Saman Khodaverdian

Robust observer-based fault detection and isolation in the standard control problem framework

Abstract This paper deals with the design of robust fault detection and isolation observers where only a single observer is employed to isolate different faults. To this end, the problem of parameterizing such observers is shown to be equivalent to designing a structurally constrained controller in the standard control problem framework. Thereby, the problem is reformulated as a well known classical control problem, which enables the use of existing tools to optimize robustness with respect to arbitrary exogenous disturbances. To account for parametric uncertainties, an approximate model matching approach is used.

Motion control for UAV-UGV cooperation with visibility constraint

The subject of this paper is a cooperative scenario where an unmanned aerial vehicle (UAV) assists an unmanned ground vehicle (UGV) to achieve a higher level task. We assume that the UAV detects the UGV within a certain range with a vision sensor. While satisfying this constraint, the UAV can provide relative information about any events of interest in its field of view. The focus is on the motion control of the UAV relative to the UGV. To this end, we present a new control law that combines an extended dynamic coverage strategy with a tracking controller. Different weighting functions to combine the controllers are evaluated. The results are further compared with a second motion strategy introduced for this purpose. Both control laws guarantee that the UAV will maintain visual contact while making use of available mobility as much as possible, allowing the UAV to gather more information and act as an anticipatory element.

Robust output synchronization for heterogeneous multi-agent systems based on input-output decoupling

This paper deals with the problem of robust output synchronization for heterogeneous multi-agent systems. We propose a method to homogenize the output dynamics of every agent, which is based on noninteracting control techniques. Hence, the problem is reduced to a simple synchronization task for homogeneous systems. Moreover, we present an approach to reformulate the homogenization purpose as a problem of finding a structurally constrained controller in the standard control problem framework, and thus use well established tools to increase the robustness of the agents against exogenous disturbances. The robust output synchronization of a heterogeneous aircraft fleet illustrates the effectiveness of the approach.

Synchronizing networks of heterogeneous linear systems via input-output decoupling

In this article, a new approach for solving the synchronization problem for networks of heterogeneous linear systems is presented. The main idea is to decouple the states to be synchronized for every agent from the rest of the system dynamics, leading to an input-output decoupling control law. We present a decentralized control strategy for homogenizing the transient behaviors of the agents. Further, for every agent the same reference signal, which can be generated by identical exosystems, is applied leading to synchronization. A numerical example illustrates the efficiency of the design procedure.

Low order controllers for output synchronization of linear heterogeneous SISO agents

This paper considers the output synchronization problem for linear heterogeneous SISO agents. The goal is not only to synchronize the agents to a given trajectory, but also to achieve this goal by distributed controllers of low order dynamics. We prove that under mild assumptions it is possible to synchronize the agents and to limit the order of the controllers to an upper bound. Moreover, a special condition is derived under which it is even possible to synchronize the agents by means of static controllers. The efficiency of the approach is illustrated by a numerical example.

A distributed analysis and design method for the synchronization of linear heterogeneous SISO systems by static state control

We present a design scheme for the synchronization of linear heterogeneous SISO systems by static feedback controllers. Extending some recent results, a distributed method is provided for synchronizing non-identical agents by a simple synchronization controller for homogeneous agents. In [4], it was shown that for a special condition heterogeneous agents can be synchronized by static controllers. In this paper, we present a constructive approach to analyze the satisfiability of this condition in a distributed manner and to design an appropriate controller.

Fast synchronization with directed small-world networks

In different articles it is shown that multi-agent synchronization can be faster under certain undirected network topologies. In this paper, we present an algorithm which generates strongly connected directed small-world networks. The main goal is a further enhancement of convergence speed. The generated graphs solve the consensus problem significantly faster than regular networks with the same number of links. However, the produced structures are generally not balanced, i.e. the out-degree remains constant, while the in-degree can vary. In this case an averaging consensus will not be reached. To solve this inconvenience, we propose a heuristic sorting algorithm to balance the generated directed small-world networks. In addition, balancing has yet another positive influence on the convergence speed. The result is a balanced, directed small-world network which solves the consensus problem faster than undirected small-world networks.

Synchronizing Linear Heterogeneous Networks by Output Homogenization

Abstract This article deals with the synchronization of heterogeneous multi-agent systems. A simple distributed control law is provided to solve the problem. The presented approach is based on the idea to ascribe the heterogeneous case to a homogeneous synchronization task, for which the solution is well known. The presented method is applicable to networks of completely different agents, even with different dimensions. Additionally, the communication effort is minimized because only the states which are to be synchronized need to be exchanged through the network. An example illustrates the efficiency of the concept.

Full and partial input-output decoupling of square and overactuated systems: A transformation-based approach

A transformation-based approach for the full and partial input-output decoupling of square and over actuated linear systems is presented. A certain transformation is introduced which highlights the coupling effects of the system. Based on the transformed system matrices, it can simply be seen how full and partial decoupling is derived. Additionally to the square case, an analysis for overactuated systems is given and it is highlighted how the degrees of freedom coming from overactuation can be used to manipulate the decoupling result. The contribution of the paper is to present an easy to handle decoupling technique which enables a simple access to input-output decoupling and visualizes all degrees of freedom.

Vollständige und teilweise Ein-/Ausgangsentkopplung: Ein Transformationsansatz

Zusammenfassung Der vorliegende Beitrag befasst sich mit der Ein-/Ausgangsentkopplung quadratischer und überaktuierter, linearer Systeme mittels statischer Zustandsrückführung. Hierfür wird eine spezielle Transformation genutzt, mit deren Hilfe sich die Koppeleffekte des Systems einfach veranschaulichen lassen. Neben einem unkomplizierten Zugang zur vollständigen und teilweisen Entkopplung kann hierdurch ein direkter Zusammenhang zwischen den Reglerparametern und den resultierenden Übertragungsfunktionen hergestellt werden. Im Falle überaktuierte Systeme, d.h. Systeme mit mehr Ein- als Ausgangsgrößen, ermöglicht die verwendete Transformation eine geschickte Veranschaulichung der zusätzlichen Freiheitsgrade und dadurch eine gezielte Manipulation der Systemdynamik über die zusätzlichen Aktoren.