Advances of semiconductor gas sensor on multi-parameter sensing and features extraction methods

Semiconductor gas sensors have undergone significant advancements in recent years, driven by the growing demand for precise and reliable gas detection in complex environments. This review systematically examines the latest progress in three critical areas: multi-parameter sensing, drift compensation, and feature extraction methodologies. First, we analyze multi-parameter data acquisition techniques, including sensor arrays, temperature modulation, and optical modulation, which have been shown to collectively enhance discrimination accuracy to over 98% for complex gas mixtures. Second, we summarize drift compensation strategies spanning traditional calibration methods to advanced machine learning algorithms, noting that deep learning approaches can maintain classification accuracy above 90% under drift conditions, thus ensuring the environmental adaptability of the sensor. Third, we critically evaluate feature extraction approaches, from conventional time-domain analysis to deep learning-driven methods, emphasizing how to achieve excellent gas recognition performance through feature extraction techniques. Finally, we outline future research directions, including the development of hybrid sensing platforms, adaptive drift-correction frameworks, and explainable algorithm models, to address emerging challenges in industrial, medical, and environmental monitoring applications.