TW384425B - Method and apparatus for solving polynomial of key function when decoding error correction codes - Google Patents
Method and apparatus for solving polynomial of key function when decoding error correction codes Download PDFInfo
- Publication number
- TW384425B TW384425B TW87100122A TW87100122A TW384425B TW 384425 B TW384425 B TW 384425B TW 87100122 A TW87100122 A TW 87100122A TW 87100122 A TW87100122 A TW 87100122A TW 384425 B TW384425 B TW 384425B
- Authority
- TW
- Taiwan
- Prior art keywords
- error
- polynomial
- generated
- value
- coefficient
- Prior art date
Links
Landscapes
- Error Detection And Correction (AREA)
Abstract
Description
A7 B7 經濟部中央標準:工消费合作杜印氧 五、發明説明(1 ) 本發明係有關於一稜解碼錯誤訂正碼(Εγγ〇γ Correcting Codes)之方法及其装置,特別是有關於在解碼 錯誤訂正碼之過程中,決定錯誤定位器多項式(Εγγ〇γ Locator Polynomials)及轉姨求值器多項式(Εγγ〇γ Evaluator Polynomials)之方法及其裝置。 經過多種不同媒介,由發信位置至目地位里之資料 傳輪,由於傳送路徑、及“介本身所造成之雜訊,會 造成傳輸資料之嫌誤H傳輸之資料不會與所接收 到之資料相同·為了判定出接收資料之错誤,已發展出 有各種方法及技術’以侦挪和訂正接收資料之錯誤。方 法之一為產生包括訊息部份(所傳送之資料)和奇偶性部 份(Parity Part,據以實施錄镇訂正之信息)的碼字(c〇de Word)。 在本文中’碼字係針對原始資料施行編碼操作而 得。碼字具有同一形式,為包括N個符號之信息,其中 前K個符號係為訊息符號,雨後Ν_κ個符貌係為奇供性 符號。 在所有知名之錯誤訂正码中,BCH瑪(Bose- Chaudhuri-Hocquenghen Codes)以及 RS 瑪(Reed-Solomon Codes)是在通訊領域和貯存器系統應用中,最廣為使用 之區堍蝎(Block Codes)。區塊蝎之數學理检基礎在「E.R. Berlekamp, Algebraic Coding Theory, McGraw-Hill, New York, 1968」以及「S. Lin and D.J. Costello,五rror Cowiro/ Coding: Fundamentals and Applications^ Prentice-Hall, 本紙張尺度適用中國國家標準(CNS ) A4规格(2丨0'〆297公釐) (請先W讀背面•之注$項再填寫本頁) ic. ?Ί -c. 鍾濟部中央樣隼局貝工消费合作杜印簟 A7 B7 . 五、發明説明(2 )A7 B7 Central Standard of the Ministry of Economic Affairs: Industrial and consumer cooperation Du Yinxuan 5. Description of the invention (1) The present invention relates to a method and a device for decoding edge correction error codes (Eγγ〇γ Correcting Codes), and in particular, it relates to decoding In the process of the error correction code, a method and a device for determining an error locator polynomial (Eγγ〇γ Locator Polynomials) and an evaluator polynomial (Eγγ〇γ Evaluator Polynomials) are determined. Through a variety of different media, from the transmission position to the data transmission wheel in the destination, due to the transmission path and the "noise caused by the media itself, it will cause the transmission of data. The transmission of the data will not be the same as the received data. Same data. In order to determine the error of receiving data, various methods and techniques have been developed to detect and correct the error of received data. One method is to generate a message part (data transmitted) and a parity part. Copy (Parity Part, according to which information is recorded and corrected) code word (here). In this article, 'code word is obtained by performing an encoding operation on the original data. The code word has the same form and includes N The information of the symbols, in which the first K symbols are information symbols, and the N_κ symbols after the rain are odd supply symbols. Among all well-known error correction codes, BCH-Chaudhuri-Hocquenghen Codes and RS-Reed -Solomon Codes) is the most widely used Block Codes in the communication field and storage system applications. The basis of the mathematical examination of block scorpions is "ER Berlekamp, Algebraic Coding The ory, McGraw-Hill, New York, 1968 "and" S. Lin and DJ Costello, Wurror Cowiro / Coding: Fundamentals and Applications ^ Prentice-Hall, this paper size applies to the Chinese National Standard (CNS) A4 specification (2 丨 0 '〆297mm) (please read the note on the back of the page and then fill in this page) ic.? Ί -c. Duyin A7 B7, Shellfish Consumer Cooperation of the Central Sample Bureau of Zhongji Department (2 )
Englewood Cliffs, NJ,1983」兩本著作均有詳細之解释說 明。 一個(N,K)BCH或RS碼具有K個訊息符號和N個 編碼符號,其中每一符號屬於一 BCH碼之GF(q)集合或 一 RS碼之GF(qm)集合。在N=2m-卜及N-K < mt之情形 下,一個二進位(N,K)BCH碼能訂正達到t個之錯誤符 號。在2t+p < N-K之情形下,一個二進位(N,K)RS碼可 以訂正t個錯誤符號及p個抹除(erasure)符號。對二進位 BCH碼而言,藉由發現錯誤符號之位置,可很簡單地訂 正一個錯誤符號。對RS碼而言,藉由發現錯誤符號之 位置及其錯誤值,即可訂正一個錯誤符號。RS碼之應用 中,抹除係定義為一已知錯誤位置上之錯誤,因此,訂 正此錯誤則簡化為發現錯誤值。 使用普遍之RS解碼器架構,應用於錯誤訂正之方 法步驟可摘要為以下四個步驟:(1)由所接收之碼字計算 出徵兆(Syndrome),(2)運算出錯誤定位器多項式及錯誤 求值器多項式,(3)發現錯誤所在位置,以及(4)運算出錯 誤值。假設錯誤和抹除均被訂正,則四個步麻修正為如 下:(1)由所接收之碼字及抹除位置計算出徵兆以及 Forney徵兆,(2)計算錯誤及抹除定位器多項式以及錯誤 及抹除求值器多項式,(3)發現錯誤所在位置,以及(4) 計算鉗誤及抹除之訂正值。 參照第la圈,其顯示一般之解碼步驟。所接收之 資料R(x)輸入徵兆計算器10以產生徵兆多項式S(x), 本紙張尺度適用中國國家標準(CNS ) A4规格(210X297公釐) ---------C------iT------c (請先«*讀背面•之注意事項再填寫本頁) A7 B7 五、發明説明(3 ) 其代表碼字之錯誤型式,藉以訂正錯誤。徵兆係僅依錯 誤型式而定,而非依傳輸之碼字而定。接著,將徵兆輸 入一鍵方程式解答器(Key Equation Solver)12,使用知名 的Berlekamp-Massey演算法以產生一錄誤定位器多項式 σ(χ)、及一錯誤求值器多項式Ω(χ)。錯誤定位器多項式 指示發生錯誤之位置,而錯誤求值器多項式可指示出錯 誤之值。下一步驟,錯誤定▲器多項式傳至Chien搜尋 器14以解出方程式之根p;1,其表示錯誤符號之位置。 錯誤求值器16接收根p;1以及錯誤求舞多項式Ω(χ),產生 相對應於根吖之錯誤值。 在實施鏈方程式解答器(上述第2步驟)時,本步驊 係有關於解出如下之健方程式: S(x) σ(χ) = Ω(χ) mod xN"K ; 其中,S(x)為徵兆多項式,σ(χ)為錯誤定位器多項式, 以及Ω(χ)為錯誤求值器多項式。當同時訂正錯誤及抹除 時,σ(χ)和Ω(χ)即分別為銪誤及抹除定位器多項式以及 錯供及抹除求值器多項式·其中,σ(χ) = λ(χ)Λ(χ),而 λ(χ)和Λ(χ)分別對應於鏘誤定位器多項式以及抹除定位 器多項式*第lb圈顧示用於錯誤及抹除訂正之一般處 理步驟。徵兆計算器20除接收R(x),也接收抹除資料, 並產生徵兆多項式S(x)和Forney徵兆多項式T(x)。