We are proud to offer the following workshop events at the 2008 Robotics: Science and Systems Conference.

Please see below for a detailed description of the workshops.

Control of Locomotion: From Animals to Robots

Duration: FULL DAY
Location: HG E1.2

Organizers: Auke Ijspeert, Paolo Dario, Sten Grillner

Abstract: The goal of the workshop is to discuss how inspiration from the animal kingdom can help improving locomotion skills in robots, and how robots can be used as scientific tools in animal motor control. All aspects of locomotion will be considered including materials, actuators, sensors, control, and learning. In particular, the workshop aims at exploring how these different aspects can collectively be designed to improve the locomotor skills of future robots.

We will discuss possible roads to tackle the challenges related to having robots get out of the labs and move in unstructured and partially unknown terrains. While robotics has clearly something to gain from biological inspiration, we will also discuss how robotics can give something back to biology, namely how robots can be used as scientific tools in animal motor control. For instance, how robots can be used to test hypotheses about neural circuits and/or biomechanical principles in invertebrate and vertebrate animals.

As speakers we intend to obtain an interesting mix of biologists and roboticists. The talks will be given by a series of invited speakers and by speakers selected after an open call for abstracts.

Underwater Robotics ... at the Microscale

Duration: FULL DAY
Location: HG E1.1

Organizers: Brad Nelson, Vijay Kumar, Sylvain Martel, Metin Sitti, Lixin Dong

Abstract: Biological motors are engineering marvels capable of providing propulsion at micro and nano scales. A variety of molecular machines swim for cell division and intracellular transport and have recently been used to power and manipulate micro-/nanoscale components. Microrobotic swimmers have also been proposed as a key technology for dramatically changing many aspects of medicine by navigating bodily fluids to perform targeted diagnosis and therapy. Microrobots, like microorganisms, swim in a low-Reynolds-number regime, requiring swimming methods that differ from macroscale swimmers. Researchers have proposed numerous microrobotic swimming methods, many biomimetic, with the vast majority utilizing magnetic fields to wirelessly power and control the microrobot. The goal of this workshop is to understand how fundamental robotics research has benefited from biology, and how biologists and robotists might benefit from one another's approach.

Topology and Minimalism in Robotics and Sensor Networks

Duration: FULL DAY
Location: HG F26.1

Organizers: Robert Ghrist, Steve LaValle, George J. Pappas

Abstract: Can we do more with less? Although obtaining reams of sensor data is cheaper than ever, unstructured environments are extremely complicated and unpredictable. Rather than trying to learn and utilize precise and complete models, this workshop explores the theme of achieving well-specified robotic and sensor network tasks with as little information as possible. The push toward minimalism offers greater robustness and simplicity, but also requires the development of powerful new mathematical techniques and algorithms. Basic tools from algebraic topology and other areas not typically associated with robotics come to the rescue. Problems approached from the miminalist perspective in this workshop include guaranteeing sensor network coverage, counting targets, time synchronization, target tracking, robot navigation, and topological map building.

Our workshop includes contributions from various leading researchers in the field. The represented organizations include the University of Illinois at Urbana Champaign, Carnegie Mellon University, the University of Pennsylvania, the University of Washington, and the University of South Carolina. The participants will present state-of-the-art results addressing the following challenging problems:

Teaching with Robots

Duration: FULL DAY
Location: HG E21

Organizers: Chris Rogers, Pedro Lima, Roland Siegwart, Illah Nourbakhsh, Aaron Dollar

Abstract: Over the past 10 years, a number of low-cost, highly capable robotic platforms have come on the market and have introducing students to math, science, and engineering at all grade levels. Our goal with this workshop is to identify innovative ideas where the hands-on, project-based nature of these platforms have successfully improved student learning at the university level. We propose to identify this in two ways: presentations in the morning and a hands-on discussion in the afternoon. The morning will be comprised of 10 min talks, with the first 1.5 hours being presentations from faculty on successes and failures in the college classroom and the second 1.5 hours being presentations on successes and failures in the college robotic competition world. In the afternoon we will offer a 2.5 hour session building a robotic zoo, where attendees will get a chance to build robots while talking amongst each other.

