Electronic nose prototype

Client project: development $12,460, parts $436 per unit. An electronic nose that tells smells apart, for example fruit and liquid samples, using an existing MEMS gas sensor array with six sensor types. The ESP32-S3 main controller reads the array, filters the signals and compensates for temperature and humidity; a pattern recognition algorithm then separates the samples, shown here as a radar plot of the sensor channels and a principal-component plot. We did the component selection, the mechanical and enclosure design, the PCB, the embedded firmware and the pattern recognition algorithm.

  • Client type: Client (not named)
  • Units: 8 units delivered
  • Development fee: $12,460
  • Parts per unit: $436 per unit
  • Stage: Client project, delivered
Original case sheet: Electronic nose prototype
Original case sheet

Overview and tech stack

Project type
Scientific & Research Equipment
Sensors
Existing MEMS gas sensor array, 6 sensor types
Controller
ESP32-S3 main controller
Signal processing
Signal filtering, temperature and humidity compensation
Algorithm
Pattern recognition (radar and principal-component plots)

What we built

  • Component selection
  • Mechanical and enclosure design
  • PCB design
  • Embedded device firmware
  • Pattern recognition algorithm

What this tech stack can also build

  • Electronic nose for food freshness
  • Gas sensor array device
  • Odor detection instrument for a lab
  • VOC monitoring device
  • Air quality monitor with multiple gas sensors
  • Smell classification with machine learning on an ESP32
  • Spoilage detector for fruit and produce
  • Fragrance and flavour test rig
  • Environmental sensing node with temperature and humidity compensation
  • Agricultural sensor for storage rooms
  • Research prototype for sensor-array data collection
  • Breath or headspace sampling device for research

Project scope

Development time
4 weeks
Development fee
USD 12,460
BOM + manufacturing
USD 436 per unit
Quantity basis
8 units delivered

Client project. A quote for your build follows a review of your files.

What to confirm before we quote

  • Target samples: which smells, fruits or liquids it must tell apart, and how many classes
  • Sensor array: the six sensor types needed, and whether you supply or approve them
  • Sampling method: how the sample is presented, chamber design and cleaning between samples
  • Environment: temperature and humidity range where it will be used
  • Algorithm and data: how you supply reference samples, and the output you want such as a label, score or plot

Start a similar project

  1. Send what you have: sketches, a reference product, files or a written idea.
  2. We review it and list the gaps and open questions.
  3. You get a written scope and quote.

All builds →

Building something similar?

Get a free project review

Or book a 30-minute call