Abstract

Motivated by recent developments in formation and cooperative control of networked multi-agent systems, the main goal of this thesis is development of efficient synchronization and formation control algorithms for distributed control of networked nonlinear systems whose dynamics can be described by Euler-Lagrange (EL) equations. One of the main challenges in the design of the formation control algorithm is its optimality and robustness to parametric uncertainties, external disturbances and ability to reconfigure in presence of component, actuator, or sensor faults. Furthermore, the controller should be capable of handling switchings in the communication network topology.

In this work, nonlinear optimal control techniques are studied for developing distributed controllers for networked EL systems. An individual cost function is introduced to design a controller that relies on only local information exchanges among the agents. In the development of the controller, it is assumed that the communication graph is not fixed (in other words the topology is switching). Additionally, parametric uncertainties and faults in the EL systems are considered and two approaches, namely adaptive and robust techniques are introduced to compensate for the effects of uncertainties and actuator faults.

Next, a distributed H_infinity performance measure is considered to develop distributed robust controllers for uncertain networked EL systems. The developed distributed controller is obtained through rigorous analysis and by considering an individual cost function to enhance the robustness of the controllers in presence of parametric uncertainties and external bounded disturbances. Moreover, a rigorous analysis is conducted on the performance of the developed controllers in presence of actuator faults as well as fault diagnostic and identification (FDI) imperfections.

Next, synchronization and set-point tracking control of networked EL systems are investigated in presence of three constraints, namely, (i) input saturation constraints, (ii) unavailability of velocity feedback, and (iii) lack of knowledge on the system parameters. It is shown that the developed distributed controllers can accomplish the desired requirements and specification under the above constraints.

Finally, a quaternion-based approach is considered for the attitude synchronization and set-point tracking control problem of formation flying spacecraft. Employing the quaternion in the control law design enables handling large rotations in the spacecraft attitude and, therefore, any singularities in the control laws are avoided. Furthermore, using the quaternion also enables one to guarantee boundedness of the control signals both with and without velocity feedback.