應用適應式卡曼濾波器估測載具氣動力係數之研究

The Estimation of Aerodynamic Coefficients Using Adaptive Kalman Filter

王修平(H.-P. Wang) ； 劉永道(Y.-T. Liu)

《中國航空太空學會學刊》 29卷3期 (1997/09) Pp. 285-291

https://doi.org/10.6124/TAASRC.199709_29(3).10

摘要

在飛彈自動駕駛儀的控制設計中，控制器常因彈體參數的變動而影響其性能，故我們對其彈體參數有著極大的興趣，於是本文的重點即是要以估測的方法得到我們想要的重要彈體參數，進而適時地修正控制器參數，以得良好的控制功效。在本文中，由於自動駕駛儀的動態是隨著時間而變化，因此我們使用一個可線性化的卡曼濾波器，也就是擴展卡曼濾波器(extended Kalman filter)來進行估測，然後再用一個修正型擴展卡曼濾波器(modified extended Kalman filter)來做估測的改善，結果證明修正型擴展卡曼濾波器能降低估測誤差。雖然上述的傳統卡曼濾波器用來估測系統參數有著不錯的效果，但是它也易於因初始值設定不當而產生較大誤差，甚至發散。因此本文提出適應式Q之修正型擴展卡曼濾波器(adaptive Q for modified extended Kalman filter)來估測彈體參數，模擬結果證明適應式Q之修正型擴展卡曼濾波器能降低估測誤差，達到我們預期的目標。

關鍵字

自動駕駛儀 ； 擴展卡曼濾波器 ； 適應式Q理論

並列摘要

The performance of autopilot controller of missile is frequently affected by the varying airframe parameters. In order to enhance the performance of the autopilot controller, we need to estimate the airframe parameters to adjust the parameters of the controller. In this paper, for the time-varying dynamics of autopilot, we linearize the dynamics for the Kalman filter. The Kalman filter so obtained is called extended Kalman filter. First, we use an extended kalman filter to estimate the airframe parameters, then we use a modified extended Kalman filter to improve the performance of estimation. The simulation results illustrate that the modified extended Kalman filter can reduce estimation errors. The above two conventional Kalman filters are good for estimating, but they often result in a large error or divergence for with a wrong initial data. Therefore, we adopt an adaptive Q for modified extended Kalman filter to estimate airframe parameters to enhance the performance of estimation. The simulation results illustrate that the adaptive Q for modified extended Kalman filter can reduce estimation errors, and it achieves the object what we expect.

並列關鍵字

autopilot ； extended Kalman filter ； adaptive Q algorithm