Joel W. Burdick - US grants

The research comprises an analytical and experimental study of hyper-redundant mechanisms. Hyper-redundant mechanisms are defined as highly articulated devices with a large or infinite number of degrees of freedom, and are analogous to snakes, elephant trunks, tentacles, etc. These devices have potential applications in robotic manipulation and locomotion, as well as many other non robotics applications such as articulated space structures, medical endoscopes, and flagellum like fluid propulsion. This investigation focuses on basic geometric, kinematic, dynamic and control issues related to hyper-redundant mechanisms. The analytical phase focuses on a differential geometry based approach to the kinematic analysis and geometric control of this class of mechanisms. In the experimental phase, a planar hyper-redundant device will be implemented in hardware and algorithms developed in the theoretical study will be tested and verified.

This project addresses experimental validation of current analytical work in robotics as well as the initiation of new work in robotics. The principal Investigator has a strong background in the mechanical aspects of robotics, including manipulator force control, redundant manipulator kinematics, the analysis of dynamically stable legged locomotion, multi fingered dexterous hand manipulation, and modular self organizing robotic systems. All of these topics are aimed at developing sophisticated multi degree of freedom machines to enhance automation. The applications are quite broad, and encompass space exploration, medicine, and hazardous waste removal. This project will implement both a prototype hyper redundant manipulation and locomotion system as well as a dynamically stable legged robotic system. They will be used to demonstrate and validate previously developed analytical models and algorithms, and to identify future research needs in manipulator kinematics and dynamics.

This research project will develop a quasistatic motion planning paradigm that is a significant extension of classical robotic motion planning paradigms. The quasistatic problem arises when significant reaction forces between the robot mechanism and its environment are required to generate and/or stabilize the robot's motion. The planning paradigm is based on the PI's recently developed configuration space mobility theory. To demonstrate the practical utility of this work, algorithms will be developed and implemented for two important quasi-static application examples: a novel workpiece fixturing algorithms that incorporates the curvature of the fixtures into the planning phase; and a proof-of-concept 2-dimensional quasistatic planning system for a highly articulated robot. This planner will be tested on an actual robot prototype in the laboratory.

97-04702 Burdick, Joel W. California Institute of Technology Postdoc: Planning for Hybrid Systems (This award supports CES associate Milos Zefran) Systems that are governed by discrete and continuous processes are known as "hybrid" systems. Such systems are described by continuous dynamical equations whose structure and/or inputs change in accordance with the state transitions of a finite automaton. Due to the widespread use of computers to control physical devices, hybrid systems abound. Successful hybrid systems must robustly combine high-level planning (planning of the finite automaton transitions) with low-level control (evolution of the dynamic equations). While the planning and control problems have been addressed separately by computer scientists and control theorists, the interaction of the discrete and continuous worlds is largely left to the ingenuity of engineers. As the complexity of the computer controlled physical systems grows, such methods break down and there is a need for rigorous theory to design, build, and evaluate hybrid systems. The proposed research will investigate how to unify discrete and continuous models to help build a foundation for hybrid system engineering. Prior work on planning for a subclass of hybrid systems will be extended in two directions. First, an optimal control framework will be extended to a game-theoretic framework so as to take uncertainty into account. Second, the hybrid control problem will be explored for systems whose dynamics exhibit some simplifying properties, such as Lie group symmetries.

This award provides funds to purchase equipment and instrumentation that will form the basis for a new experimental facility for investigating the science of neural prosthetic arms at the Engineering Research Center for Neuromorphic Engineering at the California Institute of Technology. The equipment is being used in research to investigate how the cerebral cortex plans reaching-arms movements. The goal is to understand the underlying neurobiology of the sensory-motor interface where plans first form. This understanding is the basis for cortically controlled prosthetic devices. The equipment will enable exploration of how parietal reach region neurons in the posterior parietal cortex plan real and virtual arm movements by recording electrical activity from many neurons simultaneously (using an implanted multi-electrode array) while monkeys reach to remembered visual targets. The electrical signals will then be used to drive a virtual arm in a computer model/simulation. The understanding of how the neurons work to move the arm will be the basis for control of a real prosthetic arm.

Abstract

IIS-9901056
Burdick, Joel
California Institute of Technology
$92,265 - 12 mos.

