Comparison of Registration Methods for Mobile Manipulators

Abstract

Mobile manipulators can be effective, efficient and flexible for automation on the factory floor but will need safety and performance standards for wide adoption. This paper looks at a specific area of performance standards [1] for docking and workpiece registration, with the intent of evaluating how quickly, repeatably, and accurately a mobilemanipulator end effector can be aligned with a known physical target to facilitate peg-in-hole insertion tasks. To evaluate mobile manipulator docking, we conducted experiments with an AGV-mounted arm in a laboratory space equipped with an extensive optical tracking system and a standardized test piece (artifacts) simulating an industrial assembly. We experimented with different strategies and sensors for registration and report on these approaches. * Disclaimer: NIST does not endorse products discussed within this paper nor manufacturers of these products. Products mentioned are for information purposes only and are not expressed as an endorsement for them or their manufacturer.

Towards Measurement of Advanced Mobile Manipulator Performance for Assembly Applications

“…Research and development of novel test methods for mobile manipulators include the integration and comparison of: 1) the development of an optical tracking system (OTS) measurement method [16] as a means of establishing ground truth for mobile robot and manipulator positioning uncertainty, coordination between the manipulator and the vehicle and repeatability; 2) the development of a test method simulating flexible, reconfigurable manufacturing tasks that includes a relatively inexpensive reconfigurable mobile manipulator artifact (RMMA), as compared to an OTS, for use in measuring performance of mobile robots with an onboard manipulator [16]; and 3) the development of an analytical model for characterizing the measurement uncertainty of the RMMA to measure mobile manipulator performance using the OTS measurements as the ground truth [18]. Simulated and experimental test methods of applying RMMAs were developed for testing an AGV-based mobile manipulator's ability to locate the target, with measurement of position and alignment accuracy [6][17] [26]. Computational methods for translating the target, manipulator and AGV coordinate frames to a world coordinate frame in order to assess performance were also developed and validated with errors less than 2 mm [22].…”

Model of Mobile Manipulator Performance Measurement using SysML

“…As noted in section I, previous experiments have been performed that validate the SysML models described in this paper. A mobile manipulator performance measurement concept was developed over the past two years [20][21][22]. This concept, which expanded on the experiments in [20], consisted of using a NIST-developed, reconfigurable mobile manipulator artifact (RMMA) (see Figure 1) with varying-sized fiducials to be detected by a sensor carried by the mobile manipulator to minimize the measured uncertainty.…”

Mobile manipulator performance measurement data

“…NIST has been developing a mobile manipulator, performance measurement concept over the past two years [3,4,5]. This concept, which evolved from the experiments described in [6], consists of using 1) a NIST-developed, reconfigurable, mobile manipulator artifact (RMMA) (see Figure 1) with 2) interchangeable and varying sized fiducials to be detected by a sensor carried by the mobile manipulator.…”

“…This publication is available free of charge from: https://doi.org/10.6028/NIST. TN.1965 A square search pattern provided the best search method [5] for detection of the 2 mm or smaller fiducials. An example of the square step pattern is shown in Figure 3 where the search begins away from the reflector at the chosen start point (yellow arrow dotted end).…”

“…The current location and orientation of the AGV were recorded. The arm was repositioned manually until the sensor detected alignment [3,5,6] with each of the index fiducials, and the manipulator position was recorded. This information, along with the position of the manipulator base relative to the vehicle's coordinate system, allowed the correct manipulator coordinates for the index fiducials to be calculated for an arbitrary AGV location.…”