EP1897264A2

EP1897264A2 - Frame-synchronisation method and device

Info

Publication number: EP1897264A2
Application number: EP06709524A
Authority: EP
Inventors: Sébastien c/o Get - Enst Bretagne HOUCKE; Guillaume c/o Get - Enst Bretagne SICOT
Original assignee: Get-Enst Bretagne; France Telecom SA; GET ENST BRETAGNE
Current assignee: Get-Enst Bretagne; Orange SA
Priority date: 2005-02-18
Filing date: 2006-02-20
Publication date: 2008-03-12
Also published as: US7995692B2; JP2008530933A; FR2882480B1; WO2006087497A3; JP5214977B2; US20090154623A1; FR2882480A1; WO2006087497A2

Abstract

The invention relates to a frame-synchronisation method and device. The inventive method consists in determining the start-of-frame position D. Said position D is identical to a position of a window H that slides along the length of the frame received Tr. For different possible window H positions and for different blocks B<SUB>i</SUB> of window H, the method calculates a syndrome S(i)d, and the start position D of the frame received Tr is that with the largest number of zero-valued elements of a syndrome S(i)d.

Description

FRAME SYNCHRONIZATION METHOD AND DEVICE

The present invention relates to the field of telecommunications. Within this field, the invention relates more particularly to so-called digital communications. Digital communications include in particular wireless communications whose channel of transmission is the air channel as well as wired communications.

Digital communications increasingly use more or less sophisticated channel coding. The principle of channel coding consists in adding to the information bits a controlled redundancy which will enable the receiver to detect the presence of transmission errors and possibly to correct them. Among the existing codes, the codes in blocks are more particularly distinguished: these codes correspond to a block of n _b bits of information a code word of size n _c bits with n _c > n _b . These additional bits make it possible to correct certain transmission errors; the more bits of redundancy and the better the correctness of the code. To decode and correct errors, the receiver must be able to find the beginning of the codewords in the bit sequence in order to be able to decode them correctly. This operation is commonly called frame synchronization.

The invention relates to frame synchronization techniques. Frame synchronization is generally achieved by periodically adding in the message to be transmitted an uncoded synchronization sequence known to the receiver. This sequence is then detected at the receiver thus allowing the synchronization of the frame to be decoded. Such a sequence is described in the respective reference articles: "Optimum Frame Synchronization" by James L.Massey, IEEE Transactions on Communications, Vol.Com-20, No.2, April 1972 and "Frame Synchronization Techniques" by Robert A. Scholtz, 0090-6778 / 80 / 0800-1204 $ 00.75, 1980 IEEE. The huge advances in channel coding, especially with the introduction of turbo codes, currently allow data to be transmitted with low signal-to-noise ratios, which leads to many errors in the synchronization sequence and thus unreliable synchronization. Several authors propose to insert a coded sequence; this makes it possible to partially benefit from the coding gain and consequently to increase the reliability of the synchronization. Such a sequence is described in the article entitled "Decoder-Assisted Channel Estimate and Frarne Synchronization of Turbo Coded Systems" by M.Mostofa K. Howlader, 0-7803- 7484-3 / 02 / S 17.00, 2002 IEEE, These known techniques have the disadvantage of requiring the addition of a known sequence which results in a decrease in the spectral efficiency, that is to say a decrease in the useful rate.

Also, the technical problem to be solved by the object of the present invention is to propose a frame synchronization method constituting data of code words which does not have the disadvantage of known methods and which, therefore, offers greater efficiency spectral.

