13〇〇657 九、發明說明： 【發明所屬之技術領域】 本發明係關於—種應用於正交分頻多工系統 8 1 FreqUenCy DlVlslon Multiplexing，OFDM)中的 二維通道估計方法與I置，可以藉由發射端所傳送的 …:維丨訊號，利用所估計的都普勒頻率與訊號雜訊比，適 、=地選㈣㈣通道估計方法’可準確且有效率地估計通 k、:員率響應’以進行後續等化器與檢測器之處理。除此之 •二:：由二維導引訊號的辅助，本方法與裝置在良好的通道 衣兄中的通道估計，將不需任何運算。本發明所設計之都普 力頻率估計器與訊號雜訊比估計器亦可應用於非二维式導13 〇〇 657 IX. Description of the Invention: [Technical Field] The present invention relates to a two-dimensional channel estimation method and I set in an orthogonal frequency division multiplexing system 8 1 FreqUenCy DlVlslon Multiplexing, OFDM) The estimated: Doppler frequency and signal noise ratio can be used by the transmitter to transmit the ...:Voice signal, and the (four) (four) channel estimation method can be accurately and efficiently estimated. The rate response 'for subsequent processing of the equalizer and detector. In addition to this • 2:: assisted by the two-dimensional guidance signal, the method and device in the channel estimation of the good channel brother, will not require any operation. The universal frequency estimator and signal noise ratio estimator designed by the invention can also be applied to non-two-dimensional guides.
^號之正交分頻多卫系統，並可作為細胞交遞（Handover) 及適應性調變編碼技術（Adaptive M〇dulati〇n c〇d ，A 之依據。 ’ 【先前技術】 在無線行動通訊的環境中，會受到多重路徑（施叫她） ·#干擾、無線通道衰減（Fading)、頻率偏移、相位雜訊、及 白色同斯雜 §K ( Additive White Gaussian Noise，AWGN )等 因素的影響’而導致檢測位元的錯誤1 了克服無線通道的 影響，在純端必須料通道的特性，㈣㈣位訊號處理 的技術補償無線通道的影響。此—補償的方法與裝置就是所 謂的等化器（Equalize〇。通道估計的方法，依其所計算的 領域可分為時間領域與頻率領域通道估計方法。在時間領域 上’我們要估計的是脈衝響應（ImpulseResp〇nse)，也就是 1300657 多重路徑在時間上的位置及其大小。一般是利用訓練序列 (Training Sequence )或是導引通道（pil〇t Channel)來估計 脈衝響應，前者的例子為第二代行動電話（G1〇balSystemf^ Mobile，GSM )，後者的例子為第三代行動電話（UniversaiThe Orthogonal Frequency Division Multi-Guard System of the ^ number can be used as a basis for cell handover (Handover) and adaptive modulation coding (Adaptive M〇dulati〇nc〇d, A. ' [Prior Art] in wireless mobile communication In the environment, it will be subject to multiple paths (casting her) · #干扰, wireless channel attenuation (Fading), frequency offset, phase noise, and white sK (Additive White Gaussian Noise, AWGN) The influence 'causes the error of the detection bit 1 to overcome the influence of the wireless channel, the characteristics of the channel must be expected at the pure end, (4) (four) the technology of the bit signal processing compensates for the influence of the wireless channel. This method and device for compensation is so-called equalization Equalize〇. The method of channel estimation can be divided into time domain and frequency domain channel estimation methods according to the domain it calculates. In the time domain, we want to estimate the impulse response (ImpulseResp〇nse), which is 1300657 multiple. The position of the path in time and its size. Generally, the training sequence (Training Sequence) or the pilot channel (piltt Channel) is used to estimate the impulse response. The former example is the second generation mobile phone (G1〇balSystemf^ Mobile, GSM), and the latter example is the third generation mobile phone (Universai)
Mobile Telecommimications System，UMTS)。由於在接收端 • 所接收到的訊號是發射訊號與通道脈衝響應的旋積分 ' (Convolution)，所以在估計出通道脈波響應之後，需利用 ⑩ 匹配率波器（Matched Filter)加上viterbi Decoder，以補償 通道的效應，這也就是在第二代行動電話中的作法。至於第 二代行動電話，由於它是屬於分碼多工進接技術（c〇de Division Multiple Access，CDMA)，所以它採用犁耙式接收 機的条構並利用虛擬雜訊序列（pseu(j〇 Noise，PN )的特性， 將各個多重路徑平行等化之。 在頻率領域上，我們要估計的是頻率響應（FrequenCy 春 Response )，也就是脈衝響應在頻率領域上的分佈。由於時間 7貝域上的旋積分相當於頻率領域上的乘法，所以通道頻率響 應估計出來之後，可以將接收訊號直接除以通道頻率響應， 即可消除通道的效應，這也就是所謂的Zer〇-F〇rcing技術。 至於在通道頻率響應估計部分，一般是藉由訓練序列或是放 置在正交分頻多工符元中間導引訊號的輔助。