Continuation of our R&D efforts on top of the base model of the Hanlun Tai-Chi Push-Hands Robot (the “Robot”) led to major breakthroughs in user customization and improved efficiency of the Robot-led Push-Hands exercises in improving the physical balance of the user. In terms of user customization, the Robot is now able to recognize the current user and automatically adjust the height of the Robot for a better-quality exercise. The technologies behind this breakthrough include facial recognition, 5G transmission and cloud computing plus extensive mechanical and electronic modifications to the existing Robot. The efficacy of the Robot led Push-Hands exercises in improving the physical balance of the user is further enhanced through a direct visual display of the pressure distribution of the foot against the floor while exercising. Using biofeedback, the user gradually learns how to use the foot and lower limbs to balance the moving mass of the upper body to prevent a fall.