A solution to the technical problem that is posed, according to the present invention, in that said frame synchronization method consists in determining a start position D of a received frame T ₁ , by means of a sliding window H on the received frame T. ₁ , of size Kn _c , divided into K blocks B ₁ of size n _c equal to that of a code word. The starting position D is equal to one of several possible positions of the sliding window H along the received frame T _1. The method is such that, for different possible positions of the window H, it calculates a syndrome Sj ' by block B ₁ , the starting position D of the received frame T ₁ . being that for which the number of zero elements syndromes SJ ^ ¹ calculated for this position is the largest; the method determining beforehand for each syndrome SJ the number of null elements, each syndrome being composed of elements. In addition, the subject of the invention is a device for synchronizing data frame constituting codewords, in order to determine a frame start position. The device comprises: windowing means sliding on a received frame T ₁ of size Kn ₁ , divided into K blocks Bj of size n _L equal to that of a code word, the position D of received frame start T ₁ being equal to one of several possible positions of the windowing means sliding along the received frame T ₁ . , a means for calculating a syndrome S ^ for different possible positions of the windowing means and for different blocks B ₁ of the windowing means, a means for determining a position for which the number of null elements of the SJ syndromes ^{1 1} calculated for this position is greater Ic, cε in case the calculation means determines a first number of zero elements ^syndromes} SJ, the frame start position D being equal to that position.

Thus, a method and a device according to the invention, which respectively relates to a synchronization method and a synchronization device, make it possible to determine the frame dcbtt1 position by taking it equal to the position of the sliding window which corresponds to the largest number of null elements.

The method and the frame synchronization device according to the invention solve the problem. Indeed, the synchronization is of the self-taught type; it does not require any learning sequence. Synchronization relies solely on the properties of linear block coding.

A synchronization according to the invention is particularly advantageous because the deletion of the synchronization sequence makes it possible either to increase the useful bit rate or to keep the same bit rate as before but to use codes having more bits of redundancy. that is to say to use lower efficiency codings and therefore having a greater correction power. And, as the number of redundancy bits increases, the synchronization performance of a method according to the invention increases.

A frame synchronization according to the invention does not use any known sequence inserted in the coded frame; it does not therefore require a synchronization sequence. It works with any type of block code (BCH, RS, product code, LDPC, ...). It does not require modifying the correcting code. It operates before decoding, that is, it does not take advantage of the coding gain. It is not very complex and is effective even in the presence of a high rate of binary error.

According to a preferred implementation, the steps of the method are determined by the instructions of an uncorrected frame synchronization program in an electronic circuit such as a chip itself which can be arranged in an electronic device such as a receiver. The frame synchronization method according to the invention can equally well be implemented when this program is loaded into a computing device such as a processor or equivalent whose operation is then controlled by the execution of the program.

Accordingly, the invention also applies to a computer program, including a computer program on or in an information carrier, adapted to implement the invention. This program may use any programming language, and be in the form of source code, object code, or intermediate code between source code and IeI object code only in a partially compiled form, or in any other form desirable to implement a method according to the invention. The information carrier may be any entity or device capable of storing the program. For example, little support! include storage means, such as a ROM, for example a CD ROM or a circuit ROM microclectronique, or a magnetic recording means, for example a diskette (floppy dise) or a hard disk.

On the other hand, the information medium can be a transmissible medium such as an electrical or optical signal, which can be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network.

Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question. Other features and advantages of the invention will become apparent from the following description given with reference to the accompanying figures given by way of non-limiting examples.

Figure 1 is a block diagram of the block coding for a Hamming code (7.4). Figure 2 is a diagram of the transmission delay of to bits of a transmitted frame.

Figure 3 is a diagram of the principle of a synchronization method according to the invention.

Figure 4 is a diagram of an emitted frame and a received frame. FIG. 5 represents a sliding window of a method according to the invention, on a frame received for different positions d of the window.

FIG. 6 schematically represents a first particular embodiment of a synchronization device according to the invention.

Figure 1 shows a schematic block coding scheme for a Hamming code (7.4). The principle of block coding consists in adding to the information bits B ^ b ₂ ,..., B ₄ a redundancy c _^ , c _h , c ~ controlled which will enable the receiver to detect the presence of transmission errors. and possibly correct them. The block codes correspond to a block of n _b bits of information a code mole of size n _c bits with n _t > n _h . The n _c - n _b redundancy elements are calculated as the sum of two of some bits of the information bit block. The Hamming code (7.4). in particular, has three redundancy bits such that each bit is calculated from three information bits. For example, the fifth bit of a coded codeword has \ cc a code of

Hamming (7.4) is equal to the sum modulo two of the second, third and fourth information bits:

For any block coding, there is therefore a linear relationship between the code word M ₁ and the corresponding information word m; m _c = m fi with G the generator matrix of the code word. It is a size matrix

n _b xn _c . The matrix generating the hamming code (7.4) which corresponds to

the example illustrated in FIG. 1 is equal to:

The code is said systematic if _h n first bits of code word correspond to the n _b bits of information. This is the case in our example.

