吳 虹,王 沖,劉 兵,穆巍煒,徐錫燕,馬肖旭,李欣然
(1.南開大學(xué) 電子信息與光學(xué)工程學(xué)院,天津 300071;2.天津市光電傳感器與傳感網(wǎng)絡(luò)技術(shù)重點實驗室,天津 300071)
一種改進的GFDM時頻同步算法*
吳 虹**1,2,王 沖1,2,劉 兵1,2,穆巍煒1,2,徐錫燕1,2,馬肖旭1,2,李欣然1,2
(1.南開大學(xué) 電子信息與光學(xué)工程學(xué)院,天津 300071;2.天津市光電傳感器與傳感網(wǎng)絡(luò)技術(shù)重點實驗室,天津 300071)
針對廣義頻分復(fù)用(GFDM)系統(tǒng)對符號定時同步要求較高的問題,提出了一種新的基于前綴碼的同步算法。在接收端,在獲取粗略定時信息的基礎(chǔ)上,利用前綴碼前后兩部分的相位差實現(xiàn)載波頻偏估計,并對接收序列的頻率偏移進行糾正,然后通過糾正后序列與已知發(fā)射前綴的互相關(guān)函數(shù)實現(xiàn)精確的符號定時估計。由于該前綴碼具有共軛對稱的特性,使其避免了“平頂效應(yīng)”的出現(xiàn)。結(jié)合5G中低時延高可靠場景,在頻率選擇性信道中對其進行仿真,并通過均方誤差對其性能進行了評估。理論分析及仿真結(jié)果表明,該算法相對于原算法具有更好的定時同步性能和更低的復(fù)雜度,提升了GFDM系統(tǒng)的整體性能。
廣義頻分復(fù)用;定時偏移;載波頻偏;同步算法
隨著第五代移動通信(The Fifth Generation Mobile Communication System,5G)逐漸成為全球移動通信領(lǐng)域研究的熱點,5G愿景與關(guān)鍵能力需求已基本明確。低時延高可靠場景是5G新拓展的場景,主要面向物聯(lián)網(wǎng)業(yè)務(wù),為用戶提供可靠的網(wǎng)絡(luò)連接、低的接入延遲及更好的移動支持性。文獻[1]介紹了幾種新的多載波調(diào)制方式,更好地滿足下一代移動通信的需求。廣義頻分復(fù)用(Generalized Frequency Division Multiplexing,GFDM)[2-3]是由德國5G非正交波束復(fù)形 (5G Non-Orthogonal Waveforms for Asynchronous Signalling,5GNOW)項目于2009年提出的一種多載波調(diào)制方案,由于其具有低時延、低帶外輻射等優(yōu)點,并在頻率選擇性衰落信道中具有良好的性能[4],因此將成為未來5G新場景需求的一個非常靈活的解決辦法。
多載波通信系統(tǒng)的一個重要問題即同步問題[5]。文獻[6]對GFDM在存在符號定時偏移(Symbol Timing Offset,STO)及載波頻率偏移(Carrier Frequency Offset,CFO)情況下的性能進行了分析,表明GFDM系統(tǒng)對于定時偏移比正交頻分復(fù)用(Orthogonal Frequency Division Multiplexing,OFDM)更為敏感,因此,必須對其進行非常精確的符號定時同步。與OFDM眾多的同步技術(shù)[7]相比,目前GFDM的同步方法非常少,其主要方法可分為基于數(shù)據(jù)輔助型及非基于數(shù)據(jù)輔助型兩種。文獻[8]介紹了一種非基于數(shù)據(jù)輔助的同步算法,在非常嚴(yán)重的多徑條件下,為GFDM系統(tǒng)提供了一種粗略頻偏估計的方法?;跀?shù)據(jù)輔助型算法主要是通過增加訓(xùn)練序列,并利用訓(xùn)練序列的相關(guān)性來實現(xiàn)GFDM的時頻同步[9-11],該類算法估計精度更高,應(yīng)用更廣。文獻[9]介紹了一種基于前綴碼的時頻同步算法,但其實現(xiàn)過程中會出現(xiàn)“平頂效應(yīng)”。
