TÜBİTAK 1501Completed

Greenhouse Gas Emission Monitoring and Air Quality Index Project

Air quality monitoring is often done with expensive fixed stations or single-parameter point sensors. Under this TÜBİTAK 1501-funded R&D project, HEFA Teknoloji A.Ş. developed a LoRaWAN-connected multi-parameter air quality monitoring device and web interface capable of simultaneously monitoring NO₂, SO₂, CO, CO₂, CH₄, O₃, PM₁₀, PM₂.₅ and meteorological parameters. The project output is an end-to-end system comprising hardware, embedded firmware, cloud infrastructure and an AI-assisted air quality index prediction model.

What We Built

Multi-Parameter Sensor Infrastructure

  • Electrochemical sensors: Alphasense 4-electrode electrochemical sensors providing ppb-level sensitivity for NO₂, SO₂, CO and O₃. Cross-sensitivity, linearity and temperature tests were performed and regression functions developed.
  • NDIR sensor: NDIR sensor with optical measurement principle for CO₂. Long-term stability and high accuracy were the selection rationale.
  • PM sensor: Optical particle counter module based on laser scattering for PM₁₀ and PM₂.₅.
  • Meteorology module: temperature, humidity, wind speed and direction, pressure measurement.

PCB Design — Mixed Signal Layout

  • Mixed signal layout rules were rigorously applied for co-locating analog sensor outputs, high-resolution ADC and digital communication blocks on the same board. Analog and digital sections are separated; internal layers in the multi-layer PCB were selected as solid GND planes.

Communication Infrastructure

  • LoRaWAN Class A selected as primary communication protocol. Data is transferred to the cloud via The Things Network (TTN) integration.
  • GSM module: backup communication for standalone operation where LoRaWAN gateway infrastructure is unavailable.

Web-Based Monitoring Interface

  • Dashboard supporting per-parameter graphical views, historical data queries and threshold breach notifications.

AI-Assisted Air Quality Index Prediction

  • Model taking instantaneous sensor measurements, meteorological parameters and historical data patterns as input to produce air quality forecasts.

Technical Specifications

ParameterValue
Monitored gasesNO₂, SO₂, CO, CO₂, CH₄, O₃
Particulate measurementPM₁₀, PM₂.₅
MeteorologyTemperature, humidity, pressure
Gas sensor technologiesElectrochemical (NO₂, SO₂, CO, O₃), NDIR (CO₂)
Primary communicationLoRaWAN Class A
Backup communicationGSM
Network infrastructureThe Things Network (TTN)
PCB designMixed signal layout
Data visualisationWeb-based dashboard
AI componentAir quality index prediction model
Project supportTÜBİTAK 1501

Use Cases

  • Urban Air Quality Monitoring Networks — Municipalities, environmental agencies, low-cost distributed network via LoRaWAN
  • Industrial Site Monitoring — Factory perimeter, organised industrial zones, port emission monitoring
  • Research and Measurement Campaigns — Universities, research institutions

Frequently Asked Questions

Why was LoRaWAN Class A chosen over Class C?
Class C continuously listens for downlinks, consuming high power and making it unsuitable for battery-powered end nodes. For applications like air quality monitoring nodes that send periodic data, Class A is the most power-efficient LoRaWAN operating mode.
Is TTN integration suitable for commercial applications?
TTN is excellent for rapid onboarding in project and research phases. For commercial-scale deployments requiring SLAs, transitioning to a dedicated LoRaWAN Network Server (ChirpStack, etc.) is recommended; the system is compatible with this transition.
Why was NDIR instead of electrochemical sensor used for CO₂?
CO₂ is chemically inert and therefore unsuitable for electrochemical measurement. NDIR measurement principle directly measures the infrared absorption characteristic of CO₂ molecules, which is the industry standard for CO₂ measurement in terms of high accuracy and long-term stability.
Does the AI model run in real time?
The model produces predictions from current sensor data and meteorological parameters from the nodes. Prediction outputs can be viewed on the dashboard.
How complex is it to scale the system?
LoRaWAN architecture is well-suited for scaling; adding nodes within an existing gateway's coverage area requires no additional infrastructure beyond software configuration.

This project encompasses a broad R&D process covering multi-parameter sensor integration, mixed signal PCB design, LoRaWAN/GSM communication infrastructure, cloud-connected data platform and AI model development. If you need environmental monitoring, emission tracking or multi-parameter sensor system development, we can evaluate your technical requirements together. end-to-end sensor product development service for more information.

Contact Us