一般習知的方 法包括有線性估計法（Linear Interpolation )、最大相似法 6 1300657 (Maximum Likelih〇〇d，肌）、與最小均方差法（Μ-⑽以Mobile Telecommimications System, UMTS). Since the received signal at the receiving end is the Convolution of the transmitted signal and the channel impulse response, after estimating the channel pulse response, you need to use the 10 Matching Filter plus the viterbi Decoder. To compensate for the effects of the channel, this is what is done in the second generation of mobile phones. As for the second generation mobile phone, since it belongs to the c〇de Division Multiple Access (CDMA) technology, it adopts the structure of the ploughshare receiver and utilizes the virtual noise sequence (pseu(j〇 The characteristics of Noise, PN), the equalization of each multipath is parallelized. In the frequency domain, we want to estimate the frequency response (FrequenCy Spring Response), that is, the distribution of the impulse response in the frequency domain. The upper rotation integral is equivalent to the multiplication in the frequency domain, so after the channel frequency response is estimated, the received signal can be directly divided by the channel frequency response, and the effect of the channel can be eliminated. This is the so-called Zer〇-F〇rcing technique. As for the channel frequency response estimation part, it is generally assisted by the training sequence or by placing the signal in the middle of the orthogonal frequency division multiplex symbol. Generally, the conventional method includes linear interpolation (Linear Interpolation), maximum Similarity method 6 1300657 (Maximum Likelih〇〇d, muscle), and minimum mean square error method (Μ-(10)
Mean Squared Error，MMSE)等。 由於在無線仃動通訊的環境中，會遭遇到背景雜訊舆干 擾訊號的影響，而使通道頻率響應估計不準_。除此之外， 通道頻率響應估計方法的複雜度也是設計上—項相當重要 的考里目則現有針對正交分頻多工系統所設計的通道頻率 響應估計方法與裝置’有的方法可以準確地估計通道頻率響 應但其複雜度相對地非常高，有的方法實作上較為簡單μ 低訊號雜訊比的通道中效能不好。因此，為了能兼顧高效能 與低4炅雜度的要求，雲Φ古.__ ΐτη —τ- ^而要有個可靠的通道頻率響應估計方 法與裝置以面對不同的盔繞雷瑷立、，^r上， 、， 个]的“、、踝兄，亚可減少收發機功率的 消耗並延長電池的使用時間。 【發明目的】 本發明主要的目的是針對正交分頻多工系統，提出一個 適應性、高效能、且低複雜度的通道頻率響應估計方法與裝 置。在正交分頻多工系統的發射端採用二維導引信號分配 器，將導引訊號依照-定的規則分配至時間領域及頻率領域 :。在正交分頻多工系統的接收端，包括一個訊號雜訊比估 “利用接收到的-維導引訊號以估計接收訊號的資料〜 量=訊能量的比率，一個都普勒頻率估計器利用接二 銘曰導引《以估計接收訊號在—個訊框中的都普勒頻率 里’-個適應性通道估計控制器利用都普勒頻率估計器 ^ 1300657 所估計到的都普勒頻率及訊號雜訊比估計器所估計到的訊 號雜訊比來決定通道估計方法，—個適應性通道估計器根據 適應性通道估計控制器的控制訊號進行通道估計，以及一個 -維導？H§ 5虎暫存器用以儲存前後幾個QFDM符元的導引信 號以供適應性通道估計器使用。 【發明内容】 -- 纟發明所提供之適應性二維通道估計方法與裝置包括 籲-個訊號雜訊比估計器、一個都普勒頻率估計器、一個適岸 性通道估計控制器、-個適應性通道估計器、以及—個二維 導引信號暫存器。在接收端，接收到的訊號經接收前端處理 及傅利㈣換之後，可以抽出安排在正交分頻多卫符元㈣ 導引H ^㈣。導引訊號會同時被送人二維導引件號暫 存器、都普勒頻率估計器、及訊號雜訊比估計器。二^引 信號暫存ϋ會依據不同系統的規格，儲存前後若干個正交分 頻多工符元内的導引訊號。目前正交分頻多工符元内的導引 訊號為-維導引訊號，若加上前後若干個正交分頻多工符元 •内的導引訊號則形成二維導引訊號。'維導引訊號可用以估 計訊號雜訊比，二維導引訊號可用以估計都普勒頻率。此 外，兩者可同時用來估計通道頻率響應。 【實施方式】 ^ 請參閱圖一，原始資料經過通道編碼及交錯器處理之 後’會經過調變器處理將二進位的資料調變成高階資料，例 如 QPSK(Quadrature Phase shift 與 齡Mean Squared Error, MMSE), etc. In the environment of wireless passive communication, the influence of background noise and interference signals will be encountered, and the channel frequency response estimation is not allowed. In addition, the complexity of the channel frequency response estimation method is also quite important in the design-item. The existing channel frequency response estimation method and device designed for the orthogonal frequency division multiplexing system can be accurate. The channel frequency response is estimated to be relatively high, and some methods are relatively simple to implement. The channel with low signal-to-noise ratio is not effective. Therefore, in order to balance the requirements of high efficiency and low 4炅, cloud Φ ancient.__ ΐτη —τ- ^, there must be a reliable channel frequency response estimation method and device to face different helmets, , ^,上,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, An adaptive, high-efficiency, and low-complexity channel frequency response estimation method and device are proposed. A two-dimensional pilot signal distributor is used at the transmitting end of the orthogonal frequency division multiplexing system to guide the pilot signal according to the rule. Assigned to the time domain and frequency domain: At the receiving end of the orthogonal frequency division multiplexing system, including a signal noise ratio