本文提出了一種改進的同步算法,實現(xiàn)了GFDM系統(tǒng)中符號定時及載波頻偏的準(zhǔn)確估計,并驗證了其在頻率選擇性衰落信道中的性能。
GFDM是5GNOW項目組針對下一代移動通信提出的一種新的物理層方案。與OFDM使用矩形脈沖濾波器不同,GFDM是一種采用非矩形脈沖成型的多載波調(diào)制系統(tǒng),其發(fā)送端的系統(tǒng)模型如圖1所示。
圖1 GFDM系統(tǒng)發(fā)送端框圖
Fig.1 Block diagram of the GFDM transmitter
(1)
圖2 GFDM數(shù)據(jù)塊結(jié)構(gòu)
Fig.2 Structure of the GFDM data block
定義T0為每個子符號周期,TCP為循環(huán)前綴的周期,整個GFDM的符號周期TGFDM=TCP+MT0,可以看出由于GFDM的數(shù)據(jù)塊結(jié)構(gòu),使其相對OFDM使用更少的CP,具有更高的頻譜效率。
文獻[9]介紹了一種基于前綴碼的GFDM同步方案,但由于循環(huán)前綴的影響,定時度量曲線會出現(xiàn)“平頂效應(yīng)”,需要使用文獻[12]中提出的辦法來消除其影響,導(dǎo)致系統(tǒng)復(fù)雜度的增加。本文結(jié)合OFDM中定時同步算法[13],將一種新的前綴碼引入到GFDM系統(tǒng)中。
假定GFDM的前綴碼塊由Mp=2個子符號及Kp個子載波組成。該數(shù)據(jù)塊中,一個長度為Kp且關(guān)于中心點共軛對稱的偽隨機(Pseudo-Noise,PN)序列c=(c[0],…,c[Kp/2-1],c[Kp/2],…,c[Kp-1])T被傳輸兩次,得到該前綴碼承載的數(shù)據(jù)向量
dp=(c[0],…,c[Kp-1],c[0],…,c[Kp-1])T。
時域內(nèi),該前綴碼滿足前后重復(fù)特性的同時,前后兩部分還具有關(guān)于各自中心點呈共軛對稱的特點,其時域結(jié)構(gòu)如圖3所示,其中B*與A呈共軛對稱。
圖3 前綴碼時域結(jié)構(gòu)
Fig.3 Time domain structure of the preamble
在接收端,接收序列r[n]中,發(fā)射的前綴碼必須至少被接收到一次,用于估計時間及頻率偏移。根據(jù)接收序列r[n],得到其自相關(guān)函數(shù)
(2)
進而可以得到其歸一化自相關(guān)函數(shù)
(3)
(4)
(5)
得到的頻偏信息可以用來糾正接收序列的頻率偏移,糾正后的接收序列信息為
(6)
該算法相對于文獻[9]中算法不會出現(xiàn)“平頂效應(yīng)”,但由于其訓(xùn)練序列是共軛對稱的,導(dǎo)致與正確定時位置相距N/4處出現(xiàn)兩個比較大的旁峰,影響定時的準(zhǔn)確性,需對其做進一步的優(yōu)化。本文利用糾正后的接收序列與已知發(fā)射前綴的互相關(guān)函數(shù),得到更精確的定時估計。其互相關(guān)函數(shù)為
(7)
式中:px=Adp,A為發(fā)射矩陣;px即將前綴碼通過發(fā)射矩陣后得到的序列。由于前綴碼由相同的兩部分組成,因此在ρC[n]中,與最高峰相距N/2的位置有兩個次高峰。于是將ρC[n]與μS[n]結(jié)合起來,較好地抑制了旁峰的影響:
(8)
圖4 本文算法中STO精確估計
Fig.4 The accurate estimation of STO in this paper
