An Industrial Robotic Knowledge Representation for Kit Building Applications Stephen Balakirsky, Zeid Kootbally, Craig Schlenoff, Thomas Kramer, and Satyandra Gupta Abstract— The IEEE RAS Ontologies for Robotics and Automation Working Group is dedicated to developing a methodology for knowledge representation and reasoning in robotics and automation. As part of this working group, the Industrial Robots sub-group is tasked with studying industrial applications of the ontology. One of the first areas of interest for this subgroup is the area of kit building or kitting. This is a process that brings parts that will be used in assembly operations together in a kit and then moves the kit to the assembly area where the parts are used in the final assembly. This paper examines the knowledge representations that have been developed and implemented for the kitting problem. I. INTRODUCTION Kitting is the process in which several different, but related items are placed into a container and supplied together as a single unit. Kitting itself may be viewed as a specialization of the general bin-picking problem. In industrial assembly of manufactured products, kitting is often performed prior to final assembly. Manufacturers utilize kitting due to its ability to provide cost savings [3] including saving manufacturing or assembly space [13], reducing assembly workers walking and searching times [15], and increasing line flexibility [2] and balance [9]. Several different techniques are used to create kits. A kitting operation where a kit box is stationary until filled at a single kitting workstation is referred to as batch kitting. In zone kitting, the kit moves while being filled and will pass through one or more zones before it is completed. This paper focuses on batch kitting processes. In batch kitting, the kit’s component parts may be staged in containers positioned in the workstation or may arrive on a conveyor. Component parts may be fixtured, for example placed in compartments on trays, or may be in random orientations, for example placed in a large bin. In addition to the kit’s component parts, the workstation usually contains a storage area for empty kit boxes as well as completed kits. Kitting has not yet been automated in many industries where automation may be feasible. Consequently, the cost of building kits is higher than it could be. We are addressing this problem by building models of the knowledge that will S. Balakirsky and C. Schlenoff are with the Intelligent Systems Division, National Institute of Standards and Technology, Gaithersburg, MD, USA (e-mail:stephen.balakirsky@nist.gov, craig.schlenoff@nist.gov) Z. Kootbally is with the Department of Mechanical Engineering, Univer- sity of Maryland, College Park, MD, USA (email: zeid.kootbally@nist.gov) T. Kramer is with the Department of Mechanical Engineer- ing, Catholic University of America, Washington, DC, USA (email: thomas.kramer@nist.gov) S. Gupta is with the Maryland Robotics Center, University of Maryland, College Park, MD, USA (email: skgupta@umd.edu) be required to operate an automated kitting workstation in an agile manufacturing environment. This workstation must be able to cope with variations in kit contents, kit layout, and component supply. Our models include representations for non-executable information about the workstation such as information about parts and kit designs, models of executable information such as actions, preconditions, and effects, and models of the process plan necessary for kit construction. A discussion of the functional requirements for the process plan may be found in [1]. For our automated kitting workstation, we assume that a robot performs a series of pick-and-place operations in order to construct the kit. These operations include: 1) Pick empty kit and place on work table. 2) Pick multiple component parts and place in kit. 3) Pick completed kit and place in full kit storage area. Each of these actions may be a compound action that includes other actions such as end-of-arm tool changes, path planning, and obstacle avoidance. It should be noted that multiple kits may be built simul- taneously. Finished kits are moved to the assembly floor where components are picked from the kit for use in the assembly procedure. The kits are normally designed to facil- itate component picking in the correct sequence for assembly. Component orientation may be constrained by the kit design in order to ease the pick-to-assembly process. Empty kits are returned to the kit building area for reuse. Although the knowledge requirements described in the previous paragraph have been identified for the kitting do- main, they are clearly applicable to many types of industrial robot applications. As such, we expect that these knowledge requirements will serve as the basis for the industrial robot ontology being developed in the IEEE RAS Ontologies for Robotics and Automation Working Group [11] (henceforth referred to as the IEEE WG). Throughout the process of de- veloping the kitting ontology, the group will constantly look at the applicability of the requirements outside of kitting and move the pertinent knowledge “up” the ontology (whether in the portion that models the kitting sub-domain, the industrial robot domain, or the upper ontology), as appropriate. In keeping with the standards philosophy of the IEEE WG, we require the models being developed to be as widely applicable as possible. Therefore, we have created a layered model abstraction where users may adopt as many of the layers of the abstraction as make sense for their specific application. The architecture shown in Figure 1, though developed for the kitting ontology, can be equally applicable to the implementation of any type of formal manufacturing 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems October 7-12, 2012. Vilamoura, Algarve, Portugal 978-1-4673-1735-1/12/S31.00 ©2012 IEEE 1365