estimation "Using the received - dimensional steering signal to estimate the received signal data ~ quantity = signal energy Ratio, a Doppler frequency estimator uses the second guide to "estimate the received signal in the Doppler frequency of the frame" - an adaptive channel estimation controller using the Doppler frequency estimator ^ 1300657 Calculating the channel estimation method by the measured signal noise ratio estimated by the Doppler frequency and the signal noise ratio estimator, an adaptive channel estimator performs channel estimation according to the control signal of the adaptive channel estimation controller, and One-dimensional guide? The H§ 5 Tiger Register is used to store the leading and trailing signals of several QFDM symbols for use by the Adaptive Channel Estimator. SUMMARY OF THE INVENTION - The adaptive two-dimensional channel estimation method and apparatus provided by the invention include a signal-to-noise ratio estimator, a Doppler frequency estimator, a land-based channel estimation controller, and a An adaptive channel estimator, and a two-dimensional pilot signal register. At the receiving end, after receiving the received signal and receiving the Fourier (four), the received signal can be extracted in the orthogonal frequency division multi-guard symbol (4) to guide H ^ (four). The pilot signal is also sent to the 2D guide number register, the Doppler frequency estimator, and the signal noise ratio estimator. After the signal is temporarily stored, the pilot signals in several orthogonal frequency division multiplex symbols are stored according to the specifications of different systems. At present, the pilot signal in the orthogonal frequency division multiplex symbol is a -dimensional steering signal, and if the pilot signals in the plurality of orthogonal frequency division multiplex symbols are added before and after, a two-dimensional pilot signal is formed. The 'divisional pilot signal can be used to estimate the signal-to-noise ratio, and the two-dimensional pilot signal can be used to estimate the Doppler frequency. In addition, both can be used to estimate the channel frequency response. [Embodiment] ^ Please refer to Figure 1. After the original data is processed by channel coding and interleaver, the data of the binary is converted into high-order data by the modulator processing, for example, QPSK (Quadrature Phase shift and age).
AmpHtude Modulati〇n)。之後，再將調變後的資料安排至正 • 1300657 交分頻多工系統所提供的次載波，以進行後續的頻域/時域轉 換。在這個部分，導引訊號也同時被安排至與資料不同的次 載波，以輔助接收端進行通道估計。次載波安排好之後，便 進入反快速傅利葉轉換器（Inverse Fast F〇urier 丁咖？嶋， IFFT )進行頻域/時域轉換’產生正交分頻多工符元 (S一♦接下纟，將正交分頻多工符元送入加入保護區間 方塊中’以在正交分頻多卫符元前面放置—個循環區間。此 -循環區間是從正交分頻多工符元後面的一部份所複製產 生，用以抵抗無線電通道多重路徑所造成的符元間干擾 (Inter-Symbol Interference 1ST、。田 a e，UI)。取後，經發射前端處理之 後，便由發射天線傳送出去。在接收端，經過接收天線接收 下來之後’便交由接收前端處理，包括像訊號放大、遽波、 祝框同步等處理。接下來，再移除放置在正交分頻多工符元 前面的循環區間。經過快速傅利葉轉換器㈤F_ier ^叫附）進行時域/頻域轉換之後，一方面抽出導引 訊號以進行通道頻率響應之估 〇另一方面將資料送入等化 器以利用估計所得的通道頻率響應來補償通道效應。最後， 專化之後的謂便可送人檢丨財塊進行解 解碼，以回復原來的資料。本 又，曰及 、3 β π t x月疋針對導引訊號的設計盥 通道頻率響應估計，提出一 ^ 一、 的新方法。 週應性、向效能、且低複雜度 9 1300657 凊參閱圖二，本發明所提出的適應性二維通道估計方法 與裝置中，適應性二維通道估計方法與裝置包括―個訊號雜 訊比估計112(^、-個都普勒頻率估計器2()2一個適應性通 道估計控制11204、-個適應性通道估計_5、以及—個二 維V引仏唬暫存器203。在接收端，接收到的訊號經接收前 端處理及傅利葉轉換之後，可以抽出安排在正交分頻多工符 元内的導引訊號，心⑽。導引訊號會同時被送入二維導引 ϋ暫存II2G3、都普勒頻率估計㈣2、及訊號雜訊比估計 器2 0 i。二維導引信號暫存器2 Q 3會依據不同系統的規格，儲 存前後若干個正交分頻多工符元内的導引訊號。目前正交分 頻多工符7L内的導引訊號為_維導引訊號，若加上前後若干 個正交分頻多卫符元内的導引訊號則形成二維導引訊號。— 維導引訊號可用以估計訊號雜訊比，二維導引訊號可用以估 計都普勒頻率。此外，兩者可同時用來估計通道頻率響應。 根據本發明的内容’目前正交分頻多工符元内的導引訊 號會被送入訊號雜訊比估計器2〇1以估計訊號目前的訊號雜 訊比。由於導引訊號為-已知的序列，所以將導引訊號乘上 其共輛數再求其平均值，并亚 此一平均值的平方便是目前正交分 頻多工符元的平均能量仕士+ 里估#值。若直接計算導弓丨訊號均方值 (Mean Square )，可以得丨曰乂卞丄 J目刚正父分頻多工符元與雜訊的 平均能量和。根據上述兩個估計值，我們便可以得出目前正 10 1300657 法.求大相似性法（Maximum Likelihood，ML )、一維線性内 差法（Linear Interpolation )、及二維通道估計法。在低訊號 雜訊比的通道環境中，採用高效能與高複雜度的最大相似性 法，以確保通訊的品質。在這種情況下，是用較高複雜度的 運异來換取較高效能的通道估計。在較高訊號雜訊比且高都 普勒頻率的通道環境中，可採用一維線性内差法或最大相似 性法。