Rapid Classification and Concentration Prediction of VOC Gaseous Mixtures Using Sensor Array With Multitask Learning

The sensor array is the core of the electronic nose (E-nose) system, while the pattern recognition algorithm is an important component of the system. Therefore, the sensor array optimization and the algorithm research are regarded as two key research fields. Volatile organic compounds (VOCs) mixture are ubiquitous in environmental monitoring, disease screening, industrial emissions, and natural environments. Due to the flammability and potential health impacts, rapid and accurate detection of VOCs is crucial. Currently, E-nose systems for gas detection build separate models and a random number of sensors for classification and concentration prediction, which leads to low efficiency and increased complexity. To address these issues, this study proposed a novel multitask learning (MTL) framework combined with the gated recurrent unit (GRU) and feature enhancement module (MTL-GRUA), performing both gas classification and concentration prediction tasks simultaneously. The model achieved over 99% accuracy in classification tasks for acetone, ethanol, and their mixture, and demonstrated high precision in concentration prediction tasks. By optimizing the hyperparameters of the network structure of MTL-GRUA by a particle swarm optimization (PSO) algorithm and using the first 100-s data instead of stable phase data, the detection efficiency was increased by 63%, achieving rapid detection. Furthermore, the experiment verified that even with an optimization in the number of sensors from eight to three, acceptable detection results could still be obtained. The application of MTL proposed in this work provides a more efficient method for gas detection tasks.