鍵方 程式解答器22處理S(x)和Τ(χ)以產生錯誤及抹除求值 多項式Ω(χ)、以及錯誤及抹除定位器多項式σ(χ)β鏘誤 及抹除定位器多項式輪入Chien搜尋器24以判定錯誤符 7 本纸張尺度適用中國國家橾準(CNS ) A4规格(210X 297公釐) CII (請先敗讀背面•之注$項再填寫本頁)Englewood Cliffs, NJ, 1983 "both have detailed explanations. An (N, K) BCH or RS code has K message symbols and N coding symbols, where each symbol belongs to a GF (q) set of a BCH code or a GF (qm) set of an RS code. In the case of N = 2m-bu and N-K < mt, a binary (N, K) BCH code can correct t error symbols. In the case of 2t + p < N-K, a binary (N, K) RS code can correct t error symbols and p erasure symbols. For binary BCH codes, an error symbol can be easily corrected by finding the location of the error symbol. For RS codes, an error symbol can be corrected by finding the location of the error symbol and its error value. In the application of RS code, erasure is defined as an error at a known error position. Therefore, correcting this error is simplified to find the wrong value. Using the general RS decoder architecture, the method steps applied to error correction can be summarized as the following four steps: (1) Calculate the symdrome from the received codeword, (2) Calculate the error locator polynomial and error The evaluator polynomial, (3) finds the location of the error, and (4) calculates the error value. Assuming that errors and erasures are corrected, the four steps are corrected as follows: (1) Calculate the symptoms and Forney symptoms from the received codeword and erasure position, (2) Calculate the errors and erase the locator polynomial and Error and erasure evaluator polynomials, (3) where the error was found, and (4) the correction value for the correction and erasure. Referring to circle la, it shows the general decoding steps. The received data R (x) is input into the Symptom Calculator 10 to generate the Symptom Polynomial S (x). This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) --------- C- ----- iT ------ c (please «* read the notes on the back side before filling out this page) A7 B7 V. Description of the invention (3) It represents the wrong type of code word to correct the error. Symptoms depend only on the type of error, not the codeword transmitted. Next, the symptoms are input into a Key Equation Solver 12 and the well-known Berlekamp-Massey algorithm is used to generate a recorded error locator polynomial σ (χ) and an error evaluator polynomial Ω (χ). The error locator polynomial indicates where the error occurred, and the error evaluator polynomial indicates the value of the error. In the next step, the error determinator polynomial is passed to the Chien searcher 14 to solve the root p; 1 of the equation, which represents the position of the error symbol. The error evaluator 16 receives the root p; 1 and the error polynomial Ω (χ) to generate an error value corresponding to the root acridine. When implementing the chain equation solver (the second step above), this step is related to solving the following healthy equations: S (x) σ (χ) = Ω (χ) mod xN "K; where S (x ) Is an indication polynomial, σ (χ) is an error locator polynomial, and Ω (χ) is an error evaluator polynomial. When errors and erasures are corrected simultaneously, σ (χ) and Ω (χ) are the error and erasure locator polynomials and the wrong supply and erasure evaluator polynomials, respectively, where σ (χ) = λ (χ ) Λ (χ), and λ (χ) and Λ (χ) respectively correspond to the error locator polynomial and the erase locator polynomial. The lb cycle shows the general processing steps for errors and erasure correction. In addition to receiving R (x), the indication calculator 20 also receives erasure data, and generates an indication polynomial S (x) and a Forney indication polynomial T (x). The bond equation solver 22 processes S (x) and T (χ) to generate an error and erase the evaluation polynomial Ω (χ), and an error and erase locator polynomial σ (χ) β, an error and erase locator polynomial Turn in the Chien searcher 24 to determine the wrong character 7 This paper size is applicable to China National Standard (CNS) A4 size (210X 297 mm) CII (Please read the note on the back side of the page and fill in this page)
•tT 鯉濟部中央棣準局貝工消费合作社印掣 經濟部中央樣率為貝工消费合作社印轚 A7 B7 , 五、發明说明(4 ) 號發生之位置,而錯誤及抹除求值器多項式和錯誤及抹 除位置兩者均輸入一錯誤及抹除值求值器,用以產生錯 誤及抹除值。 經常用以解出鍵方程式之技術包括Berlekamp-Massey 演算法,Euclidean 演算法,以及 Continuous-fraction 演算法 。 相較於其他兩種演 算法, Berlekamp-Massey演算法一般認為其硬體複雜度最小° Berlekamp-Massey演算法之詳細描述見於上述Berlekamp參考引 證資料之第7聿,以及J丄.Massey所發表「• tT Liye Central Bureau of Standards and Quarantine Bureau Shellfish Consumer Cooperatives India India Central Government sample rate for Shellfish Consumers Cooperatives India A7 B7, V. The location where the invention description (4) number occurred, and the error and erasure evaluator Both the polynomial and the error and erase position are input to an error and erase value evaluator to generate the error and erase value. Techniques often used to solve bond equations include Berlekamp-Massey algorithms, Euclidean algorithms, and Continuous-fraction algorithms. Compared to the other two algorithms, the Berlekamp-Massey algorithm generally considers its hardware complexity to be minimal. ° A detailed description of the Berlekamp-Massey algorithm can be found in Article 7 of the Berlekamp Reference Citation, and published by J. Massey.
