关键词: 自适应滤波, 视线角速率提取, 增益因子, 八维模型

Abstract:

Aiming at the adaptive defect problem caused by the covariance array of fixed observation noise during the filtering process, a dynamic adjustment mechanism of observation weights based on gain factor is proposed. This method uses the innovation vector to correct the prediction error covariance in real time, and balances the weights of the prior estimate and observation correction through a gain factor, effectively suppressing filter divergence. At the same time, in order to overcome the problem of insufficient accuracy of the six-dimensional state model in the line-of-sight angular rate extraction of the maneuvering target, the eight-dimensional state equation is constructed by introducing relative acceleration, and the angular rate extraction is carried out in combination with the adaptive mechanism. Simulation shows that the proposed algorithm improves the extraction accuracy of high and low angle and azimuth rate by 19.5% and 53.4% compared with the traditional extended Kalman filter method, achieves the same accuracy as the third-order volumetric Kalman filter. The eight-dimensional model has a lower root mean square error than the six-dimensional model in angular rate estimation, and has a smaller miss distance amount in interception scenarios.

Key words: adaptive filtering, angular rate extraction of line of sight, gain factor, eight-dimensional model