Dr. Paul S. Schenker
Space Robotics: Challenges to Learning and Adaptive Systems
Abstract
Space is a rich and highly variable domain for the application of robotic and automated systems. The applications environment is often unstructured, unpredictable, fast changing, and poorly suited -- in the sense of both system performance and risk -- to use of heuristics to address the unexpected. These comments become more specific in the consideration of autonomous surface mobility (e.g. rovers for science exploration) and high dexterity remote robotic operations (tele-manipulation with human in-loop commanding and time delay). We will overview the current art for these two application areas, discuss R&D examples developed to date toward more adaptive capabilities within, and outline some open research challenges to future learning and adaptive perception and control in space robotics.
Speaker Biography
Dr. Paul S. Schenker manages the Robotics Space Exploration Technology Program, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California. His responsibilities encompass the strategic development of technical capabilities for future NASA robotic and human-robotic missions. His prior assignments at JPL include management of JPL's mobility and robotics line organization of about 100 people, and supervision of two related robotics R&D groups. His research has spanned topics in robotic perception, robot control architectures, telerobotics and teleoperation, multi-sensor fusion, and most recently, multi-robot cooperation, areas to which he has contributed about 140 peer reviewed publications as well as a number of recent keynote conference presentations. He has led the development of robotic systems that include the Field Integrated Design and Operations Rover (FIDO), Planetary Dexterous Manipulator (MarsArm, microArm), Robot Assisted Microsurgery System (RAMS), Robotic Work Crew (RWC), and All Terrain Explorer (ATE/Cliff-bot), with resulting technology contributions to NASA missions that include the currently operative Mars Exploration Rovers (MER). Dr. Schenker is active in the AAAI, IEEE, OSA, and SPIE, and General Co-Chair (with Maja Mataric, USC) of the IEEE 2008 International Conference on Robotics and Automation (ICRA 2008). He has served as an elected Board member and 1999 President of SPIE, also as an elected member of the National Academy of Sciences/United States Advisory Committee to the International Commission for Optics.
Dr. Dario Floreano
How to Evolve Controllers for Truly Cooperative Robots
Abstract
Cooperation is widely spread in nature and takes several forms, ranging from behavioral coordination to sacrifice of one's own life for the benefit of the society. This latter form of cooperation is known as "true cooperation", or "altruism", and is found only in few cases in nature. Truly cooperative robots would be very useful in conditions where unpredictable events in the mission may require a cost by one or more individual robots for the success of the entire mission.
However, the interactions among robots sharing the same environment can affect in unexpected ways the behavior of individual robots, making very difficult the design of rules that produce stable cooperative behavior.
It is thus interesting to examine under which conditions stable cooperative behavior evolves in nature and how those conditions can be translated into evolutionary algorithms that are applicable to a wide range of robots. In this talk I will quickly review biological theories of evolution of cooperative behavior and focus on the theories of kin selection and group selection. I will show how these two theories can be mapped into different evolutionary algorithms and compare their efficiency in producing control systems for a swarm of sugar-cube robots in a number of cooperative tasks that vary in the degree of requested cooperation. I will then describe an example where the most efficient algorithm is used to evolve a control system for a swarm of aerial robots that must establish a radio network between persons on the ground.
In another set of experiments I describe how those evolutionary conditions can be tested for the emergence of communication where colonies of "expressive" robots are exposed to food and danger sources that cannot be uniquely be identified at distance. Here, communication of the source type brings an advantage to the colony at the expense of the individuals that decide to tell which is the food or poison. The results shed light on the conditions that may have favored the evolution of altruistic cooperation and communication.
Finally, I will describe work in progress for a real-world application of a swarm of flying robots that are expected to locate and establish an ad hoc radio network among rescuers deployed in a catastrophic scenario. The stringent mission requirements along with the unpredictable location of the rescuers on the ground made it very difficult to come up with suitable control rules. We solved the problem by using the evolutionary methods that we distilled from the previously described research in order to come up with efficient and extremely simple control systems that satisfy the basic mission requirements.
*work performed in collaboration with Sara Mitri (LIS-EPFL), Sabine Hauert (LIS-EPFL), Severin Leven (LIS-EPFL), and Laurent Keller (Department of Evolutionary Biology, University of Lausanne).
Speaker Biography
Dr. Dario Floreano is an associate EPFL professor and is founder/director of the Laboratory of Intelligent Systems at EPFL. He received an M.A. in visual psychophysics at the University of Trieste in 1988, an M.Sc. in Neural Computation from the University of Stirling in 1992, and a PhD in Cognitive Systems and Robotics from University of Trieste in 1995. He was research fellow at the National Research Council in Roma, at the Department of Computer Science and Mathematics of the University of Stirling, at the Department of Computer Science of EPFL, and at Sony Computer Science Laboratory in Tokyo. In 2000 he was awarded a Swiss NSF professorship to pursue research in bio-inspired robotics at EPFL. He has chaired several international conferences, is co-founder of the Artificial Life Society, Inc., member of the Board of Governors of the International Society for Neural Networks, and is on the editorial board of the following international journals: Neural Networks (Pergamon), Genetic Programming and Evolvable Machines (Kluwer), Adaptive Behavior (Sage), Artificial Life (MIT Press), Connection Science (Francis & Taylor), Evolutionary Computation (MIT Press), IEEE Transactions on Evolutionary Computation (IEEE), and Autonomous Robots (Kluwer).