該算法與文獻[9]中算法相比,通過引進共軛對稱序列,構(gòu)造出新的前綴碼,實現(xiàn)了符號定時及載波頻偏的準(zhǔn)確估計。從整個同步算法實現(xiàn)的過程可以看出,兩種算法均需要進行自相關(guān)函數(shù)的計算、自相關(guān)函數(shù)的歸一化處理,以及修正后序列與發(fā)射前綴互相關(guān)函數(shù)的計算,但本文算法由于不會出現(xiàn)“平頂效應(yīng)”,較原算法復(fù)雜度更低。
在理想信道情況下,由μA[n]尖銳的峰值可以得到精確的符號定時信息,但在頻率選擇性衰落信道中,初始信道的抽頭增益會受到影響,其峰值可能會比其他回波低,影響符號定時的準(zhǔn)確性。
為了在頻率選擇性衰落信道中實現(xiàn)符號定時同步,可以采用門限準(zhǔn)則的方法[14],根據(jù)設(shè)定的錯誤報警概率pFA,在找到最大峰值之前識別出其他的多徑峰值。門限值通過瑞利累積分布函數(shù)來獲取,其值如下式所示:
(9)
(10)
為了驗證本文提出算法的性能,在頻率選擇性衰落信道中,對未知量時間偏移θ及頻率偏移ε進行估計,并通過STO及CFO估計的均方誤差,對本文算法及文獻[9]算法的性能進行比較。仿真參數(shù)的設(shè)置如下:碼元符號采用16QAM;為了降低計算的復(fù)雜度,子載波數(shù)目K取256,子符號數(shù)目M取2,循環(huán)前綴長度取32;多徑搜索參數(shù)λ取32;錯誤報警概率設(shè)定為10-6。綜合考慮仿真結(jié)果的準(zhǔn)確性及系統(tǒng)的運行效率,在每個信噪比(Signal-to-Noise Ratio,SNR)值上仿真了500次,具體的仿真結(jié)果如圖5和圖6所示。
圖5 頻率選擇性信道下STO估計均方誤差曲線
Fig.5 MSE curve of the STO estimation in frequency-selective channel
圖6 頻率選擇性信道下CFO估計均方誤差曲線
Fig.6 MSE curve of the CFO estimation in frequency-selective channel
圖5給出了兩種算法符號定時同步的均方誤差曲線。由圖中可以看出,本文算法的均方誤差在信噪比變化的過程中低于文獻[9]算法,其定時同步性能更優(yōu),定時同步的準(zhǔn)確性得到了提升,解決了GFDM系統(tǒng)對時間偏移更為敏感的問題。圖6的仿真結(jié)果表明,在信噪比低的情況,本文算法中頻偏估計的均方誤差曲線較文獻[9]算法性能稍差,但隨信噪比的增加,本文算法的性能提高較為明顯,這主要是因為其頻偏估計的精度受定時同步性能的影響,隨著SNR的增加,定時同步性能明顯優(yōu)于原算法,其頻偏估計精度也明顯提高。
本文在基于前綴碼同步算法的基礎(chǔ)上,將一種具有共軛對稱性的前綴碼引入到GFDM系統(tǒng)中,介紹了一種新的時頻同步算法。仿真結(jié)果表明,該算法可以獲得與經(jīng)典的OFDM方案相近的性能,且不會產(chǎn)生原算法中出現(xiàn)的“平頂效應(yīng)”,降低了系統(tǒng)的復(fù)雜度。該算法還可較好地適應(yīng)頻率選擇性衰落信道,面臨實際復(fù)雜的信道環(huán)境,具有重要的實際意義。GFDM作為下一代移動通信熱門的候選,該算法主要解決了其同步問題,提升了其整體性能,使其可以在未來5G新場景中得到廣泛的應(yīng)用。未來計劃將與多載波同步技術(shù)結(jié)合緊密的相關(guān)技術(shù)引入GFDM系統(tǒng)中,進一步提升系統(tǒng)性能。
[1] WUNDER G,JUNG P,KASPARICK M,et al. 5GNOW:non-orthogonal,asynchronous waveforms for future mobile applications[J]. IEEE Communications Magazine,2014,52(2):97-105.
