An ambitious goal in the area of physics-based computer animation is the creation of virtual actors that autonomously synthesize realistic human motions and possess a broad repertoire of lifelike motor skills. To this end, the control of dynamic, anthropomorphic figures subject to gravity and contact forces remains a difficult open problem. We propose a framework for composing controllers in order to enhance the motor abilities of such figures. A key contribution of our composition framework is an explicit model of the "pre-conditions" under which motor controllers are expected to function properly. We demonstrate controller composition with pre-conditions determined not only manually, but also automatically based on Support Vector Machine (SVM) learning theory. We evaluate our composition framework using a family of controllers capable of synthesizing basic actions such as balance, protective stepping when balance is disturbed, protective arm reactions when falling, and multiple ways of standing up after a fall. We furthermore demonstrate these basic controllers working in conjunction with more dynamic motor skills within a prototype virtual stuntperson. Our composition framework promises to enable the community of physics-based animation practitioners to easily exchange motor controllers and integrate them into dynamic characters.

We describe a framework for cooperative control of a group of nonholonomic mobile robots that allows us to build complex systems from simple controllers and estimators. The resultant modular approach is attractive because of the potential for reusability. Our approach to composition also guarantees stability and convergence in a wide range of tasks. There are two key features in our approach: 1) a paradigm for switching between simple decentralized controllers that allows for changes in formation; 2) the use of information from a single type of sensor, an omnidirectional camera, for all our controllers. We describe estimators that abstract the sensory information at different levels, enabling both decentralized and centralized cooperative control. Our results include numerical simulations and experiments using a testbed consisting of three nonholonomic robots.

This paper serves to present our initial thoughts on the design of an architecture for highly flexible, modular, and distributed precision assembly systems. Through the use of a unified design, simulation, programming, and monitoring environment coupled with selfrepresenting cooperative agents this architecture will significantly simplify the process of factory design and deployment. Our demonstration of this architecture takes the form of a "table top" sized assembly system targeted at the partial assembly of high-density magnetic storage devices. 1 Introduction As part of a multi-million dollar four-year project funded under the NSF Multi Disciplinary Challenges component of the High Performance Computing and Communication program, this paper documents our initial thoughts on the development of a miniaturized, modular, high-precision assembly system or minifactory. We believe this type of system can only be developed through careful integration of hardware and software tools in a m...

This paper develops a method of composing simple control policies, applicable over a limited region in a dynamical system's free space, such that the resulting composition completely solves the navigation and control problem for the given system operating in a constrained environment. The resulting control policy deployment induces a global control policy that brings the system to the goal, provided that there is a single connected component of the free space containing both the start and goal configurations. In this paper, control policies for both kinematic and simple dynamical systems are developed. This work assumes that the initial velocities are somewhat aligned with the desired velocity vector field. We conclude by offering an outline of an approach for accommodating arbitrary dynamical constraints and initial conditions.

We present a new subgoal-based method for automatically creating useful skills in reinforcement learning. Our method identifies subgoals by partitioning local state transition graphs—those that are constructed using only the most recent experiences of the agent. The local scope of our subgoal discovery method allows it to successfully identify the type of subgoals we seek—states that lie between two densely-connected regions of the state space—while producing an algorithm with low computational cost.

The autonomous execution of manipulation tasks in unstructured, dynamic environments requires the consideration of various motion constraints. Any motion performed during the manipulation task has to satisfy constraints imposed by the task itself, but also has to consider kinematic and dynamic limitations of the manipulator, avoid unpredictably moving obstacles, and observe constraints imposed by the global connectivity of the workspace. Furthermore, the unpredictability of unstructured environments requires the continuous incorporation of feedback to reliably satisfy these constraints. We present a novel feedback motion planning approach, called elastic roadmap framework, capable of satisfying all of the motion constraints that arise in autonomous mobile manipulation and their respective feedback requirements. This framework is validated with simulation experiments using a mobile manipulation platform and a stationary manipulator.

We present the Incremental Focus of Attention (IFA) architecture for robust, adaptive, realtime motion tracking. IFA systems combine several visual search and vision-based tracking algorithms into a layered hierarchy. The architecture controls the transitions between layers and executes algorithms appropriate to the visual environment at hand: When conditions are good, tracking is accurate and precise; as conditions deteriorate, more robust, yet less accurate algorithms take over; when tracking is lost altogether, layers cooperate to perform a rapid search for the target in order to recover it and continue tracking. Implemented IFA systems are extremely robust to most common types of temporary visual disturbances. They resist minor visual perturbances and recover quickly after full occlusions, illumination changes, major distractions, and target disappearances. Analysis of the algorithm's recovery times are supported by simulation results and experiments on real data. In particular, ex...

RHex is an untethered, compliant leg hexapod robot that travels at better than one body length per second over terrain few other robots can negotiate at all. Inspired by

Abstract — We provide a method for planning under uncertainty for robotic manipulation by partitioning the configuration space into a set of regions that are closed under compliant motions. These regions can be treated as states in a partially observable Markov decision process (POMDP), which can be solved to yield optimal control policies under uncertainty. We demonstrate the approach on simple grasping problems, showing that it can construct highly robust, efficiently executable solutions. I.

This paper presents a class of fundamental control policies suitable for use in a novel method for designing and specifying the dynamic motion of robotic systems. Through recourse to formal stability mechanisms a hybrid control strategy (one that relies on switching of underlying continious policies) with desirable performance characteristics is demonstrated to be both stable and expressive for programming such motions. The long term intent is to utilize slightly more general control methods of this form to drastically simplify the process of integrating and programming modular automated assembly systems.