鄭恩明 宋 佳 陳新華 孫長(zhǎng)瑜 余華兵
?
基于時(shí)延差方差加權(quán)的時(shí)延差估計(jì)方法
鄭恩明*①②宋 佳①②陳新華①孫長(zhǎng)瑜①余華兵①
①(中國(guó)科學(xué)院聲學(xué)研究所 北京 100190)②(中國(guó)科學(xué)院大學(xué) 北京 100190)
信息處理;時(shí)延差估計(jì);時(shí)延差方差
針對(duì)目標(biāo)信源數(shù)未知時(shí)的時(shí)延差估計(jì),現(xiàn)有的時(shí)延差估計(jì)算法只是進(jìn)行了簡(jiǎn)單的分頻帶處理,而沒(méi)有對(duì)處理所得數(shù)據(jù)做進(jìn)一步的挖掘和應(yīng)用,已不能很好地滿(mǎn)足現(xiàn)時(shí)需求。對(duì)此,本文將文獻(xiàn)[15]所提出的基于方位方差加權(quán)的目標(biāo)檢測(cè)方法應(yīng)用到時(shí)延差估計(jì)中。本文對(duì)處理所得數(shù)據(jù)做進(jìn)一步挖掘,首先利用目標(biāo)輻射信號(hào)頻率單元對(duì)應(yīng)時(shí)延差估計(jì)比較穩(wěn)定、噪聲頻率單元對(duì)應(yīng)時(shí)延差估計(jì)比較隨機(jī)的特點(diǎn),計(jì)算各頻率單元的時(shí)延差方差,然后對(duì)各頻率單元的時(shí)延差估計(jì)結(jié)果進(jìn)行時(shí)延差方差加權(quán),突出信號(hào)所占頻率單元時(shí)延差估計(jì)結(jié)果,最后實(shí)現(xiàn)對(duì)目標(biāo)信源數(shù)目未知時(shí)的時(shí)延差估計(jì)??杀苊鈱?duì)目標(biāo)輻射信號(hào)進(jìn)行頻率搜索后再求取時(shí)延差。
水下目標(biāo)輻射信號(hào)簡(jiǎn)化形式可表示為[14]
令
圖1為接收陣接收信號(hào)示意圖。陣元接收的信號(hào)形式可表示為[14]
式(14)中,為第個(gè)陣元信號(hào),為相對(duì)參考陣元1經(jīng)延時(shí)后信號(hào),為陣間距,為目標(biāo)入射角,為有效聲速,為第個(gè)陣元接收背景噪聲,為陣元號(hào)。
令
由此可知:以單根線(xiàn)譜為例,式(1)中寬帶信號(hào)與式(4)中背景噪聲,式(1)中線(xiàn)譜信號(hào)分量與式(4)中背景噪聲的譜級(jí)比分別為
式(4)中信噪比為
當(dāng)目標(biāo)信源數(shù)未知時(shí),一般分頻帶互相關(guān)法在求取兩路信號(hào)時(shí)延差時(shí),會(huì)將所有頻率單元時(shí)延差估計(jì)結(jié)果等價(jià)地加權(quán)到時(shí)延差估計(jì)中[4,5],多目標(biāo)情況,時(shí)延差估計(jì)效果更差。為了對(duì)處理所得數(shù)據(jù)做進(jìn)一步挖掘,本文對(duì)一般分頻段法時(shí)延差估計(jì)做進(jìn)一步改進(jìn)處理。改進(jìn)算法原理如下。
當(dāng)目標(biāo)輻射信號(hào)每次均能穩(wěn)定實(shí)現(xiàn)時(shí)延差估計(jì),統(tǒng)計(jì)時(shí)間內(nèi)時(shí)延差變化緩慢時(shí),可以采用下述方法實(shí)現(xiàn)時(shí)延差估計(jì)。
如果采用一般分頻帶法進(jìn)行時(shí)延差估計(jì),最終相關(guān)譜可表示為
最后對(duì)各頻率單元時(shí)延差進(jìn)行時(shí)延差方差加權(quán)統(tǒng)計(jì),得到最終相關(guān)譜和時(shí)延差估計(jì)值。
依據(jù)圖2所示的流程圖,本算法實(shí)現(xiàn)過(guò)程可分為以下6個(gè)步驟:
圖2 本文方法的時(shí)延差估計(jì)流程圖
然后將目標(biāo)信號(hào)估計(jì)時(shí)延差方差結(jié)果累加到式(17)中,可得到
式(18)可簡(jiǎn)化為
由圖3可知:
(1)無(wú)論線(xiàn)譜譜級(jí)與干擾噪聲平均譜級(jí)比是多少,未分頻帶法已不能實(shí)現(xiàn)時(shí)延差估計(jì),原因?yàn)樽罱K相關(guān)譜是由頻帶內(nèi)所有數(shù)據(jù)均等疊加而成,受噪聲和信號(hào)混疊影響較大;
