Seite - 6 - in Proceedings of the OAGM&ARW Joint Workshop - Vision, Automation and Robotics

implemented on a computer code similar to the computation of a space truss. The single point-to-point motions of ARTS aresent toeachATEviaawirelessconnection fromamaster, which is connected to a conventional personal computer. The main problem, which is currently investigated, is based on the difference of the idealized tetrahedral mesh and the constructed geometry of the ATEs, which brings in restrictions in the motion space of the system. Promising ways to overcome these limitations have been worked out and will be presented. Fig. 2. Topview of a single ATE a) mechanical design of an ATE shown without cables and electronics: A-docking mechanism, B-male connector, C-female connector, D-actuator, E-orientation element, F-spherical joint; b) three connected ATEs forming an adaptive robot with tetrahedral structure IV. RECONFIGURATION MECHANISM Besides the mechatronic design, the control of the ATEs can be challenging, as soon as many cells are connected to each other, compare Fig. 3. In addition to the design of ARTS, we are developing several computational schemes, which define the motion of each ATE for reconfiguration from one to another shape. In order to fulfill this challenging task, the computation is split into three parts: 1) In the first part, the initial and the final mesh of the structure is computed. It is necessary that both configurations consist of a similar number of ATEs. Thesimplestway isdepicted inFig.3,where the initial configuration consists of a rectangular block. 2) The rectangular block in Fig. 3a can be understood as parking positions of the ATEs. The main task of reconfiguration, is to find according parking positions to each of the ATEs of the structure, which is a hollow sphere in the present case. The shortest ways for movement of ATEs along the surface are depicted in Fig. 3a-e. This shows how a single ATE needs to be moved. In fact, the movement strategy is done such, that an ATE which has the longest distance to the base is selected in the structure, see Fig. 3e. This ATE is moved to an available parking space at the base block, which is closest to the center. While the algorithm is computing the destruction of the hollow sphere, the steps are then applied in reversed order. 3) In the final step, the movement of the ATEs needs to be performed by means of mesh deformation. This is done such that the cells can move along the surface. c) d) e) a) b) Fig. 3. Exemplary steps to reconfigure from one to another configuration. a) parking position; a-e) the red colored surfaces mark the shortest path for movement of ATEs along the surface. Currently, this is done with manual inputs only, how- ever, an algorithm which can automatically compute this transformation is currently developed. Converting a complex structure (A) into another complex structure (B) can be performed such that between these configurations, the ATEs are transformed into a rectangular block. In this way, only the reconfiguration from a rectangu- lar block to a complex structure must be computed. V. CONCLUSIONS The single adaptive tetrahedral elements (ATEs) follow a light-weight design principle. ARTS leads to a highly redundant superstructure and has the potential for a dis- ruptive technology. Current limitations are within geometric restrictions of the workspace and the differences between an idealizedgeometricmeshand the real (constructed)geometry of ATEs. REFERENCES [1] H. Ahmadzadeh, E. Masehian, and M. Asadpour, “Modular Robotic Systems: Characteristics and Applications,” Journal of Intelligent and Robotic Systems: Theory and Applications, vol. 81, no. 3-4, pp. 317– 357, 2016. [2] B. K. An, “EM-Cube: Cube-shaped, self-reconfigurable robots sliding on structure surfaces,” Proceedings - IEEE International Conference on Robotics and Automation, pp. 3149–3155, 2008. [3] R. Belisle, C.-h. Yu, and R. Nagpal, “Mechanical Design and Loco- motion of Modular Expanding Robots,” ICRA 2010 Workshop Modular Robots: State of the Art, pp. 17–23, 2010. [4] J. Gerstmayr and M. Pieber, “Modulares, selbst rekonfigurierbares Robotersystem,” pCT/EP2016/073703, 2016. [5] M. Jorgensen, E. Ostergaard, and H. Lund, “Modular ATRON: mod- ules for a self-reconfigurable robot,” 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE Cat. No.04CH37566), vol. 2, pp. 2068–2073, 2004. [6] M. Pieber and J. Gerstmayr, “An Adaptive Robot with Tetrahedral Cells,” The 4th Joint International Conference on Multibody System Dynamics, Montral, Canada, 2016. [7] J. W. Romanishin, K. Gilpin, and D. Rus, “M-blocks: Momentum- driven, magnetic modular robots,” IEEE International Conference on Intelligent Robots and Systems, pp. 4288–4295, 2013. [8] V. Zykov, A. Chan, and H. Lipson, “Molecubes: An Open-Source Mod- ular Robotics Kit,” IROS-2007 Self-Reconfigurable Robotics Workshop, pp. 3–6, 2007. 6