Mechatronic design of a parallel robot for high- speed, impedance-controlled manipulation Luca E. Bruzzone, Rezia M. Molfino, Matteo Zoppi The traditional industrial approach of robot manufacturers and the alternative integrated mechatronic approach are compared in Fig. 1. Abstract-- The paper deals with the integration of mechanical and control aspects involved in the design of an innovative 3- dof parallel kinematics machine. Its mechanical architecture is based on the 3-PUU scheme, actuated by crank-rod mechanisms and direct-drive brushless motors. The robot architecture is conceived in order to obtain high flexibility and reconfigurability in performing impedance-controlled manipulation tasks. The (direct and inverse) kinematic and dynamic models are discussed, with special reference to their application in the control system. The application of the integrated mechatronic approach and its advantages are described. A full-scale prototype has been built; its hardware and software layouts are presented. The test phase is at present in progress; the preliminary results are satisfying. Index terms-- mechatronics, parallel kinematics machine, impedance control I. INTRODUCTION Fig. 1. Comparison between the traditional approach of robot manufacturers and the integrated mechatronic approach. One of the key-issues in the development of modern robotic systems is the mechatronic approach, i.e. the strict integration, since the earliest design phases, of mechanical, control, electrical and electronic aspects. Nevertheless, the industrial robot manufacturers rarely use this methodology, which could remarkably improve the overall performance of their products. On the contrary, they invest most of the resources to develop user-friendly man-machine interfaces, applied to tested mechanical architectures; the axes are usually separately controlled, once the trajectory is planned, by standard linear PID loops. In the first case, the mechanical architecture is designed first of all, using the multibody simulation on the basis of the desired dynamic performance (i.e. the required motion) of the machine; the detailed design of each member is checked by means of FEM analysis considering the required robot motion and applied forces. Secondly, the control system is realized using the kinematic equations, regulating separately the position of each actuator in order to impose the desired end-effector position. With the mechatronic approach mechanics and control are studied simultaneously; to this aim, not only the kinematic model, but also the dynamic model is obtained; the static model can be derived from the dynamic model, eliminating the inertial terms, or from the velocity analysis, applying the virtual works principle. The performance of the system is improved by means of proper compensation of the inertial and gravity terms; moreover, the static model allows to regulate, whenever necessary, the interactions with the environment. In most industrial cases, the robot design process is divided into two sequential phases: the mechanical design and the control system design; usually, the mechanical design is not influenced by the control system conception: the designer simply tries to obtain high structural stiffness, because a very stiff machine is easily controllable by means of decoupled linear loops applied separately to each axis. On the other hand, if the mechanical architecture is conceived in parallel with a more sophisticated, model- based control strategy, it is possible to study exhaustively the behaviour of the overall system, emphasizing the peculiar properties of the mechanical architecture itself. The advantages of this approach are particularly evident for parallel robots [1]: as a matter of fact, their high dynamic performances, due to the limited moving masses, can be emphasized by means of the inertial compensation; this compensation is more important for PKM than for serial The authors are with PMAR Lab – DIMEC - University of Genova, Via All’Opera Pia 15A, 16145 Genova, Italy.