A Sensitive and High-Accuracy Dual-Mode Wireless Sensor with MLP-Based Mutual Inductance Suppression for Ammonia Leak and Soil Environment Detection.

Real-time and accurate detection of hazardous gases such as ammonia (NH3) is essential for intelligent management in energy and chemical industries, agricultural soil monitoring, food freshness evaluation, and atmospheric pollution control. These deployment scenarios demand explosion-proofing, wire-free operation, compactness, cost-effectiveness, and self-temperature compensation, driving adoption of compact, multi-parameter wireless passive sensors. However, multi-mode passive sensors face persistent challenges with mutual inductance interference, which degrades measurement accuracy. This study develops a dual-mode inductor-capacitor sensor using a multilayer perceptron (MLP) for simultaneous NH3 and temperature measurement. Two interdigital electrodes connect to different inductor parts, enabling dual-mode sensing with reduced size/cost. Microstructure polydimethylsiloxane@graphene enhances temperature sensitivity to 205.75 kHz °C-1, tripling the non-microstructure value. Phenyl phosphonic acid modification of tungsten oxide boosts room-temperature NH3 response to 4.1%, a threefold improvement. The MLP model effectively mitigates mutual inductance interference, achieving decoupled NH3 and temperature measurements with mean squared errors of 0.73 and 0.18, respectively. The sensor demonstrates promising application for efficient NH3 leak detection in intelligent transmission line inspection and soil environment monitoring.