Related Experiment Video
Updated: Aug 5, 2026
Façade-Level Monitoring of CO2 Variability under Urban Heat Island Conditions using Low-Cost Sensor Data Loggers
Published on: December 12, 2025
Machine Learning-Enhanced High-Temperature CO2 Sensor for In Situ Exhaust Monitoring: Robust Compensation via
Quanquan Tang1, Zuorong Huang1, Bin Wu1
1State Key Laboratory of Integrated Optoelectronics, JLU Region, Key Laboratory of Gas Sensors, College of Electronic Science and Engineering, Jilin University, 2699 Qianjin Street, Changchun, Jilin Province 130012, China.
A new solid electrolyte sensor combined with random forest machine learning accurately detects automotive carbon dioxide (CO2) emissions. This system overcomes challenges like sensor drift and interference, enabling precise real-time carbon footprint tracking.
Area of Science:
- Electrochemistry
- Environmental Science
- Machine Learning
Background:
- High-temperature CO2 detection in automotive exhaust is challenging due to harsh conditions and sensor cross-sensitivity.
- Existing solid electrolyte sensors struggle with thermal drift and interference from other gases.
Purpose of the Study:
- To develop a robust sensing platform for high-fidelity carbon emission monitoring in automotive exhaust.
- To overcome the limitations of traditional solid electrolyte sensors using machine learning.
Main Methods:
- Integration of a NASICON-based solid electrolyte sensor with a random forest (RF) machine learning (ML) architecture.
- Utilized surface-optimized catalysts, stabilizers, and NASICON doping for enhanced sensor performance.
- Employed an automotive-grade data acquisition system for in situ signal capture.
- Developed a software-defined compensation strategy using vehicle operating parameters to decouple interfering factors from CO2 response.
Main Results:
- The ML-enhanced sensor achieved a coefficient of determination (R2) exceeding 0.8 in non-linear transient states.
- Cumulative mass emission error was suppressed to less than 0.4%, demonstrating metrological-grade accuracy.
- Validated sensor reliability across standardized transient cycles (WLTC, WHTC, RDE).
Conclusions:
- The synergistic approach of solid-state electrochemistry and intelligent data analytics provides a scalable, cost-effective solution for real-time carbon footprint tracking.
- This technology directly supports global initiatives for transportation decarbonization and environmental compliance.
- The developed platform offers a robust method for accurate CO2 detection in challenging automotive exhaust environments.