Page - 135 - in Proceedings - OAGM & ARW Joint Workshop 2016 on "Computer Vision and Robotics“

changeof themanipulator configurationwithout influencing theendeffector’s trajectory. Although the method of redundancy resolution is well understood for local optimization using the Jacobian matrix (see, e.g., [19]), we try to expand the knowledge of redundant robot systems onto a global view on redundancy. Based on this goal, cost functions for any type of manipulators (e.g., mobile, serial)canbeformulatedandcomputedfor theentiresystem. Moreover,acompliancecontrol scheme should be developed which uses this extensive description of the kinematic behavior. In case of a higher number of freedoms in the system (more than one) a multi-priority control can be used in ameaningfulway. Thus, theaimof thisworkpackage is to realize thecomputationofaprimary task (given trajectory of the end effector), a compliance control of the whole robot system and a desired optimization (e.g., energyminimization) basedononemathematical concept. 5. Conclusion In this article, the research project CollRob was presented proposing novel methods for human-robot collaboration. Research covers machine perception, sensitive redundant kinematic manipulation, dy- namicadaptiveplanning,human-robot interactionandinformationexchange,humanstateevaluation, and safety. To integrate different methods and system components into one working setup, a soft- ware architecture has been developed based on a publish-subscribe principle implemented in ROS. Different use cases have been identified and will be addressed for evaluation purposes of the devel- oped methods including a demo use case of joint human-robot Tangram puzzle building and various industrial applications incollaborationwith companypartners. References [1] ISO13849-1: 2015,Safetyofmachinery– Safety-related partsof control systems. [2] RobotiqCollaborativeRobotEbook. Robotiq, 2016. 6th edition. [3] JenayBeer,ArthurDFisk,andWendyARogers. Towardaframeworkfor levelsof robotautonomyinhuman-robot interaction. Journalof Human-Robot Interaction, 3(2):74, 2014. [4] CynthiaBreazeal. Toward sociable robots. Robotics andautonomous systems, 42(3):167–175, 2003. [5] Se´bastienDruon,Marie-Jose´Aldon,andAndre´Crosnier. Colorconstrainedicpforregistrationof largeunstructured 3dcolordatasets. In InformationAcquisition,2006IEEEInternationalConferenceon,pages249–255. IEEE,2006. [6] Mica R Endsley. Toward a theory of situation awareness in dynamic systems. Human Factors: The Journal of the Human Factors andErgonomics Society, 37(1):32–64,1995. [7] MarcoEwerton,GerhardNeumann,RudolfLioutikov,HeniBenAmor, JanPeters,andGuilhermeMaeda. Learning multiplecollaborativetaskswithamixtureof interactionprimitives. InRoboticsandAutomation(ICRA),2015IEEE International Conference on, pages 1535–1542. IEEE, 2015. [8] DavidFeil-SeiferandMajaJMataric´. Definingsociallyassistiverobotics. InRehabilitationRobotics,2005. ICORR 2005. 9th International Conference on, pages465–468. IEEE,2005. [9] Michael A Goodrich and Alan C Schultz. Human-robot interaction: a survey. Foundations and trends in human- computer interaction, 1(3):203–275,2007. [10] Yihsin Ho, Yoshihiro Kawagishi, Eri Sato-Shimokawara, and Toru Yamaguchi. A human motion recognition using data mining for a service robot. In Advanced Robotics (ICAR), 2011 15th International Conference on, pages 229–234. IEEE,2011. [11] Guy Hoffman and Cynthia Breazeal. Collaboration in human-robot teams. In Proc. of the AIAA 1st Intelligent Systems TechnicalConference,Chicago, IL, USA, 2004. [12] Chien-MingHuang andBilge Mutlu. Anticipatory robot control for efficient human-robot collaboration. 135