The receiver receives the transmitted code word, possibly tainted with errors due to the transmission channel. Let r be the binary word received by the receiver, r can ^be written as: r = m e _c ® relation wherein m _c is the sent code word, e is an error vector of length n _c and θ represents the sum modulo two.

From this observation r, the receiver must be able to render the n _b bits of information. This is the subject of a decoding operation. The optimal decoding on binary data is to find the remote code word of

Hamming (i.e. the number of distinct bits in a bit-to-bit comparison) of the observation r.

The principle of decoding codes in blocks is as follows. The receiver has n _c - n _h bits of redundancy to correct 2 "- 1 possible errors, so the receiver can only correct a fraction of the errors in an arbitrary way. low weight is preferred, that is the correction of the error vectors e having few non-zero elements At the generator matrix of the code, corresponds a parity matrix P of size {n _c - n _b ) xn _€ such that: GP '' = 0 (1)

The decoding method consists of calculating the syndrome $ (j) of the observation r:

S (r) = rP ^τ = mfiP ^τ ® eP ^τ

Using the relation (1), this leads to: <S (r) = eP ¹ From the matrix P, the decoding method constructs the decoding matrix which bijecitively associates a syndrome with an error word in favoring low-weight error words.

The decoding method derives from the table (2) an estimate of the error e. Then, sum the error estimate to the observation to perform the correction. If the syndrome is zero this means that r is a codeword. The Hamming code (7.4) previously retained makes it possible to illustrate by an example this method. Let an observation r equal to: r = [00 01100]

The method calculates the syndrome of r: S (r) = rP ^τ = [011] This syndrome is non-zero, so r contains at least one error. Using the table (2), the method associates with this syndrome an error vector e of weight a: e = [l000000]

The closest code word m _c is therefore: m _t = [1001100] which corresponds to the information bit vector:

M ₁ = [1001]

This decoding method assumes that the receiver knows the beginning of the code word. However, the transmission not only introduces errors in the transmitted frame but, in addition, FIG. 2 introduces a transmission delay of to bits due to the transmission time between the transmitted frame T ₁ and the received frame T ₁ . Of course, this delay is not known prior to the receiver. To calculate the syndrome S (r _k ) associated with the received code word r, the receiver must be able to estimate the transmission delay to; a synchronization method allows the receiver to estimate this delay to. A synchronization according to the invention, illustrated by FIG. 3, is performed by means of a sliding window H on the received frame T ₁ . The window is divided into K blocks B, whose size is equal to that of a mol m _tk of code, ie n _t . The size of this synchronization window is therefore equal to Kn ₁ so that once the synchronization is obtained, the window contains exactly K codewords. According to the example, K is equal to three Typically, the method calculates a syndrome Sf for each position of the window on the frame, d being able to vary from zero to n _L -1, and, for a given position, for each block B ₁ . SJ ^{ is therefore a vector of size n _c -n _b . According to an alternative, the method can perform the calculation of syndromes for some positions of the window on the frame, and for these positions, for some blocks B ₁ .

The method calculates the number of null elements in the syndromes for a given position of the window H. The method takes as an estimate of t ₀ , the position d for which the number of null elements for the syndromes associated with the same position of the window is the largest.

In the particular case where the propagation channel introduces no error, the method according to the invention gives the same success rate as a method which consists in retaining the position of the window H for which the number of syndromes S _j draws is the biggest. For the latter method, for the position d which corresponds to Z ₀ , the syndromes associated with the K blocks are all zero. In the case of the process according to the invention, the process leads to K {n _c -n _b ) zero elements.

