on the
Robotics Technology Development ProgramAbout the focus areas The U.S. Department of Energy's Office of Environmental Management has established an integrated approach for addressing waste issues based on problem, or focus areas. The focus areas are: subsurface contaminants (combining the plumes and landfill stabilization focus areas); mixed waste characterization, treatment, and disposal; high-level waste tank remediation; decontamination and decommissioning; and plutonium stabilization. Three crosscutting technology areas support the focus areas: characterization, monitoring, and sensor technology; efficient separations and processing; and robotics.
During the Seventh Annual U.S. Department of Energy Forum on Robotics for Environmental Remediation, the Robotics Technology Development Program touted its accomplishments and presented new research directions. At the forum, DOE Under Secretary Thomas Grumbly announced a $66 million Cooperative Research and Development Agreement between DOE (contributing $18 million) and SciBus Analytical, Inc. (contributing $48 million) for the continued development and commercialization of Contaminant Analysis Automation. CAA automates high-volume, repetitive laboratory analysis by bringing together nine plug-and-play Standard Laboratory Modules. CAA cuts by one-third the costs of preparing and managing samples for chemical analysis. Automation improves accuracy and reduces turnaround between sample submission and results reporting from 60 days to five days.
Integral to the CRADA is a technology transfer plan for the commercialization of CAA. As the exclusive systems integrator, SciBus Analytical is coordinating CAA project development and is developing commercial products for the analytical chemistry marketplace, based on research of CAA team members from five national laboratories and three universities: Los Alamos National Laboratory, Pacific Northwest National Laboratory, Idaho National Engineering Laboratory, Sandia National Laboratories, Oak Ridge National Laboratory, and the Universities of Florida, Tennessee, and Texas. Linton Yarbrough, program manager for RTDP, emphasized the CRADA's function in commercializing CAA. "CAA must be commercialized; otherwise it's just a nifty little toy rotting away in a DOE lab."
Yarbrough also invited private industry to join the CAA team. "If you have a system that needs to be substantially developed before it becomes a Standard Laboratory Module, contact one of the five laboratories and inquire about a Program Research and Development Announcement contract. We want as many industrial partners as we can get. Buttonhole the lab you're interested in working with."
At the forum, speakers presented information on robots developed through RTDP that are ready or will soon be ready for cleanup work at DOE sites.
Houdini –
Developed by RedZone Robotics, Inc. through funding provided by
Morgantown Energy Technology Center's Industry Program, Houdini
resembles a small bulldozer. It is a tethered, hydraulically
powered, foldable vehicle for remote work in tanks. Controlled by
an operator at a console, Houdini can be used for retrieval,
washdown and mixing, size reduction of waste, and object
manipulation. It can be outfitted with a plow and gripper used in
conjunction with a scoop, a shear, or a vacuum and a sump pump.
When deployed with a Schilling Titian manipulator, Houdini can
deliver a 240-pound payload and has a six-foot reach. When
folded, Houdini will fit through a 12-inch riser opening for
entry into a tank, where it will unfold and move either on or
under waste surfaces while tolerating hostile radiological,
chemical, and physical conditions.
As part of Oak Ridge National Laboratory's Gunite Tanks Treatability Study, Houdini was outfitted with confined sluicers as end effectors and cold tested last fall at the Robotics Technology Assessment Facility at Oak Ridge. Houdini is scheduled for hot testing at an Oak Ridge gunite tank in the October to December time frame and will also be used to remove material from a K-65 silo at Fernald. Plans for continued development of Houdini include integrating the vehicle with the Position and Orientation Tracking System to remotely orient the operator to the tank environment.
