Research / Machine smell
Research pillar 01

How machines detect chemical patterns in air.

Every environment carries an invisible chemical signature. Machine smell is the study of reading that signature — turning the reaction of a sensor array into a stable, recognizable pattern.

What we research

We do not try to identify every molecule one by one. We read the combined response of a low-cost sensor array as a fingerprint — a pattern that a model can learn, the way a person recognizes coffee without naming each compound in it.

Technical terms: volatile organic compounds (VOCs), gas sensing, electronic nose, sensor array.

Sensor array

Our electronic-nose array combines five commercially mature, low-power families. Together they produce 114 features per sample.

BME688Broad VOC pattern · 24 features
SGP40VOC index · 12 features
SHT40Humidity & temperature · 24 features
MiCS-6814Reducing / oxidizing gases · 36 features
PMS5003Aerosol / particle context · 18 features

Findings

Our analysis pipeline already separates clean air from post-use air and tells event families apart on a scenario-heldout split — while raising no false alarms on clean air. That baseline sets the bar the on-bench system has to clear.

0.90
clean vs post-use
1.00
IQOS vs vape
0.70
time bucket
0.00
false alarm

What we are figuring out next: whether these separations hold on real sensor hardware, across changing humidity and background air — the focus of the controlled chamber experiments now starting.

References

  • 01AERALYTE_RESEARCH_LAB_BRIEF.md — research pillars & technical direction.
  • 02PHASE1_CONTROLLED_ENOSE_EXPERIMENT.md — the controlled chamber experiment plan.
  • 03Kimi_Agent · Sensor Evolution & Power Trade-offs — sensor selection (BME688/690, SGP40, SHT40, MiCS-6814, PMS5003).
  • 04GitHub — XoAnonXo/aeralyte. Explore the research corpus.