Research on Malodor Component Identification Based on Sensor Array.

With the rising demand for improved living standards and environmental protection, malodor pollution has emerged as a critical concern for both the public and regulatory authorities. Accurate prediction of malodor gas composition is essential for effective environmental monitoring and safety management. However, existing online malodor detection systems often suffer from short-term sensor drift, compromising their accuracy and long-term stability. To address these challenges, this study proposes an advanced electronic nose (e-nose) detection framework based on a time series data analysis. This study presents a novel approach utilizing a multi-channel sensor array for gas sampling, which establishes a robust mapping relationship between sensor response patterns and gas concentration distributions. To address the challenges of sensor drift and enhance system stability, we propose an innovative Encoder-Decoder architecture IED-CNN-LSTM incorporating external compensation mechanisms. Experimental results demonstrate that the proposed IED-CNN-LSTM model outperforms conventional methods significantly in both prediction accuracy and long-term stability. The framework achieves enhanced feature extraction from sensor time series data, enabling more precise and reliable detection of malodorous compounds. This research contributes an effective solution for real-time environmental monitoring applications while offering substantial improvements in both performance metrics and practical implementation for industrial and regulatory scenarios.