Dynamic Modeling And Control Of Cable-driven Parallel Robots

Cable-driven parallel robots (CDPRs) are closed-loop mechanism, in which fixed and moving platforms are connected by cables. CDPRs provide large workspace potentially, have high payload-to-weight radio, and can be assembled/reassembled and reconfigured easily. However, cables are flexible and elastic, and they will sag under the effect of their own weights. Moreover, the mass of cables will change when the end effector moves. These characteristics of cables bring huge challenge to the dynamic modeling and motion control of CDPRs. In this dissertation, our research includes dynamic modeling and dynamic control of CDPRs. The main work is summarized as follows:Firstly, when the movement of CDPRs is relatively slow, the dynamic model of cables can be replaced by static model. For a6-DOF CDPR, straight line model and catenary model were built respectively in this paper. In addition, the dynamic model of the CDPR was established based on these two kinds of cable models.Secondly, it is difficult to get analytical solution of the dynamic equations since the catenary equation contains transcendental function. Hence, a nonlinear optimization method with constraints was chosen in this paper to obtain the cable tension that satisfies the force constraints. In order to guarantee the numerical solution can close to the real solution sufficiently, inverse dynamics solution of the straight model was adopted as the initial value of the iteration optimization algorithm. Compared the straight line model with the catenary model of the cable, the results demonstrate that it is necessary to build the catenary model when the cable’s weight cannot be ignored.Thirdly, when CDPRs move with high speed or acceleration, the static model of cables can’t satisfy the requirement of modeling accuracy, so that the dynamic behavior of cables needs to be considered. For this resonance, this paper applies finite element method to establish the dynamic model of cables by ignoring the transverse vibration of the cables. The simulation results indicate that the dynamical behavior of the cable would cause the vibration of the end effector.Finally, for the dynamic model of the6-DOF CDPR, the computed torque controller and PD controller were employed respectively to the motion control of the robot. In the simulation experiment, the Runge-Kutta method was used to obtain the forward dynamic solution. Simulation results indicate that the computed torque controller with feedback compensation can achieve better control performances than the PD controller.