Aiming at throttle flow sensor measurement hysteresis problem

a prediction correction method based on the GM（ 1

1） model is proposed. Firstly

measurement datas are updated in real-time according to the metabolic model

and then the new and the old datas are given different weight value by using adaptive filtering method.Finally

a GM（ 1

1） model based on these measurement data is established to predict the next data

the data to be displayed in the throttle flow sensor will be replaced by the obtained prediction data. The experimental results show that the proposed method can effectively solve the problem of throttle flow sensor measurement hysteresis

average relative error is within 0. 38 %

therefore

the method can meet the requirements of practical engineering.