Workshops at RSS 2007
List of Workshops
- Robot Manipulation: Sensing and Adapting to the Real World
- Robotic Sensor Networks: Principles and Practice
- Algorithmic Equivalences Between Biological and Robotic Swarms
- Embodied Agents for Disaster Mitigation (CANCELLED)
- Research in Robots for Education
- Interfaces, devices and methods for robots adapted to Small and Medium Enterprises (CANCELLED)
- Tutorial: Microsoft Robotics Studio (MSRS)- A Technical Introduction
Robot Manipulation: Sensing and Adapting to the Real World
Time: 9:00 - 18:00
Charles C. Kemp (Georgia Tech) (lead)
Aaron Edsinger (MIT)
Robert Platt (NASA JSC)
Neo Ee Sian (AIST)
This workshop will explore new approaches to autonomous robot manipulation that are specifically designed to handle the uncertainties of real world applications. There is a resurgence of interest in robot manipulation as researchers seek to push autonomous manipulation out of controlled laboratory settings and into applications such as domestic assistance, health care, and space exploration. This workshop will investigate promising approaches that address the challenges of autonomous manipulation within domains that require the robot to sense the world and adapt to the unexpected.
This full-day workshop will consist of talks, discussions, a poster session, and demos. It will conclude with a moderated discussion of potential future applications for autonomous robot manipulation focused on identifying tomorrow's driving applications and the research required to enable these applications.
This workshop is a successor to the RSS 2005 Workshop on Humanoid Manipulation and the RSS 2006 Workshop on Manipulation for Human Environments.
Robotic Sensor Networks: Principles and Practice
Time: 8:30 - 16:00
Gaurav S. Sukhatme (firstname.lastname@example.org)
Wolfram Burgard (email@example.com)
Sensor network research has risen to prominence in recent years. The breadth of research in the area is large. RSS 2007 will feature a one day focused workshop on *robotic* sensor networks, namely sensor networks which incorporate robotic mobility or articulation. Such systems include, e.g., a networked multi-robot group, a network of immobile computing and sensing nodes and mobile robot(s), a network of immobile nodes each with computing and actuated sensing (allowing e.g., each node to change the direction in which a sensor is pointed). The design of such systems raises algorithmic and theoretical challenges as well as challenges associated with the practicalities of of conducting real deployments. This workshop will bring together people interested in the algorithmic aspects, mathematical and statistical foundations, and experimentalists who have fielded robotic sensor networks in the context of specific applications. This synergy between theory and practice is in line with the 'realistic visions' RSS 2007 workshop theme.
The workshop will be structured as a mix of short technical (though informal) presentations, Q&A sessions, posters, demos, discussion panels, and a concluding summary session. The exact schedule will be formulated shortly.
Participation and Outcomes:
We encourage and actively seek participation by all interested members of the robotic sensor network research community. As a participant you may present a poster, give a short talk, or serve on a panel. As a presenter we simply ask that you make material relevant to your presentation (e.g., slides or poster or paper) available to the organizers two weeks before the conference for distribution to the workshop attendees. Please email the organizers with a short statement of interest if would like to participate in the workshop or if you have any questions. We are actively pursuing the possiblity of a journal special issue on this topic based on the workshop.
Algorithmic Equivalences Between Biological and Robotic Swarms
Time: 9:00 - 18:00
James McLurkin, MIT
Paulina Varshavskaya, MIT
32 Vassar St, room 32-376, Cambridge MA 02139
32 Vassar St, room 32-G585, Cambridge MA 02139
This workshop will bring together researchers in biology, robotics and computer science who study distributed physical systems: swarms, hives, colonies, and multi-robot teams. The main goal is to enable rigorous discussion of the common system constraints and algorithmic solutions employed by natural and artificial swarms.
The first requirement is to identify similarities in our respective high-level models of sensing, communication, processing, and mobility, and discuss how these models constrain the distributed algorithms employed by natural and artificial systems. What fundamental mathematical theories underly both biological models of swarm behavior and the design of robotic teams? What analytical tools are commonly used in one field, which may be of benefit to the other? What specific common constraints apply to the solutions found by nature and robot engineers? When do roboticists wish they knew more biology? When do biologists wish they knew more computer science or control theory? Are there valid equivalences between natural and robotic distributed systems at the algorithmic level? What tasks and applications for swarm technology are most like those of their natural counterparts? What is the most productive way to use "natural algorithms" in distributed robotics?.