一維線性内差法一方面會犧牲一點效能，另一方面可 以大幅降低通道頻率響應估計演算法的複雜度。在較高訊號 雜訊比且低都普勒頻率的通道環境中，採用二維通道估計 法。二維通道估計法是利用冑近正交分頻多卫符元内的導引 訊號’估計目前正交分頻多卫符元内相對應資料次載波的通 道頻率響應。若發射端所傳送的導引訊號A ±1的序列的 話，則將鄰近正交分頻多工符元内的導引訊號乘上對應的土 1即疋估计所得的通道頻率響應、。在這種情況下，—方面可 以達到與最大相似性法相近的效能，另—方面其複雜度比一 維線性内差法低。二維通道估計法除了需要記憶鄰近正交分 頻多工符元内的導引訊號外，不f要進行複雜運算。 ，根據本發明的内容，二維導引信號暫存器2G3負 鄰近正交分頻多工符元内的導引訊號，以作為都普勒頻率估 道估計法之用。二維導引信號暫存器203儲存的 U 、於正父分頻多4統的規格。二維導引信號暫存器 12 1300657 203為一個二維的暫存器，·其中一 、、隹用以代表某一個時間内 的導引訊號，其長度為正交分頻吝 刀頻夕工付兀内導弓I訊號的數 目；其中-維用以代表不同時間㈣㈣訊號，其大小為正 交分頻多工符元内導引訊號的間隔。 Μ 本發明針對正交分頻多4統，提出—個適應性、高效 能、且低複雜度的通道頻率響應估計方法與裝置。圖三是正 交分頻多工符元内次載波配置例子。由圖中我們可以知：， 每隔-段固定長度便會安插—個導引訊號，以做為通道：同 步估計之用。若㈣㈣越彡’貞彳解析度越高，通道頻率響 應的估計越準’㈣統的通透率（Th刚咖ut)也越低。反 之，若導引訊號越少，料析度越低，通道頻㈣應的估計 越不準，但系統的通透率（Throughput)也越高1以，實 際的應用要根據系統的規格與應用的環境而言曼計。在這個例 子中’為了說明的目的’我們不考慮正交分頻多工符元内兩 端的保護頻段（Guard Band)，並且將所有的次載波配置給原 始資料與導引訊號。實際應用的系統都會加入保護頻段，以 避免相鄰頻道的干擾（Adjacent Channel Interference，Aci)。 在正交分頻多工系統的發射端採用二維導引信號分配 為’將導引訊號依照一定的規則分配至時間領域及頻率領域 上。圖四在發射端分配正交分頻多工符元内二維導引訊號的 -個例子。在這個例子中，正交分頻多工符元内導引訊號的 13 !300657 間隔為8。如果正夺八此a μ 刀頻多工付元内的次載波數目為2048 的話，則正交分頻多鈐 付凡内的導引訊號的數目為256。由 圖四我們可以知道，時 /丄7、、曾 $間化+ "導引訊號的位置等於是時間 η往右循環位移1個4 #、木 + 卜 调-人载波，時間0 + 2)導引訊號的位置 等於是時間π往太低辑，^ 士 循衣位移2個次載波，以此類推。到了 時間0 +⑴的時候，道& ^ 了佚導引訊號的位置和時間”是一樣的。 循環位移的意思就是告 疋田_人載波位移超出2048時，其位置會 等於除以2CH8的铪赵山狄言w J踩数。由於導引訊號同時安排在時間與頻 員域上戶斤以稱之為二維導引訊號。藉由二維導引訊號的 、 可以估计都普勒頻率及進行二維通道頻率響應估計。 在接收端，接收到的訊號經接收前端處理及傅利葉轉換 後可以抽出安排在正交分頻多工符元内的導引訊號， P(n’k)’如圖—所示。導引訊號會同時被送人訊號雜訊比估 口十杰201、都普勒頻率估計器2〇2、及二維導引信號暫存器 2〇3 ’如圖二所示。二維導引信號暫存器203會依據不同系統 的規格’儲存前後若干個正交分頻多工符元内的導引訊號。 在^虎雜afUb估計器中2G1，利用—維導引訊號估計訊號雜 汛比’如圖五所示。假設在時間”由發射端所產生的導引 訊说序列為iP〇7j)，，，2)，... .. 則在接收端在 時間乃抽取出來的導引訊號序列心化幻為 ⑴AmpHtude Modulati〇n). After that, the modulated data is arranged to the secondary carrier provided by the 1300657 crossover multiplex system for subsequent frequency/time domain conversion. In this part, the pilot signal is also arranged to a secondary carrier different from the data to assist the receiver in channel estimation. After the subcarriers are arranged, they enter the inverse fast Fourier transform (IFFT) for frequency domain/time domain conversion to generate orthogonal frequency division multiplex symbols (S1 ♦ followed by 纟, the orthogonal frequency division multiplex symbol is sent into the guard interval block to be placed in front of the orthogonal frequency division multi-guard symbol - a loop interval. This loop interval is from the orthogonal frequency division multiplex symbol A part of the copy is generated to resist the inter-symbol interference (Inter-Symbol Interference 1ST, UI ae, UI) caused by the multipath of the radio channel. After being processed, it is transmitted by the transmitting antenna after being processed by the transmitting front end. Going out. At the receiving end, after receiving it by the receiving antenna, it is processed by the receiving front end, including processing such as signal amplification, chopping, and frame synchronization. Next, it is removed and placed in the orthogonal frequency division multiplex symbol. The previous cycle interval. After the time domain/frequency domain conversion by the fast Fourier transform (5) F_ier ^ call), the pilot signal is extracted on the one hand to estimate the channel frequency response, and the data is equalized on the other hand. Effects in order to compensate for the channel using the estimated channel frequency response obtained. Finally, after the specialization, the person can be sent to check the block for decoding and decoding to restore the original data. In addition, 曰 and 3 β π t x 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋Circumference, performance, and low complexity 9 1300657 凊 Referring to FIG. 2 , the adaptive two-dimensional channel estimation method and apparatus provided by the present invention include a “signal noise ratio” Estimate 112 (^, - Doppler frequency estimator 2 () 2 an adaptive channel estimation control 11204, - an