Design and Control of Variable Impedance Actuators for Physical Interaction of Robots with Humans and their Environment

Duration: HALF DAY (morning)
Location: HG F26.3

Organizers: Antonio Bicchi, Alin Albu Schaeffer, Bram Vanderborght

Abstract: In recent years, there has been a growing attention towards the design of actuators and transmission mechanisms that can make robots of the next generation to perform powerful, accurate and fast motion, while remaining at all times safe for humans in their vicinity or nearby. Actuators with adaptable passive impedance are also useful for tuning the mechanism to the interaction with the human or the environment - as e.g. in prosthetics or autonomous locomotion. A strong argument in favour of intrinsic joint compliance is its property to serve as an energy storage mechanism, possibly decreasing the energy consumption of the entire system during task execution. Furthermore, the stored energy can be used to considerably increase the peak link speed or force in order to approach human performance. Finally, many researchers are interested in the variable stiffness (and in particular the antagonistic) design approach from a bio-robotics point of view, with the aim of better understanding the principles of human motion and its control.

Naturally, such a fundamental design paradigm shift comes at a certain cost. The increased number of actuators and the small intrinsic damping of most implementations are certainly some of the major challenges in controlling a variable compliance joint. The expected reduction in absolute position accuracy due to the elasticity needs to be compensated by external sensing as e.g. vision. Furthermore, a lower mechanical bandwidth will be the consequence due to the generally lower joint compliance. The various existing viewpoints on the topic as well as the mentioned open problems motivate the workshop as a platform for experience exchange and discussions. In this workshop, we plan to bring together top researchers in the fundamentals of electromechanical design, planning, and control of new actuators, pHRI safety, and bio- robotics. We intend to make a point on the state of the art, initiate discussions about open problems and perspectives of variable impedance design, enable interdisciplinary and inter-group cooperation, and to further spread knowledge about the methods, the technologies, and the challenges in this field.

Robot Manipulation: Intelligence in Human Environments

Duration: FULL DAY
Location: HG E3

Organizers: Robert Platt, Sami Haddadin, Charlie Kemp, Lorenzo Natale, Neo Ee Sian

Abstract: This full-day workshop will explore new approaches to robot manipulation with an emphasis on the relationship between intelligence and successful interaction with human environments. This is the fourth in a series of manipulation-centered workshops at RSS. There has recently been renewed interest in robot manipulation as researchers seek to push autonomous manipulation out of controlled laboratory settings and into less structured domains including the factory, health care, space exploration, and the home. The problem of autonomous robot manipulation in human environments poses a number of key challenges. These include: sensing relevant information; using sparse and incomplete information efficiently; and adapting to unforeseen characteristics of the environment. These challenges connect manipulation to the larger problem of robot intelligence.

This workshop will explore these issues and feature research that addresses the challenges of autonomous manipulation in this context.

Interactive Robot Learning

Duration: FULL DAY
Location: HG F26.5

Organizers: Andrea Thomaz, Geert-Jan M. Kruijff, Henrik Jacobsson, Danijel Skocaj

Abstract: Many future applications for autonomous robots bring them into human environments as helpful assistants to untrained users in homes, offices, hospitals, and more. These applications will often require robots to flexibly adapt to the dynamic needs of human users. Rather than being pre-programmed at the factory with a fixed repertoire of skills, these personal robots will need to be able to quickly learn how to perform new tasks and skills from natural human instruction. Moreover, it is our belief that people should not have to learn a new form of interaction in order to teach these machines, that the robots should be able to take advantage of communication channels that are natural and intuitive for the human partner.

While there are numerous examples of successful applications of Machine Learning algorithms, this scenario poses new problems for the field. Learning in real-time from everyday people, unfamiliar with Machine Learning or robotics, is a new domain with new challenges for the field.

This workshop on Interactive Robot Learning will span the breadth of research questions at the intersection of Machine Learning and Human-Robot Interaction.

Inside Data Association

Duration: FULL DAY
Location: HG E41

Organizers: Udo Frese, Jose Neira, Diedrich Wolter, Jorg Kurlbaum

Abstract: Every feature based estimation algorithm faces two kinds of uncertainty. First, the continuous uncertainty originating from noise in the observations and second, the discrete uncertainty of which observation corresponds to which feature, the so-called data association. It plays a crucial role in many estimation problems, most noticeably tracking and simultaneous localization and mapping (SLAM) where wrong data-association often leads to catastrophic failure of the algorithm. Hence, we believe that a deeper theoretical understanding of data-association is a scientific merit. In this workshop we want to take a deeper look at the structure of data-association, so we explicitly encourage insightful contributions, that focus on theoretical or empirical analysis, novel views of the problem, or candidates for a gold-standard algorithm. To facilitate evaluation, we provide a SLAM data set with known data-association ground-truth and will encourage using it. This data set is preprocessed, providing extracted geometric features to allow participants to use it easily. The workshop consists of an invited talk, presentations of submitted papers with generous discussion time, and plenary discussions. The plenary discussion will be initiated with a list of "Fundamental questions in data-association". We therefor ask the participant (or anyone interested) to email such questions to the workshop organizers.