We welcome participants from the robotic fields of swarm robotics, team robotics, modular robotics and other distributed robotic systems. We welcome participation from biologists specializing in the study of group behavior and organisms comprising such groups. Prospective speakers will be asked to submit an extended abstract of their presentation, with less emphasis on their current research and a strong focus on algorithmic equivalences between biological and robotic distributed systems. Previously published research is acceptable if it fits the goals of the workshop. Accepted abstracts will be made available in a digital archive.
Embodied Agents for Disaster Mitigation (CANCELLED)
Dr. Holger Kenn, TZI Bremen (firstname.lastname@example.org)
Alexander Kleiner, University of Freiburg (email@example.com)
Dr. Sarvapali Ramchurn, University of Southampton (firstname.lastname@example.org)
In the last decade the world saw a noteworthy increase of natural disasters. Events such as the earthquake in Kobe, the Asian Tsunami, and hurricane Katrina in New Orleans, as well as man-made disasters, such as terrorist attacks in Europe, Asia, and the US demonstrated the weakness of today's society against these threats. Lessons learned from these events are that efficient first response can reduce the number of casualties significantly and facilitate recovery. However, in the case of largely devastated areas the problem is aggravated by the typically small number of first responders to search and rescue survivors at the same time. Moreover, responders have to perform a number of actions and coordinate with each other in an environment where:
1) Communication is unreliable or noisy.
2) Structures such as collapsed buildings and hazardous areas are hard to access.
3) Information is incomplete or corrupted.
Hence, the success of Disasters mitigation (also known as emergency response) depends strongly on the responders' ability to access complex and hazardous structures. Moreover, the resilience of communication structures and information fusion mechanisms will determine how efficiently the responders coordinate and choose the right coordinated course of action.
Against this background, a growing community has been researching techniques to support emergency response using robotics and multi-agent technology. In particular, researchers have been developing partially or fully autonomous systems that can access places humanly inaccessible, to develop robust probabilistic methods for data integration in an either centralized or decentralized way, and proposed efficient methods for mission planning and coordination. For such systems to be useful for disaster mitigation, these must perform their complex task on real-world data in real time which is another important research challenge.
Within this context, the workshop invites research contributions that relate to the design and implementation of multi-agent and robotic systems that can be applied in emergency response scenarios. The goal of the workshop is to provide a forum for the discussion of issues arising in designing and implementing systems that process data from disasters in real time, either directly embodied in the scenario or from the perspective of an incident commander. The workshop also aims to bring together researchers from the Robotics and Multi-Agent Systems community who have been dealing with very related issues from different perspectives.
Topics of Interest
Relevant topics include but are not limited to the following:
- Robust integration of noisy data
- Decentralized agent-based architectures
- Teamwork, Coordination, and Planning Mechanisms in dynamic and uncertain real time environments
- Mapping of complex environments (2-D, 3-D, GIS integration, ...)
- Decision making under uncertainty
- Machine Learning, ether single or multi-agent
- Autonomous robots and robot teams
- Disaster management and situation assessment
- Robot mobility (novel mechanical designs, manipulators, ...)
- Emerging technologies (sensors, power sources, micro robots, ...)
- Intelligent behaviors to improve robot performance and survivability
- Human-robot interfaces for improved remote situational awareness
- Body-worn sensors, communication systems and user interfaces for human responders
- Performance metrics and evaluation methodology for rescue systems and responders
- Perception for navigation, hazard detection, and victim identification
Papers are reviewed by 2 PC members each. The workshop proceedings will be available online. Additionally, the organizing committee plans to publish a proceedings volume with accepted full papers.
