Improvement of Ensemble Kalman Filter for Hypersonic Target Tracking

Due to the severe and complex aerodynamic effect, hypersonic targets can perform sophisticated maneuvering flight, which changes the distribution of tracking model uncertainty to be non-Gaussian. Focus on the tracking filter for non-Gaussian targets, non-Gaussian deterministic sample generation strategy and processing algorithm are designed to improve the performance of ensemble Kalman filter. According to the distribution characteristics of model error, the generalized polynomial chaos theory is introduced to determine the orthogonal polynomial, and generate one-dimensional sample set and their weights. Then, the coordinate axis sampling strategy is adopted to generate multi-dimensional sample set and their weights and avoid the dimension curse. Finally, the collocation method is improved with the orthogonal basis function matrix to process the samples and determine the coefficients of the polynomial chaos expansion. With these coefficients, the statistical characteristics of the random process are determined and more approximate to the actual probability distribution. Simulation results demonstrate that the improved ensemble Kalman filter can alleviate the calculation error of covariance to a certain extent, and enhance the stability and accuracy of hypersonic target state estimation.