劉開(kāi)華,魏沖沖,于潔瀟
?
聯(lián)合TOA和DOA的淺海多徑定位算法
劉開(kāi)華1,魏沖沖1,于潔瀟2
(1. 天津大學(xué)微電子學(xué)院,天津 300072;2. 天津大學(xué)電氣自動(dòng)化與信息工程學(xué)院,天津 300072)
針對(duì)海水媒質(zhì)分布不均勻的特性,本文提出了一種在非均勻媒質(zhì)中聯(lián)合到達(dá)時(shí)間(TOA)和到達(dá)角度(DOA)淺海多徑環(huán)境下的定位算法.該算法不僅能夠克服聲速變化引起的路徑彎曲問(wèn)題,還可以減少海水流動(dòng)對(duì)節(jié)點(diǎn)拓?fù)浣Y(jié)構(gòu)產(chǎn)生的影響.首先分別以海面和海底為反射面建立錨節(jié)點(diǎn)的虛擬節(jié)點(diǎn)(VN),把多徑環(huán)境下的非視距(NLOS)問(wèn)題轉(zhuǎn)化為視距(LOS)問(wèn)題,然后根據(jù)聲速剖面(SSP)利用費(fèi)馬原理獲得聲波曲線(xiàn)路徑的表達(dá)式.又由于海面、海底形狀的不規(guī)則性,信號(hào)在反射點(diǎn)發(fā)生散射現(xiàn)象,因此將反射點(diǎn)當(dāng)作分布式信源,利用散射信號(hào)中心波達(dá)方向作為DOA,平均波達(dá)時(shí)間作為T(mén)OA.最后利用到達(dá)時(shí)間和到達(dá)角度估計(jì)目標(biāo)的位置參數(shù).仿真結(jié)果表明,本方法在定位精度和魯棒性等方面優(yōu)于傳統(tǒng)的直線(xiàn)傳輸模型.
非均勻媒質(zhì);多徑;聲音速度剖面;費(fèi)馬原理;虛擬節(jié)點(diǎn)
近年來(lái),水聲傳感器網(wǎng)絡(luò)(underwater acoustic sensor network,UASN)在海洋環(huán)境監(jiān)測(cè)、海底資源探索、災(zāi)難預(yù)警監(jiān)測(cè)等方面應(yīng)用越來(lái)越廣泛[1-2].在UASN中,通常把傳感器節(jié)點(diǎn)布置在一定區(qū)域內(nèi)來(lái)獲取溫度、壓力等相關(guān)數(shù)據(jù),傳感器的位置信息將有助于這些數(shù)據(jù)的分析.因此,節(jié)點(diǎn)的位置獲取起到了至關(guān)重要的作用.
在淺海定位過(guò)程中,信號(hào)傳輸經(jīng)海面和海底反射,產(chǎn)生多徑現(xiàn)象,對(duì)通信質(zhì)量造成影響,由此水下多徑問(wèn)題一直是研究的熱點(diǎn)方向.室內(nèi)環(huán)境的多徑問(wèn)題已經(jīng)得到了廣泛的研究,但由于水下環(huán)境中聲速傳播慢、節(jié)點(diǎn)的移動(dòng)性以及傳播媒質(zhì)的不均勻性等特點(diǎn)[3-5],使得原有的室內(nèi)定位算法在水下不再適用.文獻(xiàn)[6]利用自相關(guān)和互相關(guān)器對(duì)TDOA進(jìn)行估計(jì),之后根據(jù)線(xiàn)性最小二乘估計(jì)目標(biāo)的深度和距離.文獻(xiàn)[7]用矢量傳感器接收多徑環(huán)境下不同信道傳輸?shù)男盘?hào),獲取因水下信道參數(shù)不同而帶來(lái)的矢量傳感器陣列的相關(guān)性等級(jí)差異.文獻(xiàn)[8]用同態(tài)解卷恢復(fù)水下信道的鏈路信息,根據(jù)鏈路的不同特征對(duì)視距和非視距路徑做分類(lèi),利用閉合形式的最小二乘方法定位目標(biāo)節(jié)點(diǎn).然而,以上文獻(xiàn)使用多錨節(jié)點(diǎn)定位,均未考慮海水的流動(dòng)性對(duì)拓?fù)浣Y(jié)構(gòu)以及定位效果產(chǎn)生的影響[9],且在UASN中為了降低能量損耗,傳感器的布置密度較低,因此僅用一個(gè)錨節(jié)點(diǎn)定位在真實(shí)的UASN中更加合理.
