1271970 九、發明說明： 【發明所屬之技術領域】 、本發明係揭露一鮮載波通訊系統及其通訊方法，尤指-種 可消除迴音的錄波通訊系統及其通訊方法。 【先前技術】 一般通訊系統可赠傳輸的錢齡為私、半雙工以及全 雙工系統。由於半雙^及全雙1韻可進行雙向傳輸，所以 下行的接收職難免會被上行的傳輪峨所預。_散多頻系 ^ (d1Scretemulti-tone system) , 利葉轉換將傳輸龍叫變域數個子載波上，然後將調變後所 產生的時域訊㈣送到傳輸介質上；當離散多頻系統欲接收一筆 貧料時’其也會將接收_時域訊號S2進行傅师轉換以產生一 解碼資料D，。然而，當離散多頻系統之類比前端續2線的合成 電路（Hybridcircuit)阻抗不匹配時，下行的時域訊號％就會被 到上行的時域訊號81所預，此即業界所習知的迴音（Μ。）。 請參閱第1圖’第1圖係為一離散多頻系統1〇的示意圖。如 圖中所示’離散多頻系統1G所設置的逆向傅利葉轉換單元22，會 用來將傳輸㈣D進行逆向傅繼觀^產生—時域訊號^，並 1271970 且輸出日守域訊號Si至傳輪濾波器26來進行濾波處理，接著，時 或札號S1便進一步地傳送給類比前端電路(丘⑽t) 28 ’最後由類比前端電路Μ將時域訊號S1送到外部傳輸介質（未 顯不）上。同樣地，離散多頻系統1〇可藉由類比前端電路28自 外部傳輸介質接收到-時域訊號&，接著，類比前端電路π便進 步地將時域訊號S2傳遞至接收遽波器％來進行後續濾波處 • 理。為了消除時域訊號81對時域訊號幻所造成的干擾（亦即時 域汛號S1的迴音），離散多頻系統1〇中設置有一迴音估測單元 • 24 ’用來依據時域訊號S1產生一迴音估測訊號Sech。，並且利用調 -整單元34將時域訊號S2減去迴音估測訊號Sech。以產生-時脈訊 號S2’，最後將時脈訊號S2,交由傅利葉轉換單$ 32進行傅利葉轉 換以產生解H料D’，如此-來，解碼資料D，即不會受到迴音的 干擾。 雖然上述的離散多頻系統1G可有效地消除系統巾的迴音，但 是由於時域訊號S1、S2皆具有較高的取樣頻率，因此在運算迴音 估測訊號sech。的過程中必須消耗相當可觀的運算資源。 【發明内容】 因此:本發明的主要目的之一在於提供一種可消除迴音的多 載波通訊系統，其可降低在消除迴音過程中所需的運算量。 8 1271970 依據本發明之實施例，其係揭露—種錄波通㈣統。該多 載波通訊系統包含有··一傳輸模組，用來依據一傳輸資料執行— 逆向傅利親算减生-第-時域峨·，—接收模組，用來接收 -第二時域訊號’並且依據鄕二時域訊錄行—侧葉運算以 產生-解碼資料；以及-迴音消除單元，墟至傳輸模組以及該 接收模組，用來依據複數個運算係數，該解碼資料以及該傳輸資 • 料產生-重建貧料，並且依據該重建資料調整該複數個運算係數。 依據本發明之實補，其另揭露—種朗於—多載波通訊系 統以肩除迴音的通訊方法。該通訊方法包含有··依據一傳輸資料 執仃逆向傅利葉運异以產生一第一時域訊號；傳送該第一時域 訊號；接收-第二時域訊號；依據該第二時域訊號執行一傅利葉 運算以產生-解碼資料；依據複數個運算係數，該解碼資料以及 鲁挪輸貝料產生-重建資料；以及依據該重建資料調整該複數個 運异係數。 “依據，發明之實施例’其亦揭露-種-種多載波通訊裝置，用 來於夕載波通訊系統中接收及傳送資料，該多載波收發器包含 有、、j域轉時域模組，用來將一欲傳送出去之第一頻域資訊轉 換為第時域訊號，以進行該多载波通訊系統中之一資料傳送 守域轉頻域模組，用來將—接收到之第二時域訊號轉換 1271970 為-第二頻域資訊，以進行該錄波通崎財之—資料接收動 作’以及-週音消除模組，減於該輯轉時域模組及該時域轉 頻麵組，依據該第二頻域資訊於頻域當中對該第一頻域資气進 行調整，以進行該多載波通訊裝置於傳送資料時之迴音消除動作。 【實施方式】 Φ 5月參閱第2 ®，第2圖係為依據本發明-實施例之多载波通 訊系統100的示意圖。本實施例中’多载波通訊系統1〇〇可為一 . 離散多頻（discrete multi-tone, DMT)系統或一正交分頻多工 (orthogonal frequenCy division multiplex< 以此為限。多載波通訊系統100中設置有一傳輸模组⑽、一接收 杈組14〇、一類比前端電路⑽以及一迴音消除單元18〇。傳輸模 組120係用來依據-傳輸資料D執行一逆向傅利葉運算以產生一 春日可域汛旒S1 ,接收模組14〇係用來接收一時域訊號S2，並且依據 Μ或碱S2執行-翻葉運算以產生—解碼資料5 ;類比前端電 路160係耦接至一傳輸介質（例如雙絞線） ，用來傳送時域訊號S1 或接收時域訊號S2 ;而迴音消除單元⑽制來依據請運算係 數w〇〜wM、傳輸資料D以及解碼資料乃產生一重建資料，使得 *建讀D’既可以反鱗碼資料β的資料内容又不會受到傳輸資 料D之迴音的干擾’其中k係代表多載波通訊系統1〇〇中所使用 的子波總數。需注意的是，雖然本實施例係以參考與子載波總 1271970 數相同數量之運算係數作為例子以進行說明，熟習此項技術者應 可理解，此雜絲之數量及運算之複雜度並非本發明之限制條 件’使用數量較少之運算係數及較簡單之運算過程亦包含於本發 明之範圍内。 一般而S，典型的傳輸模組120通常包含有執行如逆向傅利 葉轉換、循環字首（Cyclic_prefix)附加、平行轉序列、數位轉類 比及濾波等魏之硬體或軟體，或者上述功能區塊當巾之至少一 部伤’而典型的接收模組施則通常包含有執行如渡波、類比轉 數位、序_平行、循環字首移除、及侧葉轉換等功能之硬體 或軟體，A者上述功能區塊#中之至少—部份。而傳輸模組12〇、 接收权組140、及類比前端電路16〇之運作原理均為熟習此項技術 所廣泛悉知，故不在此贅述。 迴音消除單元180中另設置有一迴音估測單元182以及-調 整單元184。迴音估測單元182係用來依據複數個運算係數 對傳輸貝料D進行運算以產生—迴音估測訊動，織經由調整 單元184來將解碼資料β減去迴音估測訊號^以產生重建資料 D明/主思，本實施例中，調整單元184係由一減法器來加以實 施以調整解碼資料β。本實施例中，迴音估測單元搬另會依據重 建資料D執行一最小均方（leastmeansquare，LMS)運算來進一 11 1271970 的气，使物轉㈣可藉由調整後 與傳輪資料D產生所要的迴音估測訊勤，以 2傳輸感D所私細音。上舰騎小财縣調整運算 係數W〇〜wk4的詳細操作將 最小說明，請注意，除了使用 :# ’亦可以利用其他適應訊號處理方法，例如 RLS、Slgned LMS等，來調整運算係數i 例為限。 a 請繼續參閱第3圖，第3圖係為第2圖所示之迴音估測單元 182之第一實施例的示意圖。如目中所示，迴音估測單元182中包 =複數個運算單元192、194、、196以及一加法單元198。運 單元192、194、…、196分別用來產生迴音估測訊號6中對應不 同子載波的單頻迴音訊號在⑻〜Kn);加法單元198係用來力口總 各個單頻迴音訊號A)(η)〜(η)以產生迴音估測訊號j。 請繼續參閱第4圖，第4圖係為第3圖所示之運算單元192 之第一實施例的示意圖。由於第3圖中其他運算單元194、…、196 的木構皆與運鼻早元192相同，故不在此--^贅述。如圖中所示， 運算單元192係用來運算一單頻迴音訊號4，其中包含有複數個 有限脈衝響應（finite impulse response unit, FIR)單元 220、240、 260以及一加法單元280。由於該迴音成分有許多種成因，例如： 1271970 符元間干擾（Iito Symbol lnteffefenee，ISI)、載賴干擾（z血 carrier Interference，ICI)、或者以上兩種的混合，因此運算單元192 必須逐-估計上述的干擾贿生單頻迴音訊動。。本實施例中， 運算單元192利用k個不同的有限脈衝響應單元22〇、24〇、…、 260分別依據運算係數w〇〇〜w〇k4逐一調整子資料D〇⑻〜a 1㈤以 產生複數個運算數值4»〜4,_1(11)，其中運算數值41(11)«11) 即代表其他子載賴造成的載制干擾以及載波符元干擾（1愈 Cmiet Symbol Interference，ICSI)。