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.

Off-line simulation inspires insight: a neurodynamics approach to efficient robot task learning

There is currently an increasing demand for robots able to acquire the sequential organization of tasks from social learning interactions with ordinary people. Interactive learning-by-demonstration and communication is a promising research topic in current robotics research. However, the efficient acquisition of generalized task representations that allow the robot to adapt to different users and contexts is a major challenge. In this paper, we present a dynamic neural field (DNF) model that is inspired by the hypothesis that the nervous system uses the off-line re-activation of initial memory traces to incrementally incorporate new information into structured knowledge. To achieve this, the model combines fast activation-based learning to robustly represent sequential information from single task demonstrations with slower, weight-based learning during internal simulations to establish longer-term associations between neural populations representing individual subtasks. The efficiency of the learning process is tested in an assembly paradigm in which the humanoid robot ARoS learns to construct a toy vehicle from its parts. User demonstrations with different serial orders together with the correction of initial prediction errors allow the robot to acquire generalized task knowledge about possible serial orders and the longer term dependencies between subgoals in very few social learning interactions. This success is shown in a joint action scenario in which ARoS uses the newly acquired assembly plan to construct the toy together with a human partner.

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.

DEVELOPMENT OF A HUMANOID ROBOT ARM FOR USE IN URBAN ENVIRONMENTS

The Bucknell Humanoid Robot Arm project was developed in order toprovide a lightweight robotic arm for the IHMC / Bucknell University bipedal robot that will provide a means of manipulation and facilitate operations in urban environments. The resulting fabricated arm described in this thesis weighs only 13 pounds, and is capable of holding 11 pounds fully outstretched, lifting objects such as tools, and it can open doors. It is also capable of being easily integrated with the IHMC / Bucknell University biped. This thesis provides an introduction to robots themselves, discusses the goals of the Bucknell Humanoid Robot Arm project, provides a background on some of the existing robots, and shows how the Bucknell Humanoid Robot Arm fits in with the studies that have been completed. After reading these studies, important items such as design trees and operational scenarios were completed. The completion of these items led to measurable specifications and later the design requirements and specifications. A significant contribution of this thesis to the robotics discipline involves the design of the actuator itself. The arm uses of individual, lightweight, compactly designed actuators to achieve desired capabilities and performance requirements. Many iterations were completed to get to the final design of each actuator. After completing the actuators, the design of the intermediate links and brackets was finalized. Completion of the design led to the development of a complex controls system which used a combination of Clanguage and Java.

Biologically inspired joint control for a humanoid robot

The GuRm is a 1.2m tall, 23 degree of freedom humanoid consuucted at the University of Queensland for research into humanoid robotics. The key challenge being addressed by the GuRw projcct is the development of appropriate learning strategies for control and coodinadon of the robot’s many joints. The development of learning strategies is Seen as a way to sidestep the inherent intricacy of modeling a multi-DOP biped robot. This paper outlines the approach taken to generate an appmpria*e control scheme for the joinis of the GuRoo. The paper demonsrrates the determination of local feedback control parameters using a genetic algorithm. The feedback loop is then augmented by a predictive modulator that learns a form of feed-fonward control to overcome the irregular loads experienced at each joint during the gait cycle. The predictive modulator is based on thc CMAC architecture. Results from tats on the GuRoo platform show that both systems provide improvements in stability and tracking of joint control.

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.

Trajectory planning of a robot using learning algorithms

The authors consider the problem of a robot manipulator operating in a noisy workspace. The manipulator is required to move from an initial position P(i) to a final position P(f). P(i) is assumed to be completely defined. However, P(f) is obtained by a sensing operation and is assumed to be fixed but unknown. The authors approach to this problem involves the use of three learning algorithms, the discretized linear reward-penalty (DLR-P) automaton, the linear reward-penalty (LR-P) automaton and a nonlinear reinforcement scheme. An automaton is placed at each joint of the robot and by acting as a decision maker, plans the trajectory based on noisy measurements of P(f).

Pose estimation of a humanoid robot using images from a mobile external camera

NOGUEIRA, Marcelo B. ; MEDEIROS, Adelardo A. D. ; ALSINA, Pablo J. Pose Estimation of a Humanoid Robot Using Images from an Mobile Extern Camera. In: IFAC WORKSHOP ON MULTIVEHICLE SYSTEMS, 2006, Salvador, BA. Anais... Salvador: MVS 2006, 2006.