[2] FETTWEIS G,KRONDORF M,BITTNER S. GFDM - generalized frequency division multiplexing[C]// Proceedings of IEEE 69th Vehicular Technology Conference. Barcelona:IEEE,2009:1-4.
[3] MICHAILOW N,MATTHE M,GASPAR I S,et al. Generalized frequency division multiplexing for 5th generation cellular networks[J]. IEEE Transactions on Communications,2014,62(9):3045-3061.
[4] MICHAILOW N,MENDES L,MATTHé M,et al. Robust WHT-GFDM for the next generation of wireless networks[J]. IEEE Communications Letters,2015,19(1):106-109.
[5] BAHAI A R S,SALTZBERG B R. Multi-carrier digital communications-theory and applications of OFDM[M]. New York:Springer,2004:83-101.
[6] CHOI J H,LIM B J,KIM Y J,et al. Effect of timing and frequency synchronization errors on GFDM systems[C]// Proceedings of 2015 International Conference on Information and Communication Technology Convergence. Jeju:IEEE,2015:1322-1325.
[7] 羅仁澤. OFDM 系統(tǒng)中同步算法的分析與比較[J]. 電訊技術(shù),2006,46(4):150-155. LUO Renze. Analysis and comparison among the synchronization schemes for OFDM systems[J].Telecommunication Engineering,2006,46(4):150-155.(in Chinese)
[8] KADUR T,GASPAR I,MICHAILOW N,et al. Non-data aided frequency synchronization exploiting ICI in non-orthogonal systems[C]//Proceedings of 2014 IEEE 80th Vehicular Technology Conference. Vancouver:IEEE,2014:1-5.
[9] GASPER I S,MENDES L L,MICHAILOW N,et al. A synchronization technique for generalized frequency division multiplexing[J]. Eurasip Journal on Advances in Signal Processing,2014(1):1-10.
[10] GASPAR I,FESTAG A,FETTWEIS G. Synchronization using a pseudo-circular preamble for generalized frequency division multiplexing in vehicular communication[C]// Proceedings of 2015 IEEE 82nd Vehicular Technology Conference.Boston:IEEE,2015:1-5.
[11] GASPAR I,FETTWEIS G.An embedded midamble synchronization approach for generalized frequency division multiplexing[C]//Proceedings of 2015 IEEE Global Communications Conference. San Diego:IEEE,2015:1-5.
[12] MINN H,ZENG M,BHARGAVA V K. On timing offset estimation for OFDM systems[J]. IEEE Communications Letters,2000,4(7):242-244.
[13] PARK B,CHEON H,KANG C,et al. A novel timing estimation method for OFDM systems[J]. IEEE Communications Letters,2003,7(5):239-241.
[14] KASPARIS C. A cross-correlation approach for improved timing estimation in OFDM broadcasting systems[C]// Proceedings of 24th AIAA International Communications Satellite Systems Conference. San Diego:IEEE,2006:1039-1048.
WU Hong was born in Tianjin,in 1967. She is now a professor and also the Ph. D. supervisor. Her research concerns wireless communication techniques.
Email:wuhong@nankai.edu.cn
王 沖(1986—),男,湖北襄陽人,2007年于解放軍信息工程大學(xué)獲學(xué)士學(xué)位,現(xiàn)為碩士研究生,主要研究方向為多載波通信中的同步技術(shù);
WANG Chong was born in Xiangyang,Hubei Province,in 1986. He received the B. S. degree from PLA Information Engineering University in 2007. He is now a graduate student. His research concerns synchronization techniques in multicarrier communications.
劉 兵(1979—),男,河南人,2001年于西安郵電大學(xué)獲學(xué)士學(xué)位,現(xiàn)為博士研究生,主要研究方向為無線通信;
LIU Bing was born in Henan Province,in 1979. He received the B.S. degree from Xi'an University of Posts and Telecommunications in 2001. He is currently working toward the Ph.D. degree. His research concerns wireless communications.