Quantitative Performance Evaluation of Navigation Solutions for Mobile Robots

Duration: HALF DAY (morning)
Location: HG E22

Organizers: Raj Madhavan, Chris Scrapper, Alex Kleiner

Abstract: The development of adaptive and technically capable mobile robots that are able to safely operate in unstructured, dynamic environments is essential to the acceptance of robotic technologies, permitting collaborative operations of man and machine. Currently, there is no way to quantitatively measure the performance of these systems against user-defined requirements; and furthermore, there is no consensus on what objective evaluation procedures need to be followed to deduce the performance of these systems. The lack of reproducible and repeatable test methods have precluded researchers working towards a common goal from exchanging and communicating results, inter-comparing robot performance, and leveraging previous work that could otherwise avoid duplication and expedite technology transfer from the "drawing board" to the field. The primary focus of this workshop is to bring together what is currently an amorphous research community to define standardized methods for the quantitative evaluation of navigation solutions and robot-generated maps that will enable mobile robots to operate in dynamic unstructured environments. This workshop will seek to develop test methods to classify the performance characteristics of navigation solutions that facilitate the inter-comparison of experimental results. It will also attempt to define a de facto standard testbed for evaluation of navigation solutions that will provide a baseline for comparison and will provide the mechanisms targeting specific aspects of the system, allowing researchers to assess the performance of various systems in different scenarios and environmental conditions.

The workshop program will consist of two featured presentations, six regular presentations, and will conclude with a panel discussion involving all participants to summarize presented research and to decide future research directions and collaboration.

Experimental Methodology and Benchmarking in Robotics Research

Duration: HALF DAY (afternoon)
Location: HG E22

Organizers: Fabio Bonsignorio, John Hallam, Angel P. del Pobil

Abstract: As the complexity of current robotic and embodied intelligent systems grows, it is more and more necessary to define proper experimental approaches and benchmarking procedures. On the one hand, reliable benchmarks are called for in order to allow the comparison of the many research results in robotics research, so that their industrial application is eventually possible. On the other hand, if robotics aims to be regarded as serious science, replication of experiments deserves consciencious attention; it is necessary to be able to verify if and by which measure new procedures and algorithms proposed in research papers constitute a real advancement and can be used in new applications.

This workshop aims to provide a discussion forum on these topics and to identify guidelines for the future. It will consist of invited and regular presentations with a significant amount of additional time for discussions. It is envisioned to be useful for anyone with an interest in quantitative performance evaluation of robots and/or robot algorithms. Some controversial issues will be discussed such as: measuring autonomy or information metrics of intelligent systems, or the concept itself of replicability or benchmarking of research results in robotics.

Advances in Simulation of Robot and Task Dynamics

Duration: HALF DAY (afternoon)
Location: HG F26.3

Organizers: Evan Drumwright, Kurt Anderson, Roy Featherstone

Abstract: Simulation has proven to be highly effective for modeling the motion of robotic manipulators; however, similar success has yet to be realized in simulating robots interacting with their environments. The dominant problem has not been the accuracy of the simulation, but rather lack of robustness. Simulation instability is often encountered and arises for many reasons, including numerical integration of continuous systems, polyhedral approximations of curved surfaces, and imperfect models of physical phenomena like friction. The result of these difficulties is that robotic simulation has been almost completely limited to free-space motion, locomotion and, very recently, grasping.

This workshop aims to provide practical knowledge for better simulating robot dynamics; the primary focus will be on recent advances in robotic simulation that increase robustness, accuracy, and speed. The workshop will also consider methods for mixing simulation paradigms, like rigid body dynamics and solid mechanics, in order to model tasks that are currently not feasible to simulate (e.g., welding, drilling, etc.)

Relevant topics will include:

Grand Challenges in Microrobotics and Microassembly

Duration: FULL DAY
Location: HG G26.1

Organizers: Pierre Lambert, Stephane Regnier, Metin Sitti

Abstract: For the miniaturization of devices and machines down to nanometer sizes, micro/nanorobotic approach enabling manipulation, locomotion, and interaction at the micro and nanoscales is indispensable. Micro/Nanorobotics as an emerging robotics field is based on the micro/nanoscale physics, fabrication, sensing, actuation, system integration, and control taking the scaling effects and nanoscale physics, chemistry and biology into consideration. Micro/Nanorobotics encompasses:

The topics proposed for discussion will be mainly focused on micro/nanomanipulation, including aspects such as surface-tension-based micro and nano actuation, micro and nano devices and systems, positioning systems, biomedical microrobotics.