KittingBot: A Mobile Manipulation Robot for Collaborative Kitting in Automotive Logistics

Abstract

Individualized manufacturing of cars requires kitting: the collection of individual sets of part variants for each car. This challenging logistic task is frequently performed manually by warehouseman. We propose a mobile manipulation robotic system for autonomous kitting, building on the Kuka Miiwa platform which consists of an omnidirectional base, a 7 DoF collaborative iiwa manipulator, cameras, and distance sensors. Software modules for detection and pose estimation of transport boxes, part segmentation in these containers, recognition of part variants, grasp generation, and arm trajectory optimization have been developed and integrated. Our system is designed for collaborative kitting, i.e. some parts are collected by warehouseman while other parts are picked by the robot. To address safe human-robot collaboration, fast arm trajectory replanning considering previously unforeseen obstacles is realized. The developed system was evaluated in the European Robotics Challenge 2, where the Miiwa robot demonstrated autonomous kitting, part variant recognition, and avoidance of unforeseen obstacles.

Towards Mobile Multi-Task Manipulation in a Confined and Integrated Environment with Irregular Objects

“…Stückle et al extended this framework using a mobile manipulator in a spacious room for the Robot@Home competition [1] to grasp sausages from a barbecue using tongs, open bottles, and water plants. Pavlichenko et al developed a mobile manipulation framework for part kitting in a car manufacturing environment [16], focusing on part detection and arm trajectory optimization to increase the safety of humans in human-robot collaboration tasks. Most of these frameworks focus on specific aspects of a complex task and neglect other aspects.…”

Collaborative mobile industrial manipulator: A review of system architecture and applications

“…The operation environments are stable such as a factory or warehouse and extensive research regarding CMIM with object transferring capabilities has already been conducted. For example, the collaborative autonomous kitting logistics in a car manufacturing warehouse [33].…”

Self-supervised damage-avoiding manipulation strategy optimization via mental simulation

“…All experiments were performed using a setup of multiple similar Docker 7 [61] 5 The initial training set size should depend on the maximum tolerable time until the classifier is required for the first time. 6 Convergence of the optimization method strongly depends on the concrete implementation and application scenario and is hard to define generically as explained in the evaluation section. 7 https://docker.com containers executed in parallel, increasing the overall system efficiency even more.…”

“…The bottom-up way to deal with this problem is physical compliance of grasping systems or enhanced perception and manipulation algorithms (e.g. [1,2,3,4,5,6], amongst others). These methods commonly rely on obstacle avoidance as to not provoke any margin-violating situation in advance.…”

“…Depending on the free computing capacities on the robot, one iteration of this loop varies in runtime, hence the per-time utility improvement of the respective classifier is subject to other tasks being executed at the same time. Therefore a manipulation strategy may take a long time to converge 6 if not operated during loadfree runtimes. The prediction of manipulation sequences on real scenes, however, happens instantly in near-real time with the currently built classifier.…”