On the other hand, in the particular case where the propagation channel introduces transmission errors, the method according to the invention may be more advantageous and have a better success rate than a method which consists in retaining the position of the window H for which the number of syndromes S ^ nulls is the largest.

The implementation of the method is illustrated by FIGS. 4 and 5 which correspond to a particular example of an encoding of the information bits by means of the Hamming code (7.4). FIG. 4 represents the transmitted frame T ₁ and the received frame T ₁ , According to the example, the received frame T ₁ is delayed by one bit and contains two errors. The delay bit and the erroneous bits are shown in bold. The synchronization window is positioned by the receiver on the received frame. According to the example chosen, the size of the window is fourteen bits. The window can therefore cover two codewords.

FIG. 5 represents this window H for different possible positions which correspond to d = 0 to 6. The window H is divided into two blocks Bi and B ₂ of seven bits. From the bits of Bi, the method calculates the associated syndrome SJ ¹ . The method repeats this operation for block B ₂ to determine Sf syndrome. According to a method which comprises retaining the H window position for which the number of syndromes S] ₁ 'zero is the largest, Ie proceeds counts the number of syndromes S'',5'^'' harmed and obtains zero. According In the process, the process counts the number of null elements and obtains three. window H and repeats counting operations of null elements. For d = 1. the number of null elements is equal to four and the number of syndromes S [, <S, zero is equal to zero. For d = 2, the number of null elements is equal to three and the number of syndromes S ^, S ₂ ^{ι ~} zero is equal to zero. For d = 3, the number of null elements is equal to three and the number of syndromes d = 4, the number of null elements is equal to three and the number of syndromes ^, Sl zero is equal to zero. For d = 5, the number of null elements is equal to two and the number of syndromes S _$ , S ^ ^" 'zero is equal to zero For d = 6, the number of null elements is equal to two and the number of zero elements is equal to zero. number of syndromes S ^, S ^ zero is equal to zero, the size of a code word used is seven, therefore a seventh shift returns to a position without offset, d = 0, synchronization can not be done With a value of d varying between zero and six, in this noisy case, a method which consists of retaining the position of the window H for which the number of S zero syndromes is the largest does not make it possible to synchronize the receiver. on the beginning of a code word: whatever the position of the window on the received frame T ₁ , such a method makes it possible to obtain only non-zero syndromes, the synchronization fails. the invention advantageously makes it possible to synchronize the receiver on the beginning of a code word: for the posit Ion of the window which corresponds to d = 1 the number of null elements is of four whereas for the other positions of the window, the number of null elements is of three or two. The method therefore selects the position J = I as the synchronization position for the received frame because it corresponds to the largest number of null elements.

In the event that several different positions of the window lead to the same number of zero elements, the method chooses a position among these different positions. This selected position determines the start position of the frame. The choice can be made arbitrarily, for example it can relate to the position corresponding to the smallest value of d. Alternatively, the choice may result from the implementation of a function including tests such as. for example, if the number of positions is odd then the selected position is the central position among the different positions. Another embodiment may be to increase the size of the synchronization window.

FIG. 6 schematically represents a first particular embodiment of a synchronization device according to the invention. The device 1 comprises a windowing means 2 sliding on a received frame T _t , a means 3 for calculating a syndrome S [ ¹ ', a means 4 for determining a synchronization position D. The sliding windowing means 2 is of size Kn _c . 11 is cut into K blocks B, of size H equal to that of a code word. The starting position D is equal to one of several possible positions of the windowing means 2 sliding along the received frame T _1. According to the illustration which corresponds to a particular embodiment, this windowing means 2 can include a battery of registers forming a delay line into which the bits of the received frame enter on a serial input. The window is sized by the choice of parallel outputs of the battery. The blocks B ₁ are determined by the way in which these outputs are taken into account by the calculation means. The calculating means 3 calculates a syndrome S _d for each of the K blocks B, the means 2 of windowing and for the different positions of this means 2 of windowing. Each syndrome being composed of elements, the calculation means 3 determines for a given syndrome the number N _d of null elements of the syndrome. Typically, this calculation means 3 is a computer such as a microprocessor. The means 4 for determining a position determines the position for which the number N _d of zero elements of a syndrome is the largest. The start position D is equal to this position. This means 4 can be part of the calculation means 3. Alternatively, it may be a particular component with logical operators. A method and a synchronization device according to the invention operate with all types of codes in linear blocks. Although the described examples relate to binary data, the invention works equally well with flexible data. Flexible data means receiving data that has not yet been compared to a threshold to be declared equal to a zero or a bit. The flexible data therefore has a voltage level that can take values that are not just the voltage levels of a zero or a bit obtained after thresholding.

