Research on Intention Perception Algorithm of Exoskeleton Based on Multi-sensor Information Fusion

As a human-robot coupling system, human-robot interaction plays an important role in the power-assisted exoskeleton robot. Accurately identifying and perceiving the wearer’s motion intention is one of the research difficulties of exoskeleton robots. Aiming at the problem that the wearer’s motion intention is difficult to perceive during human-computer interaction, this paper uses the harmony search algorithm to optimize the one-versus-one support vector machines, and combines with the finite state machine to propose a new intention perception algorithm. The experimental results show that the algorithm can quickly and accurately perceive the wearer’s movement intention, and can identify the abnormal movement state transition in time to ensure the security of the system.