From First Contact to Close Encounters: A Developmentally Deep Perceptual System for a Humanoid Robot

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. The motivation for this work is simple. Training on large corpora of annotated real-world data has proven crucial for creating robust solutions to perceptual problems such as speech recognition and face detection. But the powerful tools used during training of such systems are typically stripped away at deployment. Ideally they should remain, particularly for unstable tasks such as object detection, where the set of objects needed in a task tomorrow might be different from the set of objects needed today. The key limiting factor is access to training data, but as this thesis shows, that need not be a problem on a robotic platform that can actively probe its environment, and carry out experiments to resolve ambiguity. This work is an instance of a general approach to learning a new perceptual judgment: find special situations in which the perceptual judgment is easy and study these situations to find correlated features that can be observed more generally.

Teaching an Old Robot New Tricks: Learning Novel Tasks via Interaction with People and Things

As AI has begun to reach out beyond its symbolic, objectivist roots into the embodied, experientialist realm, many projects are exploring different aspects of creating machines which interact with and respond to the world as humans do. Techniques for visual processing, object recognition, emotional response, gesture production and recognition, etc., are necessary components of a complete humanoid robot. However, most projects invariably concentrate on developing a few of these individual components, neglecting the issue of how all of these pieces would eventually fit together. The focus of the work in this dissertation is on creating a framework into which such specific competencies can be embedded, in a way that they can interact with each other and build layers of new functionality. To be of any practical value, such a framework must satisfy the real-world constraints of functioning in real-time with noisy sensors and actuators. The humanoid robot Cog provides an unapologetically adequate platform from which to take on such a challenge. This work makes three contributions to embodied AI. First, it offers a general-purpose architecture for developing behavior-based systems distributed over networks of PC's. Second, it provides a motor-control system that simulates several biological features which impact the development of motor behavior. Third, it develops a framework for a system which enables a robot to learn new behaviors via interacting with itself and the outside world. A few basic functional modules are built into this framework, enough to demonstrate the robot learning some very simple behaviors taught by a human trainer. A primary motivation for this project is the notion that it is practically impossible to build an "intelligent" machine unless it is designed partly to build itself. This work is a proof-of-concept of such an approach to integrating multiple perceptual and motor systems into a complete learning agent.

Enhancing the Dexterity of a Robot Hand Using Controlled Slip

Humans can effortlessly manipulate objects in their hands, dexterously sliding and twisting them within their grasp. Robots, however, have none of these capabilities, they simply grasp objects rigidly in their end effectors. To investigate this common form of human manipulation, an analysis of controlled slipping of a grasped object within a robot hand was performed. The Salisbury robot hand demonstrated many of these controlled slipping techniques, illustrating many results of this analysis. First, the possible slipping motions were found as a function of the location, orientation, and types of contact between the hand and object. Second, for a given grasp, the contact types were determined as a function of the grasping force and the external forces on the object. Finally, by changing the grasping force, the robot modified the constraints on the object and affect controlled slipping slipping motions.

Structural Dynamics Model of a Cartesian Robot

Methods are developed for predicting vibration response characteristics of systems which change configuration during operation. A cartesian robot, an example of such a position-dependent system, served as a test case for these methods and was studied in detail. The chosen system model was formulated using the technique of Component Mode Synthesis (CMS). The model assumes that he system is slowly varying, and connects the carriages to each other and to the robot structure at the slowly varying connection points. The modal data required for each component is obtained experimentally in order to get a realistic model. The analysis results in prediction of vibrations that are produced by the inertia forces as well as gravity and friction forces which arise when the robot carriages move with some prescribed motion. Computer simulations and experimental determinations are conducted in order to calculate the vibrations at the robot end-effector. Comparisons are shown to validate the model in two ways: for fixed configuration the mode shapes and natural frequencies are examined, and then for changing configuration the residual vibration at the end of the mode is evaluated. A preliminary study was done on a geometrically nonlinear system which also has position-dependency. The system consisted of a flexible four-bar linkage with elastic input and output shafts. The behavior of the rocker-beam is analyzed for different boundary conditions to show how some limiting cases are obtained. A dimensional analysis leads to an evaluation of the consequences of dynamic similarity on the resulting vibration.

Analysis and Control of Robot Manipulators with Kinematic Redundancy

A closed-form solution formula for the kinematic control of manipulators with redundancy is derived, using the Lagrangian multiplier method. Differential relationship equivalent to the Resolved Motion Method has been also derived. The proposed method is proved to provide with the exact equilibrium state for the Resolved Motion Method. This exactness in the proposed method fixes the repeatability problem in the Resolved Motion Method, and establishes a fixed transformation from workspace to the joint space. Also the method, owing to the exactness, is demonstrated to give more accurate trajectories than the Resolved Motion Method. In addition, a new performance measure for redundancy control has been developed. This measure, if used with kinematic control methods, helps achieve dexterous movements including singularity avoidance. Compared to other measures such as the manipulability measure and the condition number, this measure tends to give superior performances in terms of preserving the repeatability property and providing with smoother joint velocity trajectories. Using the fixed transformation property, Taylor's Bounded Deviation Paths Algorithm has been extended to the redundant manipulators.

Modeling Robot Dynamic Performance for Endpoint Force Control

This research aims to understand the fundamental dynamic behavior of servo-controlled machinery in response to various types of sensory feedback. As an example of such a system, we study robot force control, a scheme which promises to greatly expand the capabilities of industrial robots by allowing manipulators to interact with uncertain and dynamic tasks. Dynamic models are developed which allow the effects of actuator dynamics, structural flexibility, and workpiece interaction to be explored in the frequency and time domains. The models are used first to explain the causes of robot force control instability, and then to find methods of improving this performance.

