Visitor Program: Dr. Andrea Gasparri, Assistant Professor, Engineering Department, Roma Tre University, Italy
Organizer(s): Bhaskar Krishnamachari and Gaurav Sukhatme
Sponsored: Spring 2013 5/21-6/30
Contact: email@example.com Office: PHE 414
Dr. Gasparri is a young expert with rapidly growing recognition in the emerging area of distributed networked robotics, a field at the intersection of control, robotics, and networking that is currently gaining in importance. Dr. Gasparri will be visiting USC from May 21 - June 30, 2013. During his visit, he will have close research interactions with Prof. Gaurav Sukhatme and Prof. Bhaskar Krishnamachari. Because of his interest in network control, he will also likely interact with many colleagues with interests in this area: Edmond Jonckheere, Michael Neely, Rahul Jain, and David Kempe. We could also possibly use this opportunity to organize an internal USC workshop on networks and control, so that his visit may catalyze collaborations between faculty here working on various facets of this topic. He will also give two research talks to the department.
Biography: Andrea Gasparri received the Graduate degree (cum laude) in computer science in 2004 and the Ph.D. degree in computer science and automation in 2008, both from the University of Rome Roma Tre, Rome, Italy. He is currently an Assistant Professor for the Engineering Department, University of Rome Roma Tre. His current research interests include mobile robotics, sensor networks, and, more generally, networked multi-agent systems.
Research Talks by Andrea Gasparri at USC EEB 248 11:00am
June 04, 2013: Multi-Robot Systems: A Control Perspective- Tutorial Part I (PDF Presentation Slides)
June 07, 2013: Multi-Robot Systems: A Control Perspective - Tutorial Part II (PDF Presentation Slides)
June 11, 2013: Swarm Aggregation Algorithms for Multi-Robot Systems (PDF Presentation Slides)
Title: Multi-Robot Systems: A Control Perspective
Abstract: Multi-Robot Systems represent an important research field with a wide variety of topics to be addressed. In recent years a great effort has been made by the research community towards the development of decentralized techniques to provide an adequate level of robustness and flexibility to these systems. This tutorial will provide first a general overview of this research area, focusing on the most important design aspects of a multi-robot system, e.g. control and communication architecture, and illustrating the most important research problems. Then, it will focus on the control aspects of the distributed cooperation problem. In that context, the consensus problem will be first reviewed as a starting point towards the investigation of more refined techniques to achieve spatial aggregation between the robotic units. Furthermore, the connectivity maintenance problem will be introduced, a taxonomy of the approaches available at the state of the art will be derived and some relevant techniques will be described more in detail. Finally, relevant applications in the context of multi-robot systems which rely on the distributed coordination techniques previously introduced will be highlighted.
Title: Swarm Aggregation Algorithms for Multi-Robot Systems
Abstract: In this talk, a novel decentralized swarm aggregation algorithm for multi-robot systems is proposed. In this framework, the interaction among robots is limited to their visibility neighborhood, i.e., robots that are within the visibility range of each other. Furthermore, to better comply with the hardware/software limitations of mobile robotic platforms, robots actuators are assumed to be saturated. First, a decentralized control law which fulfills these requirements is derived. Successively, a more refined control law also able to i) ensure a collision-free interaction, ii) handle asymmetric saturations with respect to the forward and backward velocity, and iii) integrate an obstacle avoidance is proposed to safely move the swarm within a cluttered environment. A theoretical characterization of the main properties of the proposed swarm aggregation algorithms is provided. Simulations have been carried out to validate the theoretical results. Furthermore, experiments have been performed with a team of low-cost mobile robots to demonstrate the effectiveness of the proposed approach in a real-world environment.