Publications by authors named "Kaiyan Peng"
Article Synopsis
- This study presents a new method called time-frequency attention convolutional neural network (TFA-CNN) to improve gas type recognition and concentration prediction in electronic nose (E-nose) systems.
- The TFA-CNN employs a specialized attention block that effectively combines temporal and frequency information from E-nose signals, enhancing gas classification and concentration accuracy.
- The results show high performance, with classification accuracy reaching 100% and low mean absolute error in predictions, demonstrating the model's robustness against sensor drift and redundant information.
This work has presented gas sensors based on indium tin oxide (ITO) for the detection of SO and NO. The ITO gas-sensing material was deposited by radio frequency (RF) magnetron sputtering. The properties of gas sensing could be improved by increasing the ratio of SnO.
View Article and Find Full Text PDF