一般分頻帶法(未插值)只需要對(duì)每個(gè)頻帶做兩次FFT;而為了實(shí)現(xiàn)相關(guān)峰細(xì)化,本文方法在求取各頻率單元的時(shí)延差時(shí),采用類(lèi)內(nèi)插法對(duì)所有頻率單元互功率譜進(jìn)行先細(xì)化再求取相關(guān)峰位置,得到較高精度的時(shí)延差估計(jì)值。根據(jù)上述分析可得,本文方法相比一般分頻帶法(未插值)增加的運(yùn)算量主要包括:
對(duì)比圖5和圖6可知,未分頻帶法已不能很好地實(shí)現(xiàn)未知信源數(shù)的時(shí)延差估計(jì);對(duì)比圖5和圖7可知,一般分頻帶法未對(duì)所得數(shù)據(jù)進(jìn)一步挖掘和利用,該方法也不能很好地實(shí)現(xiàn)未知信源數(shù)的時(shí)延差估計(jì);而對(duì)比圖5和圖8可知,本文方法對(duì)所得數(shù)據(jù)做了進(jìn)一步挖掘和利用,時(shí)延差估計(jì)值可以很好地對(duì)應(yīng)上波束形成方位估計(jì)值,克服了一般分頻帶法對(duì)多目標(biāo)進(jìn)行時(shí)延差估計(jì)時(shí)的估計(jì)性能較差問(wèn)題。實(shí)驗(yàn)結(jié)果和數(shù)值仿真結(jié)果一樣表明本文方法可以很好地實(shí)現(xiàn)對(duì)未知信源的時(shí)延差估計(jì)。
圖3 不同信干比下,3種方法的時(shí)延差估計(jì)概率
圖4 實(shí)驗(yàn)線(xiàn)陣及目標(biāo)運(yùn)動(dòng)示意圖
圖5 波束形成所得方位歷程圖
圖6 未分頻帶法時(shí)延差估計(jì)結(jié)果
圖7 一般分頻帶法時(shí)延差估計(jì)結(jié)果
圖8 本文方法的時(shí)延差估計(jì)結(jié)果
[1] 姜可宇, 姚直象, 尹敬湘. 一種基于三元陣的水下目標(biāo)被動(dòng)定位方法[J]. 兵工學(xué)報(bào), 2012, 33(9): 1107-1111.
Jiang Ke-yu, Yao Zhi-xiang, and Yin Jing-xiang. A passive locating method for underwater target based on three- element-array[J]., 2012, 33(9): 1107-1111.
[2] Chen Li-jun, Gao Xiang, and An Liang. Multipathpassivelocalization in shallow water channel[J].(), 2012, 48(5): 609-615.
[3] John G, Martin S, Reid M,.. Localization of a noisy broadband surface target using time differences of multipath arrivals[J].2013, 134(1): 77-83.
[4] Alain L, Seif E, Guy P,.. Time delay estimation for acoustic source location by means of short-time cross-correlation[C]. IEEE Sensors Conference, Limerick, Ireland, 2011: 1885-1888.
[5] Wen F, Wan Q, and Liu Y. Passive time delay estimation for complex noncircular signals[C]. 10th International Symposium on Electronics and Telecommunications, Timisoara, Romania, 2012: 243-246.
[6] 馬曉紅, 陸曉燕, 殷福亮. 改進(jìn)的互功率譜相位時(shí)延估計(jì)方法[J]. 電子與信息學(xué)報(bào), 2004, 26(1): 53-59.
Ma Xiao-hong, Lu Xiao-yan, and Yin Fu-liang. Time delay estimation by using modified cross power spectrum phase technique[J].&, 2004, 26(1): 53-59.
[7] Kou W, Duan W, and Li M. An improved time delay estimation method based on cross-power spectrum phase[C]. IEEE International Conference on Signal Processing, Communication and Computing, Hong Kong, China, 2012: 686-690.
