Enhanced dual sensing capabilities of TiO₂:ZnO metal oxide sensors with machine learning approach

This study explores the development of dual-sensing technology for simultaneous gas and acceleration monitoring using TiO₂ and ZnO-based sensors with varying molar ratios. The findings indicate that the 1T:3Z (1TiO₂:3ZnO) ratio sensor demonstrates exceptional performance. For acceleration sensing, the 1T:3Z sensor achieved a sensitivity of 3.28 V/g, a maximum voltage output of 2.38 V at a 10 Hz resonant frequency, and an output voltage of 2.32 V at 1 g acceleration. In gas sensing, the same sensor exhibited a rapid response of 8.33% for CO and 20.92% for CH₄ at a 10-ppm concentration, with response times of 1.03 s and 4.01 s, and recovery times of 28.94 s and 49.06 s, respectively. The study achieved a selectivity accuracy of 95.6% between CO and CH₄ gases using the linear discriminant analysis (LDA) classification algorithm, outperforming traditional methods and other algorithms (RF, SVM and KNN). This dual-sensing capability demonstrates the potential of mixed metal oxide sensors for efficient and precise environmental monitoring. These advanced dual-sensing sensors hold significant promise for underground mining operations, where accurate monitoring of gas concentrations and vibrations is essential for worker safety and process optimization. Although achieving high selectivity for CH₄ remains a challenge, future improvements in sensor design, calibration, or data processing could further enhance CH₄ specificity.