Synthesis and BCH Decodingy IEEE Trans, on Information Theory, IT-15:122-127,1969」之掄文中。一種無反轉 (inversionless) Berlekamp-Massey 演算法為 Burton 所提 出,用以消除高成本之有限場反轉器(Finite-field inverters,FFIs) β 詳見 H.O. Burton 之檢文「Inversionless Decoding of Binary BCH Codes, IEEE Trans, on Information Theroy,IT-17:464-466,1971」e 習知技術應用傳統Berlekamp-Massey演算法,用以 計算錯誤定位器多項式和饍誤求值多項式,以及作為設 計電路之基礎·然而,每一種演算法均需要很大數董之 有限場乘法器(Finite-field multiplier,FFM),以及或許 需要有限場反轉器FFI。每一 FFM和FFI轉換成硬髏線 路並實作於積Λ鼇路中。所以,此發明之目的係為導出 一有效率之多項式解法,並且可減少實作演算法時硬髏 線路之大小(複雜度)。FFM和FFI之數目基本上是為變 本纸張尺度適用中國困家標準(CNS ) A4規格(210X297公漦) (請先ΚΓ讀背&之注意事項再填寫本頁)Synthesis and BCH Decodingy IEEE Trans, on Information Theory, IT-15: 122-127, 1969 ". An inversionless Berlekamp-Massey algorithm was proposed by Burton to eliminate high-cost Finite-field inverters (FFIs) β. For details, see HO Burton's "Inversionless Decoding of Binary BCH" Codes, IEEE Trans, on Information Theroy, IT-17: 464-466, 1971 "e. The conventional technology uses the traditional Berlekamp-Massey algorithm to calculate the error locator polynomial and the error evaluation polynomial, and as a design circuit Fundamentals. However, each algorithm requires a large number of finite field multipliers (FFMs), and possibly a finite field inverter FFI. Each FFM and FFI is converted into a hard cross-line circuit and implemented in Ji'ao Road. Therefore, the purpose of this invention is to derive an efficient polynomial solution and reduce the size (complexity) of the hard-crossed circuit when implementing the algorithm. The number of FFM and FFI is basically to change the paper size. Applicable to China Standards (CNS) A4 size (210X297 cm) (please read the precautions of κΓ before filling this page)
A7 B7五、發明説明(5 ) 數t之函數,變數t為(Ν-Κ)/2之函數。表一顧示各種不 同演算法及FFM和FFI所相對應之個數,當t等於8時。 參考演算法 FFM’s (t之函數) FFM’s數目 FFI’s數目 Berlekamp 3t 24 1 Liu 2t-l 17 1 Oh 2t 16 1 Reed 3rt+i> 27 0 表一 觑濟部中央梯率局貝工消费合作杜印装 如表一所示,僅就錯誤訂正(户非錯誤及抹除訂 正)’實作一傳統之 Beriekamp-Massey 演算法(BeriekamP 美困專利號需3t或24個FFM’s和1個FFI’s » Liu 在 r Architecture for VLSI Design of Reed-Solomon Decoders, IEEE Trans, on Computers, Vol. 33, No. 2, February 1984」一文中所示範之一演算法需2t-1個或17 個FFM’s和1個FFI’s。而美國專利號5,583,499中,Oh 等人揭露一電路需2t或16個FFM’s和1個FFI’s。 另一方面,Reed等人所揭示之演算法中,不須有 反轉(inversion),所以具有相當複雜度之FFI也不需要。 上述之演算法揭露於「KL57 〇/ vers e-Free Berlekamp-Massey Algorithm, Reed, Shin, and Truong, IEE Proceedings-E,Vol. 138, No. 5, September 1991」。然而, 即使Reed之演算法無須使用FI?I,但卻需要較大數董之 FFM’s,高達3(t+l)或27個。若應用於錯誤及抹除訂正 時,所須之FFM’s數目將會更高,通常為錯誤訂正狀況 9 (請先眸讀背面-之法項再填寫本頁) - -ο. 本纸張尺度適用中國國家梯率(CNS ) A4规格(210X297公釐) A7 B7 «濟部中夹樣率扃貝工消费合作社印氧 五、發明説明(6 ) 下之2倍。 因此,在實作演算法時,一種無須使用FFIs並可 使FFMs之數目降至最小的無反轉方法及其裝置是眾所 期望。 有鐘於此,本發明之一目的為提供一種在碼字解碼 時’用以解答鍵方程式多項式之方法及其裝置。 本發明之另一目的為提供一種植基於Berlekamp· Massey演算法之方法及其裝置,其可以最少之硬想線路 來加以實作《 本發明之又一目的為提供一種解答健方程式多項 式之方法及其裝置,其不會降低解碼器之整體解碼速 度β 簡而言之,在較佳實施例中,揭示一種在錯誤訂正 褐解考過租之鍵方程式解答步称中,用以計算嫌誤定位 器多項式和鏘誤求值器多項式之方法,經由上述方法之 若干中間步驊而產生多項式,上述中間步驛可以最小數 目之硬饉線路來加以資作。中間步驟之數目需要相對應 之數目之運算時間週期,以完成多項式之運算。然而, 依所選擇之(Ν,Κ)瑪而定’計算多項式所需之運算時間週 期數目會在計算上游資料(up-stream data)所需時間之範 团内· 特別是,有效率地蚬割小數量之有限場乘法器 (FFM’s)而無須使用有限場反轉器(FH’s),用以計算錯誤 定位器多項式和錯誤求值多項式之較佳方法也加以揭 (請先閲讀背面之注意事項再填寫本頁) AIV. 訂 CI. 本纸張尺度遥用中國國家揉準(CNS )八4規格(210X297公釐) 經濟部中央揉準局負工消费合作社印簟 A7 B7 ' ' 五、發明説明(7 ) 示。使用這些新方法,實作由無反轉Berlekamp-Massey 演算法所導出之方法,一種僅使用3個FFM’s而無須 FFI’s之具面積效率(節省線路面積)之架構亦加以揭 示。這方法及架構可廣泛地應用於各種具有適當編碼長 度之RS和BCH碼。 本發^之一優_^3提供一種方^及裝置,用^^字 本發明之另一優點為提供一種方法及裝置,基於 Berlekamp-Massey演算法,可以最少,數f之硬II線路來 ---------o^.-- (請先M讀背面之注$項再填寫本頁)A7 B7 V. Invention Description (5) The function of the number t, the variable t is a function of (N-K) / 2. Table 1 shows the various algorithms and the corresponding numbers of FFM and FFI, when t is equal to 8. Reference algorithm FFM's (function of t) Number of FFM's Number of FFI's Berlekamp 3t 24 1 Liu 2t-l 17 1 Oh 2t 16 1 Reed 3rt + i > 27 0 As shown in Table 1, only for error correction (non-error and erasure correction) 'implementing a traditional Beriekamp-Massey algorithm (BeriekamP US sleep patent number requires 3t or 24 FFM's and 1 FFI's »Liu in r One of the algorithms demonstrated in the paper "Architecture for VLSI Design of Reed-Solomon Decoders, IEEE Trans, on Computers, Vol. 33, No. 2, February 1984" requires 2t-1 or 17 FFM's and 1 FFI's. In US Patent No. 5,583,499, Oh et al. Disclosed that a circuit requires 2t or 16 FFM's and 1 FFI's. On the other hand, the algorithm disclosed by Reed et al. Does not require inversion, so it is quite complicated The FFI of the degree is not required. The above algorithm is disclosed in "KL57 〇 / vers e-Free Berlekamp-Massey Algorithm, Reed, Shin, and Truong, IEE Proceedings-E, Vol. 138, No. 5, September 1991". However, even Reed's algorithm FI? I must be used, but it requires a large number of FFM's, as high as 3 (t + l) or 27. If applied to errors and erasure correction, the number of FFM's required will be higher, usually errors Revision status 9 (please read the items on the back-before filling this page)--ο. This paper size is applicable to the Chinese National Slope (CNS) A4 specification (210X297 mm) A7 B7 «The sample rate in the Ministry of Economic Affairs Xibeigong Consumer Cooperative Co., Ltd. Printed oxygen 5. Doubled under the description of invention (6). Therefore, when implementing the algorithm, a non-inversion method and its device without using FFIs and minimizing the number of FFMs are As expected, one object of the present invention is to provide a method and apparatus for solving a polynomial of a key equation when decoding a codeword. Another object of the present invention is to provide a method based on Berlekamp · Massey calculus. Method and its device, which can be implemented with the least hard lines. Another object of the present invention is to provide a method and a device for solving the polynomial of the healthy equation, which will not reduce the overall decoding speed of the decoder. In other words, in a better implementation In the example, a method for calculating a suspected locator polynomial