最後，加法單元 28〇 來加總複 數個運算數健4»〜〇)以產生所要的單頻迴音估測訊號 A^n)，其中η係表示多載波通訊系統1〇〇之第n個傳輸時序。 請注意’有限脈衝響應單元220、24〇、26〇係具有相同的電 路架構，故以有限脈衝響應單元22〇為例。有限脈衝響應單元22〇 中設置有-延遲單元224、複數個乘法單元222、226以及一加法 單元228。有限脈衝響應單元22〇所使用的運算係數包含有w〇〇(n) 以及w〇’〇 (η) ’分別用來乘上目前傳輸時序中第〇個子載波所傳遞 的子資料_)以及前-傳輸鱗巾第G個子做所傳遞的子資料 D〇(n-l)。表示由目前傳輸時序中第〇個子載波所傳遞的 子貧料D0(n)所產生的迴音；而乃办奇〜⑽則表示由上一個傳輸 時序中第0個子做所傳翻子:紐Dg㈣因為符間干擾所產生 的迴音。因此，產生單頻迴音估測訊執⑻的詳細運算可由下列 13 1271970 方程式說明： ^o(n) = = · %, + D0i(n -1) · w0/] 方程式（一） 同理類推’由有限脈衝響應單元24〇、…、26〇所產生的運算 數值化⑻〜I⑻即代表其他子載波對k為〇的子載波所造成的 載波間干擾以及載波付元干擾（Inter carijer Symb〇l Interference， ICSI) 〇 請注意，本實施例中產生單頻迴音估測訊號Di(n)的方式不限於 僅使用兩個相鄰時序的傳輸資料Di(n)以及，亦可視系統需 求僅使用傳輸資料D#);或者，當符軒擾十分嚴重時採用兩個 以上的傳輸資料Di(n)、胁-1)、···、Di㈣來計算目前的單頻迴 音估測訊號Di(n)的，上述的實施方法均屬本發明之範疇。此外， 由於載波間干擾的影響會隨兩相互干擾的子載波的頻率差量而衰 減，因此運算單元192、194、196在計算單頻迴音訊號灿)⑻ 的過程中可以不須考慮所有子載波所造成的干擾（亦即可分別計 异出少於k個運异數值並予以相加）。例如：產生單頻 此一來便可以大付縮減有限脈衝響應單元的使用個數。 請繼續參閱第5圖，第5圖係為第3圖中運算單元194之 14 1271970 貝施例的不思圖。此一實施例係於多载波通訊系統的傳輸資料d ，、有二輛雖日禮用，亦即傳輸資料D中，第〕·個子載波傳傳遞 的子資料Dj與由第n_j個子載波傳傳遞的子資料〜互為共她 值所乂本實化係利用此一特性將子資料^分為虛部與實部來分 別運异’由於子資料Dnj的實部恰好與子資料^的實部相同，子 資料Dn«j的虛。卩恰好為子資料q之虛部的負數，因此可省略與子 資料Dni相關的運异。也因此本發明迴音估測單元中每一運算單 疋至多僅需⑤置⑽（S等於（叫1)有限脈衝響應單元，即可 利用複數個運算係數Wi(n)〜Ws⑻產生s個單頻迴音訊號 A(n) 4(n)。明’主思，在不影響本發明技術揭露之下，第5圖中 僅顯不出三個有限脈衝響應單元32〇、34〇、36〇。最後，利用加法 單元380加總運异數值&⑻〜心⑻以產生單頻迴音估測訊號 A(n)。 有限脈衝響應單元320、340、360具有相同的電路架構，因 此以下敘述係以有限脈衝響應單元32〇為例。有限脈衝響應單元 320中設置有一實部運算單元322、一虛部運算單元332、複數個 延遲單元326、336、複數個乘法單元324、328、334、338以及一 加法單元330。有限脈衝響應單元320所使用的運算係數包含有 wRi，o(n)、wR1，〇’(n)、wM1，〇(n)以及WM10’(n)，以分別乘上目前傳輸時 序中第1個子載波所傳遞之子資料的實部DR1(n)、前一傳輸時序 15 1271970 中第1個子做所傳叙子資料的實部知㈣)、目贿輸時序中 第1個子载波所傳遞之子資_虛部Dm⑻以及前—傳輪時序中 第1個子載波所傳遞之子#料的虛部，並简用加法單 兀33〇加總上述乘積以產生運算數叫⑻。因此，上述產生單頻 迴音訊號4(11)_細運討由下列輕式·: 、 方程式（二） 月心本貝細例中產生迴音估測訊號方⑻的方式不限於僅使 •用兩個相鄰時序的傳輸資料D⑼以及D㈣，亦可視系統需求僅 使用-個傳輸㈣D(n)，或者當符軒針分嚴料可採用兩個 以上相鄰時序的傳輸資料D(n)、D㈣、…、D㈣，以上實施 =觸本發明之鱗。同樣地，運算單元I92、m、m在計 #开運开數值仏⑻〜仏⑷的過程中亦可以不須考慮所有子載波所 k成的載波間干擾崎低每—運算單S巾有赚衝響應單元的使 用個數，或者直接料須考量的子載波所對應的運算倾設為零 即可降低系統的運算量。 當多載波通訊系統卿在正姻始收發資料以前，其會先利 用傳輸核組12G傳送—連串已知的訓練碼X做為傳輸資料D，以 產生正確的運异係數Wq〜I。換言之，當多載波通訊系統謂開 16 (β) 1271970 始正式收發所要資料並不使用 調整。 下列方法對運算係數WG〜Wk4進行 承上所述，當傳輸模組120傳送一連串已知的訓練碼χ 傳輸資料D ’由於此時傳輸介質上除了训練碼乂所對應的時域訊 號X以外沒有任何訊號，所以接收模、组14〇所接收到的時域訊號7 #即可視為時域訊號X所產生的迴音，因此，迴音估測訊號仏〉理論 上應雜解碼訊號相同。如此一來迴音估測單& 即可利用 .如取小均方運算’或者其他適應訊號處理方法來調整運算係數 wG〜Wk4 ’使得爾爾碼χ所魅_音钢訊辦⑻近似於訓 、、東碼ζ(”)所產生的迴音（亦即解碼訊號y⑷）。調整運算係數 的詳細操作請參考下列方程式： 方程式（三） 方程式（四） ^ («+!)=： iv, (n) + fjek (Λ).1271970 IX. Description of the invention: [Technical field to which the invention pertains] The present invention discloses a fresh-carrier communication system and a communication method thereof, and more particularly to a recording communication system capable of eliminating echo and a communication method thereof. [Prior Art] The general communication system can be used for private, half-duplex and full-duplex systems. Since the half-double and the full-double 1 rhyme can be transmitted in both directions, the downlink receiving position will inevitably be predicted by the upstream transmission. _ scatter multi-tone system ^ (d1Scretemulti-tone system), the Lie conversion will transmit the number of sub-carriers of the variable called the variable domain, and then send the time domain signal (4) generated after the modulation to the transmission medium; when the discrete multi-frequency system When it is desired to receive a poor material, it will also convert the receiving time domain signal S2 to generate a decoded data D. However, when