以追星儀配合一或二軸陀螺儀進行姿態判定之研究

Abstract

摘要 本論文發展一套以Lyapunov方法為基礎的新式估測演算法用以整合追星儀與陀螺儀的輸出資訊來進行衛星姿態判定，除了能夠估測並補償陀螺儀的訊號誤差，還能夠適應於部分陀螺儀發生故障的情況。此演算法的主要優點在於不需要衛星動態模型，不但能夠解決先前文獻所遭遇的模型誤差，還能夠提供較為簡易的運算過程以節省記憶體空間與加快運算速度。 本論文運用Matlab做系統的模擬。在陀螺儀具有訊號誤差的狀況，模擬結果指出本論文所提出的新式估測演算法能夠獲得與擴增型卡爾曼濾波器（Extended Kalman Filter）相當的估測姿態精度（5e-4 deg）。在僅使用一軸或兩軸的陀螺儀來估測衛星姿態的案例下，新式估測演算法還能運用僅剩的追星儀與陀螺儀來繼續提供精確的衛星姿態判定，以延長衛星的有效工作時間，其姿態精度分別為7e-4 deg與8e-4 deg。

Abstract A novel attitude estimation algorithm based on Lyapunov method is developed in this thesis to tackle with gyroscope failures and single drifts. This algorithm uses a star tracker and a 3-axis gyroscope to determine the satellite attitude. The advantage of this algorithm is to require no knowledge of the satellite dynamic model. As such, this attitude estimation algorithm not only eliminates the model error suffered from previous research but also provides a simple calculation process to save the memory space and reduce the calculation time. Simulations in the thesis are operated by Matlab. In the condition that the gyroscope with signal drifts, simulation results indicate that the proposed algorithm can obtain the satellite attitude with the accuracy of 5e-4 deg, which is similar to that of the extended Kalman filter. In the cases of using 1-axis and 2-axis gyroscope, the proposed algorithm can employ the star tracker and the rest of the gyroscope to prolong the effective work time and expand the work site; the attitude accuracy are 7e-4 deg and 8e-4 deg, respectively.