and a false evaluator polynomial in an incorrectly-corrected brown solution to the key equation solution step formula is disclosed, and a polynomial is generated through some intermediate steps of the above method The above intermediate steps can be funded by the minimum number of hard-wired lines. The number of intermediate steps requires a corresponding number of operation time periods to complete the operation of the polynomial. However, depending on the selected (N, K) ma, the number of calculation time periods required to calculate the polynomial will be within the range of the time required to calculate up-stream data. In particular, it is efficient. Cut a small number of finite field multipliers (FFM's) without using a finite field inverter (FH's). The better method for calculating the error locator polynomial and the error evaluation polynomial is also revealed (please read the precautions on the back first) (Fill in this page again) AIV. Order CI. The paper size is used in China National Standards (CNS) 8 4 size (210X297 mm) Central Government Standards Bureau of the Ministry of Economic Affairs Consumer Cooperatives A7 B7 '' V. Invention Explanation (7). Using these new methods, implement the method derived from the Berlekamp-Massey algorithm without inversion. A structure with area efficiency (saving circuit area) using only 3 FFM's without FFI's is also disclosed. This method and architecture can be widely applied to various RS and BCH codes with appropriate coding length. The present invention provides a method and device. Another advantage of the present invention is to provide a method and device. Based on the Berlekamp-Massey algorithm, the number of hard II lines can be minimized. --------- o ^ .-- (please read the $ item on the back before filling in this page)
本發明之又一優點為提供一種方法及裝置,用以解 ------- 答鍵方兔廛篇爲名 本發明之上述及其他特徵、優點在參照圈式及閲讀 下列本發明之詳細說明後將會更顯而易僅。式之簡單說明: 第la圈顢示解碼具錯誤訂正功能碼字時之處理方 瑰圈; 第lb圈顯示解碼具錯誤及抹除訂正功能碼字時之 處理方塊圖;以及 第2圈顯示一具3個FFM架構之較佳實施例,以 實施本發明之鍵方程式解答器。符统說明: 10、20〜徵兆計算器; 11 訂 本纸張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) A7 B7 _ 五、發明説明(8 ) 12、22〜鍵方程式解答器; 14 ' 24〜Chien搜尋器; 16〜錯誤求值器; 26〜錯誤及抹除值求值器; 30〜輸入端; 32、46、48〜有限場乘法器; 34、50〜有限場加法器; 36、40、44、54〜暫存器; 38、52〜輸出端; 42、58〜多工器; 56、60〜緩衝器; 62〜控制器;以及 CLK1、CLK2、CLK3 ~ 時脈信號》 較隹實施例: 參照在此所使用之符號,沒有〃之符號如Ω和σ 係引用原始Berlekamp-Massey演算法(具有反轉),而有 "之符號如冷,A左,Λτ;係引用無反轉演算法。習知技 術的無反轉Berlekamp-Massey演算法為一有2t步麻之 (請先虾讀背氙之注意事項再填寫本頁)Another advantage of the present invention is to provide a method and device for solving the above-mentioned and other features and advantages of the present invention. After detailed explanation will be more obvious and easy. Brief description of the formula: The first circle shows the processing circle when the decoder has the error correction function code; the first circle shows the processing block diagram when the decoder has errors and the correction function code is erased; and the second circle displays one A preferred embodiment with three FFM architectures to implement the key equation solver of the present invention. Explanation of symbols: 10, 20 ~ Symptom calculator; 11 The size of the paper is applicable to the Chinese National Standard (CNS) A4 (210X297 mm) A7 B7 _ 5. Description of the invention (8) 12, 22 ~ Key equation solver 14'24 ~ Chien searcher; 16 ~ error evaluator; 26 ~ error and erasure value evaluator; 30 ~ input end; 32, 46, 48 ~ finite field multiplier; 34,50 ~ finite field addition 36, 40, 44, 54 ~ temporary register; 38, 52 ~ output; 42, 58 ~ multiplexer; 56, 60 ~ buffer; 62 ~ controller; and CLK1, CLK2, CLK3 ~ clock SIGNAL》 Comparative Example: Referring to the symbols used here, no symbols such as Ω and σ refer to the original Berlekamp-Massey algorithm (with inversion), and symbols with " such as cold, A left, Λτ ; Refers to the non-reverse algorithm. The inversion Berlekamp-Massey algorithm of the known technology is a 2t step (please read the precautions of back-xenon before filling this page)
、1T AVT. «濟部中*標準局貝工消费合作社印裂 反復演算法,如下列所示: 初始條件: π-υ=ο; (行 1) Λ δ =1; (行2) 冷 Μ-1)⑻=^-1)⑻= Λ(χ); (行 3) = 7>+1 备浐_1>+7>分浐-1)+ . · . · (行 4) 12 本紙張尺度適用中國國家標率(CNS ) Α4規格(210Χ297公釐) A7 B7 五、發明説明(9 ) =Tp+iA〇 + Γ/,Λι + · . . + ΆΚΡ for / = /?toN-K - 1 冷(,)⑻=彡.今(卜1巾)+幺(〇x令(卜υ(χ) (行5) (行6) (行7) «濟部中央橾準局負工消费合作社4-* 么(,+ |) = r, + 2W0 + + …+Γ«-»,+ Ρ + 2 分 d (行 8) If Α(<)=0ογ2£)(μ)2/+1 (行 9) DV)=Do-t). ^ίΟ(χ) = XT^-·>(*); (行 10) else (行 H) Di0=i+\ - Dv~l)> <〇ί ^ (0(*) = ff(<_,)(^); (行 12) 其中,p為抹除(erasure)之數目,介於〇 < P S N-K ; Λ〇ο=Π(1+α沁,Λ為抹除集合(erasure set); Tj’s為 Forney 徵兆多項式T(x)之係數,其中T(x) = Λ(χ)5(χ) mod xN K ; 以0W為第i步驟錯誤及抹除定位器多項式,多項式之最 高次為Vi+p;妙4為冷(0W之係數;么(<> 為第1步驟之不符 值(discrepancy),》為先前產生之不符值,r(<)(*)為輔助 多項式,D⑴為輔助次變數。在此,演算法係提供錯誤及 抹除之訂正•若是沒有任何抹除,則P = 〇’ T(X) = S(x), 以及冷-1⑻=卜〇〇 = 1,而演算法即簡化為較簡單之形式β 由無反轉Berlekamp-Massey演算法所得之新錯誤 及抹除定位器多項式彡⑻,可用以發現和由原始 Berlekamp-Massey演算法所發現之σ(χ)相同之錄誤位 置· 如以上所顧示,第i步播無反轉Berlekamp-Massey 演算法包含以下兩組方程式(上述之行7和8): σ(,) (χ) =ί·^〇-ΐ) (*) + * r(, ·(*) (equ. 1) 13 本纸張尺度遢用中國國家標準(CNS ) A4規格(210X297公漦) {請先«讀背面之注$項再填寫本頁) C. -* C. 鑪濟部中央標丰局貝工消费合作杜印氧 A7 ___ B7 , 五、發明説明(10) 厶 << + ” = :Γ< + 2 分 P+ Γί + 1 分 V)+ …^ + 2 (eqa2) 在本發明中,我們提出如下之演算定義: σ(Γ°» for j = Q (equ. 3a) σψ = δ σ)ι'ϋ + for \^j<.Vi+P (equ. 3b) 左(’)=kS!].l + p + Γ, _ v,., + P+1 σ ry.\ VP . for j = 0 (eqa4a) Δ(J+,} = ^1-^+ for y= i^7^v/+^ (equ.^b) 其中彡y),i為彡⑺⑶之係數,而乡⑺糾0 +糾0χ +...+吟”, 巧切為沒w(jc)多項式之最高次,分是f (<)〇〇之係數。 么V+u ·ί為計算λ(< + ”之部份結果》 藉上述定義之在浐以及Α (广",在每一運算時間遇 期中,在計算;f時僅須使用2個FFMs,且在計算 么(/<+1)時僅须使用一個FFM ·使用此一方式,在一運算 時間遇期中僅須用到3個FFMs。 藉由分解原始方程式(equ. 1和2)成為較小計算量 之序列(equ. 3a, 3b,4a,及4b),可將必要FFMs之數目快 速大幅降低。然而,在任一運算時間遇期中計算每一各 別之值時’在_和衫⑼之間可能會有資料相依性(data dependency)。表二顯示此分解演算法之資料相依性: (請先M-讀背面•之注f項再填寫本頁), 1T AVT. «Repeated algorithm for the printing and cracking algorithm of the Peking Consumer Cooperative of the Ministry of Economic Affairs of the Ministry of Standards of China, as shown below: Initial conditions: π-υ = ο; (line 1) Λ δ = 1; (line 2) Cold Μ -1) ⑻ = ^-1) ⑻ = Λ (χ); (line 3) = 7 > +1 backup 浐 _1 > + 7 > minute -1) +. ·. · (Line 4) 12 sheets The scale is applicable to China's national standard rate (CNS) A4 specification (210 × 297 mm) A7 B7 V. Description of the invention (9) = Tp + iA〇 + Γ /, Λι + ·.. + ΆΚΡ for / = /? ToN-K- 1 Cold (,) ⑻ = ⑻. 