Abstract: Significance　Odorous gases, primarily emitted from unorganized sources within industrial parks, represent a significant category of atmospheric pollutants. These gases, often comprising complex mixtures like hydrogen sulfide (H2S), ammonia (NH3), methyl mercaptan (CH3SH), dimethyl sulfide (C2H6S), dimethyl disulfide (C2H6S2), and trimethylamine (C3H9N), pose substantial risks to human health and environmental quality due to their pungent smell and potential toxicity. To effectively control their emissions and scientifically evaluate the corresponding management strategies necessitates precise, real-time monitoring and accurate source apportionment of the emission fluxes and spatial distribution of these odorous gases. Traditional analytical instruments, although with high accuracy, are often limited by high cost, slow response, and lack of portability for dense spatial deployment. Consequently, miniature gas sensors have emerged as a promising solution, due to their advantages such as compact size, low cost, rapid response, and potential for integration into scalable networks. This review focuses on three pivotal types of miniature sensors, Photoionization Detection (PID), Electrochemical (ECS), and Metal Oxide Semiconductor (MOS) sensors, and evaluates their role in addressing the challenges of odor monitoring and tracing in industrial settings. Progress　This review systematically examines the working principles, structural characteristics, sensing materials, and key performance metrics of PID, ECS, and MOS sensors, as well as their practical applications and technological trajectories. PID sensors: Utilizing high-energy ultraviolet light to ionize gaseous molecules, PID sensors enable detection of volatile organic compounds (VOCs) with ionization energies below the UV photon energy (typically 9.6 eV, 10.6 eV or 11.7 eV). They are characterized by high sensitivity, fast response (within seconds), and broad-spectrum coverage, making them suitable for industrial emission monitoring and indoor air quality assessment. Recent advances have improved their portability and reliability in field applications, with calibrated systems demonstrating comparable performance to standard analytical instruments for tracking total VOCs trends. Current development focuses on miniaturization and integration for portable devices, enhancement of sensitivity and selectivity toward specific odorants, and the implementation of intelligent, networked systems for real-time data transmission and remote management. ECS sensors: Based on the principle of electrochemical reactions at electrode surfaces, ECS sensors are known for their high accuracy, low power consumption, and good selectivity for specific electroactive gases such as H2S, CO, and NH3. They are widely used in workplace safety monitoring and breath-based biomarker analysis. Ongoing research aims to enhance portability for integration into wearable devices, deepen integration with the Internet of Things (IoT) for smart monitoring systems, and explore novel materials such as conductive polymers and metal‑organic frameworks to improve sensitivity, selectivity and long‑term stability. MOS sensors: Operating on conductivity changes upon gas adsorption on metal‑oxide surfaces, MOS sensors provide high sensitivity, low cost and simple fabrication. They are employed in early warning systems for hazardous gases, ambient air quality monitoring, and breath analysis. Key challenges include humidity interference and cross‑sensitivity. Current efforts are directed toward achieving trace‑level detection (down to 10-9 volume fraction), enabling intelligent and networked operation via IoT integration, and realizing low‑power designs for sustained field deployment. Multi-sensor synergistic monitoring: Integrating PID, ECS and MOS sensors can overcome individual limitations by leveraging their complementary strengths—such as PID's high sensitivity, ECS's good selectivity, and MOS's low cost. Data‑fusion algorithms (e.g., partial least squares) play a crucial role in processing combined signals, effectively reducing cross‑sensitivity and improving overall accuracy. Deployed multi‑sensor systems have shown high correlation with human olfactory assessments and reliable long‑term performance in field applications. Remaining challenges include the complexity of data‑fusion algorithms, compatibility among different sensor types, and the relatively higher costs associated with system integration and maintenance. Conclusions and Prospects　PID, ECS, and MOS miniature sensors each offer distinct advantages and face specific limitations for odor monitoring. PID sensors provide high sensitivity and rapid response for VOCs but suffer from poor selectivity and higher cost. ECS sensors excel in accuracy and low power consumption for specific gases but have limited detection ranges. MOS sensors are highly sensitive and low-cost but are susceptible to humidity and cross-sensitivity. The ideal sensor would combine high sensitivity, strong selectivity, fast response, and robustness. Currently, key challenges include improving selectivity, mitigating cross-sensitivity, and enhancing long-term stability against environmental fluctuations, necessitating robust calibration and compensation mechanisms. The synergistic application of PID, ECS, and MOS sensors, coupled with advanced data fusion, presents a powerful strategy for comprehensive monitoring and accurate source apportionment of complex odorant mixtures in industrial parks.

Key words: odorous gas, photoionization detector, electrochemical sensor, metal oxide semiconductors, monitoring and tracing, multi-sensor fusion