Hrvatska sekcija IEEE, CS i RA odjeli, pozivaju vas na predavanje:
“Cooperative Control of Multi-agent Systems - Stability, Optimality and Robustness”
koje će Dr. Kristian Hengster Movric, znanstveni suradnik na University of Texas at Arlington, Arlington, SAD, održati u srijedu, 25.09.2013, u Bijeloj vijecnici FER-a, s početkom u 13:00 sati.
Sažetak predavanja i kratki životopis predavača nalaze se u nastavku vijesti.
Abstract:
This talk brings together stability and optimality theory to design distributed cooperative control protocols that guarantee consensus in multi-agent systems and are globally optimal with respect to a positive semi-definite quadratic performance criterion. A common problem in cooperative optimal control is that global optimization problems generally require global information, which is not available to distributed controllers. Optimal control for multi-agent systems is complicated by the fact that the communication graph topology interplays with the agent system dynamics. We present an inverse optimality approach together with partial stability to consider the cooperative consensus and pinning control. Agents with identical linear time-invariant dynamics are considered. Communication graphs are assumed directed and having fixed topology. Structured quadratic performance indices are derived that capture the topology of the graph, which allows for global optimal control that is implemented using local distributed protocols. A new class of digraphs is defined that admits a distributed solution to the global optimal control problem, namely those with simple graph Laplacian matrices.
Also, this talk is concerned with the effects of disturbances on multi-agent systems. Building on the classical results on the existence of Lyapunov functions for asymptotically stable systems and their use in assessing the effect the disturbances exert on those systems, it is possible to extend such reasoning to partially stable systems, in particular those systems reaching consensus or synchronization. With Lyapunov functions for partial stability one is able to ascertain the effect of disturbances on the partial stability of systems, and to derive conditions on those disturbances that allow for asymptotic partial stability or uniform ultimate boundedness along the target set. Asymptotic partial stability is in that sense robust to this specific class of disturbances. Furthermore, with the means to quantify the effect of disturbances one also gains the ability to compensate for it by an appropriate control law.
Curriculum vitae:
Kristian Hengster-Movric was born in Zagreb, Croatia, in 1986. Elementary and High School education was received in Zagreb. In 2004 Kristian graduated from XV. Gimnazija, mathematics and science oriented high school, as one of the best students in his generation. He enrolled in the Faculty of Electrical Engineering and Computing at The University of Zagreb in 2004 and received the M.S. degree in the field of Automatics in 2009. He received the state scholarship for being in the top 10% of his generation for the entire duration of the program. During the time in college he worked also on different projects that resulted in various publications, and served as a teaching assistant in Mathematics and Signals and Systems during his senior years. Shortly before graduation he was awarded the 2009. Rector's Prize for excellence in research work on multi-agent potential field control.
In 2009 Kristian was accepted to the University of Texas at Arlington, Electrical Engineering department for a PhD. Since the Fall semester that year he has been working towards the PhD at the UTARI institute, under the supervision of dr. Frank L. Lewis. In addition he has served in the capacity of a graduate teaching assistant for the courses in Control Systems, Circuits, Electronics, Linear Systems, Optimal Control and Distributed Decision and Control. In 2010 Kristian was inducted into The Golden Key Honor Society, for academic achievement. In 2013 he was awarded second place prize, N.M. Stelmakh Outstanding Student Research Award, for the excellence of the research work conducted while in the PhD program. He was also awarded Dean's Fellowship for summer semester 2013. Kristian successfully defended his doctoral dissertation in Summer 2013, thus completing the PhD program at the University of Texas at Arlington.
Research interests include, but are not limited to, dynamical systems and control theory applied to complex, multi-agent systems, differential geometry, topology, qualitative analysis of dynamical systems, control of physical systems, systems with distributed parameters. Recent work addressed distributed control of multi-agent systems, applied to synchronization and consensus.