the discrete multi-frequency system and the front-end 2-line hybrid circuit impedance do not match, the downlink time domain signal % is predicted by the uplink time domain signal 81, which is known in the industry. Echo (Μ.). Please refer to Fig. 1 'Fig. 1 is a schematic diagram of a discrete multi-frequency system 1 。. As shown in the figure, the inverse Fourier transform unit 22 set by the discrete multi-frequency system 1G is used to inversely transmit the transmission (4) D to the time domain signal ^, and 1271970, and output the daily SDO signal Si to the transmission filter. The filter 26 performs filtering processing, and then the time or the number S1 is further transmitted to the analog front end circuit (Qiu (10)t) 28 '. Finally, the analog front end circuit Μ sends the time domain signal S1 to the external transmission medium (not shown). . Similarly, the discrete multi-frequency system 1 can receive the time-domain signal & from the external transmission medium by the analog front-end circuit 28, and then the analog front-end circuit π progressively transmits the time domain signal S2 to the receiving chopper. For subsequent filtering. In order to eliminate the interference caused by the time domain signal 81 to the time domain signal illusion (also the echo of the instant domain nickname S1), an echo estimation unit is provided in the discrete multi-frequency system 1 • 24 ' is used to generate the time domain signal S1 An echo estimate signal Sech. And the adjustment unit 34 subtracts the echo estimation signal Sech from the time domain signal S2. To generate the -clock signal S2', finally the clock signal S2 is subjected to the Fourier transform order $32 for Fourier transform to generate the solution H', so that the data D is decoded, that is, it is not disturbed by the echo. Although the above-mentioned discrete multi-frequency system 1G can effectively eliminate the echo of the system towel, since the time domain signals S1 and S2 both have a high sampling frequency, the echo estimation signal sech is calculated. The process must consume considerable computing resources. SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a multi-carrier communication system capable of eliminating echo, which can reduce the amount of computation required in eliminating echo. 8 1271970 In accordance with an embodiment of the present invention, it is disclosed that the magnetic wave (four) system. The multi-carrier communication system includes a transmission module for performing according to a transmission data - a reverse Fourier-weighted subtraction-first-time domain 峨, - a receiving module for receiving - a second time domain signal And according to the second time domain recording line-side leaf operation to generate-decode data; and the echo cancellation unit, the market-to-transmission module and the receiving module, the decoding data and the transmission according to the plurality of operation coefficients The material is generated - the poor material is reconstructed, and the plurality of operational coefficients are adjusted according to the reconstruction data. According to the actual complement of the present invention, a communication method in which the multi-carrier communication system is used to remove echo is disclosed. The communication method includes: performing a reverse time Fourier transmission according to a transmission data to