then provide agent-wise local sufficient conditions so that the cooperative robust output regulation problem boils down to stabilizing augmented exosystem-free dynamics for each agent with a disturbance attenuation level determined by a spectral property of the graph. This enables an independent control design for each agent and, hence, a level of scalability. However, this agent-wise local design method does not incorporate transient response requirements, for instance, minimal decay rate and damping ratio, into the design. To gain the capability of improving the overall performance, a structured Lyapunov inequality, a convex formulation of the overall closed-loop stability, is further presented for the distributed dynamic state feedback control law by considering the networks of systems as a whole. The existence of a solution to this linear matrix inequality is ensured if there exist control parameters satisfying the aforementioned agent-wise local conditions for every agent. While the proposed linear matrix inequality yields performance guarantees, it lacks scalability. We conclude this talk by identifying the advantages and disadvantages of both local and global design methods, where the pros of both approaches will be attained in the following talk.
4:10 — 4:50 pm
Speaker: Ahmet Taha Koru, University of Texas at Arlington
Title: Regional Eigenvalue Assignment as a Scalable Design Method for p-copy Internal Model Approach over Fixed and Switching Graphs
Abstract: A regional eigenvalue assignment method for cooperative output regulation of heterogeneous linear multi-agent systems with a p-copy internal model-based distributed dynamic state feedback control law is a scalable one with a closed-loop performance guarantee. This method offers an agent-wise local design method to synthesize the distributed control gains while assigning, for example, the minimal decay rate and the minimal damping ratio of the overall closed-loop system as desired. Numerical examples demonstrate the efficacy of the proposed method by assigning the eigenvalues of a large-scale system to different regions, specifically, disks, shifted half-planes, and conic sectors. Furthermore, the regional eigenvalue assignment is useful to provide scalable stability analysis and design methods for the cooperative linear output regulation problem over switching graphs regarding two classes of switching signals, namely arbitrary switching signals and slow switching signals subject to an average dwell time constraint.
4:50 — 5:00 pm: Concluding Remarks
Cooperative Output Regulation Workshop at ACC 2024 