A biomimetic, ultralow-power edge-AI-empowered and self-sustaining gait analysis system
- Fuying Dong
- Chi Han
- Pengchong Xu
- Jasleen Chhatwal
- Xinnian Jiang
- Tengteng Wang
- Abigail Hsu
- Minzhu Baek
- Di Wu
- Rui Li
- Yuanwen Jiang
- Bozhi Tian
- Jason Y. Fang
- Simiao Niu
2026-08-19
Smart digital health has reshaped patient monitoring, but it faces a fundamental trade-off between device intelligence and continuous, energy-efficient monitoring. Inspired by self-sustaining intelligent biospecies, we develop a biomimetic, battery-free, and high-precision edge-AI system through a harvested-energy-constrained holistic co-design that couples ultralow-power edge-AI-empowered sensor hardware with biomechanical energy harvesting and cold-start power management. Our edge-AI-empowered motion sensor performs instantaneous, context-aware on-device inference and timely result updating from raw sensor data while consuming only 86 μW. A high-output energy harvester and tailored high-efficiency power management circuitry sustain energy levels exceeding system requirements, eliminating downtime associated with charging and enabling true 24/7, hassle-free monitoring. This breakthrough establishes a paradigm for system-level, edge-AI-empowered, and self-sustaining sensing, demonstrating that intelligence and energy autonomy can coexist within a single wearable platform and pointing to next-generation always-on, personalized digital health systems.