ACROSS–DataScience

Centre of Research Excellence for Data Science and Advanced Cooperative Systems

The ACROSS-DataScience centre fosters cooperation between the academic community and the public sector in research and development of new methods and advanced engineering approaches related to data science and cooperative systems.


Invitation to the lecture:...

ZCI-DATACROSSACROSS CoE - Centre of Excellence for Autonomous and Cooperative Robotic Systems, IEEE Croatia section, Department of Robotics and Automation and H2020 project "EXCELLABUST - Excelling LABUST in marine robotics" invite you to the lecture:

 

"Decentralized Coordination and Abstraction for Coupled Multi-Agent Systems"

 

which will be held by Dimitris Boskos, KTH Royal Institute of Technology, Stockholm, Sweden.

The lecture will take place on Tuesday, April 17th, 2018 at 10.15h, in seminar room at ZARIbuilding C, 9th floor.

You can find more about the lecturer and the seminar in the detailed news content.

Abstract:

High level planning and control of multi-agent systems constitute active research areas with various applications in robotics, power systems, sensor networks, and other engineering disciplines. For the coordination of multi-robot teams, it is additionally required that the agents satisfy relative state constraints which can guarantee collision avoidance and network connectivity based on the robots' sizes and sensing/communication ranges. Motivated by such objectives,  we present a robust connectivity maintenance control scheme, which is based on decentralized feedback laws and ensures that the network remains connected when bounded additive inputs are further assigned to the agents. These can provide to each agent the ability for high level planning, by leveraging a discrete representation, also called abstraction, of its interacting dynamics. We therefore present a framework on the derivation of distributed symbolic models for the agents, through appropriate space-time discretizations. The abstraction of each agent is based on the knowledge of its neighbors' discrete positions and the transitions are performed through hybrid control laws, which can drive the agent to its possible successor states. This approach is additionally modified towards the derivation of online abstractions, by discretizing overapproximations of the agents' reachable sets over a bounded time horizon. We finally outline how such  tools could be leveraged as building blocks for more complex multi-robot coordination scenarios including cooperative tasks such as collaborative transportation services.​

 

Brief bio:

Dimitris Boskos was born in Athens, Greece in 1981. He has received the Diploma in Mechanical Engineering from the National Technical University of Athens (NTUA), Greece, in 2005, the M.Sc. in Applied Mathematics from the NTUA in 2008 and the Ph.D. in Applied mathematics from the NTUA in 2014. Since August 2014, he is a Postdoctoral Researcher at the Department of Automatic Control, School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, Stockholm, Sweden. His research interests include distributed control of multi-agent systems, formal verification and observer design for nonlinear systems.

News list