IoT Hospital Ward Bed Occupancy and Patient-Call Monitor with Dashboard

In a busy ward, nurses learn about a waiting patient only when someone walks the corridor — call bells exist, but most demo-grade nurse-call setups are just a buzzer with no record of which bed called, when, or how long the response took. That missing record is exactly what a student build can demonstrate: per-bed occupancy sensing, addressed call buttons, response-time logging and a live ward map. This project implements it with ESP32 bed nodes at demo scale: a pressure mat under each demo bed senses occupancy, a call button with an acknowledgement LED lets the "patient" call, and nodes report over ESP-NOW to a central gateway that drives the nurse-station dashboard. The dashboard shows the ward map with live bed states, running response timers for active calls and occupancy statistics. The build is explicit about its boundary: it demonstrates ward telemetry architecture and is not a medical device, carries no patient data and makes no clinical claims.

1. Each bed node continuously reads its pressure mat for occupancy and watches its call button; state changes are debounced in firmware.
2. On a call press, the node lights its local LED, starts its local timer and transmits a call packet (bed ID, timestamp) to the gateway over ESP-NOW.
3. The gateway receives the packet, sounds the nurse-station buzzer, lights the tower LED and forwards the event over Wi-Fi to the dashboard.
4. The dashboard marks the bed tile as calling and starts the visible response timer; occupancy tiles update live as mats change state.
5. The nurse acknowledges the call from the dashboard or gateway button; the acknowledgement packet returns to the bed node and clears its LED.
6. All events — occupancy changes, calls, acknowledgements — are timestamped in the dashboard log, feeding the response-time statistics.

Tech stack:

• ESP32 bed nodes + central gateway (ESP-NOW + Wi-Fi)
• Pressure mats (occupancy) per bed
• Call buttons with acknowledgement LEDs
• Buzzer + tower LED + OLED (nurse station)
• MQTT/HTTP telemetry to dashboard
• HTML/CSS/JS ward dashboard
• Arduino IDE (C/C++ firmware)

• [Per-bed occupancy sensing] Pressure mats under each demo bed report occupied/vacant state, so the ward map always reflects the true bed picture.
• [Addressed patient-call buttons] Each bed has a call button with a local LED; pressing it raises a bed-specific call on the dashboard and starts a response timer.
• [Response-time logging] Every call records raised-at and acknowledged-at timestamps, and the dashboard reports average and longest response times per shift.
• [Live ward map dashboard] A ward layout with per-bed status tiles (vacant, occupied, calling), running call timers and occupancy statistics in one view.
• [ESP-NOW bed network] Bed nodes report to the gateway over ESP-NOW — no Wi-Fi credentials per node, low latency and simple pairing.
• [Nurse-station alert] The gateway sounds a buzzer and lights a tower LED on any new call, with the calling bed shown on its local OLED.
• [Call acknowledgement flow] The nurse acknowledges from the dashboard (or a gateway button), which clears the bed LED and stops the timer — demonstrating a complete event lifecycle.

• Working hardware prototype (4 demo bed nodes with mats and call buttons + gateway station)
• ESP32 firmware for bed nodes (debounce, ESP-NOW) and gateway (alerts, forwarding)
• Companion dashboard demo (single-file HTML ward dashboard)
• Ward layout configuration guide (mapping bed IDs to the demo floor plan)
• Wiring diagram and node enclosure notes
• Project report PDF (background, ward telemetry concepts, architecture, methodology)
• PPT presentation for final review
• Viva Q&A preparation document (ESP-NOW, debouncing, event lifecycles, non-medical scope)

• This is a teaching prototype demonstrating ward telemetry — it is not a medical device, carries no patient data and must never be presented as clinical equipment.
• Pressure mats sense presence, not identity or posture; a bag on the bed reads as occupied, which the report documents honestly.
• ESP-NOW range suits a demo ward (tens of metres); a real hospital floor would need planned infrastructure, which is outside the build.
• Response-time statistics measure the demo workflow only; they say nothing about care quality.
• The demo uses 4 beds; scaling to a full ward needs addressing, power planning and enclosure work listed as future scope.
• Call buttons are wired to the bed node — wireless patient pendants are a possible extension, not part of this build.