穆巍煒(1981—),男,遼寧錦州人,2005年于解放軍重慶通信學(xué)院獲學(xué)士學(xué)位,現(xiàn)為碩士研究生,主要研究方向為無線通信;
MU Weiwei was born in Jinzhou,Liaoning Province,in 1981. He received the B. S. degree from PLA Chongqing Institute of Communications in 2005. He is now a graduate student. His research concerns wireless communications.
徐錫燕(1991—),女,江蘇徐州人,2014年于南開大學(xué)獲學(xué)士學(xué)位,現(xiàn)為碩士研究生,主要研究方向為無線通信技術(shù);
XU Xiyan was born in Xuzhou,Jiangsu Province,in 1991. She received the B. S. degree from Nankai University in 2014. She is now a graduate student. Her research concerns wireless communications.
馬肖旭(1991—),女,重慶人,2014年于南開大學(xué)獲學(xué)士學(xué)位,現(xiàn)為碩士研究生,主要研究方向為無線通信技術(shù);
MA Xiaoxu was born in Chongqing,in 1991. She received the B.S. degree from Nankai University in 2014. She is now a graduate student. Her research concerns wireless communications.
李欣然(1990—),男,天津人,2013年于南開大學(xué)獲學(xué)士學(xué)位,現(xiàn)為碩士研究生,主要研究方向為無線通信技術(shù)。
LI Xinran was born in Tianjin,in 1990. He received the B. S. degree from Nankai University in 2013. He is now a graduate student. His research concerns wireless communications.
An Improved Timing and Frequency Synchronization Algorithm for Generalized Frequency Division Multiplexing Systems
WU Hong1,2,WANG Chong1,2,LIU Bing1,2,MU Weiwei1,2,XU Xiyan1,2,MA Xiaoxu1,2,LI Xinran1,2
(1.College of Electronic Information and Optical Engineering,Nankai University,Tianjin 300071,China; 2.Tianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology,Tianjin 300071,China)
According to the higher requirements of generalized frequency division multiplexing(GFDM) system for symbol timing synchronization,a synchronization algorithm is proposed based on a new preamble. At the receiver,after obtaining the coarse timing synchronization,the carrier frequency offset is estimated through the phase difference between the two halves of the preamble.Then the carrier frequency offset in the received signal is corrected. And a more accurate estimation of symbol timing offset is obtained by the cross-correlation with the transmitted preamble. Due to the conjugate symmetry characteristic of the preamble,this algorithm can avoid the plateau effect. In the low-latency high-reliability scenario of 5G,simulations are done in frequency-selective channels,and the performance of the scheme is evaluated in terms of mean squared error(MSE) of the frequency and time offset estimation. Theoretical analysis and simulation results show that the new algorithm has a better timing synchronization performance and can reduce the complexity of the system,thus enhancing the overall performance of system.
generalized frequency division multiplexing(GFDM);timing offset;carrier frequency offset;synchronization algorithm
10.3969/j.issn.1001-893x.2016.12.004
吳虹,王沖,劉兵,等.一種改進的GFDM時頻同步算法[J].電訊技術(shù),2016,56(12):1322-1326.[WU Hong,WANG Chong,LIU Bing,et al.An improved timing and frequency synchronization algorithm for generalized frequency division multiplexing systems[J].Telecommunication Engineering,2016,56(12):1322-1326.]
2016-05-18;
2016-07-11 Received date:2016-05-18;Revised date:2016-07-11
國家自然科學(xué)基金資助項目(61571244)
Foundation Item:The National Natural Science Foundation of China(No.61571244)
TN919.3
A
1001-893X(2016)12-1322-05
吳 虹(1967—),女,天津人,教授、博士生導(dǎo)師,主要研究方向為無線通信技術(shù);
**通信作者:wuhong@nankai.edu.cn Corresponding author:wuhong@nankai.edu.cn