An auto calibration scheme for sensor fusion

In this paper, an auto calibration scheme is proposed to determine the fusion weights and biases for a weighted sum method of multisensor system. This scheme can gradually adjust the weights and biases of all sensors based on a learning algorithm. Signals with different standard deviations and biases are treated in computer simulation as the multisensor system inputs so as to verify the effectiveness of the auto calibration scheme. It is shown with a comparison with Wiener filter, the proposed scheme can be used to minimize the fusion output's standard deviation and determine the fusion weights and biases.<>