Measurement Science for Intelligent Manufacturing Robotics and Automation Program

Seamless cooperation between humans and robots in manufacturing systems can combine tireless toil with the quick incorporation of new ideas—providing a revolution in creativity and speed on the factory floor. But creating such systems will require huge innovations for automated equipment, such as robots and machine tools, to intelligently carry out production plans and easily adapt them in response to changes in material, process, equipment, or surroundings. This project seeks new measurement science on all scales, from the macro to the nano, to explore how intelligent systems might do anything from perceiving movement in their vicinity to meeting production goals to protecting nearby humans. Such flexible systems will improve safety and make it easier and quicker to build new products.

With foreign countries able to command less expensive labor for mass production lines, U.S. manufacturers are looking for novel ways of bringing innovative, multi-featured products to the consumer quickly and cheaply.  This requires more flexible and more efficient production methods than are currently available.  Improved machines would be able to adapt more quickly, allowing for quick implementation of complex products.  Such improvement would also help robots be safe enough to work next to people -- combining dogged automated work with human adaptability.

The concept is simple, the realization less so.  Most current factory machinery simply repeats a single task over and over until told to stop, so that new production lines usually require new machines. In addition, such machinery is walled off from human workers because it is deemed unsafe to function in their presence.  Individual companies have few immediate financial incentives to improve the process, and without accurate measurement standards the building of "intelligent" robots would be very risky for any company that tried it.  To promote progress, NIST is working with industry, including airplane and car manufacturers, to provide the measurements, standards, and performance evaluation methods necessary to make robots more adaptable.  Most of the projects will take advantage of a test bed currently being constructed in the NIST shops.

One of the most crucial necessities is providing robots with the ability to accurately perceive their surroundings.  Only 10% of current manufacturing robots have any sensors at all, but a machine that can perceive complex activity going on around it could make adjustments on the fly for changes in materials, equipment, or surroundings.  This project will develop robust, quantitative tests to ensure that advanced perception systems—from the physical sensors themselves to the software programs that run them—accurately sense the three-dimensional world of moving parts in their vicinity, allowing them to alter plans when they detect an anomaly in production.

As well as providing flexibility on the shop floor, such sensors could be programmed to detect and avert situations in which humans could be harmed.  NIST will work to define methods and evaluations tests by which such an intelligent automaton can be "certified safe," ushering in an era of human/robotic cooperation that can provide U.S. industry with the high-quality, high-precision manufacturing methods it needs to maintain global competitiveness.

NIST will also tackle standards for the newest branch of automation: nanorobotics.  Nanorobotics encompasses both minuscule machines and large machines that fabricate minuscule objects.  Standards are needed to determine the accuracy of such machines—did it really insert a new part exactly 10 nanometers to the left as expected?—as well as to evaluate the precision with which tiny devices are built.  

Click here for more information

• Worked with the vendor to build a high precision system that can measure the position and orientation of a moving part at a high rate (up to 100Hz) and high precision (less than a millimeter) in a large space (80 m2), while precisely logging synchronized data from another sensor under test.
• Using the high precision system, established an initial dynamic metrology evaluation protocol for sensors that measure six degree-of-freedom position and orientation. Tested the protocol on an advanced visual servoing implementation at the Robot Vision Lab at Purdue University. This work has led to a collaboration between Purdue University, General Motors, Ford and NIST.
• The Unified System for Automation and Robot Simulation (USARSim) was significantly enhanced to support manufacturing simulation. Automated guided vehicles such as autonomous forklifts and unit loaders were modeled, as well as factory conveyor systems. USARSim enjoys wide use in the research community with over 35,000 downloads of its components from its open source website (www.sourceforge.net/projects/usarsim).
• In cooperation with the IEEE Robotics and Automation Society (RAS), organized a regional virtual manufacturing automation competition (VMAC), aimed at advancing robotic research and education through an open source high-fidelity simulation framework and competition. An international follow-on effort has been approved and will take place at the International Conference on Robotics and Automation in Kobe, Japan  in May, 2009. For more details, see IEEE Scanner, Vol. 23 No. 4 July-August 2008 edition at http://ewh.ieee.org/r2/capitalarea/eSCANNER/print.html.
  
• Drafted new language for the ITSDF B56.5 Standard that will now specify standard test pieces to be detected when using non-contact sensors on industrial automated and self guided vehicles (AGVs).
• Implemented consortium agreement with three AGV manufacturers covering research into advanced sensors for safety and obstacle detection for AGVs.
• Developed a method to embed optical fibers into the ISD/MEL MEMS nanopositioners and to generate interferometry signals to enable measurement of the motion of the nanopositioners.  Etched V-shape grooves in the supporting frame of a couple of our MEMS nanopositioners and used them to mount commercially available optical fibers.  The optical fibers are aimed at either the side surface of the moving stage of the nanopositioners or a specially constructed reflecting surface.  Infrared laser light sent through the fibers is reflected on the stage surface back into the fibers.
• Demonstrated the operation of a Frequency Modulated Optical Fiber Tip Interferometer (FM-OFTI) on a moving mirror target.  A prototype controller running on a PC, which controls the FM-OFTI system using the measurements from the interferometer was completed and tested recently.