Robotic Gait Training and Evaluation System — Five-Axis Physiological Gait Rehab

The AI1 Gait Training and Evaluation System is a hospital-grade AI rehabilitation robot for the systematic, high-volume, physiological gait training of patients with neurological, musculoskeletal, and age-related walking impairment. Based on the clinically validated principle of neural plasticity, the AI1 uses five-axis linkage robotic mechanical legs to guide patients through precisely simulated physiological gait patterns — completing the large number of repetitive, correctly-formed gait cycles required to establish new motor pathways and restore functional walking ability.

Five-Axis Linkage Technology — Autonomous, Adaptive Gait Guidance

The AI1's five-axis linkage mechanism is the engineering core of its clinical effectiveness. Five independent motorised axes — spanning hip flexion-extension, knee flexion-extension, and ankle dorsiflexion-plantarflexion across both lower limbs — operate in coordinated linkage to produce a three-dimensional, multi-joint gait pattern that faithfully replicates the biomechanics of normal human walking.

Three Training Modes — Passive, Assisted & Active (Free Switching) : 

1. Passive Mode — The robotic system guides the patient's lower limbs through the complete gait cycle without requiring any voluntary motor effort from the patient. In passive mode, the five-axis linkage provides all the mechanical force required to complete each stride — the patient's legs are moved through physiological gait while they focus on the sensory experience of normal walking mechanics. Passive mode is indicated from the earliest stages of post-stroke or post-surgical gait rehabilitation, when voluntary lower limb movement is absent or severely limited.
2. Assisted (Helpful) Mode — The system detects the patient's voluntary motor effort and provides precisely calibrated supplementary assistance to complete each gait cycle — amplifying the patient's own motor commands rather than overriding them. This mode bridges passive and fully active training: the robot follows the patient's movement intention, adding enough force to achieve a complete, correctly-formed gait cycle, while progressively reducing the assistance level as the patient's voluntary output increases.
3. Active Mode — The patient drives gait independently, with the robotic system providing safety support and guidance but no active assistance. Active mode challenges the patient to generate and control the full mechanical requirements of walking within the support environment of the AI1 — providing the progressive, high-challenge practice that drives the final stages of gait recovery toward functional independence.

Human-Computer Interactive Training — Virtual Reality Environment : The AI1 integrates virtual reality human-computer interaction with the robotic gait training platform — creating an immersive, engaging therapeutic environment that addresses one of the most significant clinical challenges in gait rehabilitation: patient motivation and active cognitive engagement during repetitive training.

AI Mirror Gait Learning Assessment : The AI1's AI mirror gait learning assessment is a machine learning-based gait analysis capability that distinguishes it from purely mechanical gait trainers. The system continuously acquires and analyses the patient's gait pattern data across training sessions.

Gait Possibility Expansion — By learning the range of gait variations within the patient's movement data, the AI generates alternative gait pathway options that expand the training stimulus beyond a single fixed template — exposing the patient's nervous system to a richer variety of physiologically valid gait patterns and stimulating broader neural pathway recruitment.

Personalised Training Optimisation — The AI model enables the system to predict and respond to the patient's likely movement at each phase of the gait cycle — adapting the assistance level, timing, and guidance force parameters based on learned patient-specific data rather than population-average defaults