generate a first time domain signal; transmitting the first time domain signal; receiving a second time domain signal; performing according to the second time domain signal a Fourier operation to generate-decode data; the decoded data and the Lura input and output data are reconstructed according to the plurality of operational coefficients; and the plurality of transport coefficients are adjusted according to the reconstructed data. According to the embodiment of the invention, a multi-carrier communication device is also disclosed for receiving and transmitting data in a night carrier communication system, and the multi-carrier transceiver includes a j-domain time domain module. The first frequency domain information to be transmitted is converted into a time domain signal for performing a data transmission domain frequency domain module in the multi-carrier communication system, which is used to receive the second time The domain signal conversion 1271970 is - the second frequency domain information, for the recording wave through the data - data receiving action 'and the weekly sound elimination module, minus the time domain module and the time domain switching surface The group adjusts the first frequency domain resource in the frequency domain according to the second frequency domain information, so as to perform an echo cancellation operation when the multi-carrier communication device transmits data. [Embodiment] Φ May see the second ®, Fig. 2 is a schematic diagram of a multi-carrier communication system 100 according to the present invention. In the present embodiment, the 'multi-carrier communication system 1' can be a discrete multi-tone (DMT) system. Or an orthogonal frequency division multiplexing (orthogonal frequenCy The multiplex control system 100 is provided with a transmission module (10), a receiving group 14〇, an analog front end circuit (10) and an echo cancellation unit 18〇. The transmission module 120 is used to The transmission data D performs a reverse Fourier operation to generate a spring time domain S1, the receiving module 14 is used to receive a time domain signal S2, and performs a leaf-turning operation according to the base or base S2 to generate-decode data 5; The analog front end circuit 160 is coupled to a transmission medium (for example, a twisted pair) for transmitting the time domain signal S1 or the receiving time domain signal S2; and the echo cancellation unit (10) is configured to calculate the coefficient w〇~wM and transmit the data. D and the decoded data are generated as a reconstruction data, so that *D' can not interfere with the data content of the scale data β and is not interfered by the echo of the transmission data D. Where k represents the multi-carrier communication system The total number of wavelets used. It should be noted that although this embodiment uses the same number of operation coefficients as the number of sub-carriers total 1271970 as an example for explanation, those skilled in the art should It is understood that the number of such wires and the complexity of the operation are not the limiting conditions of the present invention. The use of a smaller number of operating coefficients and a simpler computing process are also included in the scope of the present invention. Generally, S, a typical transmission module 120 usually includes a hardware or software that performs such as reverse