Real-Time Adjustment of Gait Training Patterns Based on Muscle Activation for Lower Limb Exoskeleton

Lower limb exoskeleton (LLE) has been widely studied recently. While how to adjust the subject-specific gait training patterns for LLE is one of the key problems that have not been well addressed. In this paper, a real-time method based on the muscle activation extracted from sEMG signals is proposed to adjust the gait training patterns, to maintain the given level of muscle activation and to enhance the subjects' engagement during gait training. Firstly, the muscle activation is calculated based on the previous research results. Then, in order to simplify the joint trajectory representation, the hip, knee, and ankle joint trajectories are parameterized based on finite Fourier series (FFS). Therefore, the gait training patterns can be represented by the coefficients and periods. Moreover, a real-time strategy is proposed to adjust the parameters of FFS based on the feedback of subjects' muscle activation during gait training. Finally, the performance of the proposed method is demonstrated by the experiment.