Enhanced Selectivity Electronic Nose Systems for Agricultural Ammonia Gas Detection via a co-designed WO3-ZnO Sensor Array and Convolutional Neural Networks
Bibliographic record
Abstract
Electronic noses (e-noses) offer a practical solution for real-time monitoring of ammonia (NH 3 ) in agricultural environments, where NH 3 often coexists with interfering gases such as CO 2 , CH 4 , and H 2 S. However, semiconductor-based gas sensors commonly used in e-nose systems suffer from inherent cross-sensitivity, which reduces measurement accuracy. This study investigates the cross-sensitivity of NH 3 detection and introduces a mitigation strategy through convolutional neural networks (CNNs) for sensor data fusion. Experimental results show that WO 2 -based sensors exhibit strong NH 3 selectivity, with response ratios of 7.3:1 against CH 4 and 17.8:1 against H 2 S. Density functional theory (DFT) analysis confirmed that the WO 3 sensor exhibited strongest NH 3 binding energy (− 1.45 eV), compared to SnO 2 (− 1.10 eV), explaining the observed selectivity. Measurement uncertainties (± 8%) were quantified under varying humidity (30–90% RH) and temperature (10–40 °C) using a weighted least squares error propagation model. A quasi-2D sensor array improved NH 3 classification accuracy to 96.4% (7.2% increase) while reducing concentration errors by 50.8%, as validated by linear discriminant analysis. Long-term stability tests demonstrated that SnO 2 sensors maintained a low baseline drift of 0.18%/day over 180 days, outperforming CH 4 (0.31%/day) and ZnO (0.42%/day) sensors. Furthermore, the CNN model, trained on multi-sensor time-series data, achieved 91.7% accuracy in mixed-gas environments by capturing non-linear response patterns, ensuring reliable NH 3 quantification despite interferents. These findings highlight the promise of CNN-enhanced e-nose systems for precise NH 3 monitoring in complex agricultural settings, addressing key challenges of cross-sensitivity and environmental stability
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".