Case Study

Wearable IoT Gas Meter Development

An in-depth look at the design and prototyping of a full-stack IoT system for first-responder safety.

role

Lead Engineer (Hardware, Firmware & Cloud Design)

technologies

KiCad, nRF9151 (LTE-M), Zephyr RTOS, C/C++, MQTT, OCI, TimescaleDB, Docker, Node.js

1. The challenge

To conceptualize, design, and prototype a complete, end-to-end IoT system from scratch. The goal was a portable, low-power, multi-sensor gas and particle meter to enhance first-responder safety, capable of transmitting real-time data to a cloud platform.

2. My solution

I developed a full-stack prototype demonstrating broad capabilities across the entire IoT technology stack. The solution includes custom hardware design, embedded firmware, and a cloud backend for data visualization.

3. Technical development breakdown

This project spanned four primary technical areas:

3.1. Hardware design & prototyping

The hardware development cycle included:

Sensor selection (gas, LEL - lower explosive limit, and particulate matter) and the design of schematics and multi-layer PCBs using KiCad.
Integration of the core processing and communication component: a Nordic Semiconductor nRF9151 SiP (system-in-package), optimized for low-power cellular IoT (LTE-M and GNSS) applications.
Prototype assembly and initial hardware validation.

3.2. Embedded firmware development

Significant effort was dedicated to the firmware, developed in C/C++ on the Zephyr RTOS. The scope encompassed:

Interfacing with multiple complex sensors.
Managing device power states for low-power operation.
Implementing the LTE-M cellular communication stack, including module control via AT commands.
Handling data packetization using protocols like MQTT over TCP/IP.

3.3. Cloud backend & data processing

A complete cloud-based data pipeline was architected and deployed:

Ingestion: Configuring a HiveMQ MQTT broker to receive real-time data from the devices.
Storage: Utilizing a TimescaleDB time-series database, hosted on Oracle Cloud Infrastructure (OCI).
Processing: Developing server-side logic within the OCI environment to process, analyze, and potentially trigger alerts based on the incoming sensor readings.

3.4. Frontend (dashboard) & DevOps

Visualization: To make the data useful, a web-based dashboard and landing page were developed using Node.js. This interface consumed the APIs to plot sensor readings and device locations.
DevOps: Implemented DevOps practices using Git for version control and Docker for containerization. The backend services (HiveMQ, TimescaleDB, Node.js) were deployed and tested on Oracle Cloud Infrastructure (OCI) for greater scalability.