[8] 張學(xué)森, 馮海泓, 黃敏燕, 等. 兩步法提高時(shí)延差估計(jì)精度的分析和實(shí)驗(yàn)驗(yàn)證[J]. 聲學(xué)學(xué)報(bào), 2013, 38(2): 153-159.
Zhang Xue-sen, Feng Hai-hong, Huang Min-yan,.. Analysis and experimental verification on improving time-delay difference estimation accuracy with two-step algorithm[J]., 2013, 38(2): 153-159.
[9] 蔣伊琳, 司錫才. 基于互相關(guān)和MUSIC算法的時(shí)延估計(jì)[J].彈箭與制導(dǎo)學(xué)報(bào), 2009, 29(5): 208-211.
Jiang Yi-lin and Si Xi-cai. Time delay estimation based on cross-correlation and multiple signal classification[J]., 2009, 29(5): 208-211.
[10] Liu W. A novel robust adaptive time delay estimation method[C]. 2011 International Conference on Electronic & Mechanical Engineering and Information Technology, Harbin, China, 2011: 12-14.
[11] Jonathan L O, Jeffrey L K, and Jeffrey S R. Maximum-likelihood spatial spectrum estimation in dynamic environments with a short maneuverable array[J]., 2012, 133(1): 311-322.
[12] McDonough R N and Whalen A D. Detection of Signals in Noise[M]. USA, 2nd Edition, California, Academic Press, 1995: 247-287.
[14] 李啟虎, 李敏, 楊秀庭. 水下目標(biāo)輻射噪聲中單頻信號(hào)分量的檢測(cè):理論分析[J]. 聲學(xué)學(xué)報(bào), 2008,33(3): 193-196.
Li Qi-hu, Li Min, and Yang Xiu-ting. The detection of single frequency component of underwater radiated noise of target: theoretical analysis[J]., 2008, 33(3): 193-196.
[15] 陳新華, 鮑習(xí)中, 李啟虎, 等. 水下聲信號(hào)未知頻率的目標(biāo)檢測(cè)方法研究[J]. 兵工學(xué)報(bào), 2012, 33(4): 471-475.
Chen Xin-hua, Bao Xi-zhong, Li Qi-hu,.. Research on detection of underwater acoustic signal with unknown frequency[J].2012, 33(4): 471-475.
[16] Hou H, Sheng G, andJiang X. Robust time delay estimation method for locating UHF signals of partial discharge in substation[J]., 2013, 28(3): 1960-1968.
[17] Valero X and Alias F. Narrow-band autocorrelation function features for the automatic recognition of acoustic environments[J]., 2013, 134(1): 880-890.
鄭恩明: 男,1985年生,博士生,研究方向?yàn)樾盘?hào)與信息處理、水下聲學(xué)定位系統(tǒng).
陳新華: 男,1978年生,副研究員,主要研究方向?yàn)殛嚵行盘?hào)處理.
孫長(zhǎng)瑜: 男,1954年生,研究員,博士生導(dǎo)師,主要研究方向?yàn)殛嚵行盘?hào)處理.
余華兵: 男,1975年生,副研究員,碩士生導(dǎo)師,主要研究方向?yàn)殛嚵行盘?hào)處理.
Weighted Time Delay Difference Estimation Method Based on Its Variance
Zheng En-ming①②Song Jia①②Chen Xin-hua①Sun Chang-yu①Yu Hua-bing①
①(,,100190,)②(,100190,)
A weighted estimation method based on the Time Delay Difference (TDD) variance is proposed with regard to the problem of TDD estimation of unknown source. This method utilizes the TDD of the target radiation signal frequency unit and the TDD of the noise frequency unit with respective characteristics of being stable and random to weight TDD estimation results of each frequency unit, enhancing the TDD estimation results of signal frequency unit, and achieving the TDD estimation of unknown source. The simulation results show that, compared with the conventional cross-correlation method, the estimation performance of this method is improved 3 dB. The theoretical analysis and simulation results both show that the robustness of this method is better than the conventional cross-correlation method.
Information processing; Time Delay Difference (TDD) estimation; Time delay difference variance
TB565
A
1009-5896(2014)06-1362-06
10.3724/SP.J.1146.2013.01164
鄭恩明 zhengenmingioa@163.com
2013-08-02收到,2013-12-11改回
國(guó)家海洋公益性行業(yè)科研專(zhuān)項(xiàng)項(xiàng)目(201005001)資助課題