ParaDex: A Novel 6-DOF Parallel Manipulator

ParaDex (PARAllel DEXterous) is a novel robotic architecture that offers unprecedented dexterity in interacting with its environment. Originally developed for surgical applications under a Cooperative Research and Development Agreement (CRADA) between MicroDexterity Systems (MDS) and Sandia National Laboratories' Intelligent Systems and Robotics Center (ISRC), the ParaDex mechanical architecture was designed to minimize all forms of friction, inertia, and backlash. Coupling this architecture with state-of-the-art control hardware and advanced impedance-based servo algorithms allows for quick and precise control of both movements and interactive forces. While these capabilities are essential to enhance surgeons' dexterity in the operating room, they also create new opportunities for industrial robotics.

Today's robots can move large payloads very quickly from one place to another or move along a programmed path (for example, in welding applications) very adeptly, but are quite poor at controlling interaction forces within their workspace. Attempts to adapt these types of robots to control interaction forces often result in slow and sluggish performance (primarily due to the inherent friction, stiction, and moving mass), which disqualifies adaptation of these devices to the factory floor. ParaDex will enable delicate assembly operations such as gear meshing and bearing insertion to be automated, freeing factory workers from many repetitive and injury-laden operations, thereby reducing costs and improving the quality of assembled products.

The ParaDex mechanical architecture is completely scalable, allowing for large payloads while still maintaining the design parameters for optimal control of interaction forces. Prototypes built to date operate within workspaces ranging from roughly the size of a tennis ball to that of a beach ball, although larger workspaces can be accommodated by attaching ParaDex to a standard industrial manipulator for gross positioning. Payload capabilities of the current prototypes range from 5 to 75 pounds. The first prototype unit (with a softball sized workspace) has a positional repeatability of better than 5 microns. In addition, custom algorithms control interaction forces during assembly, either under autonomous control or in a bilateral, force-reflecting telemanipulation mode. This latter mode allows a human operator to perform a task remotely and provides an intuitive method for a factory technician to "teach" the robot how to perform an assembly task that does not require programming.

ParaDex is the cornerstone technology of the Flexible Robotic Assembly for Powertrain Applications (FRAPA), a joint venture between MDS, Ford Motor Company, Progressive Tool & Industries, Perceptron, and the National Center for Manufacturing Sciences. The project is funded both by private funds of the joint venture partners and matching funds from the National Institute of Standards and Technology under an Advanced Technology Program Cooperative Agreement. As a CRADA partner with MDS, Sandia serves as a subcontractor to the FRAPA joint venture. Under this program, Sandia is working with industry and other technology providers (including Case Western Reserve University) to develop ParaDex as the primary enabling technology to solve complex force-controlled assembly tasks.

Automotive powertrain assembly is an ideal candidate for automated assembly demonstration. The application is widespread within industry, current manual assembly results in high rates of repetitive stress injuries due to the relatively heavy components involved, very delicate force control is required to avoid damaging components, and the actual required workspace is relatively small (about 30 degrees in orientation, about 1 foot in x and y, and about 3 feet in z). Insertion of torque converters, clutch packs, and gear and spline meshing will be demonstrated in this project.

The first generation ParaDex robot built specifically for powertrain assembly has been shipped to Case Western Reserve University, where further software development is under way. Sandia, MDS, Ford, and Case Western will collaborate on the final force-controlled assembly algorithms necessary for the robot to perform automatic transmission assembly tasks.