應用RS碼之慢跳頻╱多階頻移鍵反干擾系統之設計

Abstract

跳頻是一種常用的電子反反制(electronic counter-counter measure, ECCM)方法；一般可分成慢跳頻和快跳頻兩種系統。因為快跳頻系統具有 多向度(diversity)的特性而有編碼增益 (coding gain)，使其抗干擾的 能力通常比慢跳頻系統好。為了加強慢跳頻系統抗干擾之能力，往往需要 加入正向改錯碼(FEC)。本論文便針對一利用多階頻移鍵(MFSK)調變和RS 改錯碼的慢跳頻系統，分析其反干擾能力並討論相關的系統設計問題。在 此，由於複雜度的問題，我們不考慮完全的軟式解碼(soft-decision decoding)，而是使用錯誤和擦去式(error-erasure)解碼。本文所考慮的 干擾方式，包括部分頻帶干擾(partial band noise jamming)和帶頻式多 波道單音干擾 (band multitone jamming)。同時我們並討論二種偵察不 可靠符元的方法來增加改錯能力。其中一種方法我們稱為直接測試法( direct test,DT)。另一種則為Viterbi's所提出的比例門檻測試法( ratio thresholdtest, RTT)。在本論文中另一個討論要點是考慮有限交 錯(interleaving)長度。我們列出了數值結果來比較DT及RTT的由有效性 及跳頻率、交錯器大小和碼率(coding rate)之間的關係以提供系統設計 者參考。 Frequency hopping systems are often divided into slow FH and fast FH systems. The latter class has a hopping rate of $k$ hops/symbol where $k > 1$ while for the former class $k \le 1$. The anti-jam (AJ) capability of a fast FH system is superior to that of a slow FH system for it has an implicit coding gain through the use of diversity. To enhance a slow FH system's AJ capability forward error-control (FEC) coding is employed. It is well known that soft-decision decoding hard-decision decoding in an AWGN channel by an average margin of 2 dB. But the soft-decision decoding gain is often much more impressive when operates in a non-AWGN channel. This thesis studies a special slow FH system, one that uses MFSK modulation and RS code. We do not consider a full soft-decision decoder due to the complexity consideration. Instead, we use a single-pass errors-and-erasures decoder. Two jamming scenarios are studied: partial-band noise jammer (PBNJ, in Chapter 3) and band multitone jammer (BMTJ, in Chapter We examine two schemes for generating a decision to erase an unreliable symbol (erasure insertion) so that the error correcting capability can be increased. One of them is called the direct test (DT) and the other is borrowed from Viterbi's ratio threshold test (RTT). Another issue under investigation is the effect of finite interleaving length, which was usually neglected in performance analysis. Although we consider block interleaver only our results can easily be used for systems using convolutional interleaver by simple modifications on parameter values. Numerical results are presented to compare the effectiveness of DT and RTT and show the relationships amongst the hopping rate, the interleaver size and the coding rate. Most of our analysis concentrates on the case when the signal size ($M$) and the codeword symbol field size ($|GF(q)|=2^{\ell}$) are equal. An alternative design option $M \neq q$, $\ell/ (log_2 M)=$ ~integer is briefly addressed (Section 3.3.5).