Self-tuning descriptor Kalman filter with correlated noise

For the descriptor system with correlated noise and unknown noise variances, the self-tuning descriptor Kalman filter is presented. Firstly, the fused estimates of these unknown noise variances can be obtained by the correlation method. Then, substituting these fused estimates into the recursive descriptor Kalman filter yields the self-tuning descriptor the Kalman filter and its filtering error variance. A simulation example of a descriptor system verifies its effectiveness.