Improving 3D Vision-Robot Registration for Assembly Tasks

Published

April 3, 2020

Author(s)

Marek Franaszek, Geraldine S. Cheok, Karl Van Wyk, Jeremy A. Marvel

Abstract

Quality registration between the coordinate frames of a perception system and a robot is important for the efficient operation of autonomous systems in vision-guided assembly lines. Rigid-body registration, which is based on the measurement of corresponding points (fiducials)in both frames, is a commonly used method. Noise and possible bias in the measured points degrade the quality of registration as gauged by the Target Registration Error (TRE). This paper presents a procedure which leads to decreased TRE when the bias is larger than the noise. Experiments performed with a motion tracking system and robot arm show that the Root Mean Squared target errors (RMST) can be reduced by as much as 84 % by placing fiducials in carefully selected locations in the working volume and through the use of the Restoration of Rigid Body Condition (RRBC) method.

Citation

NIST Interagency/Internal Report (NISTIR) - 8300

Report Number

8300

NIST Pub Series

NIST Interagency/Internal Report (NISTIR)

Pub Type

NIST Pubs

Download Paper

DOI Link

Keywords

rigid-body registration, autonomous robotic systems, bias

Citation

Franaszek, M. , Cheok, G. , Van, K. and Marvel, J. (2020), Improving 3D Vision-Robot Registration for Assembly Tasks, NIST Interagency/Internal Report (NISTIR), National Institute of Standards and Technology, Gaithersburg, MD, [online], https://doi.org/10.6028/NIST.IR.8300 (Accessed October 27, 2025)