Lower limb rehabilitation training trajectory optimization method with generalization performance

The invention discloses a lower limb rehabilitation training trajectory optimization method with generalization performance, which comprises the following steps of: designing an initial state of a lower limb exoskeleton, wearing the lower limb exoskeleton by a healthy subject, driving the lower limb exoskeleton to surround on a cross section, and collecting a motion trajectory through an IMU (Inertial Measurement Unit); the collected motion trail is separated into two vectors (# imgabs0 # and # imgabs1 #), the # imgabs2 # vector and the # imgabs3 # vector are clustered respectively, and different clusters are obtained; an interpolation algorithm is adopted for each cluster, so that the clusters are consistent in time length; performing DBA analysis on the interpolated data to obtain a center vector of each cluster; selecting feature parameters of a healthy subject, and training a feature classifier with center vectors of the # imgabs4 # vector and the # imgabs5 # vector respectively; and inputting the charact