An all-position type control strategy of the haptic interactive virtual training system based on cable-driven

Abstract

The virtual microgravity training system based on cable drive usually uses a force-position hybrid control strategy which has following problems: the force control method is sensitive to load disturbances, variable stiffness characteristics of cables reduce the control accuracy of PID controllers, and the expected tension fluctuations are large. These will affect the control accuracy, and further affect the tactile sensation and training effectiveness of astronauts. For the above problems, an all-position type control strategy is proposed to improve the system control accuracy. This strategy uses a compliant control method. In this method, elastic elements are connected in cables, the conversion model of tension and displacement is established, and the tension control is realized by the displacement control which has characteristics of high control accuracy and strong resistance to load disturbance. The PID controller is replaced by the active disturbance rejection controller. In this controller, the tracking differentiator is used to reduce high frequency noises of the input signal, and the extended state observer is used to estimate and compensate the error caused by the change of the cable stiffness. A tension distribution method is designed to make expected cable tensions approach the average tension to reduce the tension fluctuation. The experimental results show that compared with the force-position hybrid control strategy, the all-position type control strategy reduces the tension error and speed error by about 51% and 33% respectively.