This thesis addresses situated, embodied agents interacting in complex domains. It focuses on two problems: 1) synthesis and analysis of intelligent group behavior, and 2) learning in complex group environments. Basic behaviors, control laws that cluster constraints to achieve particular goals and have the appropriate compositional properties, are proposed as effective primitives for control and learning. The thesis describes the process of selecting such basic behaviors, formally specifying them, algorithmically implementing them, and empirically evaluating them. All of the proposed ideas are validated with a group of up to 20 mobile robots using a basic behavior set consisting of: safe--wandering, following, aggregation, dispersion, and homing. The set of basic behaviors acts as a substrate for achieving more complex high--level goals and tasks. Two behavior combination operators are introduced, and verified by combining subsets of the above basic behavior set to implement collective flocking, foraging, and docking. A methodology is introduced for automatically constructing higher--level behaviors

... This article illustrates the potential learning capabilities of purely local learning and offers an interesting and powerful approach to learning with receptive fields.

This paper proposes the concept of basis behaviors as ubiquitous general building blocks for synthesizing artificial group behavior in multi--agent systems, and for analyzing group behavior in nature. We demonstrate the concept through examples implemented both in simulation and on a group of physical mobile robots. The basis behavior set we propose, consisting of avoidance, safe--wandering, following, aggregation, dispersion, and homing, is constructed from behaviors commonly observed in a variety of species in nature. The proposed behaviors are manifested spatially, but have an effect on more abstract modes of interaction, including the exchange of information and cooperation. We demonstrate how basis behaviors can be combined into higher--level group behaviors commonly observed across species. The combination mechanisms we propose are useful for synthesizing a variety of new group behaviors, as well as for analyzing naturally occurring ones. Key words: group behavior, robotics, eth...

Nearest-neighbour interpolation algorithms have many useful properties for applications to learning, but they often exhibit poor generalization. In this paper, it is shown that much better generalization can be obtained by using a variable interpolation kernel in combination with conjugate gradient optimization of the similarity metric and kernel size. The resulting method is called variable-kernel similarity metric (VSM) learning. It has been tested on several standard classification data sets, and on these problems it shows better generalization than back propagation and most other learning methods. An important advantage is that the system can operate as a black box in which no model minimization parameters need to be experimentally set by the user. The number of parameters that must be determined through optimization are orders of magnitude less than for back-propagation or RBF networks, which may indicate that the method better captures the essential degrees of variation in learni...

Discovery involves collaboration among many intelligent activities. However, little is known about how and in what form such collaboration occurs. In this paper, a framework is proposed for autonomous systems that learn and discover from their environment. Within this framework, many intelligent activities such as perception, action, exploration, experimentation, learning, problem solving, and new term construction can be integrated in a coherent way. The framework is presented in detail through an implemented system called LIVE, and is evaluated through the performance of LIVE on several discovery tasks. The conclusion is that autonomous learning from the environment is a feasible approach for integrating the activities involved in a discovery process. 1 Introduction Learning from the environment requires integration of a variety of activities. A learning system must be able to explore, to plan, to experiment, to adapt, and to discover. These activities should be studied together in ...

. We discuss the tradeoffs involved in control of complex articulated agents, and present three implemented controllers for a complex task: a physically-based humanoid torso dancing the Macarena. The three controllers are drawn from animation, biological models, and robotics, and illustrate the issues of joint-space vs. Cartesian space task specification and implementation. We evaluate the controllers along several qualitative and quantitative dimensions, considering naturalness of movement and controller flexibility. Finally, we propose a general combination approach to control, aimed at utilizing the strengths of each alternative within a general framework for addressing complex motor control of articulated agents. Key words: articulated agent control, motor control, robotics, animation 1. Introduction Control of humanoid agents, dynamically simulated or physical, is an extremely difficult problem due to the high dimensionality of the control space, i.e., the many degrees of freed...

We describe and compare two implemented controllers for Adonis, a physically simulated humanoid torso, one based on joint-space torques and the other on convergent force-fields applied to the hands. The two come from different application domains: the former is a common approach in manipulator robotics and graphics, while the latter is inspired by biological limb control. Both avoid explicit inverse kinematic calculations found in standard Cartesian control, trading generality of motion for programming efficiency. The two approaches are compared on a common sequential task, the familiar dance "Macarena" and evaluated based on ease of generating new behaviors, flexibility, and naturalness of movement; we also compare them against human performance on the same task. Finally, we discuss the tradeoffs and present a more general framework for addressing complex motor control of simulated agents. 1 Introduction Control of humanoid agents, dynamically simulated or physical, is an extremely ...

We introduce a constructive, incremental learning system for regression problems that models data by means of spatially localized linear models. In contrast to other approaches, the size and shape of the receptive field of each locally linear model as well as the parameters of the locally linear model itself are learned independently, i.e., without the need for competition or any other kind of communication. This characteristic is accomplished by incrementally minimizing a weighted penalized local cross validation error. As a result, we obtain a learning system that can allocate resources as needed while dealing with the bias-variance dilemma in a principled way. The spatial localization of the linear models increases robustness towards negative interference. Our learning system can be interpreted as a nonparametric adaptive bandwidth smoother, as a mixture of experts where the experts are trained in isolation, and as a learning system which profits from combining independent expert knowledge on the same problem. It illustrates the potential learning capabilities of purely local learning and offers an interesting and powerful approach to learning with receptive fields.

The cerebellum has long been known to be associated with the coordination of the human motor system. Contemporary research indicates that this is one product of the cerebellum's true function: the generation of dynamic models of systems both within and without the body. This thesis describes the instantiation of one such model on the humanoid robot Cog, developed at the MIT Artificial Intelligence Laboratory. The model takes the form of an adaptive mapping of head movements into anticipated motion in the visual field. This model is part of a visual subsystem which allows Cog to detect motion in the environment without being confused by movement of its own head. The author hopes that this work will be the first step in creating a generalized system for generating models between sensorimotor systems, and that such a system will be the first step in the development of a fully-functional artificial cerebellum for Cog. Thesis Supervisor: Rodney A. Brooks Title: Professor of Electrical Eng...

optimization algorithms for dynamics-based robot movement generation. The robots that we consider are modeled as rigid multibody systems containing multiple closed loops, active and passive joints, and redundant actuators and sensors. While one can, in principle, always derive in analytic form the equations of motion for such systems, the ensuing complexity, both numeric and symbolic, of the equations makes classical optimization-based movement-generation schemes impractical for all but the simplest of systems. In particular, numerically approximating the gradient and Hessian often leads to ill-conditioning and poor convergence behavior. We show in this paper that, by extending (to the general class of systems described above) a Lie theoretic formulation of the equations of motion originally developed for serial chains, it is possible to recursively evaluate the dynamic equations, the analytic gradient, and even the Hessian for a number of physically plausible objective functions. We show through several case studies that, with exact gradient and Hessian information, descent-based optimization methods can be forged into an effective and reliable tool for generating physically natural robot movements. Index Terms—Closed chain, movement optimization, multibody system dynamics, Newton’s method, redundant actuation, robot dynamics. I.