Robust standard hardware, open architecture.
From the sensor in the room to the dashboard: every layer built on proven, maintainable components — privacy-compliant and offline-capable.
From the sensor in the room to the dashboard: every layer built on proven, maintainable components — privacy-compliant and offline-capable.
| Layer | Component | Role |
|---|---|---|
| Wearable | Smartwatch (model-agnostic) | Heart rate, SpO₂, movement — over Bluetooth Low Energy |
| Radio bridge | ESP32-S3 (1 per room) | Reads the watch over BLE, captures WiFi-CSI, sends via WiFi |
| Server | Local appliance (mini-PC) | Storage, analysis, notification logic, dashboard |
| Interface | Web dashboard | Per-resident status, roles, notifications |
One ESP32-S3 module works in each room. It combines Bluetooth 5 (more stable connection to wearables) with WiFi and enough memory for CSI buffering and simple analysis right at the edge of the network.
Proven and maintainable — no exotic special stack:
CSI uses the subtle changes in existing WiFi signals to detect movement in a room. We use it where it works reliably today.
Based among others on Espressif's ESP-CSI toolkit and publicly published CSI sensing research.
Data encrypted at rest and in transit (TLS on the local network).
Processing inside the facility. Cloud only optional and encrypted.
Personal identity kept separate from measurement data where feasible.
Separate access rights for care, physicians and administration.
CSI uses radio only — no image or audio recording.
Data processing agreements and impact assessment as part of rollout.
A fast, responsible market entry — with clear wording at every stage.
Notification system for care staff. No diagnosis, no therapy. This enables a fast pilot operation.
Quality management to ISO 13485, software lifecycle to IEC 62304, risk management to ISO 14971.
Conformity under the EU Medical Device Regulation as a later stage for selected functions.
Accompanying studies and data analysis on effectiveness, together with pilot partners.