Control of Autonomous Multi-Agent Systems
Prof. Francesco Delli Priscoli and Prof. Giuseppe Oriolo
Dipartimento di Ingegneria Informatica, Automatica e Gestionale
Sapienza Università di Roma
Information
| schedule | 26 Feb - 1 Jun 2018
see websites for Part 1 and Part 2 for detailed schedule
|
| office hours | Delli Priscoli: send e-mail
Oriolo: after class
|
| e-mail | dellipri [at] diag [dot] uniroma1 [dot] it
oriolo [at] diag [dot] uniroma1 [dot] it |
Audience
Students
of the Master in Control Engineering at Sapienza University of Rome.
Objective
This 6-credit
course presents the basic methods for modeling, analyzing and
controlling multi-agent systems, with special emphasis on distributed
strategies. Applications will be presented in the control of
communication and electrical networks as well as of multi-robot
systems. The student will be able to analyze and design architectures,
algorithms, and modules for controlling multi-agent systems.
Syllabus
The
course is organized in 2 parts:
- Part
1 (Francesco Delli Priscoli, Feb-Apr 2018): Examples of
multi-agent scenarios in the communication and energy networks.
Centralized vs. decentralized architectures. Multi-agent networks with
limited or no information exchanges among agents. Overview of learning
methodologies (in particular, Reinforcement Learning), game theory,
negotiation, auctions, Kalman filtering. Behavior of these
methodologies in distributed multi-agent frameworks (e.g., behavior of
Reinforcement Learning techniques in a distributed framework,
Distributed Kalman Filtering). Application to specific multi-agent
scenarios (e.g, Reinforcement Learning techniques for Quality of
Experience control, Distributed Kalman Filtering for sensor networks).
- Part
2 (Giuseppe Oriolo, Apr-May, 2018): Examples of applications of
multi-robot systems. Centralized vs. decentralized architectures.
Mathematical tools: Adjacency graph and matrix; Laplacian; Connectivity
and Consensus; Passivity and Lyapunov stability; Interconnection of
mechanical systems. Application to multi-UAV systems: Formation control
with time-varying topology; Formation control with connectivity
maintenance; Steady-state behaviors; Bearing-based formation control.
Application to multi-UGV systems: Cooperative Mobile Manipulations; Cooperative exploration of unknown
environments; Mutual localization with anonymous measurements; Target
localization and encircling.
For
additional details and material, access the webpages
of the individual parts by following the above links.
Grading
To
obtain 6 credits for this course it is necessary to
complete separate activities for Parts 1 and 2. In particular, for each part the student must carry out a small project which typically requires reading some technical papers and performing some simulations. To register the final grade, sign-up via Infostud is required.
Questions/comments: oriolo [at] diag [dot] uniroma1 [dot] it