Luis Alvergue

Output consensus control for heterogeneous multi-agent systems

We study distributed output feedback control of a heterogeneous multi-agent system (MAS), consisting of N different continuous-time dynamic systems. Conditions on the existence of distributed output feedback control protocols are derived and a solution is proposed to synthesize the stabilizing and consensus control protocols over directed graphs. The results in this paper complement the existing ones, and are illustrated by a numerical example.

Consensus Control for Heterogeneous Multiagent Systems

We study distributed output feedback control for a heterogeneous multi-agent system (MAS), consisting of N different continuous-time linear dynamical systems. For achieving output consensus, a virtual reference model is assumed to generate the desired trajectory that the MAS is required to track and synchronize. A distinct feature of our results lies in the local optimality and robustness achieved by our proposed consensus control algorithm. In addition our study is focused on the case when the available output measurements contain only relative information from the neighboring agents and reference signal. Indeed by exploiting properties of strictly positive real (SPR) transfer matrices, conditions are derived for the existence of distributed output feedback control protocols, and solutions are proposed to synthesize the stabilizing and consensus control protocol over a given connected digraph. It is shown that design techniques based on the LQG, LQG/LTR and H-infinity loop shaping can all be directly applied to synthesize the consensus output feedback control protocol, thereby ensuring the local optimality and stability robustness. Finally the reference trajectory is required to be transmitted to only one or a few agents and no local reference models are employed in the feedback controllers thereby eliminating synchronization of the local reference models. Our results complement the existing ones, and are illustrated by a numerical example.

Secure networked control systems against replay attacks without injecting authentication noise

This paper studies detection of replay attacks on networked control systems, assuming that the actuation and sensing signals are transmitted over an additive white Gaussian noise channel. It explores the use of the spectral estimation technique to detect the presence of a replay attack without injecting authentication noise to the control signal at the plant input. Our proposed detection method is applicable to networked feedback systems equipped with stable controllers and designed with classical and modern control system techniques. A numerical example from the literature is used to illustrate the detection results for replay attacks on networked control systems.

Feedback Stabilization of a Reduced-Order Model of a Jet in Crossflow

Airfoils in gas turbine engines used for aeropropulsion are actively cooled by film cooling; and, to cost-effectively optimize or control the film-cooling behavior, reduced-order models combined with a suitable control strategy are needed. Motivated by this film-cooling application, a linear state feedback controller for stabilization of a proper orthogonal decomposition-based reduced-order model of a film-cooling jet injected over the airfoil surface into a hot crossflow is presented. A full-order direct numerical simulation of the jet in crossflow is used to generate the data required to build the reduced-order model. The control objective is to prevent the mixing of the coolant jet and hot gases from the combustor so that the coolant remains close to the blade surface. The kinetic energy of the turbulent flow is used as a measure of mixing, which in turn is used to guide the design of a desired steady-state condition with low kinetic energy. The linear controller guarantees that the nonlinear reduced-o...

A consensus approach to dynamic state estimation for smart grid power systems

We propose a distributed dynamic state estimator based on the multi-agent system (MAS) framework and on the availability of supervised control and data acquisition (SCADA) systems. We assume that the power system is modeled by a linear DC power flow model and is operating in a quasi-steady state. The phase angle at the network buses is modeled as a heterogeneous MAS with discrete time dynamics. Conditions on the existence of a distributed estimator are derived and a solution is proposed to construct the estimator gain.

Feedback control for quadratic nonlinear systems

We propose two state feedback control strategies based on quadratic stability for quadratic nonlinear systems, specifically reduced order fluid flow models. Stabilizing linear controllers are designed at each operating point, but unlike traditional linear designs, the nonlinear dynamics are taken into account during the design process by representing them as elements of a bounding set. By solving a set of linear matrix inequalities an estimate of the region of attraction at each operating point of the closed loop system is presented. The overall design consists of a sequence of linear controllers that regionally stabilize and enlarge the region of attraction of the desired operating points. Under the assumption that the system admits a centered-ε-cover, a switching strategy is proposed that achieves semiglobal stabilization of the nonlinear system.

Output consensus control for multi-agent systems in the presence of gap metric uncertainties

Consensus control is studied for discrete-time multi-agent systems involving gap metric uncertainties, aimed at synchronizing all its outputs to the desired trajectories generated by a given reference model. Both distributed robust stabilization and robust consensus are formulated and tackled. Results in this paper include derivation of the explicit conditions on the agent dynamics, on the bound of the gap metric uncertainty, and on the feedback graph, and derivation of the synthesis procedure for designing the distributed robust feed-forward and feedback controllers that achieve robust stability and consensus in the presence of bounded gap metric uncertainties.

Stabilization of feedback control systems over signal-to-noise ratio constrained channels

In this paper, we study feedback stabilization of a linear time-invariant (LTI) multi-input/multi-output (MIMO) discrete-time system with information constraints in both the input and output channels, characterized by the additive white Gaussian noise (AWGN). A signal-to-noise ratio (SNR) constraint model is adopted in studying its communication capacity limits for the underlying MIMO feedback control system. It has been proven that the computation of the minimum channel capacity is equivalent to a constrained ℋ2 control problem. We propose an iterative algorithm to compute the required channel capacity for MIMO feedback stabilization and the corresponding MIMO stabilizing feedback controller. The convergence of the proposed iterative algorithm is shown and demonstrated by a numerical example.

A generalized sector bound approach to feedback stabilization of nonlinear control systems

This paper proposes and develops a generalized sector bound approach to feedback stabilization of nonlinear control systems described by state-space models. This approach is inherited from the methodology of the sector bounded or passive nonlinearities, and influenced by the concept of absolute stability. It aims not only to regionally stabilize the nonlinear dynamics asymptotically but also to maximize the estimated region of quadratic attraction and to ensure nominal performance at each equilibrium. Similar to gain scheduling control, a path of equilibria is programmed based on the assumption of centered-ϵ-cover which leads to a sequence of linear controllers that regionally stabilize the desired equilibrium asymptotically. Simulation results are worked out to illustrate our proposed design method.