Current evidence and clinical applications of robotic gait training devices for motor function recovery post-stroke are increasingly available. Although existing research demonstrates that robotic gait training can improve patients' gait and balance, there remains a lack of in-depth investigation into its specific mechanisms of action concerning central nervous system (CNS) reorganization - notably, changes in activity within the motor cortex and associated neural networks. The intrinsic changes within the CNS have received insufficient attention, limiting a comprehensive and profound understanding of the rehabilitation outcomes. Therefore, this study aims to elucidate the potential mechanisms underlying robotic gait training-induced neuroplasticity by integrating functional near-infrared spectroscopy (fNIRS) technology with multi-dimensional lower limb motor function assessment tools (such as FAC, BBS, AMEDA, 10MWT, 6MWT, TUGT). It will systematically investigate the effects of robotic gait training on both the central nervous system and lower limb motor function in stroke patients. Furthermore, the study will compare the differences in functional recovery efficacy between robotic gait training and conventional rehabilitation therapies.