DE102005021321A1 - Error correction symbol`s index determining method for portable device, involves determining indexes by identifying region of code symbols, which can be requested in form of correction symbol to reconstruct source symbols, by two parameters - Google Patents

Abstract

The method involves generating coded code symbols using a generator matrix of block codes form a number of source symbols. The code symbols are transmitted from a transmitting device to a receiving device, where transmission error occurs in the received code symbols. Indexes of error correction symbols that are to be transmitted are determined by identifying a region of the code symbols by two parameters. The region can be requested in the form of the correction symbol by the receiving device from the transmitting device for reconstructing the source symbols in an error-free manner. An independent claim is also included for a device for executing a method of determining error correction symbols.

Description

The The invention relates to a method according to the preamble of the claim 1 and a device according to the preamble of claim 5.

at the transmission of data packets, such as Audio or video data from a transmitter to a receiver it comes because of transmission errors for incorrect reception of these data packets. Especially for download services It may be necessary for all data packets to be reconstructed without error at the receiver can be.

at MBMS services (MBMS - Multimedia Broadcast / Multicast Service), which, for example, in 3GPP (3GPP - Third Generation Partnership Project) become data packages transmitted from one transmitter to multiple receivers. It can at different receivers different data packets erroneously reach the respective recipient.

Around to ensure a fault-free reception could transmit the data packets multiple times become so that a faulty reception is very unlikely. However, this approach is extremely inefficient, because all recipients, independently if there is an incorrect reception, the data packets sent repeatedly become.

Besides can additionally error protection in the form of parity data for the data packets transmitted to the recipient become. This allows a maximum number of erroneous data packets be corrected, however, so that a faultless reception is not be guaranteed.

alternative or additionally can be a receiver after incorrect reception of one or more data packets a point-to-point connection with the transmitter to request the faulty data packets. The more data packets are requested, the more bandwidth will be for the MBMS service needed. Further increase the transmission costs with the number of requested data packets, because more bandwidth is needed becomes.

in the In general, a data packet, a parity packet and / or an error correction packet be formed of one or more symbols. Include a data packet source symbols, a parity packet parity symbols and an error correction packet error correction symbols.

Of the Invention is based on the object, a method and an apparatus specify a selection with a minimum number of error correction symbols achieved for error-free reconstruction of source symbols.

These Task is by the independent claims solved. This allows the inventive method to determine a minimum number of error correction symbols that needed for error-free reconstruction of the source symbols. This is the for the transfer need the error correction symbols Bandwidth reduced. Further, by a minimum number of Error correction symbols achieved a reduction in the transmission time since in real transmission systems is the transmission bandwidth limited. Furthermore, can by the inventive method transmission costs be saved because only a minimum number of error correction symbols transmitted must become.

The The invention also relates to a device for carrying out the inventive method. Thus, the inventive method in a device, in particular a portable mobile device, implemented and executed.

other Further developments of the invention are given in the dependent claims.

The Invention and its developments are described below with reference to Figures closer explained.

It demonstrate:

1 an embodiment of the inventive method with a transmitting and receiving device and a faulty transmission channel;

2 a flow diagram of the method according to the invention;

3 Determination of coded code symbols in the case of several, consecutively nested systematic block codes;

4 systematic construction of a systematic Raptor code;

5 Application of a device according to the invention in a portable device.

Elements with the same function and mode of action are in the 1 With 5 provided with the same reference numerals.

With the help of 1 and 2 the process of the invention is explained in more detail. In 1 a transmission device SV, for example an MBMS transmission device (MBMS - Multimedia Broadcast / Multicast Service), can be seen. This transmission device SV encodes source symbols Q.sub.s with the aid of a generator matrix G of a block code BC. These source symbols Q represent, for example, a compressed voice or image file. For example, the compressed voice file has been encoded according to the AMR standard (AMR - Adaptive Multirate) and the compressed image file has been encoded according to the JPEG standard (JPEG - Joint Picture Expert Group). The source symbols Q may generally describe any data that may be compressed or uncompressed. For further execution, the vector or matrix notation is used.

