Search

We propose a theoretical framework for joint system identification and control on a class of stochastic linear systems. We investigate optimization algorithms for inferring endogenous and environmental parameters from data, part of which are used for control purposes. A number of non-trivial interplays among stability and performance, as well as computational challenges and fundamental limits in identification rate emerge. Our results are validated via simulation example on a quadcopter control problem.

This work investigates performance and optimal distributed controller design of power networks with time-delay (latency). Performance of the network is measured by the square of the system’s H2-norm. The focus of this paper is on allocating weights to interconnections in order to enhance the performance of the time-delay network and these weights serve as feedback gains in the system. We consider a stylized power system model that inspires our analysis. First, we provide an exact expression for the time-delay margin. Then, we find the performance formula as a function of the spectrum of the Laplacian matrix of the coupling graph, time-delay, and the distributed controller gains. Moreover, we formulate the weight allocation problem as an optimization problem and compare the cases where time-delay is considered and time-delay is neglected in the design process. Lastly, we show that in the presence of time-delay increasing connectivity does not necessarily improve the performance.

A systemic risk framework is proposed for robustness analysis in linear post-fault dynamics of a power grid. We investigate scenarios where perturbations, induced by exogenous sources of disturbances, steer the power network away from operation equilibria. In addition, feedback control laws that are prone to communication time-delays are implemented as a means to stabilize the network against perturbations. Systemic risk methods are then employed to assess the possibility of the network to deviate into undesirable modes of operation due to both the additive noise and the possibly negative effect of the time-delay implemented feedback control. Numerical explorations reveal possible limitations regarding the efficiency of the closed-loop system, as a result of network connectivity, control gains, time-delay, and noise.

We investigate the stability of a class of nonlinear flocking schemes of Cucker-Smale type executed on a finite population of autonomous agents. We use algebraic graph theory tools and derive sufficient conditions for asymptotic convergence for velocity coordination while the flock remains sufficiently connected. In addition, the positions of the agents converge to some predefined relative distances. A simulation example is provided to illustrate our theoretical findings.

This work studies a class of linear first order and non-linear second order static distributed consensus networks with time-varying multiple propagation delays, in continuous time. We provide conditions for convergence of to a common constant value, under an increased connectivity condition. The results are delay-independent in the sense that they hold for arbitrary bounded delays. Our approach makes use of fundamental concepts from the Non-Negative Matrix Theory in a fairly elementary way.

We consider a generic non-linear consensus model and prove convergence results to a common value together with prescribed rate of convergence. Instead of a Lyapunov approach we consider a functional metric space and make a fixed point theory argument using contraction mappings. We are restricted to the case of static networks.