Nathan Schurr (University of Southern California, USA)
Daniele Nardi (Universita di Roma "La Sapienza", Italy)
Itsuki Noda (NIAIST, Japan)
Raj Madhavan (NIST, USA)
Alex Rogers (University of Southampton, UK)
Nick Jennings (University of Southampton, UK)
Alessandro Farinelli (University of Southampton, UK)
Claude Sammut (University of New South Wales, Australia)
(more to be announced soon)
Tentative list of speakers
Dr. Sarvapali Ramchurn, University of Southampton (email@example.com)
Dr. Stephen Balakirsky, NIST (firstname.lastname@example.org)
Prof. Dr. Michael Lewis, University of Pittsburgh (email@example.com)
Prof. Dr. Michael Lawo, TZI Bremen
Research in Robots for Education
Time: 9:00 - 18:00
Doug Blank (Bryn Mawr College)
Maria Hybinette (University of Georgia)
Keith O'Hara (Georgia Tech)
Daniela Rus (MIT)
The landscape of robots in education has continued to change since the 2005 RSS Robotics Education Workshop. Over the last two years, there has been a noticeable spike in interest in the use of robots in education. For example: robots are discussed as platforms for education at leading conferences and workshops such as SIGCSE and AAAI; Universities are integrating robots into their classrooms; Robot- centered competitions like FIRST, BotBall and RoboCup continue to flourish. Industry is interested as well: iRobot recently announced the Roomba Create; LEGO has updated their popular Mindstorm robot; And Microsoft has funded two centers to explore the utility of robots for computer science education.
What is the basis for this excitement? What is the evidence that robots in the classroom advance education? The focus of this workshop is to provide a venue for presentation of the research supporting (or contradicting) the effectiveness of robots in education, and to help shape future research in this area.
In particular, the workshop will explore how robots are used differently as educational tools, in terms of hardware, software, pedagogy, and assessment, in different disciplines (e.g. ME, EE, CE and CS) and why certain types of robots may be more effective for different purposes. As an example, many teachers take a constructionist approach in which students build their own robots, while others provide students with a working platform that they should not change.
The workshop will also explore new curricula and robot platforms and the research behind them. The objective of this workshop is to re-evaluate the state of the art of robotics education and discuss how to continue the broad adoption of tools and materials in the classroom. As part of this discussion, we will explore what areas remain unsolved, and which are immediately available for realistic use. Moreover, we hope to create a community beyond the workshop for future exchange of ideas.
The full-day workshop will include: about 10 formal talks, discussion panels, lightning talks (5 minute talks to promote posters), and a poster/demo session in which participants will be invited to provide hands-on demonstrations their robots and/or curricular materials. The talks will focus on both demonstration and research results.
Interfaces, devices and methods for robots adapted to Small and Medium Enterprises (CANCELLED)
J. Norberto Pires
With this session we expect to expose results from the SMErobotTM European project, but also open some space to external contributions from experts and submitted papers. The aim of the SMErobotTM project is to develop devices, methods and robots specially adapted to Small and Medium Enterprises needs. The workshop will also constitute an evaluation of the current status of the project, exposing to an international audience the sustainable results, test-beds and achievements. The main focus will be on:
- Devices for robot programming and HMI;
- Plug-and-produce technologies and test-beds;
- Parallel robot for SME force control applications developed in the project;
- Low cost force torque sensor developed in the project;
- Light-weight robot for SME applications;
- High-level programming concepts;
- Robots without fences.
Tutorial: Microsoft Robotics Studio (MSRS)- A Technical Introduction
Time: 9:00 - 10:30
Stewart Tansley (firstname.lastname@example.org)
Joseph Fernando (Microsoft Robotics Group)
To further accelerate the growth of the emerging field of robotics, Microsoft has created a new software development kit for the robotics community - the Microsoft Robotics Studio (MSRS). MSRS provides a software platform that enables development of a wide variety of applications which can be used across a variety of hardware, applicable to a wide audience of users.
The robotics session will provide both a high level overview of the MSRS tools and runtime. The content is suited to both novices as well as advanced robotics professionals.
Microsoft Robotics Studio delivers in three areas of software and all will be described in this technical introduction:
- A scalable, extensible concurrency and distributed runtime architecture that can span a wide variety of hardware and devices. The programming interface can be used to create applications to drive robots using 8-bit or 16-bit processors (from a connected PC) as well as 32-bit systems with multi-core processors; and devices from simple touch sensors to laser distance finding devices.
- A set of tools that make programming and debugging robot applications scenarios easier. These include a high resolution visual simulation environment that integrates 3D software physics supplied by the Ageia Technologies PhysX engine. While Microsoft Robotics Studio can be used with programming languages such as those included in Microsoft Visual Studio and Microsoft Visual Studio Express, also included is a new visual programming language that enables the creation of applications using a simple drag-and-drop interface.
- A set of useful technology libraries (services) to help developers get started with writing robot applications, and tutorials which illustrate the basics of how to get started in a variety of programming languages.