Graduation Year

2020

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Mechanical Engineering

Major Professor

Tansel Yucelen, Ph.D.

Committee Member

Eric Butcher, Ph.D.

Committee Member

Rajiv Dubey, Ph.D.

Committee Member

Kyle Reed, Ph.D.

Committee Member

Yasin Yilmaz, Ph.D.

Keywords

Cooperative Output Regulation, Global and Local Objectives, Internal Model, Synchronization Roles

Abstract

The overarching objective of this work is to propose solutions to quite a few distributed control problems arising from networks of heterogeneous agents or the heterogeneous nature of multiagent systems. Each problem with its solutions is concisely summarized below.

We consider the cooperative output regulation problem of heterogeneous linear multiagent systems over fixed directed communication graphs. The purpose of this problem is to design a distributed control law such that the overall closed-loop stability is ensured and the tracking error of each agent converges to zero asymptotically for a class of reference inputs and disturbances generated by a so-called exosystem. We investigate the solvability of the problem with internal model-based distributed control laws, namely dynamic state feedback, dynamic output feedback with local measurement, and dynamic output feedback. The approach is twofold: First, the overall closed-loop stability (i.e., global property) is assumed and it is shown, under mild assumptions, that the problem is solved. Second, an agent-wise local sufficient condition is derived to guarantee the global property under standard assumptions.

Then, we update the definition of the linear cooperative output regulation problem to allow not only common output synchronization among agents but also an additional output synchronization among a proper subset of the agents for a distributed dynamic state feedback control law that does not exchange its state variables through a communication graph. Similar to the above-mentioned approach, its solvability is investigated by making use of the internal model design from the linear output regulation theory and a small-gain theorem for large-scale interconnected systems.

This dissertation also focuses on distributed control of linear multiagent systems with both global and local objectives over fixed directed communication graphs. While the global objective is achieving leaderless synchronization (i.e., consensus) or synchronization to a leader, local objectives for a subset of agents are tasks determined by agent-specific dynamical systems around the synchronization mapping of the global objective. Our main goal is to design a distributed control law such that each agent obeys the global objective when it is not assigned the local task and performs its own local objective otherwise. To this end, we construct reference models for all agents via two existing synchronization results, introduce two classes of distributed controllers, and formally define the considered problems. Then, they are solved by utilizing the converging-input converging-state property for a class of linear systems and the feedforward design approach from the linear output regulation theory.

Share

COinS