今 (卜 1 巾) + 幺 (〇x 令 (卜 υ (χ) (Line 5) (Line 6) (Line 7) 4- * ((, + |) = r, + 2W0 + +… + Γ «-», + Ρ + 2 points d (line 8) If Α (<) = 0ογ2 £) (μ) 2 / + 1 (Line 9) DV) = Do-t). ^ ΊΟ (χ) = XT ^-· >(*); (line 10) else (line H) Di0 = i + \-Dv ~ l) > < 〇ί ^ (0 (*) = ff (< _,) (^); (line 12) where p is the number of erasure, between 〇 < PS NK; Λ〇ο = Π ( 1 + α Q, Λ is erasure set; Tj's is the coefficient of the Forney symptom polynomial T (x), where T (x) = Λ (χ) 5 (χ) mod xN K; 0W is the ith Wrong steps And erase the locator polynomial, the highest degree of the polynomial is Vi + p; Miao 4 is cold (a coefficient of 0W; (< > is the discrepancy value of step 1), "is the discrepancy value generated previously, r (<) (*) is the auxiliary polynomial, and D⑴ is the auxiliary variable. Here, the algorithm provides corrections for errors and erasures. • If there is no erasure, then P = 〇 'T (X) = S ( x), and cold-1⑻ = 卜 〇〇 = 1, and the algorithm is simplified to a simpler form. β The new error and erasure locator polynomial 所得 obtained from the non-inversion Berlekamp-Massey algorithm can be used to find The same recording error position as σ (χ) found by the original Berlekamp-Massey algorithm. As shown in the above, the i-th step without inversion Berlekamp-Massey algorithm contains the following two sets of equations (the above line 7 and 7) 8): σ (,) (χ) = ί · ^ 〇-ΐ) (*) + * r (, · (*) (equ. 1) 13 This paper uses China National Standard (CNS) A4 (210X297 public money) {Please read «Note $ on the back side before filling in this page) C.-* C. Duoyang A7 ___ B7, the co-operation between shellfish and consumer goods of the Central Biaofeng Bureau of the Ministry of Economic Affairs, V. Description of the invention (10 )厶 < < + ”=: Γ < + 2 points P + Γί + 1 point V) +… ^ + 2 (eqa2) In the present invention, we propose the following calculus definition: σ (Γ °» for j = Q (equ. 3a) σψ = δ σ) ι'ϋ + for \ ^ j < .Vi + P (equ. 3b) Left (') = kS!]. l + p + Γ, _ v,., + P +1 σ ry. \ VP. For j = 0 (eqa4a) Δ (J +,) = ^ 1-^ + for y = i ^ 7 ^ v / + ^ (equ. ^ B) where 彡 y), i is The coefficient of 彡 ⑺⑶, and the township correction 0 + correction 0 χ + ... + yin ", is cut to the highest degree without w (jc) polynomial, and the point is a coefficient of f (<) 〇〇. Then V + u · ί is the partial result of calculating λ (< + ”.” By the above definitions of 浐 and Α (广 ", in each calculation time period, only 2 FFMs, and only one FFM is required in the calculation (/ < +1). Using this method, only three FFMs are required in a calculation time period. By decomposing the original equations (equ. 1 and 2 ) Becomes a sequence of smaller calculation amount (equ. 3a, 3b, 4a, and 4b), which can quickly and drastically reduce the number of necessary FFMs. However, when calculating each individual value in any calculation time period, 'in_ There may be a data dependency between the shirt and the jacket. Table 2 shows the data dependency of this decomposition algorithm: (please read M-note on the back of the first page before filling in this page)
Cycle V ♦” σ^{χ) j = 〇 j = l j = 2 ^ (〇 = ^ (v?. , + P + Γ, _ V, ., + χ, + ! σ ί; : ,*> p ^V + ,) = Tt,2^^〇 种)=》·釣-〇 +奶约-ι> σψ = ί·σ§-° + j=Vi+ ^ V( : y = ^ v, ; y. Η· T,.„ _p+ 3σ %p., • 镰· 表二 14 本纸張尺度逍用中國國家揉準(CNS )八4規格(210X29*7公釐) 鍾濟部中央樑準局貝工消费合作社印褽 A7 B7 > 五、發明説明(11) 如表二所示,於運算時間週期j計算幺穴”需要 (^仏和之值’其於運算時間週期j_l已計算出。同 理’於運算時間週期j計算备需要妒和巧-〇,其於運 算時間遇期0、及(i-Ι)步驟中已分別計算出。附錄A顯 示使用較佳實施例之演算法產生錯誤及抹除求值器多 項式和錯誤及抹除定位器多項式之處理步驟。 使用上述之分解演算法,使得以一 3_FFM實施之 無反轉Berlekamp-Massey演算法作為鍵方程式解答器成 為可能,且如第2圈所示。一第一 FFM 32,一第一有限 場加法器34(FFA)’及暫存器36係用以計算不符值。 在第i步驟之第j運算時間週期操作中,FFM 32接收 Forney徵兆值Ti.j+3作為第一輸入並接收於⑻之第(jq) 係數<^2,作為第二輸入。FFA 34和暫存器36累積乘算結 果β關於輸出端38,當錯誤及抹除求值器多項式谷⑻之 係數計算完成時’係數之值將會提供至輸出端38。 FFMs 46和48、以及FFA 50計算鏘誤及抹除定位 器多項式&W之係數。FFM 46接收不符值A⑴作為一輪 入,另一輸入為祀“乘法器58和緩衝器6〇允許枚0之 選择及貯存。緩衝器56和60貯存由前一步驟所得彡⑻和 户(X)之係數’乘法器58選擇衧-〇之新值。ffM 48接收一 輪入外及另一輸入》。乘法器42和暫存器44斟酌》之 選擇和貯存。由FFMs46和48之輸出由FFA加蟪起來, 產生岬-»。衫也貯存於暫存器54並回授至緩衝器56 和FFM 32。如果A<0 = 0或是2Ζ^-”2,·+ι,那麼印=对-1>且 15 U張^^用中國S家揉準(CNS ) A4〇^ ( 21GX297公釐) ---~ --------JQ------ir------Ό (請先聞讀背面之注意事項再填寫本頁) «濟部中*標率扃貝工消#合作社印釁 A7 ______B7 五、發明説明(12) 乡保持不變;否則,令p =&(丨-〇且》=含(〇 〇暫存器在此係作 為延遲元件,由控制器62產生之不同之時脈信號 CLK1、CLK2、及CLK3所控制。暫存器44在每一步驟 之第一運算時間週期將其内部值加以更新。暫存器36 在每一步驟之第二運算時間週期中其内部值歸零。經過 2t個步驟後,輸出值&可由輸出端52獲得。 這一架構可使用於錯誤訂正或是錯誤及抹除訂 正。相較於先前提出之架構其需4t至6t個FFM’s以實 施鏘誤及抹除訂正或是需2t至3t個FFM,s以實施錯誤 訂正,本發明之較佳資施例可以大大地減少硬艟複雜度 至僅需3個FFM’s❶然而,為了完成第i步驟演算法, 較佳實施例之架構需要Vi+ p+l個運算時間遇期,但是 習知技術之架構僅需要2至3個運算時間週期。 使用本發明之架構,用以產生資料所須增加之額外 時間並不會減慢系統整艟之處理速度。原因之一為習知 技術架構並未將時間和破Λ之使用加以同步。難然習知 技術在任一級之計算結果可快速獲得,但是為了使任何 資料被接收和處理,卻必须等待由上游步驟所得之結 果。 此外,本發明之方法及其裝置,藉由使用舆計算 行(X)相同之線路,而將硬Λ加以最小化。計算錄誤及抹 除求值器多項式Ω〇〇之傳統方式為舆σ(ΛΓ)之計算作平行 之運算·使用Berlekamp-Massey演算法,這程序包括用 以計算Ω(〇(χ)之一 2t步驊反ft演算法。然而,如果最高次 16 本纸張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請尤閲讀背*.