摘要: The multi-rate Kalman filter can be used for the data fusion of displacement and acceleration, which were sampled at different frequencies. However, the noise covariance matrices, especially the process noise covariance matrix, are usually unavailable in the practical applications. With inappropriate noise covariance matrices, the state estimates of multi-rate Kalman filter is suboptimal. In this paper, a new adaptive multi-rate Kalman filter, which is based on the autocovariance least-squares method, is proposed. For a given set of displacement and acceleration data sampled at different frequencies, the data fusion problem is formulated as the single-rate Kalman filter rather than the multi-rate Kalman filter. And the correlations between the innovations were used to establish a relationship to the unknown parameters about the noise covariance matrices. Therefore, the unknown parameters can be estimated by solving the least-squares problem. The validity of the proposed method is demonstrated by a numerical example and an earthquake engineering test from the Large High-Performance Outdoor Shake Table.
|
