摘 要
擴頻通信由于具有抗干擾能力強、截獲率低、良好的碼分多址能力等優(yōu)點，被廣泛地應用于移動通信、雷達、導航和定位等領域。在通信偵察和頻譜監(jiān)測等非協(xié)作通信領域中，由于信噪比低和先驗知識條件缺乏，以至直接序列擴頻信號的檢測與參數(shù)估計難于實現(xiàn)，使之成為當前這一領域中的重要研究課題。
由于直接序列擴頻信號的帶寬遠大于基帶信號帶寬，能量分布于更寬的頻帶，且功率譜密度很低，以至于通常淹沒在噪聲中。正是由于這些特點使得直接序列擴頻信號難于檢測，且在偽隨機序列未知的前提下，即使檢測到了也難以恢復待傳輸?shù)男畔ⅲ瑢е鲁Ｒ?guī)處理方法在此情況下將失效。目前，對于直接序列擴頻信號的檢測和參數(shù)估計已有一些方法，這些方法對于單一參數(shù)具有良好的檢測結果，但在低信噪比情況下性能趨于惡化。
本文主要研究了時域相關檢測、延時相乘、相關累積、循環(huán)譜檢測等方法，在歸納前人理論的基礎上，采用了改進方法，主要研究成果如下：
1. 在直擴信號檢測及偽碼周期的參數(shù)估計方面，分析常規(guī)時域相關檢測方法的基礎上，采用了時域相關二階矩檢測方法，對信號進行分段自相關運算，然后均方迭加平均相關數(shù)據(jù)。此法有效地抑制噪聲，實現(xiàn)了在低信噪比的條件下直接序列擴頻信號偽碼周期的準確估計。
2. 在延時相乘法的基礎上，分析了在相關域上可檢測出偽碼的周期特性，在頻域上可檢測出偽碼速率和載頻參數(shù)的特點，并結合自適應噪聲抵消器、相關累積方法和頻譜校正方法，組成了時域延時相關檢測系統(tǒng)。在低信噪比條件下，可實現(xiàn)對直接序列擴頻信號的偽碼周期、偽碼速率和載頻的準確估計。
3. 為實現(xiàn)更低信噪比條件下的直接序列擴頻信號檢測與參數(shù)估計，在循環(huán)譜理論的基礎上，分析了循環(huán)統(tǒng)計量抑制平穩(wěn)噪聲的能力，采用了基于Welch法的集平均循環(huán)譜方法，對信號分段使用頻域平滑循環(huán)周期圖算法后，進行迭加平均。此法可有效利用循環(huán)譜包絡估計出直接序列擴頻信號的偽碼速率、載頻參數(shù)。
實驗證明，以上三種改進方法在低信噪比條件下具有良好的估計效果，對于直接序列擴頻信號的盲解擴具有一定的意義。

關鍵詞 直擴信號；時域相關二階矩；延時相乘；改進循環(huán)譜；參數(shù)估計





Abstract
The spread spectrum communication depended on good anti-interference ability, low probability of interception and the advantages of CDMA, was widely used in mobile communication, radar, navigation, orientation and other fields. In the area of non-cooperative communication such as communication reconnaissance and spectrum monitoring, owing to low SNR and lacking of priori knowledge, direct sequence spread spectrum(DSSS) signal detection and parameters estimation were difficult to achieve those have become an important issue.
Since the bandwidth of DSSS signal was much larger than the bandwidth of baseband signal, thus energy of DSSS signal was distributed in the much wider bandwidth, power spectrum density was very low so as to submerge in the noise. These features made DSSS signal difficult to detect, or it was difficult to restore the information which was transmitted in the premise of the unknown pseudo-random(PN) sequence. This made the conventional approach invalid at low SNR. At present, there had been some methods for DSSS signal detection and parameters estimation. These methods were good for test results of single parameters. However, for DSSS signal, detection performance tended to deteriorate at low SNR.
In this dissertation, time-domain correlation detection, delay-multiply, correlation cumulation, cyclic spectrum and other methods were considered. On the base of predecessors’ studies, the improved methods were presented here.
1. In the dissertation, in order to detect DSSS signal and estimate the period of PN, time-domain second-order moment detection was proposed based on time-domain correlation detection. The DSSS signal was cut up into several segments in this method, their correlation functions were obtained, and then the mean for superposition of square of correlation data was calculated. This method is able to suppress the additive white Gaussian noise so as to achieve an accurate estimation of the period of PN at low SNR.
2. Based on delay-multiply detection, it indicated that the period of PN showed on the correlation domain, in addition, chip rate and carrier frequency displayed on the frequency domain. This method combined the formers with adaptive noise cancellation, correlation cumulation and spectrum correction in order to compose a detection system of time-domain delay correlation. It can estimate the period of PN, chip rate and carrier frequency accurately.
3. In order to achieve DSSS signal detection and parameters estimation at lower SNR, cyclic statistics had the ability to suppress stationary noise based on cyclic spectrum theory. Improved set-average cyclic spectrum based on Welch method was proposed in this paper. DSSS signal that was divided into several sections used frequency smoothed cyclic periodogram algorithm, then the results for computing the mean were added up. This method used the envelope of cyclic spectrum to estimate the chip rate and carrier frequency, and it had high accuracy.
The simulation showed that the aboved methods can achieve high precision at the low SNR in non-cooperative communication, and had important significance for blind despreading.

Keyword..