Fourier transform, Cyclic_prefix addition, parallel rotation sequence, digital conversion analogy, and filtering, or the above functional block is at least one injury of the towel. The receiving module usually includes a hardware or a software that performs functions such as a wave, an analog-to-digital number, a sequence_parallel, a cyclic prefix removal, and a side leaf conversion, and at least one of the above functional blocks # Part. The operation principles of the transmission module 12A, the receiving right group 140, and the analog front end circuit 16 are well known to those skilled in the art, and therefore will not be described herein. An echo estimation unit 182 and an adjustment unit 184 are additionally provided in the echo cancellation unit 180. The echo estimation unit 182 is configured to calculate the transmission material D according to the plurality of operation coefficients to generate an echo estimation signal, and the adjustment unit 184 subtracts the decoded data β from the decoded data to generate the reconstructed data. In the present embodiment, the adjustment unit 184 is implemented by a subtractor to adjust the decoded data β. In this embodiment, the echo estimation unit moves to perform a least mean square (LMS) operation based on the reconstruction data D to enter a gas of 11 1271970, so that the object rotation (4) can be generated by adjusting and transmitting the data D. The echo is estimated by the telecom, with a transmission of D. The detailed operation of adjusting the calculation coefficient W〇~wk4 on the small ship riding Xiaocai County will be the minimum description. Please note that in addition to using: # ', other adaptive signal processing methods such as RLS, Slgned LMS, etc. can be used to adjust the operation coefficient i. Limited. a Please continue to refer to FIG. 3, which is a schematic diagram of a first embodiment of the echo estimation unit 182 shown in FIG. As shown in the figure, the echo estimation unit 182 includes a plurality of arithmetic units 192, 194, and 196 and an adding unit 198. The transport units 192, 194, ..., 196 are respectively used to generate the single frequency echo signals corresponding to different subcarriers in the echo estimation signal 6 at (8) ~ Kn); the adding unit 198 is used to force each single frequency echo signal A) (η)~(η) to generate an echo estimation signal j. Please refer to FIG. 4, which is a schematic diagram of the first embodiment of the arithmetic unit 192 shown in FIG. Since the wooden structures of the other arithmetic units 194, ..., 196 in Fig. 3 are the same as those of the Yunan early element 192, they are not described here. As shown in the figure, the arithmetic unit 192 is used to calculate a single-frequency echo signal 4, which includes a plurality of finite impulse response unit (RF) units 220, 240, 260 and an adding unit 280. Since the echo component has many kinds of causes, for example: 1271970 Interference Symbol (ISI), ISIS (ISI), or a mixture of the two, the arithmetic unit 192 must It is estimated that the above-mentioned interference bribes are single-frequency echo. . In this embodiment, the operation unit 192 adjusts the sub-data D〇(8)~a1(5) one by one according to the operation coefficients w〇〇~w〇k4 by using k different finite impulse response units 22〇, 24〇, . . . , 260 respectively to generate a complex number. The operand values 