Trajectory planning and control of cable-driven parallel robots

Abstract

The aim of this work is to investigate on trajectory planning and control of cable-driven parallel robots to improve the system performance. Stiffness and dexterity are the performance indices widely used in design and control of robotic systems. No previous work on adaptive cable-driven systems has discussed how to control the position of the pulley blocks to achieve optimal dexterity and stiffness. Considering a quasi-static motion of the end-effector, we neglected the active stiffness of the system and proposed pulley blocks trajectory planning strategies that maximize dexterity and elastic stiffness indices simultaneously for some cases of adaptive cable-driven designs by taking advantage of the increased redundancy. For non-adaptive design of cable-driven parallel robots, it is impossible to change the dexterity and elastic stiffness indices for a certain position of end-effector due to fixed orientation and length of cables; however, active stiffness can be modified by changing the tension in cables. Tension increment can be desirable due to stiffness augmentation, higher trajectory tracking performance, more precise motion and disturbance rejection; however, it can increase power consumption, and saturation in actuators may occur. Usually, cable tension distribution methods work based on a fixed minimum tension in cables. Such values are chosen through experiments to gain the desired trajectory tracking performance of the system, considering capability of actuators at the same time. To improve the system performance we proposed Dynamic Minimum Tension Control (DMTC) method. In this approach, the minimum tension is changing on-the-fly according to stiffness, dynamics of the system, and error values as feedback. We used a simple test bed to compare traditional fixed minimum tension utilization, and the proposed approach. Experimental results showed that the DMTC is more efficient than traditional approaches in terms of accuracy and energy consumption. Also an appropriate control algorithm can improve the system performance. The linear quadratic optimal control can play an important role in controlling cable-driven parallel robots by providing all the states of the system for the feedback, including velocity and position, in addition to optimal results. A linear quadratic optimal controller was designed and tested. The significant experimental results are presented and discussed.