之注意事項再填寫本頁) 訂 經濟部中央揉準扃貝工消费合作杜印* A7 B7 五、發明説明(13) 為V+P之σ(χ)已先得到,則由鍵方程式和牛頓等式 (Newton’s identity)可得: Ω(Λ:)=θ(χ)σ(χ)ηκχΙ^_Λ:Cycle V ♦ ”σ ^ {χ) j = 〇j = lj = 2 ^ (〇 = ^ (v ?., + P + Γ, _ V,., + Χ, +! Σ ί;:, * > p ^ V +,) = Tt, 2 ^^ 〇 species) = "· fishing -〇 + milk about -ι > σψ = ί · σ§- ° + j = Vi + ^ V (: y = ^ v,; y Η · T ,. „_p + 3σ% p., • Sickle · Table 2 14 This paper size is free to use Chinese National Standard (CNS) 8-4 (210X29 * 7 mm) Industrial and Consumer Cooperatives Seal A7 B7 > V. Description of the Invention (11) As shown in Table 2, calculating the acupoints at the calculation time period j "requires (^ 仏 和 的 值 ', which has been calculated at the calculation time period j_l. Same as The calculation of calculations in the calculation time period j requires jealousy and Q--0, which have been calculated separately in the calculation time period 0 and (i-1) steps. Appendix A shows that using the algorithm of the preferred embodiment generates an error And erasing the evaluator polynomial and the error and erasing the locator polynomial. Using the above-mentioned decomposition algorithm, it is possible to use a 3_FFM non-inversion Berlekamp-Massey algorithm as a key equation solver, and such as As shown in lap 2. FFM 32, a first finite field adder 34 (FFA) 'and a register 36 are used to calculate the discrepancy value. In the jth operation time period operation of the i-th step, the FFM 32 receives the Forney symptom value Ti.j + 3 as the first input and received from the (jq) coefficient < ^ 2 as the second input. The FFA 34 and the register 36 accumulate the multiplication result β. With regard to the output 38, the error and erasure evaluator When the calculation of the coefficient of the polynomial valley is completed, the value of the coefficient will be provided to the output terminal 38. FFMs 46 and 48, and FFA 50 calculate the coefficient of the error and erase the locator polynomial & W. FFM 46 receives the non-compliance value A⑴ as One round in, the other input is "the multiplier 58 and the buffer 60 allow the selection and storage of 0. The buffers 56 and 60 store the coefficient 'multiplier 58' of the unit X (H) obtained in the previous step. Select the new value of 衧 -〇. FfM 48 receives one round of input and another input. The multiplier 42 and the register 44 consider the selection and storage. The outputs from FFMs 46 and 48 are added by FFA to generate the cape- ». Shirts are also stored in register 54 and fed back to buffer 56 and FFM 32. If A < 0 = 0 or 2Z ^-" 2, · + Ι, then India = right -1 & 15 U Zhang ^^ using Chinese S family standard (CNS) A4〇 ^ (21GX297 mm) --- ~ -------- JQ --- --- ir ------ Ό (Please read the notes on the back before filling out this page) «Jibuzhong * standard rate 扃 贝 工 消 # Cooperative Society Printing A7 ______B7 V. Description of Invention (12) Township Remain the same; otherwise, let p = & (丨 -〇 and》 = (00) register is used here as a delay element, the different clock signals CLK1, CLK2, and CLK3 generated by the controller 62 control. The register 44 updates its internal value during the first operation time period of each step. The internal value of the register 36 returns to zero during the second operation time period of each step. After 2t steps, the output value & can be obtained from the output terminal 52. This structure can be used for error correction or error and erasure correction. Compared with the previously proposed architecture, it needs 4t to 6t FFM's to implement correction and erasure correction or 2t to 3t FFM, s to implement error correction. The preferred embodiment of the present invention can greatly reduce the hard艟 Complexity to only 3 FFM's❶ However, in order to complete the i-th step algorithm, the architecture of the preferred embodiment requires Vi + p + 1 computation time periods, but the architecture of the conventional technology requires only 2 to 3 Operation time period. With the architecture of the present invention, the additional time required to generate data does not slow down the system's overall processing speed. One of the reasons is that the conventional technology architecture does not synchronize the use of time and time. It is difficult to know that the calculation results of the technology at any level can be obtained quickly, but in order for any data to be received and processed, it is necessary to wait for the results obtained by the upstream steps. In addition, the method and apparatus of the present invention minimize the hard Λ by using the same line as the calculation line (X). The traditional way of calculating errors and erasing the evaluator polynomial Ω〇〇 is to calculate σ (ΛΓ) in parallel. Using Berlekamp-Massey algorithm, this program includes one of the calculations of Ω (〇 (χ) 2t step inverse ft algorithm. However, if the highest 16 paper sizes are applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the notes on the back *. And then fill out this page) Order the Ministry of Economic Affairs The Central Government's Consumer Cooperation Du Yin * A7 B7 V. Description of the Invention (13) σ (χ) for V + P has been obtained first, then it can be obtained from the bond equation and Newton's identity: Ω ( Λ:) = θ (χ) σ (χ) ηκχΙ ^ _Λ:
Q=iS+i〇ir+-...4i5q; i=Ql,...,v+/>-L 也就是說在得出σ(χ)之後,實施Ω〇〇之計算可以直接且更 有效率。如Reed等人所示範,利用無反轉Berlekamp-Q = iS + i〇ir + -... 4i5q; i = Ql, ..., v + / > -L That is, after σ (χ) is obtained, the calculation of Ω〇〇 can be directly and more effectiveness. As demonstrated by Reed et al., The use of inversion-free Berlekamp-
Massey演算法,备(x) = CCT(x);因此,藉由直接計算,可 得出下列結果: ό (x) = S(x)a (x) mod xN~K > =CQ(x). 使用Forney演算法,顢示冷(x)和6⑻可產生相同之錯誤 及抹除訂正值 ^ 4m^)=e/. 〇 各m c&(pri) 此外,可看出合《•之計算和之計算相似。因此, 在得出冷w後,用以計算# w之相同硬艎可再建構用以計 算6〇c)。么可以如下方式計算: 6^ = 5, + ,0-0 ,forj = 0 特別地,再參照第2圈,FFM 32、FFA 34和暫存 器36係用以計算。為求第i係數之第j運算時間週 期中,FFM32接收徵兆值Si#作為一輸入,冷W的第j 係數心作為另一輸入。FFA 34和暫存器36用以累積乘 算結果。當經2t個步驟而得出々(X)時,其係數孑/將會 被存入緩衝器56。經由設定緩衝器60或暫存器40之輸 17 本紙張尺度適用中國國家標準(CNS ) A4規格(2丨0·〆297公釐) --------ο------ix------C, (請$讀背面之注意事項再填寫本頁) A7 ______B7 五、發明説明(14) 出為0,以及暫存器44之輸出為1,緩衝器56之輸出可 被循環並回授至FFM 32之輸入。直接計算之輸出 (di),可在輸出端38得到。 在考量使用本實施例之3-FFM架構,計算汐(JC)和 占⑻所須之運算時間週期之總數,其對系統整艘效能表 現之濟在影響是我們所關心的。從反復演算法中,嫌明 W⑻之最高次(degree)在每一重覆過程中最多增加1β 因此,方程式v^vu丨+1用以投定Vj+p之上限。 (1)鏘誤訂正,和(2)錯誤及抹除訂正之結果分別示 範如下。若是只需訂正錯誤,則2t < N - K,計算A(0)只 須一運算時間遇期,而且V,Q·,/仰0U以,及 νβί,/οτΉΚΜ »計算汐⑻所須之運算時間遇期數目為: 21-1 f-1 2<-1 ^ 1 Σ(ν<+1)$Σ(ί+1)+Σ(ί+1)=4ί2+晏^。 /*〇/«〇/«# 2 2 計算ά⑻所須之運算時間遇期數目為: =τί2+-τί ° 因此,所須運算時問週期之總數小於2t2 + 2t+l 如果錯誤和抹除兩者均訂正,則2t + p<==N-K, 計算初始Δ<Λ須p + 1運算時間週期,且 ν,^ρ + i, for 0^i<t » v,^p+tf for ⑸<2i。計算彡⑻所須運 算時間遇期之數目為: 1) = 1^+(2/,+2)/ 2i-l r-1 2i-l Σ(ν,+ι)<Σ(ρ+ι+1)+ΐ>+ί+ i«0 <*0 計算ά⑻所須運算時間週期之數目為: 18 本紙張尺度適用中國國家搮準(CNS ) A4说格< 2丨OX297公釐) 經濟部中央梯準局負工消費合作社印«. Λ 7 Β7 五、發明説明(15) 因此,所須運算時間週期之總數小於2t2 + (3p + 2)t + (l/2)p(p + 1) + ρ+1»因為t和p均為整數,所以在2t+p <=N-K之限制條件下,沒有會使運算時間週期之總數 成為最大之(t, p)閉销型公式(closed-form formula)。