For the following embodiment, binary source symbols Q are used, such as, for example. Q = [1 1 0 1] T , (1) wherein a first number K = 4 of source symbols Q are processed.

The generator matrix G has, for example, the following structure:

By this generator matrix G K = 4 source symbols QN = 7 coded code symbols CC are formed. Since this generator matrix G describes a systematic block code BC, K = 4 coded code symbols CC are identical to the K = 4 source symbols Q and the remaining NK = 3 coded code symbols CC correspond to redundancy symbols R. The reference symbol N is designated as the second number N. Representatives of systematic block codes BC are, for example, an LDGM code or an LPDC code (Low Density Generator Matrix (LDPC) - Low Density Parity Code). The coded code symbols CC are calculated as: CC = G × Q = [1,1,0,1,0,0,1] T (3)

During the transmission of the coded code symbols CC from a transmitting unit SE of the transmitting device SV via a transmission channel UE to a receiving unit EE of a receiving device EV, transmission errors UEF occur. In this case, the receiving device is, for example, an MBMS receiver. In particular, in the case of transmission via wireless transmission channels UE, for example via a mobile radio channel operating according to the GSM standard (GSM - Global System for Mobile Communications) or UMTS standard (UMTS - Universal Mobile Telecommunications System), transmission errors UEF come to light. Furthermore, errors can also occur in the case of wired transmission paths, such as, for example, in the case of IP over LAN (IP - Internet Protocol, LAN - Local Area Betwork). In this case, coded code symbols CC can be received in a corrupted manner or coded code symbols CC are deleted during the transmission and do not reach the receiving unit EE. In the present invention, coded code symbols CC reach the receiving device EV as incorrectly received code symbols C. In 1 are received code symbols C, which are faulty due to, for example, a falsification or deletion, marked with a reference 'X'. In 2 is the reception of the Code symbols C are reproduced in step 21.

In the present embodiment, the received code symbols C have the following appearance: C = [X, 1, X, 1, X, X, 1] T (4)

Consequently the received code symbols C are received with an index 2, 4, 7 or with the index 1, 3, 5, 6 received incorrectly.

In a further processing step S22, a first and matrix M1 are formed. The first matrix M1 is formed from the generator matrix G of the block code BC such that those lines of the generator matrix (G) are generated which generate those code symbols (C) which have been received without errors. In an alternative embodiment, the generator matrix G may be copied into the first matrix M1, and thereafter those lines ZX which produce those code symbols C which have been erroneously received are removed. Domit the lines ZX of the generator matrix G are not included in the first matrix M1, which generate the code symbols C with the index 1, 3, 5, 6. This corresponds to the first, third, fifth and sixth rows ZX of the generator matrix G. Thus, the first matrix M1 results

The columns of the first matrix M1 are marked in step S23 with a respective column index SI, the respective column index SI corresponding to the column number of the generator matrix G. In order to distinguish between the coefficient and the row index SI, the respective column index SI is placed in quotation marks. The marked first matrix M1 is thus:

In the first matrix M1 are natural Numbers used to describe the column indices. in the In general, markers that are used as a column index allow a clear identification of the marked column. Next natural Can pay also letters or memory addresses are used.

• (I) Addition eines Vielfaches einer Zeile zu einer anderen Zeile
• (II) Vertauschen zweier Zeilen
• (III) Multiplikation einer Zeile mit einem Skalar λ ≠ 0

In a further step S24, the first matrix M1 is transformed by elementary line conversion and / or column exchange into the second matrix M2, resulting in RG independent rows ZU, where RG corresponds to a rank of the first and second matrix M1, M2. Under elementary line transformation, the following operations are known according to [1] page 61:

• (I) Addition of a multiple of a line to another line
• (II) Swapping two lines
• (III) Multiplication of a line with a scalar λ ≠ 0

• a) Vertauschen der ersten und dritten Zeile
  
• b) Vertauschen der zweiten und dritten Zeile
  
• c) Vertauschen der dritten und vierten Spalte
  

In order to determine the second matrix M2, the matrix M1 is first copied and then the following work steps are carried out, the intermediate result of the second matrix M2 being indicated below for each work step:

• a) Swap the first and third lines
  
• b) Swapping the second and third lines
  
• c) Swapping the third and fourth columns
  

in this connection It should be noted that the indexes are also exchanged when the column is exchanged become.

