While oceans cover the majority of our planet, these vast expanses remain relatively unexplored. Among the most interesting parts of the ocean are the shallow reef systems, which contain a huge amount of the planet’s biodiversity. The Sustained Engagement Autonomous Tracking of Underwater RepTiLEs or SEATURTLE is a low cost Autonomous Underwater Vehicle designed to carry out missions in these shallow environments. Its small displacement and precise movement make it ideal for navigating tight spaces, and its package of sensors make it easily adaptable to a variety of missions. For this project the vehicle was configured to autonomously track tagged objects underwater, using image recognition and the April Tags system.

<p> A linear finite element procedure is presented for modeling a single flexible link robotic manipulator system. The model takes into consideration the DC motor, torsional springs/dampers, cantilever beam and the accelerometer at the tip. The finite element method for system modeling is validated through comparison to experimental data using the frequency response. With a validated model for the system, feedback control is implemented for vibration reduction and accurate positioning of the beam tip. The accuracy of the controlled model is compared to experiment and a discrete time transfer matrix method of modelling. The closed-loop finite element method model exhibited an accurate response, related to experiment, with encoder feedback and accelerometer feedback using digital compensators</p>

For robotic systems to realise lifelong autonomy they must be able to navigate accurately in changing environments. In this thesis we describe, implement and validate a new approach to the problem of long-term navigation. To begin, we present our stereo visual odometry system which provides highly accurate pose estimation. Our approach combines several techniques found in existing implementations and a recently published image descriptor that simplifies the solution architecture. The performance and versatility of our system is demonstrated through testing on multiple datasets. Equipped with our visual odometry system, we describe a new approach to the problem of lifelong navigation. We learn a model whose complexity varies naturally in accordance with the variation of scene appearance. As the robot repeatedly traverses its workspace, it accumulates distinct visual experiences that, in concert, implicitly represent the scene variation - each experience captures a visual mode. When operating in a previously visited area, we continually try to localise in these previous experiences while simultaneously running the visual odometry. Failure to localise in a sufficient number of prior experiences indicates an insufficient model of the workspace and instigates the laying down of the live image sequence as a new distinct experience. In this way, over time we can capture the typical temporally varying appearance of an environment and the number of experiences required tends to a constant. Although we focus on vision as a primary sensor, the ideas we present here are equally applicable to other sensor modalities. We demonstrate our approach working on a road vehicle operating over a three month period at different times of day, in different weather and lighting conditions.

In this thesis, the synthesis of a grasping and manipulation controller of the Barrett hand, which is an archetypal example of a multi-fingered robotic hand, is investigated in some detail. This synthesis involves not only the dynamic modelling of the robotic hand but also the control of the joint and workspace dynamics as well as the interaction of the hand with object it is grasping and the environment it is operating in. Grasping and manipulation of an object by a robotic hand is always challenging due to the uncertainties, associated with non-linearities of the robot dynamics, unknown location and stiffness parameters of the objects which are not structured in any sense and unknown contact mechanics during the interaction of the hand’s fingers and the object. To address these challenges, the fundamental task is to establish the mathematical model of the robot hand, model the body dynamics of the object and establish the contact mechanics between the hand and the object. A Lagrangian based mathematical model of the Barrett hand is developed for controller implementation. A physical SimMechanics based model of the Barrett hand is also developed in MATLAB/Simulink environment. A computed torque controller and an adaptive sliding model controller are designed for the hand and their performance is assessed both in the joint space and in the workspace. Stability analysis of the controllers are carried out before developing the control laws. The higher order sliding model controllers are developed for the position control assuming that the uncertainties are in place. Also, this controllers enhance the performance by reducing chattering of the control torques applied to the robot hand. A contact model is developed for the Barrett hand as its fingers grasp the object in the operating environment. The contact forces during the simulation of the interaction of the fingers with the object were monitored, for objects with different stiffness values. Position and force based impedance controllers are developed to optimise the contact force. To deal with the unknown stiffness of the environment, adaptation is implemented by identifying the impedance. An evolutionary algorithm is also used to estimate the desired impedance parameters of the dynamics of the coupled robot and compliant object. A Newton-Euler based model is developed for the rigid object body. A grasp map and a hand Jacobian are defined for the Barrett hand grasping an object. A fixed contact model with friction is considered for the grasping and the manipulation control. The compliant dynamics of Barrett hand and object is developed and the control problem is defined in terms of the contact force. An adaptive control framework is developed and implemented for different grasps and manipulation trajectories of the Barrett hand. The adaptive controller is developed in two stages: first, the unknown robot and object dynamics are estimated and second, the contact force is computed from the estimated dynamics. The stability of the controllers is ensured by applying Lyapunov’s direct method.

