Fabrication and optimization of passive flexible ammonia sensor for aquatic supply chain monitoring based on adaptive parameter adjustment artificial neural network (APA-ANN)

Abstract

Traditional rigid sensors and imperfect calibration methods are difficult to meet the demand of food quality and safety monitoring in aquatic products supply chain, which requires redesign and optimization. This paper designed and optimized a passive flexible ammonia sensor based on adaptive parameter adjustment artificial neural network (APA-ANN) for aquatic monitoring. The proposed sensor could acquire ammonia information and transmit to the wireless reader by radio frequency identification (RFID) with 13.56 MHz. Through the energy harvesting module, which has a harvesting efficiency of about 63.07%, the passive flexible ammonia sensor could harvest RF energy and supply continuous power for the normal operation of sensor without battery. Oysters and abalone were selected for the sensor output calibration and verification in waterless and watery environments, respectively. The sensor output had a good fitting effect in the sensing range (R2 > 0.99932) and the APA-ANN could improve and optimize the sensor output accuracy under the multiple interference and the impact of cross-sensitivity. The final result showed that all regression coefficients were>0.997 and the accuracy of the optimization model was higher than 89.5%, which proved the applicability of the passive flexible ammonia sensor and provided a optimization method for sustainable monitoring in agriculture.