RESUMEN
Motivation and voluntary drive of patients can be improved by applying biofeedback during robot-assisted rehabilitation trainings. Biofeedback systems were traditionally based on theoretical assumptions. In this paper, we present a novel approach to calculate biofeedback during robot-assisted gait training. Our method was based on empirical data that were obtained from healthy subjects when simulating distinctive degrees of walking performance during robot-assisted gait training. This empirical data-based biofeedback (EDBF) method was evaluated with 18 subjects without gait disorders. A higher correlation between the subjects' walking performance and biofeedback values was found for the EDBF method compared to a theory-based biofeedback approach.