Rosie II – Through a Program Research and Development Announcement contract, Carnegie Mellon University and RedZone Robotics, Inc. have developed Rosie, a heavy-duty, mobile work platform for nuclear facility decontamination and dismantlement. Rosie can deploy a full range of tools and sensors for D&D. Its heavy manipulator has a 30-foot reach. Finer operations are handled by a dexterous manipulator on the end of the heavy manipulator. Rosie's telerobotic operations are made more efficient by Artisan, a graphics control package developed by RTDP. An operator remotely controls the boom, locomoter, system power, tether, and cameras through a console equipped with joysticks and levers. Rosie will undergo testing at Oak Ridge's Robotics Technology Assessment Facility in preparation for its planned use in the dismantlement of the Chicago Pile No. 5 reactor at Argonne National Laboratory. During the third day of the forum when participants toured Sandia's robotics labs, they got a chance to be teleoperators of a jackhammer-equipped Rosie.
Dual Arm Work Platform – The Dual Arm Work Platform will be deployed by a polar crane over Argonne National Laboratory's CP-5 reactor and rigged with appropriate tools for dismantlement of the reactor. The project is one of the three large-scale demonstrations being undertaken by the D&D focus area.
Modified Light Duty Utility Arm – The Oak Ridge National Laboratory will use a Modified Light Duty Utility Arm to retrieve waste from the gunite tanks using a waste dislodging and conveyance end effector. Plans call for Spar Aerospace Limited to deliver the MLDUA to ORNL during the final quarter of fiscal year 1997. Like the LDUA, which is scheduled for characterization work at DOE's Hanford Site, the MLDUA is a dual-section, telescoping vertical mast, a five-link arm, and a deployment and containment system that has remote control capabilities. However, the MLDUA will be capable of carrying heavier payloads and has a longer reach.
RTDP is funding several enhancements to the control system, including a flexible structure damping algorithm currently under development at Pacific Northwest National Laboratory and ORNL, which will improve position control by compensating for structural deflections of the manipulator system. Washington University is developing a function-based sharing control scheme that will allow an operator to modify a robot's trajectory in real-time based on hand controller input and then return to automatic mining. A new sensor-based control method for controlling impact and regulating force will be additional safety features on the MLDUA.
During a presentation and tour of robotic activities at Sandia National Laboratories, forum participants got a chance to hear and see how developing technologies will be integrated to more efficiently and safely control environmental restoration operations. The major thrust of robotic research activities is the continued development and evaluation of operator interface and controls capabilities.
Virtual Collaborative Control – Network-based intelligent systems will allow teams of operators and observers to discuss and plan strategies for safe robotic operations in a virtual collaborative control environment. Virtual tools developed at Penn State University, with support from Sandia National Laboratories through their university programs, will facilitate virtual collaborative control. Using a mouse, the operator selects a tool from the tool box panel. With a cyberglove, the operator orients and places the tool into the three-dimensional scene and communicates to the robot how the tool is to be used. For routine operations, a single operator at the site can use a few basic virtual tools to specify tasks. Virtual tools will also allow multiple experts connected by a network to collaborate in decisions about how to approach and accomplish challenging tasks, such as extracting pump housings buried within a storage tank. Once satisfied with the effects of using the virtual tools, the operator can then use the same series of motions to drive the actual robot to complete the task.
Collision Avoidance – Through a CRADA with Merritt Systems Incorporated, Sandia National Laboratories is working on ultrasonic and capacitive sensor systems for collision avoidance. Sensors within the arm of a robot detect obstacles and a collision avoidance algorithm directs the arm away from obstacles while moving toward or maintaining the position of the end effector.
Real World Modeling – To safely operate remote robots, operators depend on true graphical representations of the robot's world that allow operators to test and practice a task before actually performing it in real-time. Sandia and Los Alamos National Laboratories are working on systems that will rapidly convert sensor data into three-dimensional computer graphics. Sandia's structured light system uses lasers to map the surface topography of a robot's environment, providing x, y, and z coordinates. Current methods of constructing world models from sensor data present the following problems.
The researchers at Sandia and Los Alamos are using an approach known as model-based segmentation to overcome these problems. The graphical program recognizes and treats as independent objects data points which lie within defined configurations, such as planes and cylinders. Rapid world modelling will allow the model to be quickly updated to reflect the changing environment, while the robot continues its work.
For an agenda, list of attendees, and abstracts of papers presented at the Robotics Forum, visit Sandia's Web site at http://www.sandia.gov/ cc_at/Forum96/1Forum96.html.
