The application belongs to the technical field of wearable
exoskeleton device control, and provides a wearable
exoskeleton device control method based on
muscle coordination, which comprises the following steps: acquiring electromyographic signals of key
muscle groups of bilateral lower limbs in real time and extracting normalized
muscle activation; inputting the normalized
muscle activation vector into a pre-trained
Gaussian mixture model, calculating the likelihood probability of belonging to a normal muscle coordination mode, and determining the most matched
Gaussian component; calculating a comprehensive
abnormality degree E based on the likelihood probability; generating an auxiliary torque of each joint based on the comprehensive
abnormality degree E and an auxiliary
intensity coefficient of the corresponding side, wherein the auxiliary
intensity coefficient is distributed according to a bilateral
motor ability score, and a side with weaker
motor ability obtains a larger coefficient; and driving the movement of each joint by the auxiliary torque. The application quantifies the
abnormality degree of muscle coordination from the neuromuscular
control level, and adaptively generates an auxiliary torque according to the same, so as to guide the muscle coordination mode of a child patient to be reshaped in a normal direction.