The Optimal Selection of Robot Modules for Space Manipulators

by Robert Ambrose, MITRE Corp, Houston, United States,
Delbert Tesar, MITRE Corp, Houston, United States,



Document Type: Proceeding Paper

Part of: Robotics for Challenging Environments

Abstract:

Monolithic robot systems are poorly suited to the broad requirements and uncertainties of space automation applications. Designers have considered building many robots, each of which can do one thing well, or one robot which can do many things poorly. These solutions are unattractive for operations in low Earth orbit, and might prove infeasible for tasks on the Lunar or Martian surfaces. Instead of building specific robots, it is proposed to design a set of robot modules (joints, links, controllers, and sensors) that will have a wider task spectrum, at a lower overall weight than an equivalent set of monolithic machines. Three space station tasks were selected from a recent NASA automation study and used as the basis for this design problem. These tasks included ORU handling, external inspection and robot repair. Three sets of arm performance goals were then identified in terms of a common set of criteria, with each goal's importance weighted for that task. A set of 4 joint modules and 9 link modules have been built for the UT Modular Robotics Testbed. These modules were designed for experimentation and not commercialization, and therefore are more versatile than optimal. Experiments with these joints and links, isolated on the test stand as described in, have measured their stiffness, backlash and friction, which will be combined with inertia, payload, speed, range of motion and weight specifications to provide a module data base. A design method is proposed to search this data base for the best combination of modules for a specified task. While the available joints and links are admittedly limited (in number and performance) the incorporation of experimentally identified module data has been shown to allow for a confident configuration (arm) performance prediction. Furthermore, the design method is sufficiently general so as to allow the addition of new modules to the existing set, and thus broadening the module set's application spectrum.



Subject Headings: Building design | Joints | Databases | Robotics | System analysis | Space stations | Uncertainty principles | Orbits | Utah | United States

Services: Buy this book/Buy this article

 

Return to search