海水的非均勻性還會(huì)導(dǎo)致聲音速度在不同水深發(fā)生變化,從而使得信號(hào)傳播路徑發(fā)生彎曲[10-11],因此在變聲速的海洋環(huán)境中傳統(tǒng)的直線(xiàn)傳播模型會(huì)加大目標(biāo)的定位誤差.文獻(xiàn)[12]分析了同一時(shí)間延時(shí)和同一速度模型下直線(xiàn)模型和曲線(xiàn)模型的定位表現(xiàn).文獻(xiàn)[13]分析了等梯度速度刨面(sound speed profile,SSP)下的定位情況.之后,文獻(xiàn)[14]分析了在多個(gè)梯度層下的定位誤差.文獻(xiàn)[15]根據(jù)海水深度信息和聲速剖面圖估計(jì)出水下聲波傳播時(shí)間和傳播距離關(guān)系,建立了查詢(xún)表.文獻(xiàn)[16]在非均勻的水下媒質(zhì)中分析了聲波傳輸損失,并且利用接收信號(hào)強(qiáng)度RSS估計(jì)兩點(diǎn)之間曲線(xiàn)距離.文獻(xiàn)[17]利用分層補(bǔ)償?shù)姆椒ㄌ幚砗K謱有?yīng),之后通過(guò)牛頓-拉普森迭代算法求解曲線(xiàn)路徑下目標(biāo)源的位置.
本文將曲線(xiàn)傳輸模型應(yīng)用到水下多徑環(huán)境中,提出了一種在非均勻媒質(zhì)中聯(lián)合TOA和DOA的水下多徑環(huán)境下的定位算法,且該算法僅用一個(gè)錨節(jié)點(diǎn)即可完成定位.首先利用幾何學(xué)的方法建立錨節(jié)點(diǎn)的虛擬節(jié)點(diǎn)(virtual node,VN),構(gòu)建虛擬直射路徑,然后根據(jù)聲速模型利用費(fèi)馬原理求出曲線(xiàn)傳輸路徑,最后利用錨節(jié)點(diǎn)測(cè)到的多條路徑到達(dá)時(shí)間和到達(dá)角度估計(jì)目標(biāo)的位置坐標(biāo).仿真結(jié)果表明,該定位算法在定位精度上優(yōu)于傳統(tǒng)的直線(xiàn)模型算法,并且這種優(yōu)勢(shì)隨著目標(biāo)距離的增大而更加明顯.
圖1?水下節(jié)點(diǎn)分布示意
信號(hào)在反射面反射時(shí),由于海底和海面的不規(guī)則性,反射后的信號(hào)會(huì)出現(xiàn)一簇散射多徑,因此可把海面和海底的反射點(diǎn)等效為分布式信源.鑒于信號(hào)的分布對(duì)稱(chēng)特性,取中心波達(dá)方向作為到達(dá)角[18];又因同一簇的不同散射信號(hào)具有近似相同的波達(dá)時(shí)間,故取一簇散射信號(hào)波達(dá)時(shí)間的均值為信號(hào)到達(dá)時(shí)間[19],估計(jì)參數(shù)表示為
???(1)
???(2)
海洋環(huán)境聲音速度剖面對(duì)于定位精度起到了至關(guān)重要的作用.本文使用文獻(xiàn)[20]中的速度模型進(jìn)行分析,如圖2(a)所示,虛擬節(jié)點(diǎn)所處空間的速度模型關(guān)于對(duì)稱(chēng)面對(duì)稱(chēng),如圖2(b)所示.
圖2?水下聲音速度隨深度變化剖面
本文目標(biāo)節(jié)點(diǎn)定位算法流程如算法如下.
for=1 to 3 do
end
步驟2?根據(jù)錨節(jié)點(diǎn)及建立的虛擬節(jié)點(diǎn)位置分別獲取目標(biāo)節(jié)點(diǎn)的位置坐標(biāo).
for=1 to 3 do
end
步驟3?賦予不同坐標(biāo)信息不同的權(quán)重,并進(jìn)行融合.