adaptive channel estimate _5, and - a two-dimensional V 仏唬 register 203. After receiving the front-end processing and Fourier transform, the received signal can be extracted from the pilot signal (10) arranged in the orthogonal frequency division multiplex symbol. The pilot signal is simultaneously sent to the two-dimensional guide. Save II2G3, Doppler frequency estimation (4) 2, and signal noise ratio estimator 2 0 i. 2D pilot signal register 2 Q 3 will store several orthogonal frequency division multiplexers before and after according to different system specifications. The pilot signal in the element. At present, the pilot signal in the orthogonal frequency division multiplex symbol 7L is a _ dimensional pilot signal, and if the pilot signals in several orthogonal frequency division multi-guard symbols are added before and after, the second signal is formed. Dimension guidance signal. - Dimensional guidance signal can be used to estimate the signal noise ratio The two-dimensional pilot signal can be used to estimate the Doppler frequency. In addition, both can be used to estimate the channel frequency response. According to the content of the present invention, the pilot signal in the current orthogonal frequency division multiplex symbol is sent. The signal noise ratio estimator 2〇1 estimates the current signal-to-noise ratio of the signal. Since the pilot signal is a known sequence, the pilot signal is multiplied by its total number of vehicles and then averaged. The average convenience of the average value is the average energy of the current orthogonal frequency division multiplex symbol. The value of the average value is 1. If you directly calculate the mean square value (Mean Square), you can get 丨曰乂卞丄J According to the above two estimates, we can get the current 10 1300657 method. Maximum Likelihood (ML), one-dimensional linearity Linear Interpolation and two-dimensional channel estimation method. In the channel environment with low signal noise ratio, the maximum similarity method of high efficiency and high complexity is adopted to ensure the quality of communication. Is using a higher complexity In exchange for higher-efficiency channel estimation. In the channel environment with higher signal-to-noise ratio and high Doppler frequency, one-dimensional linear internal difference method or maximum similarity method can be used. One-dimensional linear internal difference method will be used. Sacrificing a bit of performance, on the other hand, can greatly reduce the complexity of the channel frequency response estimation algorithm. In the channel environment with higher signal noise ratio and low Doppler frequency, the two-dimensional channel estimation method is adopted. The pilot signal in the near-orthogonal crossover multi-guard symbol is used to estimate the channel frequency response of the corresponding data subcarrier in the current orthogonal frequency multi-guard symbol. If the pilot signal A is transmitted by the transmitting end, If the sequence of 1 is used, the pilot signal in the adjacent orthogonal frequency division multiplex symbol is multiplied by the corresponding channel 1 and the estimated channel frequency response. In this case, the effect can be achieved similar to the maximum similarity method, and the complexity is lower than the one-dimensional linear internal difference method. In addition to the need to memorize the pilot signals in adjacent orthogonal frequency division multiplex symbols, the two-dimensional channel estimation method does not require complex operations. According to the content of the present invention, the two-dimensional pilot signal register 2G3 is negatively adjacent to the pilot signal in the orthogonal frequency division multiplex symbol for use as a Doppler frequency estimation method. The U-stored by the two-dimensional pilot signal register 