Towards Measurement of Advanced Mobile Manipulator Performance for Assembly Applications

Abstract

Advanced mobile manipulators offer the manufacturing industry the potential of a highly adaptive system to perform precision assembly tasks in agile environments. Developing performance measurement capabilities for mobile manipulator systems will support the advancement of manipulator-vehicle coordination, precision, accuracy, and robustness. The National Institute of Standards and Technology (NIST) developed a measurement methodology for advanced mobile manipulators, which are mobile robots with an onboard robot arm, each from a different manufacturer and with a different controller. The measurement methodology is designed to assess the performance of mobile manipulators for manufacturing applications such as assembly, where position and orientation accuracy are critically important. The proposed methodology used an artifact designed at NIST to simulate an assembly task station. An optical tracking system measured the location of the targets, the mobile base position and orientation, and the position of a manipulator end-of-arm tool. The experimental data collected, using three different system computers, were timestamped using local clocks synchronized via the NIST internal and local area networks to align the data streams. The mobile manipulator system, the experimental methodology, the data analysis, and the results of the measurement methodology uncertainty and mobile manipulator positioning uncertainty using the novel artifacts are described in this paper. For the system tested, the mobile manipulator detection of 2 mm diameter fiducials can be achieved after registration. With validation from the optical tracking system, the study demonstrated that the Reconfigurable Mobile Manipulator Artifact (RMMA) has potential for serving as a novel standalone, low-cost test method for measuring the positioning uncertainty of mobile manipulators within a measurement uncertainty of 2 mm.

Performance measurement of a mobile manipulator-on-a-cart and coordinate registration methods for manufacturing applications

“…The performance assessment of mobile manipulators has focused on measures reflective of "safety, precision, and accuracy" [6]. In the past, research predominantly emphasized manipulator tipover stability and the enhancement of coordination between the vehicle and robotic arm in various environments [6][7][8]. It was also noted that measurement systems existed to assess the individual sub components of a mobile manipulator.…”

“…It was also noted that measurement systems existed to assess the individual sub components of a mobile manipulator. Examples of such systems include laser interferometers (non-contact) and artificial cues (contact-based) used to assess the performance of manipulator arms [6,9,10]. However, the performance measurement of mobile manipulators as a complete system represented a relatively new area of research [3].…”

“…However, the performance measurement of mobile manipulators as a complete system represented a relatively new area of research [3]. As the potential application domains continue to expand, it becomes more important for users and manufacturers to ensure mobile manipulator accuracy, repeatability, and coordination [6,11].…”

“…Past work included the development of the Re-configurable Mobile Manipulator Artifact (RMMA), which introduced a cost-effective, and flexible artifact for measuring mobile manipulator performance across numerous simulated manufacturing tasks and the use of an optical tracking system (OTS) to establish a ground truth reference of comparison to the RMMA [9,15,16]. Both measurement techniques were applied towards measuring both a mobile manipulator that consisted of an AGV and an autonomous mobile robot (AMR) for the mobile platform [6,15,[17][18][19]. Additionally, the RMMA was used to compare coordinate registration methods for use with the mobile manipulators [20].…”

“…3.1.3 -3.1.2) for localizing retro-reflective targets placed on the RMMA, as well as visual fiducials tracked via an augmented reality toolkit. More recently, experiments on the AMR used the same laser bisect method [6]. For the second part of the mobile manipulator-on-a-cart project, the primary interest was to introduce an additional, alternative technique for rapid coordinate registration to the mobile manipulator-on-a-cart system.…”