Page - 143 - in Proceedings - OAGM & ARW Joint Workshop 2016 on "Computer Vision and Robotics“

6. Conclusion In thispaperanautonomousorderpickingsystemwaspresented. Inorder tokeepthesystemmodular and portable a 3-TIER architecture was developed. The planning layer utilized an AI planner, which uses a PDDL description of the planning problem. The planner received the description of system skills, as well as start and goal state and provided a list of planned skills. Each skill is composed of skill primitives. These skill primitives are needed to address manipulation of the box, grasping items and perceptual tasks. The decomposition of skills into primitives enriched with monitoring and recovering capabilities contribute to the dependability of the system. The proposed system was implementedas aprototypeusing the robotBaxter andstandard robotics software libraries. Using this prototype implementation the concept of the skill primitives was evaluated. Although mostprimitivesworkedquitewell, theevaluationpointedoutsomeproblemsof thisproof-of-concept system. Within most primitives, the major problem was that the execution of planned trajectories was aborted because Baxter was not able to execute them precisely enough. However these errors weredetectedbyoursystemandreported to thehigh-levelcontroller. For futureworkamorereliable executionof armmotions byBaxterneeds tobeaddressed. References [1] S. Chitta, E.G. Jones, M. Ciocarlie, and K. Hsiao. Mobile manipulation in unstructured en- vironments: Perception, planning, and execution. IEEE Robotics & Automation Magazine, 19(2):58–71, June2012. [2] SachinChitta, IoanSucan,andSteveCousins. Moveit! IEEERobotics&AutomationMagazine, 1(19):18–19, 2012. [3] MateiCiocarlie,KaijenHsiao,EdwardGilJones,SachinChitta,RaduBogdanRusu,andIoanA S¸ucan. Towardsreliablegraspingandmanipulationinhouseholdenvironments. InExperimental Robotics, pages241–252. Springer, 2014. [4] Malik Ghallab, Dana Nau, and Paolo Traverso. Automatedplanning: theory&practice. Else- vier, 2004. [5] Chih-Wei Hsu, Benjamin W Wah, Ruoyun Huang, and Yixin Chen. Handling soft constraints and goals preferences in SGPlan. In International Conference on Automated Planning and Scheduling (ICAPS), 2006. [6] J.Huckaby, S. Vassos, andH.I. Christensen. Planningwith a task modeling framework inman- ufacturing robotics. In IEEE/RSJ InternationalConference on Intelligent Robots and Systems (IROS), pages5787–5794, Nov2013. [7] D. Mcdermott, M. Ghallab, A. Howe, C. Knoblock, A. Ram, M. Veloso, D. Weld, and D. Wilkins. PDDL - The Planning Domain Definition Language. Technical Report TR-98-003, Yale Center for Computational VisionandControl, 1998. [8] RobinMurphy. Introduction toAI robotics. MITpress, 2000. [9] Matthias Nieuwenhuisen, David Droeschel, Dirk Holz, Jorg Stuckler, Alexander Berner, Jun Li, Reinhard Klein, and Sven Behnke. Mobile bin picking with an anthropomorphic service robot. In IEEE International Conference on Robotics and Automation (ICRA), pages 2327– 2334. IEEE, 2013. 7 143