203 is divided into the specifications of the positive-father frequency division. The two-dimensional pilot signal register 12 1300657 203 is a two-dimensional register, one of which is used to represent the pilot signal in a certain time, and the length thereof is an orthogonal frequency division. The number of the inner guide I signals; the dimension is used to represent the different time (four) (four) signals, and the size is the interval of the pilot signals in the orthogonal frequency division multiplex symbol. Μ The present invention proposes an adaptive, high-performance, and low-complexity channel frequency response estimation method and apparatus for orthogonal frequency division multiple systems. Figure 3 is an example of subcarrier configuration in a crossover multiplex symbol. As can be seen from the figure, a fixed-length signal is inserted every other segment to serve as a channel: synchronization estimation. If (4) (4), the higher the resolution, the more accurate the channel frequency response is, and the lower the penetration rate (Th just coffee ut). Conversely, if the pilot signal is less, the lower the resolution, the more accurate the channel frequency (4) should be, but the higher the system's throughput (1), the actual application should be based on the system specifications and applications. The environment is manned. In this example, for the purpose of illustration, we do not consider the guard bands at both ends of the orthogonal frequency division multiplex symbol, and configure all the secondary carriers to the original data and pilot signals. The actual application system will be added to the protection band to avoid interference (Adjacent Channel Interference, Aci). At the transmitting end of the orthogonal frequency division multiplexing system, the two-dimensional pilot signal is allocated as 'the pilot signal is allocated to the time domain and the frequency domain according to certain rules. Figure 4 shows an example of distributing a two-dimensional pilot signal in an orthogonal frequency division multiplex symbol at the transmitting end. In this example, the 13 !300657 interval of the pilot signals in the orthogonal frequency division multiplex symbol is 8. If the number of subcarriers in the octave multiplex multiplexer is 2048, the number of pilot signals in the Orthogonal Frequency Division is 256. From Figure 4 we can know that the time / 丄 7, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , The position of the pilot signal is equal to the time π to too low, ^ 士衣衣shift 2 subcarriers, and so on. When time 0 + (1) is reached, the position of the track & ^ 佚 pilot signal is the same as the time. The meaning of the cyclic displacement is to tell the field _ when the carrier displacement exceeds 2048, its position will be equal to 2CH8. Zhao Shan Di Yan w J step count. Since the pilot signal is arranged in the time and frequency domain, it is called a two-dimensional pilot signal. By two-dimensional guidance signal, the Doppler frequency can be estimated. The two-dimensional channel frequency response estimation is performed. At the receiving end, the received signal can be extracted by the receiving front end processing and the Fourier transform to extract the pilot signal arranged in the orthogonal frequency division multiplex symbol, P(n'k)' Figure - The pilot signal will be sent at the same time as the signal noise ratio estimator 10 Jie 201, Doppler frequency estimator 2 〇 2, and 2D pilot signal register 2 〇 3 ' The two-dimensional pilot signal register 203 will store the pilot signals in several orthogonal frequency division multiplex symbols before and after the storage according to the specifications of different systems. In the ^hu afUb estimator 2G1, use-dimensional guidance The quotation number estimates the signal noise ratio as shown in Figure 5. Assume at time Exit end of the guide information produced by said sequence iP〇7j) ,,, 2), ... .. at the receiving end is the time the pilot signal extracted sequences to the heart of the magic ⑴
