A method for fault detection sensor information fusion Kalman filter

The present invention discloses a method of fault detection sensor information fusion Kalman filter, comprising the following steps: Step 1: The state equation and the observation equation of the Kalman filter theory, a linear dynamic system; Step II: The step gains an observation equation, using the least squares method of obtaining state estimation mean square error and the corresponding array, the innovation sequence; step three: using known innovation sequence, different channels to obtain a normalized new sequence of information; and the composition matrix innovation m parallel channels sensor; step four: the innovation obtained in step three matrix, the matrix of the innovation acquired spectral norm and the mean number of spectral norm; step five: Kalman filtering information for fault detection sensor fusion; present invention by mathematical and statistical estimation interval, it simplifies complex calculations greatly improves the speed of fault detection.