取而 代之,各種不同之(N,K)RS碼,其N-K之編碼長度範圍 介於4至16者,運算時間週期所須之總數已計算出並 圈示於表三。若是N大於所須之運算時間週期數目,則 本發明之方法及其裝置因此可用以減少硬體複雜度,而 仍保持整艟之解碼速度。 , N-K t p cycles 4 2 讎 13 4 1 2 16 6 3 - 25 6 1 4 31 8 4 - 41 8 2 4 51 10 5 - 61 10 2 6 76 12 6 • 85 12 3 6 106 14 7 - 113 14 3 8 141 16 8 - 145 16 4 8 181 表三 在通訊和貯存系统中,BCH和RS碼有多種之應 用,其均可以由本發明之方法及其裝置獲益。例如,數 位影音光碟(digital versatile disks ; DVDs)係使用 RS 產 生碼,其在列方向為(182,172),而在攔方向為(208,192); 19 --------ο------II------C (請先閲讀背面之注意事項再填寫本頁) 本纸張尺度適用中國國家揉準(CNS 规格(210X297公釐) A7 B7 五、發明説明(16) 數位電視廣播使用(204,188)之RS碼;CD-ROM使用多 组較小之RS碼,包括(32,28)及(28,24);無線通訊中, AMPS蜂巢式行動電話系統使用(40,28)及(48,36)之二進 位BCH碼,其均為(63,51)碼之縮減碼。可訂正2個錯誤 (N-K=12,m=6)之(63,51)碼,需小於12運算時間週期 (t=2,表三之第一列)》所有諸如此類和其他方面之應 用,均可從本發明之方法及其裝置有所獲益。 雖然本發明已以較佳實施例揭露如上,然其並非用 以限定本發明,任何熟悉本項技藝者,在不脫離本發明 之精神和範園内所提出之修改和潤飾,均係涵括在本發 明之保護範圍内,而本發明之保護範圍視後附之申請專 利範圍所界定者為準》 --------Q------1T------ο, (請先閲讀背*·之注意事項再填寫本頁) 經濟部中央標準局貝工消费合作社印製 本纸張尺度適用中國國家揉準(CNS ) A4規格(2丨0X297公釐) 經濟部中央標率局Λ工¾费合作社印«. A7 B7 五、發明説明(17)Massey algorithm, prepare (x) = CCT (x); Therefore, by direct calculation, the following results can be obtained: ό (x) = S (x) a (x) mod xN ~ K > = CQ (x ). Using the Forney algorithm, showing that cold (x) and 6⑻ can produce the same error and erase the correction value ^ 4m ^) = e /. 〇 mc & (pri) In addition, it can be seen that the calculation of "• And the calculation is similar. Therefore, after the cold w is obtained, the same hard frame used to calculate #w can be reconstructed to calculate 60c). It can be calculated as follows: 6 ^ = 5, +, 0-0, forj = 0. In particular, referring to the second circle, FFM 32, FFA 34, and register 36 are used for calculation. To find the j-th operation time period of the i-th coefficient, FFM32 receives the symptom value Si # as an input, and the j-th coefficient center of the cold W as another input. The FFA 34 and the register 36 are used to accumulate multiplication results. When 々 (X) is obtained after 2t steps, the coefficient 孑 / will be stored in the buffer 56. Input 17 via setting buffer 60 or register 40 This paper size applies Chinese National Standard (CNS) A4 specification (2 丨 0 · 〆297 mm) -------- ο ------ ix ------ C, (Please read the notes on the back and fill in this page again) A7 ______B7 V. Description of the invention (14) The output is 0, and the output of register 44 is 1, and the output of buffer 56 is 1. Can be looped and fed back to the input of FFM 32. The directly calculated output (di) is available at output 38. When considering the use of the 3-FFM architecture of this embodiment, calculating the total number of calculation time periods required for the JC and the occupants, their impact on the overall performance of the system is of concern to us. From the iterative algorithm, it is suspected that the highest degree of W⑻ will increase by 1β at most during each iteration. Therefore, the equation v ^ vu 丨 +1 is used to determine the upper limit of Vj + p. The results of (1) error correction, and (2) error and erasure correction results are shown below. If it is only necessary to correct the error, then 2t < N-K, the calculation of A (0) only needs a calculation time period, and V, Q ·, / 扬 0U and νβί, / οτΉΚΜ » The number of calculation time encounters is: 21-1 f-1 2 < -1 ^ 1 Σ (ν < +1) $ Σ (ί + 1) + Σ (ί + 1) = 4ί2 + Yan ^. / * 〇 / «〇 /« # 2 2 The number of calculation time periods required to calculate ά⑻ is: = τί2 + -τί ° Therefore, the total number of calculation time periods required is less than 2t2 + 2t + l If both are corrected, then 2t + p < == NK, calculate the initial Δ < Λ shall be p + 1 operation time period, and ν, ^ ρ + i, for 0 ^ i < t »v, ^ p + tf for ⑸ < 2i. The number of time periods required to calculate 彡 ⑻ is: 1) = 1 ^ + (2 /, + 2) / 2i-l r-1 2i-l Σ (ν, + ι) < Σ (ρ + ι +1) + ΐ > + ί + i «0 < * 0 The number of calculation time periods required to calculate ά⑻ is: 18 This paper size applies to China National Standards (CNS) A4 parlance < 2 丨 OX297 mm) Printed by the Central Laboratories of the Ministry of Economic Affairs and Consumer Cooperatives «. Λ 7 Β7 V. Description of Invention (15) Therefore, the total number of calculation time periods required is less than 2t2 + (3p + 2) t + (l / 2) p (p + 1) + ρ + 1 »Because t and p are integers, under the constraint of 2t + p < = NK, there is no (t, p) closed-pin formula that will maximize the total number of operation time periods. (Closed-form formula). Instead, for various (N, K) RS codes, the N-K encoding length ranges from 4 to 16, and the total number required for the calculation time period has been calculated and shown in Table 3. If N is greater than the required number of operation time periods, the method and device of the present invention can therefore be used to reduce hardware complexity while still maintaining a tidy decoding speed. , NK tp cycles 4 2 雠 13 4 1 2 16 6 3-25 6 1 4 31 8 4-41 8 2 4 51 10 5-61 10 2 6 76 12 6 • 85 12 3 6 106 14 7-113 14 3 8 141 16 8-145 16 4 8 181 Table 3. In communication and storage systems, BCH and RS codes have multiple applications, all of which can benefit from the method and device of the present invention. For example, digital versatile disks (DVDs) use RS to generate codes, which are (182,172) in the column direction and (208,192) in the block direction; 19 -------- ο-- ---- II ------ C (Please read the precautions on the back before filling this page) This paper size is applicable to the Chinese national standard (CNS specification (210X297 mm) A7 B7 V. Description of the invention (16 ) Digital TV broadcasting uses (204,188) RS code; CD-ROM uses multiple smaller RS codes, including (32,28) and (28,24); In wireless communication, AMPS cellular mobile phone system uses (40 , 28) and (48,36) are binary BCH codes, which are reduced codes of (63,51) codes. (63,51) codes of 2 errors (NK = 12, m = 6) can be corrected, Requires less than 12 computing