Engineering and Computer Science at M.I.T. and a member of the Artificial Intelligence Laboratory where he leads the mobile robot group. He has authored two books, numerous scientific papers, and is the editor of the International Journal of Computer Vision. There is an alternative route to Artificial Intelligence that diverges from the directions pursued under that banner for the last thirty some years. The traditional approach has emphasized the abstract manipulation of symbols, whose grounding, in physical reality has. rarely been achieved. We explore a research methodology which emphasizes ongoing physical interaction with the environment as the primary source of constraint on the design of intelligent systems. We show how this methodology has recently had significant successes on a par with the most successful classical efforts. We outline plausible future work along these lines which can lead to vastly more ambitious systems. 1.

firestorm of interest in autonomous robots with the introduction of the Subsumption architecture 1 [Brooks86]. At the time, the dominant view in the AI community was that a control system for an autonomous mobile robot should be decomposed into three functional elements: a sensing system, a planning system, and an execution system [Nilsson80]. The job of the sensing system is to translate raw sensor input (usually sonar or vision data) into a world model. The job of the planner is to take the world model and a goal and generate a plan to achieve the goal. The job of the execution system is to take the plan and generate the actions it prescribes. The sense-plan-act (SPA) approach has two significant architectural features. First, the flow of

. Specifying a reactive behavioral configuration for use by a multiagent team requires both a careful choice of the behavior set and the creation of a temporal chain of behaviors which executes the mission. This difficult task is simplified by applying an object-oriented approach to the design of the mission using a construction called an assemblage and a methodology called temporal sequencing. The assemblage construct allows building high level primitives which provide abstractions for the designer. Assemblages consist of groups of basic behaviors and coordination mechanisms that allow the group to be treated as a new coherent behavior. Upon instantiation, the assemblage is parameterized based on the specific mission requirements. Assemblages can be re-parameterized and used in other states within a mission or archived as high level primitives for use in subsequent projects. Temporal sequencing partitions the mission into discrete operating states with perceptual triggers causing tra...

This thesis presents a perceptual system for a humanoid robot that integrates abilities such as object localization and recognition with the deeper developmental machinery required to forge those competences out of raw physical experiences. It shows that a robotic platform can build up and maintain a system for object localization, segmentation, and recognition, starting from very little. What the robot starts with is a direct solution to achieving figure/ground separation: it simply `pokes around' in a region of visual ambiguity and watches what happens. If the arm passes through an area, that area is recognized as free space. If the arm collides with an object, causing it to move, the robot can use that motion to segment the object from the background. Once the robot can acquire reliable segmented views of objects, it learns from them, and from then on recognizes and segments those objects without further contact. Both low-level and high-level visual features can also be learned in this way, and examples are presented for both: orientation detection and affordance recognition, respectively.

Complex systems and complex missions take years of planning and force launches to become incredibly expensive. The longer the planning and the more expensive the mission, the more catastrophic if it fails. The solution has always been to plan better, add redundancy, test thoroughly and use high quality components. Based on our experience in building ground based mobile robots (legged and wheeled) we argue here for cheap, fast missions using large numbers of mass produced simple autonomous robots that are small by today's standards (1 to 2 Kg). We argue that the time between mission conception and implementation can be radically reduced, that launch mass can be slashed, that totally autonomous robots can be more reliable than ground controlled robots, and that large numbers of robots can change the tradeoff between reliability of individual components and overall mission success. Lastly, we suggest that within a few years it will be possible at modest cost to invade a planet with millions of tiny robots. 1.

Recently there has been a good deal of interest in using techniques developed for learning from reinforcement to guide learning in robots. Motivated by the desire to find better robot learning methods, this thesis presents a number of novel extensions to existing techniques for controlling exploration and inference in reinforcement learning. First I distinguish between the well known exploration-exploitation trade-off and what I term exploration for future exploitation. It is argued that there are many tasks where it is more appropriate to maximise this latter measure. In particular it is appropriate when we want to employ learning algorithms as part of the process of designing a controller. Informed by this insight I develop a number of novel measures of the agent's task knowledge. The first of these is a measure of the probability of a particular course of action being the optimal course of action. Estimators are developed for this measure for boolean and non-boolean processes. These...

Formal models of animal sensorimotor behavior can provide effective methods for generating robotic intelligence. In this article we describe how schema-theoretic models of the praying mantis derived from behavioral and neuroscientific data can be implemented on a hexapod robot equipped with a real-time color vision system. This implementation incorporates a wide range of behaviors, including obstacle avoidance, prey acquisition, predator avoidance, mating, and chantlitaxia behaviors that can provide guidance to neuroscientists, ethologists, and roboticists alike. The goals of this study are threefold: to provide an understanding and means by which fielded robotic systems are not competing with other agents that are more effective at their designated task; to permit them to be successful competitors within the ecological system and capable of displacing less efficient agents; and that they are ecologically sensitive so that agent-environment dynamics are well-modeled and as predictable ...

The aim of this paper is to discuss and present some experimental data on the use of optical, chemical and ultrasound sensors to control the motion of single and cooperating robots. The distinctive feature is that in both cases the motion of the robots is not driven by a localized target (e.g. a light, or a sound) but by a concentration gradient such as the one generated by a gas leak. Experiments using a simple optical sensor reading a variable density pattern of dark spots painted on the ground plane, is presented along with experiments performed using simple chemical sensors measuring the local gradient of a substance. Different strategies will be discussed based upon different configurations of the sensing devices and reflexive motor controls. The goal of the experiment is to analyze the performance of simple autonomous robots in reaching the point of highest concentration (i.e. the gas leak). This performance has been analyzed for single robots and for cooperating robots using a v...

In practical applications of robotics, it is usually quite difficult, if not impossible, for the system designer to fully predict the environmental states in which the robots will operate. The complexity of the problem is further increased when dealing with teams of robots which themselves may be incompletely known and characterized in advance. It is thus highly desirable for robot teams to be able to adapt their performance during the mission due to changes in the environment, or to changes in other robot team members. In previous work [40, 44], we introduced a behavior-based mechanism --- called the ALLIANCE architecture --- that facilitates the fault tolerant cooperative control of multi-robot teams. However, this previous work did not address the issue of how to dynamically update the control parameters during a mission to adapt to ongoing changes in the environment or in the robot team, and to ensure the efficiency of the collective team actions. In this paper, we address this iss...

Computer programming environments for learning should make it easy to create worlds of responsive and autonomous objects, such as video games or simulations of animal behavior. But building such worlds remains difficult, partly because the models and metaphors underlying traditional programming languages are not particularly suited to the task. This dissertation investigates new metaphors, environments, and languages that make possible new ways to create programs -- and, more broadly, new ways to think about programs. In particular, it introduces the idea of programming with "agents" as a means to help people create worlds involving responsive, interacting objects. In this context, an agent is a simple mechanism intended to be understood through anthropomorphic metaphors and endowed with certain lifelike properties such as autonomy, purposefulness, and emotional state. Complex behavior is achieved by combining simple agents into more complex structures. While the agent metaphor enables...