Rp(n，k、= H(n，k、P(n，k) + W(n，k)，\m K 14 1300657 其中TV代表導引訊號的週期，Μ代表平均的長度， &㈨代表相隔5個正交分頻多工符元内導引訊號的差異 陡代表導引訊號的平均能量。經過數學的推導，可以 #到都普勒頻率的估計值如下 — Yti:TsTpNJ° ~⑻)， (8) 其中L代表取樣週期，#是正交分頻多i符元内次載波 的數目，V是 zeroorder Bessel funeti()n 〇f 加 fim “Μ 的 反函數1可以由查表得到。為了縮短儲存導引訊號的長 度’並降低運算的複雜度，本發明提出下列簡&⑺式的 ⑼ 其中〇<r<i是用以控制羋仏且由以 .f ,]千均長度的洩漏係數（Leaky c 0Γ)。當DC〇)不大的時候，（9)與 在 /、 式為一對一的關 係。所以，在實際應用上 外信品 不而要求出貫際都普勒頻率的估 前都普勒頻率的高低。 由⑼式的結果來判斷目 根據估計所得的訊號雜訊比 估計控制H2G4決定目前 W，適應性通道 木取何種通道頻率響應估計方 16 1300657 法。圖七是適應性通道估計控制器204運作流程圖。首先， 在接收端經過快速傅利葉轉換之後，可以取出安置在正交分 頻多工符元内的導引訊號。之後，根據在訊號雜訊比估計器 201中所得到的訊號雜訊比值進行通道頻率響應估計方法之 選擇。如果目前通道環境屬於低訊號雜訊比（例如小於i 5Rp(n,k,=H(n,k,P(n,k) + W(n,k),\m K 14 1300657 where TV represents the period of the pilot signal, Μ represents the average length, & (9) The difference between the pilot signals representing the five orthogonal frequency division multiplex symbols is the average energy of the pilot signal. After mathematical derivation, the estimated value of the # to Doppler frequency is as follows—Yti:TsTpNJ° ~(8) ), (8) where L represents the sampling period, # is the number of subcarriers within the orthogonal frequency division multi-symbol, V is zeroorder Bessel funeti()n 〇f plus fiim "Μ inverse function 1 can be obtained by looking up the table In order to shorten the length of storing the pilot signal' and reduce the complexity of the operation, the present invention proposes the following (9) (9) where 〇<r<i is used to control 芈仏 and is controlled by .f,] The leakage coefficient of the length (Leaky c 0Γ). When DC〇) is not large, (9) has a one-to-one relationship with the /, and therefore, in practical applications, the foreign letter does not require a continuation. The Pu'er frequency is estimated by the level of the Doppler frequency. The result of the equation (9) is used to determine the target noise-to-noise ratio estimation control H2G4 decision. W, which channel frequency response estimation method 16 1300657 method is adopted. Figure 7 is a flow chart of the operation of the adaptive channel estimation controller 204. First, after the fast Fourier transform at the receiving end, it can be taken out and placed in orthogonal points. The pilot signal in the frequency multiplex symbol. After that, the channel frequency response estimation method is selected according to the signal noise ratio obtained in the signal noise ratio estimator 201. If the current channel environment is a low signal noise ratio ( For example less than i 5
dB )，則採用咼效能與高複雜度的最大相似性法，以確保通 訊的品質。如果目前通道環境屬於高訊號雜訊比（例如大於 15 dB )，則繼續進行下一階段的判斷。利用都普勒頻率估 計器202中所得到都普勒頻率估計值，可以判斷目前通道變 化的速度。如果都普勒頻率較高（例如大於l〇〇Hz)，則 表示目前通道變化較快，所以可採用—維線性内差法或最大 相似〖生法 維線性内差法一方面會犧牲一點效能，另一方 面可以大幅降低通道頻率響應估計演算法的複雜度。如果都 曰勒頻率車乂低（例如小於i 〇〇 Hz )，則表示目前通道變化 較慢’所以可採用：維通道估計法。二維通道估計法是利用 鄰近正交分頻多工符元内的導引訊號，估計目前正交分頻多 '夺元内相對應貝料次載波的通道頻率響應，如圖四中的箭 頭所示。若發射端所傳送的導引訊號$ ±1㈣列的話，則 將鄰近正交分T „ 、 付元内的導引訊號乘上對應的± 1即 是估計所得的通道頻率響應。在這種情況下，一方面可以達 到與最大相似性法相w a k的效能，另一方面其複雜度比一維線 17 1300657 了需要記憶鄰近正交分頻多 進行複雜運算。最後，利用 通道效應，也就是所謂的等 性内差法低。二維通道估計法除 工符元内的導引訊號外，不需要 估計所得的通道頻率響應來補償 化處理（Equalization )。 導引信號儲存在一個二維的暫 J ’存為中，其中一維用以代 表某一個時間内的導引訊號，复 ，、長度為正交分頻多工符元内 導引訊號的數目；其中一維用以 卞用以代表不同時間内的導引訊dB), the maximum similarity between 咼 performance and high complexity is used to ensure the quality of communication. If the current channel environment is a high signal noise ratio (eg greater than 15 dB), the next phase of the decision is continued. Using the Doppler frequency estimate obtained in the Doppler frequency estimator 202, the speed of the current channel change can be determined. If the Doppler frequency is high (for example, greater than l〇〇Hz), it means that the current channel changes faster, so the linear-internal difference method or the