METHOD FOR CALIBRATING A MULTI-SENSOR SYSTEM USING AN ARTIFICIAL NEURAL NETWORK

The invention relates to a method for calibrating a sensor (2) with respect to a reference trajectory (T) indicating a number of timely successive data values using an encoder/decoder neural network (10) with an encoder function (g) and a successive decoder function (f), comprising the steps of: - Training in a first training phase the encoder function (g) and the decoder function (f) of the encoder/decoder neural network (10) with respect to at least one reference trajectory (T) so that a reconstruction error of the reference trajectory (T) at an input of the encoder function (g) and a reconstructed reference trajectory (T) at an output of the decoder function (f) is minimized, - Training in a second training phase a calibration vector (Wt ) with respect to a sensor trajectory (T) so that a reconstruction error between an output sensor trajectory (T, T, T) and a corresponding reference trajectory (T) is minimized, wherein the output sensor trajectory (T, T, T) is obtained by applying the calibration vector (Wt ) on the sensor trajectory (T, T, T) and by applying the encoder function (g) and decoder function (f) on the the calibrated sensor trajectory; - Applying the calibration vector (Wt ) to calibrate the sensor.