時間序列在過去二十年當中，受到熱列地討論，而絕大多數的文獻都是研 究線性時間序列模式。但在現實生活中，很多時間序列並不符合線性的假 設，因此近十年來很多學者致力研究非線性時間序列模式。其中有一種雙 線性模式，因其性質與線性模式類似，故引起了廣泛注意。在本篇文章中 我們是採用Subba Rao 和Gabr(1984)提出的迭代等式以及高斯－賽德迭代 法估計參數，再配合 Subba Rao(1981)提出的巢狀搜尋程序，來選取雙線 性模式的階數。將其選模結果與AIC、BIC以及修正後的 PKK 選模法比較 。

In this thesis a stepped-frequency Inverse Synthetic Aperture Radar (ISAR) is employed to develop two-dimensional range-Doppler images of a small ghter aircraft which exhibits three dimensional rotational rotation. The simulation is designed such that the target can exhibit yaw, pitch and roll motions at the same time. First, radar returns from prominent scatterers of various parts of the target are processed and displayed using conventional Fourier transform. The eects of dierent complex motion types and scenarios are observed and discussed. Then, several linear and bi-linear time-frequency distributions including shorttime Fourier transform, Wigner-Ville, pseudo Wigner-Ville, smoothed pseudo Wigner-Ville, Choi-Williams, Born-Jordan and Zhao-Atlas-Marks distributions are applied to the same target and scenarios. The performance of the transforms is compared for each scenario. The reasons for success of the distributions are discussed thoroughly.