Upcoming IEEE CSS Lunch Meeting

When: 19 November 2010, 11:00 – 1:00

Where: Gibson’s Bar-B-Q, 3319 South Mermorial Parkway (Just East of “Useless” Overpass North of Airport Road)

Guest Speaker: Dr. Vadim Azhmyakov, CINVESTAV (The Mexican Center of Advanced Study and Research), Department of Automatic Control, Mexico

Topic:

A Hybrid Linear Quadratic Approach to Optimization of some Classes of Multi-agent Systems: Applications to the Safety Control

Abstract

The increasing complexity associated with many real-world engineering applications, including flight control, autonomous robots and vehicles guidance, automation of complex technological processes, power electronics, process control in sensor-rich environments, and control of biological systems, has far-reaching implications for modern systems design. As an example, switched/hybrid and multiagent systems, interacting among themselves and remote users over control/communication networks, introduce a whole new set of system-level challenges. In this situation the main classic control design objectives such as stability and performance are being complemented with a number of new technical modifications and extensions. In general, a sophisticated interconnection of the complex nonlinear dynamical objects such as autonomous robots/vehicles, airplanes, satellites in the presence of the associated communication networks is a prevailing attribute of many modern applied control systems. Therefore, the necessary analysis and adequate (optimal) design procedures for these systems has been recognized as major challenging problems in the control engineering. Our talk is devoted to a specific optimal control problem associated with a class of multi-agent dynamic system. The interest is placed on minimization of the tracking error in the so called multi-agent leader-follower model. We apply the hybrid approach to the initial path following optimization problem related to a group of dynamic agents. By this means, the initial problem is replaced by a specific hybrid optimal control problem. In particular, we consider multi-agent control systems with monotonically increasing dimensions of the state vector. The change of the state dimension has the character of a state jump and is modeled by an impulsive hybrid system. We also discuss the possible computational procedure associated with the above optimal tracking multi-agent control problem in the specific setting. The theoretical and numerical approaches presented in our talk are applied to some practically motivated examples.

Bio-sketch. Having graduated with concentration in Aerospace Control from Moscow Technical University, Russia, in 1990, Vadim Azhmyakov went through graduate studies in Control Theory at Institute of Control Sciences, Moscow, Russia with PhD degree in 1994, and Habilitation (the Postdoctoral thesis) in Applied Mathematics, University of Greifswald, Germany in 2006. In 1999-2005 he was with University of Greifswald, Department of Numerical Analysis, Germany; in 2005-2007 he was a Senior Researcher at the Departments of Electrical Engineering, the Universities of Magdeburg and Berlin, Germany. Since 2007 he is with CINVESTAV (The Mexican Center of Advanced Study and Research), Department of Automatic Control, Mexico, where he holds a rank of Professor. His research interests include optimal control, numerical methods in optimization/optimal control, hybrid/switched, stochastic and variable structure dynamical systems, networked control methodologies, communication and control over networks.