4»~4,_1(11), where the operand value 41(11)«11) represent the carrier interference caused by other subcarriers and the carrier symbol interference (ICSI). Finally, the adding unit 28 adds a plurality of operands 4»~〇) to generate the desired single-frequency echo estimation signal A^n), where η represents the nth transmission of the multi-carrier communication system 1〇〇 Timing. Please note that the finite impulse response units 220, 24, and 26 have the same circuit architecture, so the finite impulse response unit 22 is taken as an example. The finite impulse response unit 22A is provided with a delay unit 224, a plurality of multiplication units 222, 226, and an addition unit 228. The operational coefficients used by the finite impulse response unit 22〇 include w〇〇(n) and w〇'〇(η)' respectively used to multiply the sub-data transmitted by the third subcarrier in the current transmission timing _) and before - Transfer the sub-data D〇(nl) passed by the Gth child. Represents the echo generated by the sub-lean D0(n) transmitted by the third subcarrier in the current transmission sequence; and the odd-numbered (10) indicates that the 0th sub-transfer of the previous transmission sequence is used: New Dg (4) Because of the echo generated by the inter-symbol interference. Therefore, the detailed operation of generating a single-frequency echo estimation signal (8) can be explained by the following 13 1271970 equation: ^o(n) = = · %, + D0i(n -1) · w0/] Equation (1) Similarly analogy The computational digitizations (8)~I(8) generated by the finite impulse response units 24〇,...,26〇 represent inter-carrier interference and carrier-frequency interference caused by other subcarriers for subcarriers with k being 〇 (Inter carijer Symb〇l Interference, ICSI) Please note that the method for generating the single-frequency echo estimation signal Di(n) in this embodiment is not limited to the transmission data Di(n) using only two adjacent timings, and only the transmission is used depending on the system requirements. Data D#); or, when the Fuxuan is very serious, use two or more transmission data Di(n), threat-1), ···, Di(4) to calculate the current single-frequency echo estimation signal Di(n) The above-described implementation methods are all within the scope of the present invention. In addition, since the influence of inter-carrier interference is attenuated by the frequency difference between the two mutually interfered subcarriers, the arithmetic unit 192, 194, and 196 can calculate all the subcarriers in the process of calculating the single-frequency echo signal (8). The resulting interference (or less than k different values and added separately). For example, generating a single frequency can reduce the number of finite impulse response units used. Please continue to refer to FIG. 5, which is a diagram of the example of the operation of the arithmetic unit 194 in FIG. In this embodiment, the data d is transmitted in the multi-carrier communication system, and the second data is transmitted in the data D, and the sub-data Dj transmitted by the sub-carriers is transmitted and transmitted by the n_j sub-carriers. The sub-data ~ mutual mutual value of her 乂 实 实 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用The same, the sub-data Dn«j virtual.