Estimations of signal and parameters using covariance information in linear continuous systems

✍ Scribed by Seiichi Nakamori

• Publisher: Elsevier Science
• Year: 1992
• Tongue: English
• Weight: 850 KB
• Volume: 16
• Category: Article
• ISSN: 0895-7177
• DOI: 10.1016/0895-7177(92)90056-q

✦ Synopsis

The estimation problem of a signal is considered for the white Gaussian observation noise in linear continuous systems. At first, the recursive fixed-point smoother and filter are designed using the covariance information. The observation equation is given by y(t) = z(t) + v(t), z(t) = H(t)z(t), where y(t),z(t),v(t) and H(t) denote the observed value, the signal to be estimated, the white Gaussian observation noise and the observation matrix, respectively. It is assumed that the observed value, the autocovariance Kz(t, s) of z(t) and the variance R(t) of v(t) are known beforehand. Also, the spectral factorization problem is discussed on the system matrix F(t), the input matrix G(t) for the white Gaussian noise and the observation matrix H(t), and the parameter estimation algorithms for G(t) and H(t) are developed by using covariance information.