Task-Level Robot Learning

We are investigating how to program robots so that they learn from experience. Our goal is to develop principled methods of learning that can improve a robot's performance of a wide range of dynamic tasks. We have developed task-level learning that successfully improves a robot's performance of two complex tasks, ball-throwing and juggling. With task- level learning, a robot practices a task, monitors its own performance, and uses that experience to adjust its task-level commands. This learning method serves to complement other approaches, such as model calibration, for improving robot performance.

Design and Implementation of a Flexible Robot

This robot has low natural frequencies of vibration. Insights into the problems of designing joint and link flexibility are discussed. The robot has three flexible rotary actuators and two flexible, interchangeable links, and is controlled by three independent processors on a VMEbus. Results from experiments on the control of residual vibration for different types of robot motion are presented. Impulse prefiltering and slowly accelerating moves are compared and shown to be effective at reducing residual vibration.

On Probabilistic Strategies for Robot Tasks

Robots must act purposefully and successfully in an uncertain world. Sensory information is inaccurate or noisy, actions may have a range of effects, and the robot's environment is only partially and imprecisely modeled. This thesis introduces active randomization by a robot, both in selecting actions to execute and in focusing on sensory information to interpret, as a basic tool for overcoming uncertainty. An example of randomization is given by the strategy of shaking a bin containing a part in order to orient the part in a desired stable state with some high probability. Another example consists of first using reliable sensory information to bring two parts close together, then relying on short random motions to actually mate the two parts, once the part motions lie below the available sensing resolution. Further examples include tapping parts that are tightly wedged, twirling gears before trying to mesh them, and vibrating parts to facilitate a mating operation.

A Distributed Model for Mobile Robot Environment-Learning and Navigation

A distributed method for mobile robot navigation, spatial learning, and path planning is presented. It is implemented on a sonar-based physical robot, Toto, consisting of three competence layers: 1) Low-level navigation: a collection of reflex-like rules resulting in emergent boundary-tracing. 2) Landmark detection: dynamically extracts landmarks from the robot's motion. 3) Map learning: constructs a distributed map of landmarks. The parallel implementation allows for localization in constant time. Spreading of activation computes both topological and physical shortest paths in linear time. The main issues addressed are: distributed, procedural, and qualitative representation and computation, emergent behaviors, dynamic landmarks, minimized communication.

Passive and Active Grasping with a Prehensile Robot End-Effector

This report presents a design of a new type of robot end-effector with inherent mechanical grasping capabilities. Concentrating on designing an end-effector to grasp a simple class of objects, cylindrical, allowed a design with only one degree of actuation. The key features of this design are high bandwidth response to forces, passive grasping capabilities, ease of control, and ability to wrap around objects with simple geometries providing form closure. A prototype of this mechanism was built to evaluate these features.

The PHD: A Planar, Harmonic Drive Robot for Joint Torque Control

This thesis details the development of a model of a seven degree of freedom manipulator for position control. Then, it goes on to discuss the design and construction of a the PHD, a robot built to serve two purposes: first, to perform research on joint torque control schemes, and second, to determine the important dynamic characteristics of the Harmonic Drive. The PHD, is a planar, three degree of freedom arm with torque sensors integral to each joint. Preliminary testing has shown that a simple linear spring model of the Harmonic Drive's flexibility is suitable in many situations.

Robot Motion Vision by Fixation

In many motion-vision scenarios, a camera (mounted on a moving vehicle) takes images of an environment to find the "motion'' and shape. We introduce a direct-method called fixation for solving this motion-vision problem in its general case. Fixation uses neither feature-correspondence nor optical-flow. Instead, spatio-temporal brightness gradients are used directly. In contrast to previous direct methods, fixation does not restrict the motion or the environment. Moreover, fixation method neither requires tracked images as its input nor uses mechanical tracking for obtaining fixated images. The experimental results on real images are presented and the implementation issues and techniques are discussed.

Heterogeneous Multi-Robot Cooperation

This report addresses the problem of achieving cooperation within small- to medium- sized teams of heterogeneous mobile robots. I describe a software architecture I have developed, called ALLIANCE, that facilitates robust, fault tolerant, reliable, and adaptive cooperative control. In addition, an extended version of ALLIANCE, called L-ALLIANCE, is described, which incorporates a dynamic parameter update mechanism that allows teams of mobile robots to improve the efficiency of their mission performance through learning. A number of experimental results of implementing these architectures on both physical and simulated mobile robot teams are described. In addition, this report presents the results of studies of a number of issues in mobile robot cooperation, including fault tolerant cooperative control, adaptive action selection, distributed control, robot awareness of team member actions, improving efficiency through learning, inter-robot communication, action recognition, and local versus global control.

Embodiment and Manipulation Learning Process for a Humanoid Hand

Babies are born with simple manipulation capabilities such as reflexes to perceived stimuli. Initial discoveries by babies are accidental until they become coordinated and curious enough to actively investigate their surroundings. This thesis explores the development of such primitive learning systems using an embodied light-weight hand with three fingers and a thumb. It is self-contained having four motors and 36 exteroceptor and proprioceptor sensors controlled by an on-palm microcontroller. Primitive manipulation is learned from sensory inputs using competitive learning, back-propagation algorithm and reinforcement learning strategies. This hand will be used for a humanoid being developed at the MIT Artificial Intelligence Laboratory.