A low power ammonia sensor node embedded with a light weight non-linear analytics

Abstract

A wireless gas sensor node with nanosensor for rapid in-situ detection of ammonia gas is developed and an analytical model for precise and accurate quantification of ammonia in any industrial or closed environment has been proposed. A nanosensor is preferred due to its highly selective and sensitive properties for target gases. In this work, a traditional sensing circuit is replaced with a single sensor based voltage divider sensing circuit for signal feeding to microcontroller. As the room temperature operated chemiresistive ammonia sensor has eliminated the use of micro heater, overall power consumption is reduced. A low sensing power consumption of 56mW was observed and the performance of non-linear model based gas detection algorithm was analysed using Relative Prediction Error (RPE), Root Mean Square Error for Cross Validation (RMSECV) and percentage recovery (% recovery). Best performance metrics values (RPE=0.04%, RMSECV=6.128, % recovery=99.273) were obtained. In addition to this, the wireless sensor node is designed with power harvesting techniques to overcome the issues on node lifetime.