Improving Complex Task Performance in Powered Upper Limb Exoskeletons With Adaptive Proportional Myoelectric Control for User Motor Strategy Tracking

Powered exoskeletons have emerged as promising tools with applications in assistance, augmentation, and rehabilitation. However, the realization of their full potential hinges on the accurate classifications of user intent. Traditional proportional myoelectric controllers with fixed thresholds require users to develop the necessary motor program – the optimal coordination of movements with the exoskeleton – prior to effective operation. Novices, who may not have mastered this coordination, often experience decreased accuracy in intention classification, leading to a trade-off between ease of static and dynamic tasks: easier movement initiation typically results in less stable holding, and vice versa. This study introduced a novel proportional myoelectric controller with real-time adaptive thresholds designed to continuously track the user's evolving motor program to enhance intent classification for both movement and holding. In an elbow target position matching task with twelve participants, this controller showed reductions in both intention classification error magnitudes and muscular effort during movement initiation compared to the traditional fixed thresholds method. Nonetheless, participants did not perceive significant improvements, suggesting the need for continued enhancements. This letter presents an innovative approach to leveraging the user's current motor program for determining intention classification parameters, moving beyond the limitations of fixed or manually-tuned settings.