Abstract: Hand exoskeleton robots are a kind of hand rehabilitation training device. Traditional link-type hand exoskeleton robots usually only consider the endpoint trajectory coincidence of the distal interphalangeal joint, without considering the endpoint trajectory of the palm-phalangeal and interphalangeal joints. As a result, they cannot guarantee that the fingers can bend and stretch naturally, which poses problems in terms of comfort and safety. Based on the human-machine fusion exoskeleton structure synthesis method, the constraint equation of the human-machine structure is established, and a human-machine fusion design method is proposed, which considers the exoskeleton robot as a whole body structure with the fingers after being worn. An adaptive angle acquisition device was designed for data collection experiments, and two joint correlation functions were fitted and introduced into the human-machine structural constraint equation. The scale parameters were solved and optimized through particle swarm algorithm to complete the design of the human-machine coexisting exoskeleton robot and simulate its workspace. A prototype of the exoskeleton robot was built and experiments were conducted, showing that the machine has the characteristics of human-machine fusion, and it can fit the trajectory of the three joints well, proving the effectiveness of the design method.

Key words: human-machine fusion, link-type hand exoskeleton robot, end-point trajectory of joints, correlation function of finger joints, workspace