⚡ electrical engineering

Development and Validation of a PredictiveMathematical Model for Cognitive Performanceand Occupational Stress Based on Indoor AirQuality: A Low-Cost Implementation UsingEmbedded Systems

This study presents and validates a low-cost, ESP32-based embedded system that utilizes a novel mathematical model to predict cognitive performance and occupational stress levels in real-time based on indoor air quality parameters, demonstrating high accuracy and practical utility for preventive occupational health management.

Original authors: Lucas Lima Freitag

Published 2026-07-01

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Lucas Lima Freitag

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine your brain is a high-performance race car engine. It needs clean air and the right temperature to run smoothly. If you fill the engine with dirty air or let it overheat, the car slows down, sputters, and eventually breaks down.

This research paper is about building a low-cost "dashboard" for your brain that lives inside your office. Instead of just telling you the temperature or how much smoke is in the air, this device predicts exactly how those conditions are making your brain feel and how well you can think.

Here is the breakdown of how it works, using simple analogies:

1. The Problem: The "Stuffy Room" Effect

We spend most of our time indoors. When a room gets crowded, the air gets "stale." Two main things happen:

Too much CO2: You and your coworkers breathe out carbon dioxide. If the room is small and closed, it builds up like a thick fog.
Too many VOCs: These are invisible gases from furniture, cleaners, and electronics. Think of them as invisible "smog" that irritates your brain.

When these build up, people start feeling tired, irritable, and unable to focus. Usually, we don't realize this is happening until we feel terrible. This study wanted to catch it before you feel bad.

2. The Solution: A "Brain Weather Station"

The researchers built a tiny, cheap computer (costing less than $30) that acts like a weather station, but for your brain's performance.

The Hardware: They used a small chip called an ESP32 (the brain of the device) connected to two sensors:
- One sensor smells the air (detecting CO2 and VOCs).
- One sensor feels the air (temperature and humidity).
The "Magic" Math: The device doesn't just show numbers. It runs a special mathematical recipe. It takes the air data and asks: "If the air is this bad, how much is the brain slowing down?"

3. How the Math Works (The "Degradation Index")

The researchers created a formula that acts like a stress thermometer.

The Inputs: It looks at how much CO2 is in the room, how hot it is, and how hard your brain is working (e.g., are you just sitting there, or are you in a high-stress crisis?).
The Calculation: It calculates a "Degradation Index." Imagine this as a score of how much your brain is "rusting" due to bad air.
The Output: It turns that rust score into two easy numbers:
1. Cognitive Efficiency: How sharp are you? (e.g., 98% sharp vs. 62% sharp).
2. Stress Level: How stressed is your body?

4. The "Crystal Ball" Feature

The coolest part of this device is that it doesn't just look at the now; it looks at the future.

The Scenario: Imagine it's 2:00 PM. The device sees the air getting bad. It calculates that if nothing changes, by 6:00 PM, your brain efficiency will drop to a critical low, and you'll be extremely stressed.
The Alert: It sends a warning to the manager: "Alert! In 4 hours, the team will be in 'Severe Stress' territory. Open a window now!"
The Result: In their test, the device predicted that by 2:00 PM, the team's thinking ability would drop from 98% (very sharp) to 62% (very foggy). When they opened the windows, the numbers went back up.

5. Did It Work?

The researchers tested their "recipe" against four different scientific studies from around the world.

The Score: Their predictions were incredibly accurate, with an error margin of less than 0.2%. It's like guessing the weight of a person and being off by less than a single paperclip.
The Cost: Traditional air quality machines cost thousands of dollars. This one costs about $30 (roughly the price of a nice lunch).

Summary

This paper presents a cheap, smart device that translates invisible air pollution into a clear report on how well your brain is working. It acts like a "check engine light" for your office environment, telling you exactly when the air is about to make your team tired and stressed, so you can fix it before the damage is done.

What the paper does NOT claim:

It does not claim to cure diseases.
It does not claim to replace doctors.
It does not claim to work for every single person individually (it uses an average "brain" model).
It does not claim to measure dust or smoke particles (it focuses on gases like CO2 and VOCs).

It is strictly a tool for predicting how air quality affects mental sharpness and stress levels in a workplace.