The invention applies to any type of wired or wireless transmission system including the transmitting data are discrete, but may s ^is additionally applied to any storage system information.

Claims

1. Method of synchronizing data frame constituting code words, by means of a sliding window H on a received frame T ₁ , of size Kn _c , divided into K blocks B, of size n _c equal to that of a code word, consisting in determining a start position D of the received frame T ₁ , this starting position D being equal to one of several possible positions of the sliding window H along the received frame T ₁ . , such that, for different possible positions of the window H, the method calculates a syndrome S ^ per block B ₁ , a syndrome being composed of elements, and determines for each syndrome Sj the number of null elements, the position

D of the beginning of the received frame T ₁ , being that for which the number of null elements of the syndromes Sj 'calculated for this position is the largest.

2. Method of synchronizing data frame constituting code words according to the preceding claim, wherein the calculation of syndromes S ^ is performed for each possible position of the sliding window H along the received frame T _r and for each of these positions, the calculation of syndromes S _j is performed for each K blocks B ₁ of the window H.

3. Method of synchronizing data frame constituting code words according to one of the preceding claims, wherein the start position of the received frame T ₁ . is arbitrarily chosen from among several possible positions obtained after calculating syndromes.

4. Method of synchronizing data frame constituting code words according to one of the preceding claims, wherein the data is of the flexible data type.

5, method for synchronizing data frame constituting code words according to one of the preceding claims, wherein the data is of the binary data type.

6 method for receiving data constituting codewords, comprising a frame synchronization method according to one of the preceding re verbications

7. Device (1) for synchronizing data frame constituting codewords, for determining a frame start position, cai "actérisé in that said device comprises: a means (2) of windowing sliding on a received frame T ₁ , size

Kn _c , cut into K blocks B ₁ of size n _c equal to that of a code word, the position D of received frame start T ₁ being equal to one of several possible positions of the means (2) of windowing sliding the along the received frame T ₁ , - means (3) for calculating a syndrome S _d composed of elements for different possible positions of the windowing means (2) and for different blocks B ₁ of the windowing means (2) , the calculating means (3) determines a number of null elements of the syndrome S _j , a means (4) for determining a position for which the number of null elements of the syndromes S _d calculated for this position is the highest. large, the starting position D being equal to this position.

8. Codeword component data receiver, comprising a data frame synchronization device (1) for determining a frame start position D according to one of claims 7.

9. Computer program on an information carrier, said program comprising program instructions adapted to the implementation of a method according to any one of claims 1 to 6, when said program is loaded and executed in a electronic device (1).

A digital signal for use in an electronic device for synchronizing data constituting codewords, by means of a sliding window H on a received frame T ₁ of size Kn ₁ , divided into K blocks B, of size n _c equal to that of a codeword, consisting in determining a position D of beginning of the received frame T ₁ . , this starting position D being equal to one of several possible positions of the sliding window H along the received frame T ₁ , the digital signal comprising at least codes for the execution by the electronic device of the step next :

for different possible positions of the window H and for different blocks B, of the window H. the electronic device calculates a syndrome SJ ¹ 'composed of elements and determines for each syndrome .VV' the number of null elements of the syndrome SJ, the starting position D of the received frame T being that for which the number of null elements of a syndrome Sj 'is the largest.

An information carrier having program instructions adapted to implement a method according to any one of claims 1 to 6 when said program is loaded and executed in an electronic device.