This paper presents the results on developing an ensemble machine learning model to combine commercial gas sensors for accurate concentration detection. Commercial gas sensors have the low-cost advantage and become key components of IoT devices in atmospheric condition monitoring. However, their native coarse resolution and poor selectivity limit their performance. Thus, we adopted recurrent neural network (RNN) models to extract the time-series concentration data characteristics and improve the detection accuracy. Firstly, four types of RNN models, LSTM and GRU, Bi-LSTM, and Bi-GRU, were optimized to define the best-performance single weak models for CO, O₃, and NO₂ gases, respectively. Next, ensemble models which integrate multiple single weak models with a dynamic model were defined and trained. The testing results show that the ensemble models perform better than the single weak models. Further, a retraining procedure was proposed to make the ensemble model more flexible to adapt to environmental conditions. The significantly improved determination coefficients show that the retraining helps the ensemble models maintain long-term stable sensing performance in an atmospheric environment. The result can serve as an essential reference for the applications of IoT devices with commercial gas sensors in environment condition monitoring.

To improve the weak piezoelectric excitation of lateral field excited (LFE) acoustic wave sensors in air, a novel bulk acoustic wave device, named pseudo LFE acoustic wave sensor, is proposed in this paper. The commercially available finite element software, COMSOL Multiphysics, is adopted to analyze pseudo LFE acoustic wave gas sensors and further calculate their sensitivities to electric conductivity variation of the polyisobutylene (PIB) sensing film. An AT-cut quartz plate with a thickness of 185 ?m is chosen as the piezoelectric substrate due to its tiny temperature coefficient to frequency for thickness shear mode (TSM). After the mesh convergence analysis, a mesh level is decided for following calculations. In frequency response analysis, a bare pseudo LFE sensor shows a pure TSM and a stronger excitation than a bare LFE sensor. The experiments also verify this result. In sensitivity analysis, a gradual conductivity increase of the sensing film from 0 to 3 S/m causes a significant resonance frequency decrease of the pseudo LFE sensor. Besides, the pseudo LFE sensor exhibits an excellent sensitivity to the film conductivity variation, about 30 times as high as the QCM sensor. Finally, the electrode dimensions of pseudo LFE sensors are discussed for optimizing device performances. One can find ringed floating electrodes exhibit greater sensitivity than circular ones. Moreover, a larger gap width between the LFE electrodes also contributes to larger sensitivity but weakens piezoelectric excitation. The results conclude that the pseudo LFE sensor exhibits a satisfactorily large admittance and an excellent sensitivity to external electrical fluctuation.

The control of humidity is required in various areas such as the smart living space, industry, agriculture, and medicine. To reduce cost and increase lifetime, a wireless humidity sensor without an additional power supply should be developed to fulfill their needs. Recently, there are more and more efforts on surface acoustic wave (SAW) based radio frequency identification (RFID) system. The RFID system primarily consists of a reader and transponders, composed of SAW tags and antennas. Due to the capabilities of passive operation and wireless connection, SAW tags are also suitable for remote sensing. Therefore, in this study, a wireless humidity sensor is accomplished by integrating a 433MHz 128°YX-LiNbO3 SAW tag with a resistive humidity sensor. The SAW tag is designed as an impedance loaded sensor. Moreover, because nanostructure sensing materials possess high surface-to-volume ratio, large penetration depth, and fast charge diffusion rate, CSA-doped polyaniline nanofibers are synthesized by the interfacial polymerization method and further deposited on the resistive humidity sensor as sensitive film to enhance the sensitivity. Finally, the humidity sensor was constructed and measured. Results indicate that our proposed sensor exhibits not only good linearity but also high sensitivity. Furthermore, it is indeed capable of passive remote sensing.