Research On Master-Slave Control For The Underactuated Prosthetic Hand

Anthropomorphic prosthetic hands as a typical inter-disciplinary research field which integrates the subjects of biology, electromechanical and computer together, the study of prosthetic hand has been a popular research field of robotic area. but its development has been limited to the limitation of biological/electromechanical interface technology, which has few motion modes and worse intuitive control feelings. It is a very difficult and challenging task to design an autonomous prosthetic hand due to the dynamic, complex and unknown environment with the technology constraint of human intelligence and computer supporting techniques. We use human’s higher level of intelligence, rich experience, powerful ability of study and deal practical problems to finish the high-rise decision-making which the prosthetic hand difficult to achieve, and finish the real-time grasp task by prosthetic hand, it is the organic combination of human intelligence and mechanical properties of prosthetic, and this study will provide the foundation for future humanoid myoelectric control.The present research status of prosthetic hand are describled in detail, base on the characteristics and develop tendency of humanoid prosthetic hand, a multifunctional three-jointed underactuated humanoid hand with coupling and self-adaptive functions is researched.We deduce the kinematic model of prosthetic hand firstly, We derived the dynamical equations with the method of Lagrange equation and virtual spring approach together, the constraint pairs are replaced by virtual spring and it was included in dynamic equations automatically, and contains no coupling terms. The3D prototype was introduced into ADAMS with the dynamics simulation, to verify the correctness and reasonableness.According to the structural features of human hand and prosthetic hand, its master-slave control scheme based on the data glove was constructed, establish the motion mapping relation via the kinetic model. We use the computed torque method design a master-slave control rules base on its dynamic model, and analysis its control effectiveness and robustness with the tracking control simulation. Finally, we put up hardware control platform of master-slave control based on data glove, finish the data acquisition of finger joints via the data glove and conduct the real-time control on virtual instrument platform. We develop the software control system based on C++and Lab VIEW. Finally, the control experiment of master-slave tracking control was conducted, and the result verifies the effectiveness of control methods.