1. Work
  2. About
  3. Experience
  4. Highlights
  5. Contact
View résumé

All work

Case study · 04 · IoT safety system

SafeSense: a fire alarm designed to reach the person who can act.

An ESP8266‑based fire and smoke detection system that combines a local alarm with GSM and Wi‑Fi notifications. I led the four‑person team that designed, built and presented the working prototype.

Category
IoT, embedded systems, safety
Stack
ESP8266 · NodeMCU · Arduino IDE · GSM · Wi‑Fi · Flame sensor · Smoke sensor
Timeline
2023
Recognition
CIIA finalist
Creative Ideas & Innovations in Action
Aditya presenting the SafeSense scale-model home at the CIIA exhibition

Overview

A conventional alarm is only useful if somebody can hear it. That is no help if you are at work, asleep two rooms away, or the building is empty. SafeSense detects a potential fire locally and reaches registered contacts remotely, including over GSM when there is no internet.

We installed flame and smoke sensors across a scale model of a home. When readings crossed calibrated thresholds, the NodeMCU raised a local alarm and triggered both remote paths at once. The zoned demonstration model let judges trigger detection live, and the project was selected as a finalist at the Creative Ideas & Innovations in Action exhibition.

How it works

  • Sensing. An MQ‑series smoke sensor and IR flame sensors feed analog and digital signals into the ESP8266.
  • Thresholding. Readings are filtered and compared against calibrated thresholds, so cooking smoke does not raise a false alarm.
  • Dual alert channels. A GSM module sends SMS to registered numbers while Wi‑Fi pushes a parallel notification. Neither path depends on the other.
  • Local response. A buzzer and indicator LEDs give an immediate on‑site warning.
  • Zoned demo rig. A scale model home with separate sensor zones made the behaviour testable in front of judges.

Impact

3
alert paths: local, GSM, Wi-Fi
4
person team, led from build to demo
Finalist
CIIA exhibition

What I learned

Real-world systems live on redundancy, calibration and clear failure behaviour. A safety system cannot assume one network path will always be available, which is why GSM and Wi-Fi fire in parallel rather than as a fallback chain.

Leading four people through a build and a live demo also taught me to design for the worst demo conditions, not the best ones.