在建立虛擬節(jié)點(diǎn),引入虛擬直射路徑之后,對(duì)于每條路徑分析過(guò)程相同,故以DP為例進(jìn)行分析.定位示意如圖3所示.
信號(hào)沿曲線(xiàn)傳輸?shù)恼鎸?shí)時(shí)間可由曲線(xiàn)積分得到
???(3)
???(5)
???(6)
???(7)
?(8)
即上式的原函數(shù)是一個(gè)常數(shù),用公式表示為
???(9)
轉(zhuǎn)換形式為
???(10)
???(11)
???(12)
???(13)
???(14)
經(jīng)過(guò)3條路徑估計(jì)出的3個(gè)坐標(biāo)用加權(quán)平均法進(jìn)行信息融合,最后得到的坐標(biāo)即為目標(biāo)位置.由于信號(hào)傳輸可信度與傳輸路徑長(zhǎng)度有關(guān),可用傳輸時(shí)間的長(zhǎng)短來(lái)判斷權(quán)值的大小,即
???(15)
則最終目標(biāo)估計(jì)位置坐標(biāo)表示為
???(16)
???(17)
似然函數(shù)為
???(19)
?????(20)
???(21)
???(22)
???(23)
根據(jù)式(11)和式(12),能得到
?????(24)
???(25)
???(26)
?????(27)
式中:
???(28)
???(29)
???(30)
對(duì)目標(biāo)位置估計(jì)的均方根誤差(root mean square error,RMSE)定義為
???(31)
所以其理論下界,即克拉美羅界可以表示為
???(32)
圖4?不同角度方差下的定位誤差
圖5?不同時(shí)間方差下的定位誤差
圖6?不同距離下定位誤差
圖7?CDF曲線(xiàn)
針對(duì)淺海多徑環(huán)境中海水媒質(zhì)分布不均勻的特性,本文提出了一種非均勻媒質(zhì)中聯(lián)合TOA和DOA的定位算法.該算法只用單獨(dú)錨節(jié)點(diǎn)即可完成定位,并能有效地減少移動(dòng)的拓?fù)浣Y(jié)構(gòu)對(duì)定位產(chǎn)生的影響.首先對(duì)錨節(jié)點(diǎn)關(guān)于反射面建立虛擬節(jié)點(diǎn),之后獲得到達(dá)時(shí)間和到達(dá)角度,并利用費(fèi)馬原理求出不同路徑的曲線(xiàn)模型,最后根據(jù)測(cè)得數(shù)據(jù)估計(jì)目標(biāo)節(jié)點(diǎn)的位置.仿真結(jié)果表明,該算法在定位精度中優(yōu)于傳統(tǒng)的直線(xiàn)傳輸模型,并且具有較好的魯棒性.
[1] 黎作鵬,蔡紹濱,張?菁,等. 水聲傳感器網(wǎng)絡(luò)節(jié)點(diǎn)定位技術(shù)綜述[J]. 小型微型計(jì)算機(jī)系統(tǒng),2012,33(3):12-17.
Li Zuopeng,Cai Shaobin,Zhang Jing,et al. Survey on node localization technology of underwater acoustic sensor networks[J].,2012,33(3):12-17(in Chinese).
[2] Tan H P,Diamant R,Seah W K G,et al. A survey of techniques and challenges in underwater localization[J].,2011,38(14/15):1663-1676.
[3] Akyildiz I F,Pompili D,Melodia T. Underwater acoustic sensor networks: Research challenges[J].,2005,3(3):257-279.
[4] Stojanovic M,Preisig J. Underwater acoustic communication channels:Propagation models and statistical characterization[J].,2009,47(1):84-89.
[5] Cheng W,Teymorian A Y,Ma L,et al. Underwater localization in sparse 3D acoustic sensor networks [C]//27. Arizona,USA,2008:236-240.
[6] Yuan Y X,Carter C,Salt J E. Near-optimal range and depth estimation using a vertical array in a correlated multipath environment[J].,2000,48(2):317-330.
[7] Abdi A,Guo H. Signal correlation modeling in acoustic vector sensor arrays[J].,2009,57(3):892-903.
