MULTIMODAL SENSING FOR USER INTENTION PREDICTION IN UPPER LIMB EXOSKELETONS

The upper limb exoskeleton holds promising benefit to decrease negative physical outcomes for individuals in the workspace. Utilizing a human centric design to upper limb exoskeleton technology can remarkably improve the compatibility and user adaptation of the exoskeleton which positively affect the work performance. In this paper, we present artificial intelligence based multimodal sensing modality for active upper limb exoskeletons to predict motion of exoskeleton user. Our approach involves the integration of electromyography, motion, and ultrasound sensors to detect and analyze the precise movements associated with specific work tasks. Our data analysis revealed distinct patterns corresponding to each task, suggesting the promising utility of these patterns for neural network-based user intention detection.