Experiments, Modeling, and Automatic Planning for Fixturing and Gripping

This is the first year funding of a three-year continuing award. The research is to investigate analytically and experimentally into problems of workpiece fixturing and related problems of object gripping. Prior efforts in automated fixture planning have largely been based on rigid-body models or ad-hoc models of compliant contact. These models are too inaccurate for manufacturing operations that involve large forces and precision deflection tolerances. The first goal of this research is to extend work on compliant contact modelling to include other important contact phenomena, such as the effect of friction in nonlinearly compliant contacts, and global deformations in workpieces with slender internal structures. Further, the PI plans to integrate the
relevant contact phenomena into a single effectively computable model. A second major goal is to construct a comprehensive experimental system that will be the first to measure the complete range of physical phenomena underlying the fixturing problem. These experiments should not only validate and refine the models, but also help convince the community of fixturing researchers and practitioners that nonlinear compliance models are to be preferred over currently used ad-hoc models. A third goal is to develop automated fixturing and gripping synthesis procedures that are based on the extended models. These algorithms will be able to automatically generate fixture arrangements that guarantee meaningful performance criteria, such as deflection tolerance, in the presence of externally applied loads. The PI also wishes to include design parameters, such as surface curvature and the minimal number of fixtures needed to guarantee task
performance criteria, that have not previously been addressed in the fixture synthesis literature. In the long term, the PI's team plan to use these results to study active control of the fixture loading process so as to inherently ensure accurate part localization. Since fixturing is closely related to gripping, these results should have a longer-term impact on the deployment of more sophisticated gripping devices.

This project, a collaborative with 03-24864 Papanikolopoulos at University of Minnesota and 03-24977 Kostas Daniilidis at University of Pennsylvania, and 03-25017, addresses research issues key to an important application of robot teams and information technology (emergency response in hazardous environments for various tasks). The research sets 6 goals:
Development of new algorithms that enable collaborative sensing.
Development of distributed localization/mapping methods that leverage capabilities of the heterogeneous robots.
In-depth study of communication issues with emphasis on transparent integration of ultra wideband communication methodologies.
Development of methods for team coordination and dynamic distribution of tasks to robots.
Creation of algorithms for the presentation of sensory information to users.
Experimental validation of the scalability of the aforementioned algorithms and techniques.
The PIs use the Scout and MegaScout robotic platforms designed at the University of Minnesota along with other testbeds at CalTech and U Penn to conduct the research. The project integrates the algorithms with first responder teams, emphasizing realistic scenarios; mentors students from underrepresented groups in order to retain them in CS/EE programs; conducts outreach activities through demonstrations at local schools and youth groups; conducts workshops that emphasize cross-disciplinary interaction; creates web resources; innovates classroom uses of multi-robot teams; and includes parts of the research in design projects for seniors. The project also includes international collaboration with groups at NTUA (Greece) and the University Louis Pasteur-Strasbourg I (France).

Abstract
0428075
Joel Burdick
Caltech

This project will develop basic theory, algorithms, and experimental demonstrations of networks of autonomous mobile sensory platforms. Mobile sensor units enable a sensor network to transiently focus on events or locations that might be interesting or important while at the same time maintaining awareness of the overall environment. Particular focus on tasks where mobile sensory-motor platforms operate in human environments and interact with human operators. The consideration for theory and algorithms include: (1) use mobility to improve network performance; (2) facilitate the interaction between sensor networks and humans; and (3) improve the ability of networks to detect learned categories or events. New unsupervised learning methods will form the basis for object and event detection. We will demonstrate our methods on the Caltech Multi-Vehicle Wireless Testbed (MVWT), an experimental testbed at Caltech consisting of fan-driven and wheeled vehicles operating in a common environment endowed with a variety of sensors and communication schemes.

This is a project supported under the Sensors Initiative NSF 04-522.

A novel pulsatile jet propulsion scheme for low speed maneuvering of small underwater robots is developed, demonstrated, and characterized. This propulsion scheme is loosely analogous to that used by squid and jellyfish. The potential for pulsatile jet propulsion is explored by first optimizing the design of a pulsatile jet actuator and associated actuation concepts. Next, a vehicle-level fluid dynamical model is developed in order to capture the interaction of the pulsatile jet flows with the primary flow past the vehicle. Prior development in nonlinear averaging-based vehicle feedback control schemes is adapted to this technology using such models. The pulsatile jet prototypes and control scheme is integrated into a prototype underwater vehicle, whose performance is characterized. The suggested propulsion scheme has very few moving parts, has no protruding components that increase drag, and takes up relatively little volume.

