An adaptive estimation of forecast error covariance matrices is proposed for Kalman filtering data assim- ilation. A forecast error covariance matrix is initially estimated using an ensemble of perturbation forecasts....An adaptive estimation of forecast error covariance matrices is proposed for Kalman filtering data assim- ilation. A forecast error covariance matrix is initially estimated using an ensemble of perturbation forecasts. This initially estimated matrix is then adjusted with scale parameters that are adaptively estimated by minimizing -2log-likelihood of observed-minus-forecast residuals. The proposed approach could be applied to Kalman filtering data assimilation with imperfect models when the model error statistics are not known. A simple nonlinear model （Burgers＇ equation model） is used to demonstrate the efficacy of the proposed approach.展开更多
Feature initialization is an important issue in the monocular simultaneous locahzation ana mapping （SLAM） literature as the feature depth can not be obtained at one observation. In this paper, we present a new featu...Feature initialization is an important issue in the monocular simultaneous locahzation ana mapping （SLAM） literature as the feature depth can not be obtained at one observation. In this paper, we present a new feature initialization method named modified homogeneous parameterization （MHP）, which allows undelayed initialization with scale invariant representation of point features located at various depths. The linearization error of the measurement equation is quantified using a depth estimation model and the feature initialization process is described. In order to verify the performance of the proposed method, the simulation is carried out. Results show that with the proposed method, the SLAM algorithm can achieve better consistency as compared with the existing inverse depth parameterization （IDP） method.展开更多
基金The study has been continued under the support of the Foundation for Research Science and Technology of New Zealand under contract C01X0401
文摘An adaptive estimation of forecast error covariance matrices is proposed for Kalman filtering data assim- ilation. A forecast error covariance matrix is initially estimated using an ensemble of perturbation forecasts. This initially estimated matrix is then adjusted with scale parameters that are adaptively estimated by minimizing -2log-likelihood of observed-minus-forecast residuals. The proposed approach could be applied to Kalman filtering data assimilation with imperfect models when the model error statistics are not known. A simple nonlinear model （Burgers＇ equation model） is used to demonstrate the efficacy of the proposed approach.
文摘Feature initialization is an important issue in the monocular simultaneous locahzation ana mapping （SLAM） literature as the feature depth can not be obtained at one observation. In this paper, we present a new feature initialization method named modified homogeneous parameterization （MHP）, which allows undelayed initialization with scale invariant representation of point features located at various depths. The linearization error of the measurement equation is quantified using a depth estimation model and the feature initialization process is described. In order to verify the performance of the proposed method, the simulation is carried out. Results show that with the proposed method, the SLAM algorithm can achieve better consistency as compared with the existing inverse depth parameterization （IDP） method.