time periods (t = 2, the first column of Table III). All applications such as these and other aspects can benefit from the method and device of the present invention. Although the present invention has been better implemented The example is disclosed as above, but it is not intended to limit the present invention. Any person familiar with the art, without departing from the spirit and scope of the present invention, proposes modifications and retouching. Within the scope of protection of the present invention, and the scope of protection of the present invention shall be determined by the scope of the attached patent application "-------- Q ------ 1T ------ ο, (Please read the notes on the back * · before you fill out this page) The paper size printed by the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative, is applicable to the Chinese National Standard (CNS) A4 specification (2 丨 0X297 mm) Standards Bureau Λ Labor ¾ Fees Cooperatives Seal «. A7 B7 V. Description of Invention (17)
附錄A 無反棘Berlekamp-Massev分解演翼法 第I部份:(計算冷⑷)Appendix A Berlekamp-Massev Decomposition and Wing Decomposition Method Part I: (Calculate Cold Heading)
Dip~l)=〇, »=l; /* 初始條件 */ ^ = Tp^K.*TPK^ *Tthp for ϊ = /? to N - K - 1 begin /* 外部迴圈開始 */ 鉍、於r) / if i != p 在⑴=M?.i + p + r卜、,…AH; Δ(〇+,)= 0 for j = 1 to v,+p begin/* 内部迫圈開始 */ 々卜沾+妒⑽ Α〇+» Α«+·) ^ ^ Λ<〇 **> = Δ/-ι + Γι-/+3 tfy-i end loop /*結束内部迴圈*/Dip ~ l) = 〇, »= l; / * initial conditions * / ^ = Tp ^ K. * TPK ^ * Tthp for ϊ = /? To N-K-1 begin / * external loop start * / bismuth, In r) / if i! = P in ⑴ = M? .I + p + r BU ,, ... AH; Δ (〇 +,) = 0 for j = 1 to v, + p begin / * internal forced circle starts * / 々 卜 沾 + ⑽⑽ Α〇 + »Α« + ·) ^ ^ Λ < 〇 ** > = Δ / -ι + Γι-/ + 3 tfy-i end loop / * End internal loop * /
If Δί0 =0 ογ2£>(Μ) ϋ + 1 />ω = £^W) =xf)⑻; elseIf Δί0 = 0 ογ2 £ > (Μ) ϋ + 1 / > ω = £ ^ W) = xf) ⑻; else
2)(0 = ί + 1-办卜”,含=幺。,仝。(jc)=>(x); end loop /*結束外部迴圈 V •分+办r 〜···+cv” 21 本紙張尺度適用中困國家標率(CNS ) Α4规格(210Χ297公釐) --------Q------1T------C, (請讀背面之·ά意事項再填将本頁) A7 B7 五、發明説明(is) 会存〇 + dl JC + …+ θν+/> 。 第II部份:(計算一〇) Ω^0) = Si σο for i = 1 to υ+ρ-l begin tL、= for j = 1 to i begin end loop end loop ft ⑻=d+.··+〇-丨 Δ Λ。+ Al + ... + Av + p-l jpV+P 1 e -- (請4閲讀背v#之注$項再填寫本頁) 訂 -C1. 經濟部中央樣率局貝工消费合作社印裝 本纸張尺度適用中國國家標率(CNS ) A4規格(210X297公釐)2) (0 = ί + 1-do Bu ", including = 幺., The same. (Jc) = >(x); end loop / * End the external loop V • min + do r ~ ·· + cv ”21 This paper size is applicable to the national standard rate (CNS) Α4 specification (210 × 297 mm) -------- Q ------ 1T ------ C, (Please read the back · Please fill in this page and fill in this page) A7 B7 V. The invention description (is) will be stored 〇 + dl JC +… + θν + / >. Part II: (Calculate 1) Ω ^ 0) = Si σο for i = 1 to υ + ρ-l begin tL, = for j = 1 to i begin end loop end loop ft ⑻ = d +. · ++ 〇- 丨 Δ Λ. + Al + ... + Av + pl jpV + P 1 e-(Please read the note of v # and fill in this page again 4) Order-C1. Printed copy of the shellfish consumer cooperative of the Central Sample Rate Bureau of the Ministry of Economic Affairs Paper size applies to China National Standards (CNS) A4 specifications (210X297 mm)
Claims (1)
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/914,421 US6197302B1 (en) | 1996-07-18 | 1997-08-19 | Method of stimulating T cells |
Publications (1)
Publication Number | Publication Date |
---|---|
TW384425B true TW384425B (en) | 2000-03-11 |
Family
ID=25434341
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW87100122A TW384425B (en) | 1997-08-19 | 1998-01-06 | Method and apparatus for solving polynomial of key function when decoding error correction codes |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW384425B (en) |
-
1998
- 1998-01-06 TW TW87100122A patent/TW384425B/en not_active IP Right Cessation
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6374383B1 (en) | Determining error locations using error correction codes | |
EP0114938B1 (en) | On-the-fly multibyte error correction | |
US6119262A (en) | Method and apparatus for solving key equation polynomials in decoding error correction codes | |
US6347389B1 (en) | Pipelined high speed reed-solomon error/erasure decoder | |
TW297190B (en) | ||
US8335808B2 (en) | Method and apparatus for processing multiple decomposed data for calculating key equation polynomials in decoding error correction code | |
Chang et al. | New serial architecture for the Berlekamp-Massey algorithm | |
JPH0728227B2 (en) | Decoding device for BCH code | |
JP4134029B2 (en) | Soft decision decoding method for Reed-Solomon code | |
JPH10112659A (en) | Error correction decoding device | |
JP3345385B2 (en) | Chain search device | |
US9191029B2 (en) | Additional error correction apparatus and method | |
TW451124B (en) | Method and apparatus for decoding an error correction code | |
TW566008B (en) | Apparatus for solving key equation polynomials in decoding error correction codes | |
Zhang et al. | On the high-speed VLSI implementation of errors-and-erasures correcting Reed-Solomon decoders | |
US20080140740A1 (en) | Systems and methods for processing data sets in parallel | |
TW384425B (en) | Method and apparatus for solving polynomial of key function when decoding error correction codes | |
TW311189B (en) | The error-corrected decoding method and its apparatus for Reed-Soloman code | |
JP3343857B2 (en) | Decoding device, arithmetic device, and methods thereof | |
JP3233502B2 (en) | Decryption device | |
JPH09307458A (en) | Polynomial evaluation device for error correction | |
Lee et al. | Algebraic decoding of quasi-reversible BCH codes using band matrices | |
US7693927B2 (en) | Data processing system and method | |
JP3398560B2 (en) | Shortened error correction decoding device | |
JP3230888B2 (en) | Euclidean circuit |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GD4A | Issue of patent certificate for granted invention patent | ||
MK4A | Expiration of patent term of an invention patent |