The invention discloses an
exhaust gas sensor current detection method based on
machine learning, and relates to the technical field of
exhaust gas sensors. The method aims at solving the problems that an existing diesel vehicle
waste gas sensor is large in current measurement error and poor in batch consistency due to the
ceramic substrate insulativity, materials, interdigital
electrode distance difference and temperature influences, and a traditional method cannot cope with the long-term stability problem caused by sensor aging. The method comprises the steps of
data acquisition, compensation current calculation, training
data set construction, data preprocessing,
machine learning model training and
embedded system integration. The BP neural network is adopted for intelligent prediction, current detection errors can be accurately compensated, and the detection precision is improved; the constructed training
data set considers multiple influence factors, and the model adaptability and generalization ability are high; and the integrated
online learning module can dynamically update the model to adapt to sensor aging, so that long-term stability is ensured, the batch consistency of the sensors is improved, and the production and use cost is reduced.