maximum similarity method can be used to sacrifice some performance. On the other hand, the complexity of the channel frequency response estimation algorithm can be greatly reduced. If the 曰 频率 frequency is low (for example, less than i 〇〇 Hz ), it means that the current channel changes slowly ‘ so the channel estimation method can be used. The two-dimensional channel estimation method uses the pilot signals in the adjacent orthogonal frequency division multiplex symbol to estimate the channel frequency response of the corresponding orthogonal subcarriers in the current orthogonal frequency multi-band, such as the arrow in FIG. Shown. If the pilot signal transmitted by the transmitting end is $±1 (four), then the adjacent channel T „ and the pilot signal in the pay element are multiplied by the corresponding ± 1 to estimate the channel frequency response. On the one hand, it can achieve the performance of the maximum similarity method wak, on the other hand its complexity is more complex than the one-dimensional line 17 1300657 requires memory adjacent orthogonal frequency division. Finally, the channel effect is also called The equal-internal difference method is low. The two-dimensional channel estimation method does not need to estimate the channel frequency response to compensate for the equalization signal in the worker symbol. The pilot signal is stored in a two-dimensional temporary J. In the 'save as one, one dimension is used to represent the number of pilot signals in a certain time, complex, and the length is the number of pilot signals in the orthogonal frequency division multiplex symbol; one dimension is used to represent different Time guidance
號，其大小為正交分頻冬JL ίΧ i Hr·、# 又刀夕貝夕工付7L内導引訊號的間隔。 【特點及功效】 本么明針對正父分頻多工系統提出一種適應性二維通 道估計方法與裝置。此—方法與裝置可以根據無線電通道的 環境，在適應性通道估計控制器的控制下，適應性地切換不 同的通道估計方法，以提昇訊號檢測品f並降低接收機複雜 度。在系統複雜度與品質之間，取得最佳的平衡點。 【圖式簡單說明】 為了使貴審查委員能進一步了解本發明之特點及技 術内容，請參閱以下有關本發明之附圖及詳細說明，然而所 附圖式僅提供參考與說明用，並非用來對本發明加以限制。 有關該實施例之附圖為： 圖一為正交分頻多工系統的方塊圖； 圖二為適應性二維通道估計方法與裝置方塊圖； 圖三為正交分頻多工符元分配圖； 18 1300657 圖四為二維導引信號分配圖； 圖五為訊號雜訊比估計器方塊圖； 圖六為都普勒頻率估計器方塊圖；以及 圖七為適應性通道估計控制器運作流程圖。 【主要元件符號說明】 201 訊號雜訊比估計器 202 都普勒頻率估計器 203 二維導引信號暫存器 204 適應性通道估計控制器 205 適應性通道估計器No., the size of the orthogonal crossover winter JL Χ H i Hr·, # 刀 夕 夕 夕 夕 夕 夕 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 [Features and Efficacy] This paper proposes an adaptive two-dimensional channel estimation method and device for the positive-father frequency division multiplexing system. The method and the device can adaptively switch different channel estimation methods under the control of the adaptive channel estimation controller according to the environment of the radio channel to improve the signal detection product f and reduce the receiver complexity. Get the best balance between system complexity and quality. BRIEF DESCRIPTION OF THE DRAWINGS In order to provide a further understanding of the features and technical contents of the present invention, the following drawings and detailed description of the present invention are provided. The invention is limited. The drawings of the embodiment are as follows: Figure 1 is a block diagram of an orthogonal frequency division multiplexing system; Figure 2 is a block diagram of an adaptive two-dimensional channel estimation method and apparatus; Figure 3 is an orthogonal frequency division multiplexing symbol allocation Figure 18; 1300657 Figure 4 is a two-dimensional pilot signal distribution diagram; Figure 5 is a signal noise ratio estimator block diagram; Figure 6 is a Doppler frequency estimator block diagram; and Figure 7 is an adaptive channel estimation controller operation flow chart. [Main component symbol description] 201 Signal noise ratio estimator 202 Doppler frequency estimator 203 Two-dimensional pilot signal register 204 Adaptive channel estimation controller 205 Adaptive channel estimator