卩 happens to be the negative of the imaginary part of the sub-data q, so the transport related to the sub-data Dni can be omitted. Therefore, in the echo estimation unit of the present invention, at most 5 (10) (S is equal to (called 1) finite impulse response unit, each operation unit can generate s single frequencies by using a plurality of operation coefficients Wi(n) to Ws(8). Echo signal A(n) 4(n). It is clear that, without affecting the disclosure of the present invention, only three finite impulse response units 32 〇, 34 〇, 36 显 are shown in Fig. 5. Finally The summation unit 380 adds the total value & (8) to the heart (8) to generate a single frequency echo estimation signal A(n). The finite impulse response units 320, 340, 360 have the same circuit architecture, so the following description is limited The impulse response unit 32 is exemplified. The finite impulse response unit 320 is provided with a real part operation unit 322, an imaginary part operation unit 332, a plurality of delay units 326, 336, a plurality of multiplication units 324, 328, 334, 338 and a The adding unit 330. The operating coefficients used by the finite impulse response unit 320 include wRi, o(n), wR1, 〇'(n), wM1, 〇(n), and WM10'(n), respectively, to multiply the current transmission. The real part DR1(n) of the sub-data transmitted by the first subcarrier in the sequence, a transmission sequence 15 1271970, the first sub-subsidiary of the sub-subsidiary data (4)), the sub-carrier _ imaginary part Dm (8) transmitted in the first sub-carrier, and the first sub-carrier in the pre-transmission timing The imaginary part of the passed child #料, and simply add the above product by the addition unit 33〇 to generate the operand called (8). Therefore, the above-mentioned single-frequency echo signal 4(11)_ is discussed in the following light form:: Equation (2) The method of generating the echo estimation signal (8) in the case of the monthly heart is not limited to only two The transmission data D(9) and D(4) of adjacent time series can also use only one transmission (four) D(n) depending on the system requirements, or can transmit data D(n), D(4) of more than two adjacent timings when the symbol is strictly divided. , ..., D (four), the above implementation = touch the scale of the invention. Similarly, the arithmetic unit I92, m, m can also calculate the inter-carrier interference of all subcarriers in the process of counting the values 仏(8)~仏(4). The number of uses of the impulse response unit, or the calculation of the subcarrier corresponding to the subcarrier to be considered directly to zero can reduce the amount of computation of the system. When the multi-carrier communication system sends and receives data before the marriage, it will first transmit the transmission core group 12G—a series of known training codes X as the transmission data D to generate the correct migration coefficient Wq~I. In other words, when the multi-carrier communication system is pre-opened 16 (β) 1271970, the official data is sent and received without adjustment. The following methods are used to calculate the operating coefficients WG~Wk4. When the transmission module 120 transmits a series of known training codes, the transmission data D' is due to the time domain signal X corresponding to the training code on the transmission medium. There is no signal, so the time domain signal 7 received by the receiving module and group 14 can be regarded as the echo generated by the time domain signal X. Therefore, the echo