
Advanced Sensing and Control Techniques to Facilitate Semi-Autonomous Decommissioning of Hazardous Sites - Final Report
Author(s) -
Robert J. Schalkoff
Publication year - 2000
Language(s) - English
Resource type - Reports
DOI - 10.2172/775369
Subject(s) - computer science , rendering (computer graphics) , artificial intelligence , computer vision , robot , control point , computer graphics (images) , robustness (evolution) , biochemistry , chemistry , gene
This report summarizes work after 4 years of a 3-year project (no-cost extension of the above-referenced project for a period of 12 months granted). The fourth generation of a vision sensing head for geometric and photometric scene sensing has been built and tested. Estimation algorithms for automatic sensor calibration updating under robot motion have been developed and tested. We have modified the geometry extraction component of the rendering pipeline. Laser scanning now produces highly accurate points on segmented curves. These point-curves are input to a NURBS (non-uniform rational B-spline) skinning procedure to produce interpolating surface segments. The NURBS formulation includes quadrics as a sub-class, thus this formulation allows much greater flexibility without the attendant instability of generating an entire quadric surface. We have also implemented correction for diffuse lighting and specular effects. The QRobot joint level control was extended to a complete semi-autonomous robot control system for D and D operations. The imaging and VR subsystems have been integrated and tested