DFT式跳頻信號擷取器之反干擾性能分析

Abstract

隨著我們進入資訊時代，通訊安全的重要性更勝以往。在數位化戰場上，軍事通訊系統正面對著電子戰的威脅，其中通訊干擾便是最常見的電子戰威脅之一。展頻通訊系統具有對抗蓄意干擾的抗干擾能力是眾所週知的。在本論文中，我們首先回顧衛星通訊及通道的一些基本知識，我們以一些例子說明在受到干擾的環境下Ka 頻段的鏈路分析。接著我們提出三種DFT式跳頻信號擷取器架構，其擷取過程是在時域及頻域中同時進行，此架構能夠於一個跳頻序列週期內計算出所有的測試統計量（test statistics）。文中並分別推導此架構在理想的加成性高斯雜訊通道（即雜訊功率已知），非理想的加成性高斯雜訊通道（雜訊功率需經由估測而得），以及非理想的加成性高斯雜訊通道加上寬頻帶雜訊干擾及部分頻帶雜訊干擾下使用不同的信號擷取器架構所產生之偵測機率及誤警機率的數學表示式。此外，文中同時提出了一些方法使信號擷取器在蓄意干擾的環境下能夠估計雜訊功率並將其正規化（normalization）。我們採用 Debruijn 序列作為擬隨機（pseudo random）模式的產生序列，這是由於Debruijn 序列是一個長度為2的冪次的序列，這一點使我們可以有效率地運用快速傅立葉轉換演算法完成計算。文中並以各擷取架構之平均擷取時間相互比較。最後，對於此架構在前述不同場景下，提供了數值模擬結果。此外我們亦分別提供了使用一個週期較長及另一個週期較短的跳頻模式在理想的加成性高斯雜訊通道下的數值模擬結果。

As we are entering the information age, the communication security has become a more important issue than ever. Military communication systems are faced with the electronic warfare (EW) threat in the digitized battlefield. Communication jamming is one of the most-encountered EW threat. It is well known that the spread spectrum communications have the anti-jam capability against intentional jamming. In this thesis, we first review some basics of the satellite communications. Examples for Ka band link budget analysis in the adverse environment are given. We propose three different DFT-based discrete acquisition schemes for the acquisition of the frequency-hopped signals. The acquisition process proceeds in both time and frequency domain. The proposed schemes make acquisition in just one period become possible. Mathematical expressions for the detection probability and false alarm probability using various acquisition schemes in the ideal AWGN channel (noise power perfectly known), the non-ideal AWGN channel（noise power is estimated）, the non-ideal AWGN channel with broad-band noise jamming and partial-band noise jamming are derived. We also propose methods for the estimation and normalization of the noise and the noise-plus-jamming power in the jamming environment. The Debruijn sequence is adopted as the pseudo random hopping pattern generating sequence since it has a code length of power of 2. This allows an efficient computation for FFT algorithm. Finally, numerical results of the system performance for the aforementioned scenarios are presented. Besides, we also give numerical results using hopping patterns with a longer and a shorter period.