Data from PhD thesis

The ocean quahog, Arctica islandica is the longest-lived non-colonial animal known to science. A maximum individual age of this bivalve of 405 years has been found in a population off the north western coast of Iceland. Conspicuously shorter maximum lifespan potentials (MLSPs) were recorded from other populations of A. islandica in European waters (e.g. Kiel Bay: 30 years, German Bight: 150 years) which experience wider temperature and salinity fluctuations than the clams from Iceland. The aim of my thesis was to identify possible life-prolonging physiological strategies in A. islandica and to examine the modulating effects of extrinsic factors (e.g. seawater temperature, food availability) and intrinsic factors (e.g. species-specific behavior) on these strategies. Burrowing behavior and metabolic rate depression (MRD), tissue-specific antioxidant and anaerobic capacities as well as cell-turnover (= apoptosis and proliferation) rates were investigated in A. islandica from Iceland and the German Bight. An inter-species comparison of the quahog with the epibenthic scallop Aequipecten opercularis (MLSP = 8-10 years) was carried out in order to determine whether bivalves with short lifespans and different lifestyles also feature a different pattern in cellular maintenance and repair. The combined effects of a low-metabolic lifestyle, low oxidative damage accumulation, and constant investment into cellular protection and tissue maintenance, appear to slow-down the process of physiological aging in A. islandica and to afford the extraordinarily long MLSP in this species. Standard metabolic rates were lower in A. islandica when compared to the shorter-lived A. opercularis. Furthermore, A. islandica regulate mantle cavity water PO2 to mean values < 5 kPa, a PO2 at which the formation of reactive oxygen species (ROS) in isolated gill tissues of the clams was found to be 10 times lower than at normoxic conditions (21 kPa). Burrowing and metabolic rate depression (MRD) in Icelandic specimens were more pronounced in winter, possibly supported by low seawater temperature and food availability, and seem to be key energy-saving and life-prolonging parameters in A. islandica. The signaling molecule nitric oxide (NO) may play an important role during the onset of MRD in the ocean quahog by directly inhibiting cytochome-c-oxidase at low internal oxygenation upon shell closure. In laboratory experiments, respiration of isolated A. islandica gills was completely inhibited by chemically produced NO at low experimental PO2 <= 10 kPa. During shell closure, mantle cavity water PO2 decreased to 0 kPa for longer than 24 h, a state in which ROS production is supposed to subside. Compared to other mollusk species, onset of anaerobic metabolism is late in A. islandica in the metabolically reduced state. Increased accumulation of the anaerobic metabolite succinate was initially detected in the adductor muscle of the clams after 3.5 days under anoxic incubation or in burrowed specimens. A ROS-burst was absent in isolated gill tissue of the clams following hypoxia (5 kPa)-reoxygenation (21 kPa). Accordingly, neither the activity of antioxidant enzymes superoxide dismutase (SOD) and catalase (CAT), nor the specific content of the ROS-scavenger glutathione (GSH) was enhanced in different tissues of the ocean quahog after 3.5 days of self-induced or forced hypoxia/anoxia to prepare for an oxidative burst. While reduced ROS formation compared to routine levels lowers oxidative stress during MRD and also during surfacing, the general preservation of high cellular defense and the efficient removal and replacement of damaged cells over lifetime seem to be of crucial importance in decelerating the senescent decline in tissues of A. islandica. Along with stable antioxidant protection over 200 years of age, proliferation rates and apoptosis intensities in most investigated tissues of the ocean quahog were low, but constant over 140 years of age. Accordingly, age-dependent accumulations of protein and lipid oxidation products are lower in A. islandica tissues when compared to the shorter-lived bivalve A. opercularis. The short-lived swimming scallop is a model bivalve species representing the opposite life and aging strategy to A. islandica. In this species permanently high energy throughput, reduced investment into antioxidant defense with age, and higher accumulation of oxidation products are met by higher cell turnover rates than in the ocean quahog. The only symptoms of physiological change over age ever found in A. islandica were decreasing cell turnover rates in the heart muscle over a lifetime of 140 years. This may either indicate higher damage levels and possibly ongoing loss of functioning in the heart of aging clams, or, the opposite, lower rates of cell damage and a reduced need for cell renewal in the heart tissue of A. islandica over lifetime. Basic physiological capacities of different A. islandica populations, measured at controlled laboratory conditions, could not explain considerable discrepancies in population specific MLSPs. For example, levels of tissue-specific antioxidant capacities and cell turnover rates were similarly high in individuals from the German Bight and from Iceland. Rather than genetic differences, the local impacts of environmental conditions on behavioral and physiological traits in the ocean quahog seem to be responsible for differences in population-specific MLSPs.

Inverse kinematics of a humanoid robot with non-spherical hip: a hybrid algorithm approach

This paper describes an approach to solve the inverse kinematics problem of humanoid robots whose construction shows a small but non negligible offset at the hip which prevents any purely analytical solution to be developed. Knowing that a purely numerical solution is not feasible due to variable efficiency problems, the proposed one first neglects the offset presence in order to obtain an approximate “solution” by means of an analytical algorithm based on screw theory, and then uses it as the initial condition of a numerical refining procedure based on the Levenberg‐Marquardt algorithm. In this way, few iterations are needed for any specified attitude, making it possible to implement the algorithm for real‐time applications. As a way to show the algorithm’s implementation, one case of study is considered throughout the paper, represented by the SILO2 humanoid robot.

Internet-Based Supervisory Teleoperation of a Virtual Humanoid Robot

This paper describes the approach used by the Sarbot-Team for controlling the Atlas humanoid robot during the DARPA Virtual Robotics Challenge that took place in June 2013. Herein we present a proposal for overcoming the restrictions on performance caused by limited bandwidth, high latency and the effects of signal degradation induced by beyond line of sight (BLOS) conditions, RF interference, and other related circumstances. Experimental evaluation confirmed the effectiveness of our approach and present an alternative for coping with constrained communication conditions during the control of humanoid robot deployed at unattended areas.