GripSync AI

What it does

GripSync AI is a wearable rehab kit pairing a soft-pneumatic glove with a hydraulic sensing ball. Multimodal data and cloud analysis auto-adjust assistance, gamify tasks and give instant feedback, offering stroke patients light, affordable home rehab.

Your inspiration

After my grandfather was discharged following a stroke, our home lacked both professional guidance and suitable devices; left alone with a printed exercise sheet, he saw no progress and soon quit. I realised successful rehab must merge expert feedback, home-friendly gear and visible results. While working in a university soft-robotics lab, I found that lightweight pneumatic “muscles” paired with low-cost sensors could assist recovery; adding a squeezable force-ball to capture multimodal hand motions and cloud-based AI assessment sparked GripSync AI, turning monotonous therapy into daily training where progress is felt and seen.

How it works

GripSync AI is a home-rehab trio: a hand-worn device with pneumatic actuators , a hydraulic squeeze ball, and a desk station for game-style guidance. Two ESP32-S3 boards (Espressif & Xiao) power sensing and control. The wearable has snap-off air tubes; the air cushions gently bend or open the fingers. Bend strips plus a motion chip track every angle in real time. The silicone ball, filled with safe fluid, is wrapped by thin pressure pads and hosts its own motion chip. Squeezing or twisting shifts inner pressure, letting the system read grip force and coordination. The book-sized station holds a quiet pump, ten fast valves, a display and a control board. Via low-energy Bluetooth it links with the other parts. It turns each exercise into a small game, shows progress live, and rates every move. After each session, data pass a GSC scoring model, sync to a cloud AI helper that sends a report, tweaks the next plan, and chats encouragement.

Design process

Focusing on home stroke recovery, we interviewed therapists and patients and used SAPAD + AHP to rank four priorities: interaction accuracy, data feedback, first-time guidance and usability. The solution merges a soft-pneumatic glove, a hydraulic pressure ball and a desktop hub: the glove supplies joint assist and angle data, the ball captures grip strength and coordination, and the hub aggregates signals and links to a cloud AI. A Xiao ESP32-S3 in the glove streams bend, IMU and pressure values; an ESP32-S3 hub with pump, valve manifold and BLE/Wi-Fi drives the actuator and relays data. We 3D-printed ball and hub shells, used paper mock-ups to refine one-hand donning, then laser-cut a dual-layer perforated-leather glove; a twist-lock pneumatic connector further simplifies setup. Wear tests with motion capture cut don time and improved detection stability. The final design uses single-shot moulded shells and off-the-shelf pneumatics for scalable manufacture, while a three-step app—quick setup, live training, data review—delivers seamless, measurable and motivating rehabilitation at home.

How it is different

AI-Agent Integration: A cloud AI agent fuses all sensor streams, judges motion quality and fatigue, and instantly adjusts each exercise while offering expert advice and motivating chat. Gamified Interaction: Hand-worn tracking plus a hydraulic FSR ball recognize grasp, pinch and twist, turning them into block-picking, balloon-squeezing or can-opening games; the hub gives live cues while algorithms score pressure curves and accuracy for both fun and precise feedback. End-to-End Rehab: Glove, ball and hub are swappable; as recovery advances, users can detach the pneumatic assist and move from passive to active training. All sessions log to the cloud, producing stage reports for clinicians and closing the rehab loop from hospital to home.

Future plans

Future Plan: Phase 1 – refine glove materials, pressure-ball sensors and pneumatic fittings, run durability trials and earn core IEC/ISO safety approvals. Phase 2 – partner with rehab clinics for a 30–50-patient pilot, logging motion and adherence to fine-tune the AI model and validate outcomes. Phase 3 – add AR/VR visual guidance, richer haptics and clip-on actuators for wrist and elbow therapy, turning the kit into a full upper-limb platform. Commercially we will launch a lease-plus-subscription model for clinics and content bundles for home users, while completing medical-device registration and scaling production within three years.

Awards

First Prize of Excellent Graduation Design of Henan University of Science and Technology