Accuracy Analysis of a Robotic Radionuclide Inspection and Mapping System for Surface Contamination
- Department of Mechanical Engineering, University of Nevada, Las Vegas, NV 89154-4027 (United States)
The mapping of localized regions of radionuclide contamination in a building can be a time consuming and costly task. Humans moving hand-held radiation detectors over the target areas are subject to fatigue. A contamination map based on manual surveys can contain significant operator-induced inaccuracies. A Fanuc M16i light industrial robot has been configured for installation on a mobile aerial work platform, such as a tall forklift. When positioned in front of a wall or floor surface, the robot can map the radiation levels over a surface area of up to 3 m by 3 m. The robot's end effector is a commercial alpha-beta radiation sensor, augmented with range and collision avoidance sensors to ensure operational safety as well as to maintain a constant gap between surface and radiation sensors. The accuracy and repeatability of the robotically conducted contamination surveys is directly influenced by the sensors and other hardware employed. This paper presents an in-depth analysis of various non-contact sensors for gap measurement, and the means to compensate for predicted systematic errors that arise during the area survey scans. The range sensor should maintain a constant gap between the radiation counter and the surface being inspected. The inspection robot scans the wall surface horizontally, moving down at predefined vertical intervals after each scan in a meandering pattern. A number of non-contact range sensors can be employed for the measurement of the gap between the robot end effector and the wall. The nominal gap width was specified as 10 mm, with variations during a single scan not to exceed {+-} 2 mm. Unfinished masonry or concrete walls typically exhibit irregularities, such as holes, gaps, or indentations in mortar joints. These irregularities can be sufficiently large to indicate a change of the wall contour. The responses of different sensor types to the wall irregularities vary, depending on their underlying principles of operation. We explored capacitive, ultrasound, and optical Laser range sensors. The unshielded capacitive range sensors were found to be influenced by objects in their vicinity, and to have comparatively low sensitivity. Therefore they were not used for gap measurement. We did, however, use them successfully to detect obstacles in the field of motion of the sensor head. The four sensors pointing diagonally from four corners of the sensor head are capacitive range sensors, which stop the robot motion any time a pre-set threshold signal level is exceeded. Ultrasound range sensors were found to have good resolution. The ultrasound beam was frequently deflected sideways by the wall's roughness, resulting in a spurious signal peaks. On the other hand, ultrasound signals are inherently insensitive to variations of the optical surface properties. Laser range sensing proved to be generally less noisy than ultrasound measurements. The aforementioned sensitivity to the target surface's optical properties was not evident on grey, non-reflective surfaces. On such surfaces, laser range sensing was found to be superior to ultrasound measurements. Occasionally, however, surface reflectivity and specularity caused erroneous measurements. The computational burden of measuring and filtering the sensor data in real time made it impractical to control the robot directly based on sensor feedback. The robot performs a first horizontal surface range scan merely to gage the surface profile. During the subsequent passes, the robot records the radiation measurements, while recording the range data for the next pass, with the range sensor positioned below the radiation sensor. The decoupling of measurements from robot control prevents the robot controller from responding to spurious sensor signals, ensuring steadier and safer radiation surveying. The controller is configured to accept either optical or ultrasound range signals, so as to respond better to user-specific requirements. A detailed analysis of non-contact range sensors and control for the automated mapping of surface contamination has been presented. The Contamination Survey Machine was tested and is operational.
- Research Organization:
- American Nuclear Society, 555 North Kensington Avenue, La Grange Park, Illinois 60526 (United States)
- OSTI ID:
- 21144238
- Resource Relation:
- Conference: DD and R 2007: ANS Topical Meeting on Decommissioning, Decontamination, and Reutilization 2007, Chattanooga, TN (United States), 16-19 Sep 2007; Other Information: Country of input: France; Related Information: In: Proceedings of the 2007 ANS Topical Meeting on Decommissioning, Decontamination, and Reutilization - DD and R 2007, 336 pages.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Sensing and Sampling of Trace Contaminations by a Dexterous Hexrotor UAV at Nuclear Facilities - 18600
Purpose-Built UAVs for Physical Sampling of Trace Contamination at the Portsmouth Gaseous Diffusion Plant - 17331