Robust distributed control of multi-agent systems with application to electrical networks

Abstract: The increasing diffusion of artificial intelligence, i.e., embedding microprocessors and software in commonly used devices, and the understanding that emerging behaviours arise in the presence of interconnected systems are the motivations of the increasing interest of researchers on control, estimation and robustification of complex systems. These systems are often modelled by a set of dynamical agents connected by links representing their interactions. Such a modelling naturally leads to the graph theory as the main tool for both defining the properties of the overall system and designing effective approaches to the control and estimation of multi agent systems (MASs). The distributed control scheme for MASs has been recognised as an effective compromise between the centralised and fully decentralised ones. In this approach the local controllers/estimators are connected just with some neighbourhoods such that the amount of transferred data on the network is decreased and the control is not completely lost even in the case of communication faults. An usual approach to control and estimation in MASs is based on consensus, where the control law is designed to achieve specific tasks (e.g., steering the state of each agent to the same value, flocking, leader following) and it relies on proper interaction rules among neighbourhoods. One interesting application is the control of electrical networks which require that the frequency is everywhere the same and voltages in the nodes must fulfil specific requirements. Because of the intrinsic uncertainty in modelling as well as the presence of external unknown disturbances, the integration of the consensus approach with control techniques having intrinsic robustness property is welcome. Since Variable Structure Control with Sliding Modes has well stated robustness properties, in this talk some of the control and estimation schemes based on this technique are presented and discussed to highlight their pros and possible cons. In particular, the approach based on consensus will be discussed and an application to the control of electrical networks will be presented. Some hints for discussion and open problems will be given as well.

Bio: Elio USAI is currently full professor at the Department of Electrical and Electronic Engineering (DIEE) at the University of Cagliari where he teaches basic and advanced automatic control. He received the Laurea degree in Electrical Engineering at the University of Cagliari in 1985. He worked for international companies first as a research engineer for the development of communication cables, then as process engineer and finally as production manager of the auxiliary services plants in a petrochemical factory. In September 1994 he reached the DIEE as an assistant professor in the Automatic Control Group. His first research interest was on parameter identification in nonlinear systems and on numerical algorithms for optimal control problems. Then he focused on variable structure control theory, specifically on the higher-order sliding mode approach, and on its applications to dynamical systems with uncertainty. The developed original control algorithms were successfully tested on some laboratory prototypes. Current interests are in output-feedback control, state estimation and FDI via higher order sliding modes in linear, nonlinear, infinite dimensional and multi-agent systems. Elio USAI is co-author of more than 200 articles in international journals and proceeding of international conferences with review. He was member of the International Program Committee of international conferences, was General Chairperson of the 9th Int. Workshop on Variable Structure Systems (VSS’06), and he currently serves as an Associate Editor for Asian J. of Control, Journal of the Franklin Institute and IEEE Trans. on Systems, Man and Cybernetic: Systems. He served as Associate Editor for IEEE Trans. on Control System Technology from 2014 to 2019. He is/was the leader of research projects on process control, on control of mechanical systems with uncertainty and on model based fault detection funded by EC, Italian Government, University of Cagliari, and RAS (local regional administration). He is currently the Rector’s Delegate for the Quality of processes and services at the University of Cagliari.

VSS_2022_Plenary_Elio

VSS_2022_Plenary_Elio