Research On Grasping Performance Optimization And Grasp Force Control For The Underactuated Prosthetic Hand

The underactuated prosthetic hand with anthropomorphic motion feature can adaptively grasp objects with different sizes and shapes. Its number of actuators is less than that of degrees of freedom, which not only decreased the weight and size without reducing hand dexterity, but also reduced control difficulty. So it became a hot topic in the study of a multi-freedom prosthetic hand. However, ejection phenomenon may occur in the process of grasping, which reflexes the essence of underactuation and influences the grasp stability seriously. In this paper, grasping performance of underactuated prosthetic hand is optimized, and a grasp force control method based on its grasping characteristic is proposed.First, the research status of the underactuated prosthetic hand and the existing grasping performance optimization methods are introduced in detail. Based on that, the deficiencies of these methods are sumerized. The development status of prosthetic hand grasp force control is also reviewed.The static model of the prosthetic finger with muti-links mechanism is established by using the principle of virtual work. According to the closed kinematic chains of the finger, the dynamic model is established using Lagrange equation based on the virtual spring approach for the underactuated prosthetic finger. The dynamic model was verified by grasping simulation.The grasping characteristics of the underactuated finger are studied. The influence that the coupling link spring stiffness has on its grasping performance is discussed. Optimization criterions are developed in three aspects (i.e. the underactuated finger adaptability, grasping stability and grasping force distribution). Then the underactuated finger grasping performance is optimized.In order to further reduce the ejection of underactuated prosthetic hand in the process of grasping, a grasp force control method which can prevent ejection is proposed. Whether the ejection will occur can be judged by detecting changes in the location of the base joint. Thus ejection is prevented through modifying the reference force. A fuzzy PI controller is also designed as force controller. The grasping simulations are carried out to evaluate the grasp force control strategy.Eventually, the grasping performance of the underactuated hand and its grasp force control experiments are done. Experimental results show that the grasping performance of finger is improved with the optimized coupling link spring and the force control method has good control effect.