[8] Emokpae L E,Dibenedetto S,Potteiger B,et al. UREAL:Underwater reflection-enabled acoustic-based localization[J].,2014,14(11):3915-3925.
[9] Han G,Zhang C,Shu L,et al. Impacts of deployment strategies on localization performance in underwater acoustic sensor networks[J].,2015,62(3):1725-1733.
[10] Berger C R,Zhou S,Willett P,et al. Stratification effect compensation for improved underwater acoustic ranging[J].,2008,56(8):3779-3783.
[11] Liu J,Wang Z,Cui J H,et al. A joint time synchronization and localization design for mobile underwater sensor networks[J].,2016,15(3):530-543.
[12] Ameer P M,Jacob L. Localization using ray tracing for underwater acoustic sensor networks[J].,2010,14(10):930-932.
[13] Ramezani H,Jamali-Rad H,Leus G. Target localization and tracking for an isogradient sound speed profile [J].,2013,61(6):1434-1446.
[14] Ramezani H,Leus G. Accurate ranging in a stratified underwater medium with multiple isogradient sound speed profile layers[J].,2012,45(5):146-151.
[15] Casalino G,Turetta A,Simetti E,et al. RT 2:A real-time ray-tracing method for acoustic distance evaluations among cooperating AUVs[C]//. Sydney,Austra-lia,2010:1-8.
[16] Poursheikhali S,Zamiri-Jafarian H. Ranging in underwater wireless sensor network:Received signal strength approach[C]//. Doha,Qatar,2016:1-6.
[17] Zhang B,Wang H,Xu T,et al. Received signal strength-based underwater acoustic localization considering stratification effect [C]//. Shanghai,China. 2016:1-8.
[18] Yu J,Zhang L,Liu K. Coherently distributed wideband LFM source localization[J].,2015,22(4):504-508.
[19] 王永良,陳?輝,彭應(yīng)寧,等. 空間譜估計(jì)理論與算法[M]. 北京:清華大學(xué)出版社,2004.
Wang Yongliang,Chen Hui,Peng Yingning,et al.[M]. Beijing:Tsinghua University Press,2004(in Chinese).
[20] Kim E,Lee S,Kim C,et al. Floating beacon-assisted 3-D localization for variable sound speed in underwater sensor networks[C]//. Basel,Switzerland,2010:682-685.
(責(zé)任編輯:王曉燕)
Joint TOA and DOA Localization Algorithm in Shallow Ocean Multipath Environment
Liu Kaihua1,Wei Chongchong1,Yu Jiexiao2
(1.School of Microelectronics,Tianjin University,Tianjin 300072,China;2.School ofElectrical and Information Engineering,Tianjin University,Tianjin 300072,China)
To overcome the problem caused by the inhomogeneous water medium,a joint time-of-arrival(TOA) and direction-of-arrival (DOA) localization algorithm considering the inhomogeneous water medium in shallow ocean multipath environment is proposed. This algorithm can not only avoid the slanted path induced by the various sound speeds,but also reduce the influence of dynamic underwater network topology. Firstly,the virtual node (VN) of the anchor node was set up based on the surface and the bottom of the sea to convert the non-line-of-sight (NLOS) problem into a line-of-sight (LOS) problem. Then,according to the known sound speed profile (SSP),the numerical path of signal was derived by using Fermat’s principle. Furthermore,due to the irregularity of sea surface and seabed,the signal was scattered at the reflection point. So the reflection point was regarded as a distributed source,and the central angle and the average arrival time was regarded as DOA and TOA,respectively. Finally,the location parameter of target was estimated by TOA and DOA. The simulation results indicate that the proposed method is superior to conventional straight line propagation model in location accuracy and robustness.
inhomogeneous medium;multipath;sound speed profile;Fermat’s principle;virtual node
10.11784/tdxbz201703045
TP391.9
A
0493-2137(2018)02-0129-06
2017-03-19;
2017-07-07.
劉開(kāi)華(1956—??),男,博士,教授,liukaihua@tju.edu.cn.
于潔瀟,yjx@tju.edu.cn.
國(guó)家自然科學(xué)基金資助項(xiàng)目(61501322).
the National Natural Science Foundation of China(No.,61501322).