In the section of my presentation describing the newest research directions I am pursuing, I will discuss the fundamental opportunities and limitations that arise as we evolve robots and robotic systems that are much larger or much smaller than everyday scale. My talk will focus especially on what tentatively emerge as the two most important issues to consider in terms of scaling relations: the distance that a mobile robot can travel as a function of its size, and the range of communication between robots or between robots and their base stations as a function of their sizes.

For the miniaturization of devices and machines down to atomic and molecular sizes, a micro/nano-robotic approach enabling precision manipulation, manufacturing, and interaction at the micro- and nano-scales is indispensable.  Micro/Nano-robotics as an emerging field is based on micro/nano-scale physics, fabrication, sensing, actuation, system integration, and control taking the scaling effects into consideration.  This presentation will focus on current and near-future micro/nano-robotics research activities at the NanoRobotics Laboratory such as miniature surgical micro-robots, surface-climbing space robots, robots walking on water, and micro/nano-assembly systems.

A new invention at the Robotics Institute, the Sonic Flashlight^TM, enables doctors to see where they are inserting a needle or a scalpel, by superimposing an ultrasound image on the direct line of site in real time.  Extensions of this technology to non-medical technology will allow fire-fighters to see through smoke and scuba divers to look through murky water.

Within our lifetimes, teams of robots will assemble structures in space. The tasks are expected to be complex enough  that humans and robots will need to share control to be efficient and robust. In this talk I will discuss preliminary work in the assembly of a structure by a team composed of three robots and a human operator.

Agents are computer programs that act on behalf of some entity, be it an individual, organization, or another agent. The key characteristics of software agents are situatedness, autonomy, interactivity, and adaptivity. By situatedness we mean that an agent is in an environment, it receives sensory inputs from the environment, and it acts to change the environment in some way. By autonomy, we mean that an agent is goal-directed, and has control over its own actions and internal state, i.e. it can reason and act without direct intervention of others (humans or agents). By interactivity, we mean that an agent interacts in peer-to-peer fashion with other agents or humans in multiparty interactions in order to accomplish its various goals. By adaptivity we mean that an agent exhibits proactive and reactive behavior to changes in its environment. The confluence of three technologies, namely Agents, Web Services, and the Semantic Web give rise to many opportunities for further automating business relationships on the Internet and for ubiquitous connectivity and computing. In this talk, I will expand on this vision, discuss challenges, and present technical work in my lab on Semantic Web Services that could bring about this vision.

Effective allocation of tasks (and resources) is a key capability in multi-robot systems. A nasty fly in this ointment, which makes the problem complicated, is that a robot's cost (or value) of taking on a task generally depends on what other tasks the robot will get. I will discuss an approach for tackling this problem that has emerged from 14 years of my research on allocation mechanisms for parties with such combinatorial preferences. I will start from peer-to-peer negotiation with combinatorial contract types, and then present a paradigm shift to a more efficient approach: a combinatorial exchange with a mediated clearing point (such as a chosen robot) that runs a market clearing algorithm. I will also describe how optimal combinatorial exchanges can be used without requiring the robots to compute their costs (local plans) for all combinations of tasks. This is accomplished by using selective incremental preference elicitation by the clearing point.

As autonomous robots become more prevalent, there will be an increasing need for people, especially novices, to interact with them. To be accepted in society, we believe that it is necessary for the robots to behave in a socially acceptable manner -- the robots should behave more like people, so that people do not have to behave robotically. This talk will present research that we have been doing in human-robot social interaction, both in spatial interaction (traveling in hallways, standing in line) and in conversational interaction. We will describe the research in the context of two socially interactive robots -- Grace, a mobile robot developed for the AAAI Robot Challenge, and Valerie, the roboceptionist in Newell Simon Hall.

After several years of researching effective robot soccer teams, in this talk I will analyze the main challenges we have solved and the ones that are still well open to be solved.

Advances in sensing technologies now provide unprecedented ability to obtain and maintain real-time information on the location and status of materials and resources over time. In application areas as diverse as supply chain management, transportation planning, and health care delivery it is now truly possible to base planning, scheduling, and logistics decisions on actual execution states, offering the possibility for much more informed and responsive organizational decision-making. Yet, most contemporary planning and scheduling technologies are not designed for such continuous, “real-time” operations and are not able to take advantage of execution data streams. In this talk, we describe work toward a new class of execution–-driven planning and scheduling technologies. In contract to classical batch-oriented planning and scheduling tools, these technologies rely on dynamic, incremental problem solving techniques, which enable planning to keep pace with execution and promote early, informed responses to evolving execution circumstances. Other synergistic advantages of this incremental approach include flexible integration with human decision-making and natural support for collaborative multi-agent planning. I’'ll demonstrate the concepts of execution-driven planning and scheduling through examples of systems developed for large-scale military planning and logistics applications, discuss the broader applicability of this core technology, and identify outstanding technical challenges.

The TechBridgeWorld initiative was created in 2003 with the goal of investigating the role technology can play in enabling and enhancing sustainable development around the globe. The different programs under this initiative aim to encourage relevant discourse and understanding within the academic community of the positive intersection between technology and development, and of the importance of training technology-aware citizens, and civic-minded technologist, as future leaders.

In the Bio-Robotics Lab, we are developing hyper-redundant robots which are snake-like devices that can thread through tightly packed volumes accessing locations that people and machinery otherwise cannot. These devices also have the feature that they can access such locations in a minimally invasive fashion without disturbing the surrounding areas. Applications for these devices include urban search and rescue, inspection, and surgery.