This thesis presents trajectory-tracking control of flexible-link robots, overhead cranes, and multiple mobile robots in formation. Firstly, a new model-based trajectory control is proposed for the control of one-link flexible robots. The proposed control guarantees asymptotic stability with all internal signals bounded. Next, a distributed-parameter dynamic model, consisting of two ordinary differential equations and one partial differential equation, is derived using the extended Hamilton's principle for a two-link rigid/flexible robot. A collocated trajectory-tracking control scheme is designed based on the distributed-parameter dynamic model. With only two joint actuators, the proposed control guarantees stability throughout the entire trajectory control and asymptotic stability at desired goal positions. The proposed control is free from the so-called spillover instability Secondly, a sliding-mode anti-swing trajectory control is proposed for overhead cranes with high-speed load hoisting. In association with a new anti-swing motion-planning scheme, the proposed control realizes a typical anti-swing trajectory control in practice, allowing high-speed load-hoisting motion and sufficient damping of load swing. Lyapunov stability theorem is applied for the stability analysis of all of the above theoretical results, and the effectiveness is evaluated with control experiments Thirdly, a new formation control scheme is proposed for a group of mobile robots based on multi-objective potential forces. The angle of the potential force, with respect to the global coordinate system, is used to generate trajectories for the navigation of a group of nonholonomic mobile robots. A smooth and continuous control law is designed to reduce the global orientation error asymptotically to zero while maintaining proper formation for a target configuration. Finally, a trajectory-tracking control is designed for a group of nonholonomic mobile robots in a virtually structured formation. A real-time trajectory modification scheme is presented such that the center of the mobile robots tracks a desired trajectory. Lyapunov stability theorem is applied as the mathematical design tool. Moreover, an obstacle avoidance mechanism is designed based on a formation controller such that a mobile robot can escape from concave-shaped obstacles / acase@tulane.edu

Thesis (Ph.D.) -- University of Texas at Arlington, 2008.

<p> In this work, a framework for a system which will intelligently assign an edge detection filter to an image based on features taken from the image is introduced. The framework has four parts: the learning stage, image feature extraction, training filter creation, and filter selection training. Two prototypes systems of this framework are given. The learning stage for these systems is the Berkeley Segmentation Database coupled with the Baddelay Delta Metric. Feature extraction is performed using a GIST methodology which extracts color, intensity, and orientation information. The set of image features are used as the input to a single hidden layer feed forward neural network trained using back propagation. The system trains against a set of linear cellular automata filters which are determined to best solve the <i> edge image</i> according to the Baddelay Delta Metric. One system uses cellular automata augmented with a fuzzy rule. The systems are trained and tested against the images from the Berkeley Segmentation Database. The results from the testing indicate that systems built on this framework can perform better than standard methods of edge detection on average across many types of images.</p>

<p> In this thesis, the author presents a feasibility study of methods to modernize the port crane systems with the application of haptic (force) feedback assistive technology to assist the crane operator in the container handling process. The assistive technology provides motion guidance to the operator that could help increase the safety and productivity of the system. The technology of haptic feedback is successful in applications such as gaming, simulators etc., and has proven quite efficient in alerting the user or the operator. This study involves the implementation of haptic feedback as an assistive mechanism through a force-feedback joystick used by the operator to control the motion of a scaled port crane system. The haptic feedback system has been integrated to work with the visual feedback system as part of this study. The visual feedback system shares information needed to trigger the haptic (force) feedback display on the joystick. The force feedback displayed on the joystick has been modeled on Hooke's law of spring force. The force feedback and the visual feedback form a motion guidance system. The integrated system has been implemented and tested on a lab-scale testbed of a gantry crane. For experimental purposes, this concept has been tested on a PC-based Windows platform and also on a portable single board Linux-based computer, called the Beagleboard platform. The results from test runs on both the platforms (PC and Beagleboard using ARM processor) are reported in this study. 2</p>

<p>Time-lapse live cell imaging has been increasingly employed by biological and biomedical researchers to understand the underlying mechanisms in cell physiology and development by investigating behavior of cells. This trend has led to a huge amount of image data, the analysis of which becomes a bottleneck in related research. Consequently, how to efficiently analyze the data is emerging as one of the major challenges in the fields. </p><p> Computer vision analysis of non-fluorescent microscopy images, representatively phase-contrast microscopy images, promises to realize a long-term monitoring of live cell behavior with minimal perturbation and human intervention. To take a step forward to such a system, this thesis proposes computer vision algorithms that monitor cell growth, migration, and differentiation by detecting three cellular events&mdash;mitosis (cell division), apoptosis (programmed cell death), and differentiation&mdash;and tracking individual cells. Among the cellular events, to the best our knowledge, apoptosis and a certain type of differentiation, namely muscle myotubes, have never been detected without fluorescent labeling. We address these challenging problems by developing computer vision algorithms adopting phase contrast microscopy. We also significantly improve the accuracy of mitosis detection and cell tracking in phase contrast microscopy over previous methods, particularly under non-trivial conditions, such as high cell density or confluence. We demonstrate the usefulness of our methods in biological research by analyzing cell behavior in scratch wound healing assays. The automated system that we are pursuing would lead to a new paradigm of biological research by enabling quantitative and individualized assessment in behavior of a large population of intact cells. </p>

<p> The goal of this project was to design, build, program, upgrade, and remotely control an existing industrial-grade robotic manipulator on a budget that was equal to a graduate student's salary.</p><p> Inverse kinematic equations were developed to model the CRS arm using Denavit-Hartenberg convention. After converting these equations into a system of Python code, the equations were verified via graphical simulation.</p><p> The CRS arm's obsolete motor driver box was completely removed and replaced with cutting-edge, low-cost microcontrollers. The wiring, mechanics, and controls for the five degrees of freedom (DOF) were reverse engineered based on very limited datasheets, including the arm's motors, brakes, encoder pin-outs, gripper, and fixed global reference frame movement.</p><p> The microcontrollers were then programmed in C/C++ to allow a user to control every electrical and mechanical aspect of the arm. The kinematic equations were implemented on a Python server, which commands how each joint in the robotic arm must move to reach a desired point in space. The user interface was developed in conjunction with Evan Boldt to allow for remote control of the robotic arm and monitoring through webcams, which includes twisting and tilting a tablet.</p>