Smell SensorsPrototype

Visualization of Smell

We are currently researching and developing Smell Sensors: a QCM-type sensor that identifies odors at high concentrations and a MEMS semiconductor-type sensor that detects abnormalities from changes in conditions at low concentrations.

Advantages & Features

Compact and Multifunctional for a Wide Range of Applications

Various smells can be identified by mounting a sensor that can respond to a variety of different smell molecules.
The miniaturized sensor module can be used for a variety of applications, from fixed sensors to mobile applications.

Anomaly detection: Reports abnormalities even in unmanned environments, Inspection robots: Finds problem areas by smell Quality control: Visualizes freshness and spoilage, Exhaled air (pre-disease detection): Manages physical condition by smell, Visualization of smells: How people perceive smells visualize

Smell Identification by Machine Learning

Similarly for human smell identification, the system learns in advance what a certain smell is.
The system remembers the learned results to identify and recognize the smell from memory when it senses the smell.
In machine learning, we have developed an algorithm with good identification accuracy to visualize smells.

Measurement system (reproducibility of measurements): Development of measurement methods with good reproducibility / Machine learning (odor identification): Development of algorithms with good identification accuracy /  Transducer (sensor sensitivity): High sensitivity sensor module development / Sensitive membrane (odor selection system): Development of a sensitive membrane that can detect various odors

Smell Sensor System Structure

This system is composed of a sensing unit and a PC that controls the sensor.
There are two types of connections between the Smell Sensors and the PC:
wired (USB cable) and wireless (Bluetooth or proprietary 2.4GHz communication).
As an option, it can be synchronized with environmental sensors, enabling simultaneous
sensing with temperature/humidity sensors, CO₂ sensors, and dust sensors.
By creating a learning model from the measurement data and applying it to a microcontroller,
it is also possible to output judgment results instead of raw measurement values.

Related Videos

Download Documents

Product Trend / Advantages / Characteristics

Odor recognition with multi-sensor

Can we see smells?

Technology of smell visualization is under development! Smell analysis technology accelerated by AI

Conveying smells through a TV set?!

Feasibility of smell generation technology that will change common sense

Product and Solution Inquiries