estimation signal 理论> theoretically has the same decoding signal. Such a round-trip estimate list & can be used. For example, take the small mean square operation ' or other adaptive signal processing method to adjust the operation coefficient wG ~ Wk4 'make the err code χ _ _ steel steel office (8) similar to training The echo generated by the east code (") is the decoded signal y(4). For the detailed operation of adjusting the operation coefficient, please refer to the following equation: Equation (3) Equation (4) ^ («+!)=: iv, ( n) + fjek (Λ).
〜⑻(η) 其中，k代表多載波通訊系統1〇〇所使用的第]^個子載波，以及 代表第η個傳輸時序。 如刖所述之多載波通訊系統與其通訊方法係利用一迴音估測 單兀於頻域（frequencyd〇main)運算出一迴音估測訊號。由於頻 域訊號的取樣頻率較時序訊號為低，因此可大幅降低多載波通訊 系統的運算量’同時上述多載波通訊系統與其通訊方法於估測迴 17 1271970 音估測訊號時’顿可考轴不同子載波的通道響應，更可 效地抑制姐間谓干酬產生㈣音，使得 f為車石盒。 貝针 以上所述僅為本發明之錄實施例，凡依本發明巾請專利範 圍所做之均等變化與修飾，皆應屬本發明之涵蓋範圍。 【圖式簡單說明】 第1圖為為習知離散多頻系統的示意圖。 苐2圖為本發明多載波通訊系統的示意圖。 第3圖係為第2圖中迴音估測單元之一實施例的示意圖。 第4圖係為第3圖中運算單元之第一實施例的示意圖。 弟5圖係為第3圖中運算單元之第二實施例的示意圖。~(8)(η) where k represents the ?^ subcarrier used by the multicarrier communication system 1 and represents the nth transmission timing. The multi-carrier communication system and its communication method as described above use an echo estimation unit to calculate an echo estimation signal in the frequency domain (frequencyd〇main). Since the sampling frequency of the frequency domain signal is lower than that of the timing signal, the calculation amount of the multi-carrier communication system can be greatly reduced. Meanwhile, the multi-carrier communication system and the communication method thereof are used to estimate the sound of the 17 1271970 sound estimation signal. The channel response of different subcarriers can more effectively suppress the generation of (4) sounds in the inter-sister, so that f is a car stone box. The above descriptions are only examples of the invention, and all variations and modifications made to the scope of the invention according to the invention are intended to be within the scope of the invention. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic diagram of a conventional discrete multi-frequency system. Figure 2 is a schematic diagram of a multi-carrier communication system of the present invention. Figure 3 is a schematic diagram of one embodiment of the echo estimation unit of Figure 2. Fig. 4 is a schematic view showing the first embodiment of the arithmetic unit in Fig. 3. The fifth diagram is a schematic diagram of the second embodiment of the arithmetic unit in Fig. 3.
18 1271970 【主要元件符號說明】 10、100 多載波通訊系統 22 逆向傅利葉轉換 XiO — 早兀 24、182 迴音估測單元 26、36 濾波器 28、160 類比前端電路 32 傅利葉轉換單元 34、184 調整單元 120 傳輸模組 140 接收模組 180 迴音消除單元 192、194、···、 運算單元 220、240、260、 有限脈衝響應單 196 320、340、360 元 222、226、 324 、 328 、 334 、 338 乘法單元 224、326、336 延遲單元 198 、 228 、 280、330、380 加法單元 322 實部運算單元 332 虛部運算單元 19 ⑧18 1271970 [Description of main component symbols] 10, 100 multi-carrier communication system 22 Reverse Fourier transform XiO - early 24, 182 echo estimation unit 26, 36 filter 28, 160 analog front end circuit 32 Fourier transform unit 34, 184 adjustment unit 120 transmission module 140 receiving module 180 echo cancellation unit 192, 194, ..., operation unit 220, 240, 260, finite impulse response list 196 320, 340, 360 yuan 222, 226, 324, 328, 334, 338 Multiplication unit 224, 326, 336 delay unit 198, 228, 280, 330, 380 addition unit 322 real operation unit 332 imaginary unit operation unit 19 8