Microsoft Builds AI Nose to Detect Smells: A Breakthrough in Smart Sensing

Microsoft Conducted an "Artificial Nose" Experiment to Revolutionize Scent Recognition.

Posted: Sep 18 2023

Updated: Nov 8 2023

Home Blog ASOWorld News Microsoft Builds AI Nose to Detect Smells: A Breakthrough in Smart Sensing

AI Nose Project "Artificial Nose"

In a remarkable DIY project, Benjamin Cabé, a program manager at Microsoft, has developed an innovative device known as the "Artificial Nose".
This smart device is designed to recognize a wide range of smells using a simple gas sensor and a microcontroller.

Cabé's motivation for creating this technology stemmed from a desire to enhance his breadmaking skills during the COVID-19 pandemic. He aimed to find a more precise way to determine when his sourdough bread starter was ready to bake.

The Concept

Cabé's idea revolved around using AI to accurately gauge the readiness of his bread dough. To achieve this, he envisioned building a device capable of collecting data on the gas composition of specific smells and training an AI model to identify them.

This concept marked the birth of the Artificial Nose.

How It Works

The Artificial Nose relies on a neural network to correlate the concentration of gases in the air to distinct smell categories.

Once connected to an IoT platform, this device can be applied in various scenarios, from real-time food spoilage alerts to the detection of specific gases in the atmosphere.

From Gas to Smell

The device's operation involves gas sensors measuring the concentration of gases emanating from a substance. The AI model then extracts key characteristics from these gases, enabling it to infer the corresponding smell.

The result is displayed to the user and can be sent to Azure IoT for further analysis or to contribute to model improvement.

Technical Details

Artificial Nose Architecture
(Source: Microsoft)

Building the Artificial Nose commences with basic hardware components: an off-the-shelf gas sensor and a microcontroller.

The sensors, with the ability to detect gases such as Carbon Monoxide (CO), Nitrogen Dioxide (NO2), Ethyl Alcohol (C2H5OH), and Volatile Organic Compounds (VOCs) in the surrounding environment, are connected to a microcontroller. This microcontroller is tasked with collecting gas sensor data and feeding it into the AI model.

The heart of the Artificial Nose lies in its neural network, which correlates gas concentrations with specific smell categories.

To train the model, gas sensor readings for each smell category were captured, and their "olfactory fingerprints" were extracted by analyzing factors like minimum, maximum, and average gas concentrations.

These characteristics were used as inputs to a neural network, which established correlations between olfactory fingerprints and smell categories.

Optional Features

The Artificial Nose can be customized further with a 3D nose enclosure to house the gas sensor and microcontroller. Leveraging the microcontroller's Wi-Fi module, it can become an IoT Plug and Play device, connecting to platforms like Azure IoT.

This enables real-time alerts and reporting, such as automatic notifications for gas leaks on a factory floor or alerts for spoiled food in a consumer's fridge.

Click "Learn More" to drive your apps & games business with the ASO World app promotion service now.

ASO Topic News,

All content, layout and frame code of all ASOWorld blog sections belong to the original content and technical team, all reproduction and references need to indicate the source and link in the obvious position, otherwise legal responsibility will be pursued.

Start Your Store Ranking Acquisition Capability Detection

Step by step to measure all of your app store optimization efforts with our ASO tools.