Development and Applications of a Robot Tracking System for NIST Test Methods

Open Access
Author:
Pangborn, Herschel Cyrus
Area of Honors:
Mechanical Engineering
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
  • Dr. Sean N Brennan, Thesis Supervisor
  • Karl Martin Reichard, Thesis Supervisor
  • Henry Joseph Sommer III, Honors Advisor
Keywords:
  • robot
  • mechatronics
  • image processing
  • fiducial identification
  • MATLAB
  • Python
  • NIST
  • object identification
  • emergency response
Abstract:
To investigate the standardization of ground robot endurance as defined by the National Institute of Standard and Technology (NIST), this thesis presents both the fabrication of a NIST testing arena and the development of a novel robot tracking system that uses several overhead cameras to record the number of laps and the distance traveled by a robot over the duration of a test. The computer algorithms employed perform four primary functions: 1) image acquisition and correction for camera barrel distortion, 2) localization of the robot through fiducial identification, 3) lap counting between user-defined “end-zones,” and 4) conversion of the path traversed from pixels to real-world distances via user-conducted calibrations. Analyses of the precision and accuracy of this system, as well as expected sources of error, are provided. Two applications relevant to robot endurance are discussed using data from three separate testing events. The first evaluation is the consistency of laps completed (the current NIST method of estimating the distance traversed), comparing distance and time across different robots and operators. The second evaluation considers trends in operator performance over time for the duration of a test.