Undergraduate engineering students are heavily involved in the design, fabrication, and testing of the pulsatile jet actuators and underwater vehicle prototypes proposed in this project. These activities provide excellent hands-on engineering experiences for the participating students.

The objective of this analytical and experimental research project is to develop new nonlinear modeling methods for the physics of contact in multi-body grasps and fixtures, and to verify these models on a novel experimental fixturing system. The analytical portion of the project will formulate lumped parameter nonlinear compliance laws that more accurately describe the physics of multi-body contact involving both friction and compliance. The experimental component of this project will create a novel experimental system that can accurately measure the forces involved in compliant and frictional fixturing arrangements. The analytical models will be verified and refined on this experimental fixturing system. The resulting nonlinear models will be used to develop novel fixturing and grasping algorithms that accept task specifications such as workload level and immobilization tolerance, and the adaptively select the simplest possible fixture arrangement that satisfies both the task constraints and the governing physics of the contacts.

Physical arrangements that involve contact between multiple bodies are common in industrial fixturing and robotic grasping applications. The success of fixture, grasp, and manipulation planning algorithms may crucially depend on the accuracy of the underlying contact models upon which the plans are based. If successful, the nonlinear contact models that will be developed in this research project will allow entirely new fixture planning paradigms. Current fixturing paradigms rely on conservative rigid body models to ensure proper workpiece immobilization. But many light-to-moderate duty applications need not be gripped by as many fixels as are required by these classical paradigms in order to achieve task specifications. The improved friction-compliance laws that will be created in this project will allow more intelligent and adaptive fixturing strategies that should yield simple yet robust fixturing and gripping systems. This research will also lead to new robotic manipulation algorithms that avoid catastrophic phenomena such as wedging and jamming.

DESCRIPTION (provided by applicant): Of the approximately 10 million people in the US living with paralysis, 15,000 are the result of spinal cord injury each year. The first year of car can range from $322,000-$986,000, with lifetime costs of $1.4-4M for someone injured at 25 years of age. In addition to potentially devastating sensorimotor disturbances, there is a huge financial cost, estimated to be $13.55B in medical care, therapy, and lost productivity nationwide. Until very recently, the recovery from spinal cord injury (SCI) was bleak, with little hope of restoring motor function. To address this we have demonstrated that the physiological state of the spinal circuitry of rats and cats can be modulated with epidural stimulation to generate voluntary limb motor function over a range of speeds, loads, and directions, a finding we have extended to humans. Three years post-injury, a motor complete spinal cord injured human subject was implanted with an epidural electrode array over the lumbosacral spinal cord. In less than one month after implantation, the subject could stand independently, and after 7 months of daily epidural stimulation and motor training, voluntary control of both legs was evident in the presence of epidural stimulation, whereas complete paralysis remained in absence of epidural stimulation. We will advance these discoveries with the use of non-invasive stimulation of the cervical cord to improve arm and hand function following SCI. Central to this proposal is our discovery of a painless electrical single-channel (stimulation of one part of the spinal cord) and dual-channel (stimulation of two different parts of the cord) paradigm that can be applied to the surface of the skin, termed transcutaneous electrical stimulation (TES), bypassing the need for a surgically-implanted electrode array. In the first phase of this proposal we will demonstrate proof-of-principle that stimulation of the cervical spinal cord can improve motor function by: 1) Testing responses to transcutaneous electrical stimulation in subjects with spinal cord injury; and 2) defining the operational parameters of electrical stimulation that that are most effective using a machine-learning protocol, and 3) produce a dual-channel commercial prototype. This commercial product will undergo testing similar to the proof- of-principle device. This device will then be tested in subjects with cervical spinal cord injury and evaluated with a machine-learning protocol. This Phase I proposal will deliver a device that can painlessly and non-invasively aid in the recovery of SCI by delivering a specific electrical stimulation paradigm to the cervical cord that improves use of the arms and hands.