Technical Summary: Development and Validation of a Predictive Mathematical Model for Cognitive Performance and Occupational Stress Based on Indoor Air Quality

Problem Statement
Indoor air quality (IAQ) is a critical determinant of cognitive performance and psychological well-being in occupational settings. While poor IAQ is linked to anxiety, fatigue, and reduced decision-making capabilities, current tools fail to bridge the gap between real-time environmental sensing and actionable psychological impact metrics. Existing solutions fall into three limited categories: epidemiological studies that offer statistical associations but no real-time prediction; sensor networks that monitor air parameters without quantifying human impact; and controlled chamber experiments that are not deployable in operational environments. Furthermore, conventional monitoring stations are prohibitively expensive (thousands of dollars), limiting widespread adoption.

Methodology
The study proposes a low-cost, embedded system solution that integrates a calibrated mathematical model with hardware sensors to predict cognitive efficiency and occupational stress in real-time.

Mathematical Framework: The core of the system is a composite degradation index ( $I_{deg}$ ) derived from three layers:
1. Mass Balance & Thermodynamics: Models contaminant dynamics using a continuity differential equation, incorporating room volume, ventilation rates, occupant count, and thermal/humidity correction factors ( $f_T$ , $f_H$ ).
2. Biological Absorption: Calculates cumulative pollutant doses for CO2 ( $A_{CO2}$ ) and VOCs ( $A_{VOC}$ ) based on respiratory minute volume and available air volume per individual.
3. Psychological Transfer: Maps the degradation index to cognitive and stress outcomes using sigmoid logistic functions. Key variables include the Oxygen Displacement Hypoxia Factor ( $f_{O2}$ ), Hyperventilation Metabolic Rate ( $M_{met}$ ), and Organic Clearance Constant ( $\lambda_{wash}$ ).
- The model outputs Cognitive Efficiency ( $P_{con}$ ) and Psychophysiological Stress ( $I_{estresse}$ ), with $P_{con}$ ranging from 0% to 100%.
Hardware Implementation: The system is built on the ESP32 microcontroller, utilizing:
- ENS160 Sensor: For equivalent CO2 (eCO2), Total VOCs (TVOC), and Air Quality Index (AQI).
- BME280 Sensor: For temperature, humidity, and barometric pressure.
- Connectivity: The ESP32 operates as a standalone Wi-Fi access point, hosting a web server that serves real-time environmental reports and JSON data without requiring external internet or cloud subscriptions.
- Cost: The total component cost is estimated between USD 22 and USD 35.
Experimental Protocol: The system was deployed in a 120 m³ office with 10 occupants over an 8-hour workday. Data was recorded every 5 seconds, with the firmware performing numerical integration of doses and extrapolating 4-hour predictions to alert managers of impending critical stress levels.

Key Results

Model Validation: The model was cross-validated against four independent international studies (Satish et al., Allen et al., Cedeño Laurent et al., and Künn et al.). The results demonstrated a Root Mean Square Error (RMSE) of 0.18 percentage points and a maximum absolute error of 1.5 percentage points, confirming high predictive accuracy across heterogeneous datasets.
Operational Performance: In the field deployment, the system successfully tracked cognitive degradation. At 08:00, cognitive efficiency was 98.2% with 1.8% stress. By 14:00, as CO2 rose to 1,680 ppm and TVOC to 210 ppb, efficiency dropped to 62.4% and stress rose to 37.6%.
Predictive Capability: The system generated 4-hour projections that correctly identified a trajectory toward critical stress levels (projected 61.8% stress), triggering a "SEVERE ALERT" recommendation for forced ventilation.
Sensitivity Analysis: CO2 concentration was found to contribute 62% of the variance in cognitive efficiency, followed by exposure duration (22%). The model proved robust to ±10% variations in calibration constants.

Significance and Claims
The paper claims to provide a practical, accessible tool for preventive occupational health management by transforming IAQ monitoring from descriptive logging into prescriptive management. Its primary contributions are:

Integration: It uniquely combines real-time sensing, a physiological/mathematical model, and predictive analytics into a single deployable device.
Accessibility: By reducing the cost of monitoring from thousands of dollars to under USD 30, it enables distributed monitoring for Small and Medium Enterprises (SMEs) and portable applications.
Preventive Action: The system's ability to project future stress levels allows managers to intervene before psychological damage consolidates, addressing the gap between environmental data and human impact.

The authors note limitations, including the use of MOS-based eCO2 sensors (which may have uncertainty in low VOC-CO2 correlation environments) and the assumption of uniform pollutant distribution. Future work is suggested to include particulate matter monitoring, multi-sensor coverage, and longitudinal clinical validation.