After the steps a) to c) RG = 3 independent lines ZU are present in the second matrix M2. In this case, RG corresponds to a rank RG of the second matrix M2. Further, by the RG = 3 columns of the RG = 3 independent rows ZU, a square third matrix M3 having a dimension RG × RG having a row-step shape is formed. In equation (9), this third matrix M3 is formed by the first three columns and rows. To define a matrix in line-level form, see page 23 in [2]. Thus, the third matrix M3 is:

Further, in step S24, the columns of the RG = 3 independent lines ZU that are outside the third matrix M3 are combined into a fourth matrix M4. In the present example, this fourth matrix M4 comprises only the column '3' marked in the second matrix M2. The fourth matrix M4 is thus:

In a further step S25 are indicated by the column index or columns SI of the fourth matrix M4 indices of those coded code symbols CC described in the form of error correction symbols FKS the receiving device EV for error-free reconstruction of all Source symbols Q requested and transmitted by the transmitting device SV become. In the present embodiment is a single error correction symbol FKS, namely the coded one Code symbol CC with the index '3', to request to be able to reconstruct the source symbols Q without errors. in this connection becomes the reconstruction of the code symbols C with the index 1 and 3 only a single error correction symbol FKS, which the coded Source symbol CC corresponds to the index 3, repeated by the Transmitting device SV to the receiving device EV transmitted. The receiving device EV is thus capable of both faulty source symbols Q under Use the code symbols C and the requested error correction symbol FKS error free to reconstruct. On the reconstruction will not further discussed, since the procedure known to a person skilled in the art is.

The inventive method has been described with reference to the first to fourth matrices. In general For example, the invention may be practiced with fewer dies. For example, in the generator matrix G all lines are deleted, with which code symbols C that have been received erroneously generated become. After pasting The column indexes become one by transformation within the generator matrix square matrix formed with row-step shape. In connection at this point, the column indices of the transformed generator matrix become not taken through the columns of the square matrix are occupied. These extracted column indices are in the form of requested at least one error correction symbols. A number taken column indices equals a number of requested ones Error correction symbols.

In an extension, the method according to the invention can also be used in the case of several concatenated systematic block codes BC1, BC2. 3 shows two concatenated systematic block codes BC1, BC2. The coded code symbols CC are formed on the basis of the source symbols Q as follows: CC = G2 × G1 × Q = G × Q (12)

Thus, the generator matrix G is determined by multiplication of the second and first generator matrix G2, G1. In general, the generator matrix G is generated in the case of linked systematic block codes as follows: G = Gp × ... × G1 (13)

there the p-generator matrix Gp represents the last generator matrix, which the coded code symbols CC at the output of the p-th block code Outputs BCp.

The method according to the invention has been described on the basis of binary symbols. In general, the method according to the invention can be used with binary values or values of a Galois field GF, such as in the Galois field (2 ⁸ ).

The method according to the invention is suitable for determining a minimum number of error correction symbols FKS to be requested in the case of a systematic Raptor code. It corresponds, as in 4 the non-systematic block code BCN represents the inner code of the Raptor code and the first and second block codes BC1, BC2 represent the respective outer codes of the Raptor code. The inner code is also known as LT code (LT - Luby Transform). In addition, in non-systematic Raptor codes, a correction matrix G0 of a non-systematic block code BC0 may be inserted before the second systematic block code BC2 to create systematic coded symbols CC. This creates a systematic Raptor code and systematically coded code symbols CC can be requested as error correction symbols FGS using the method according to the invention. The term non-systematic block code can also be understood to mean systematic block codes in the context of this invention.

The inventive method is feasible with a device V, wherein the receiving unit EE Step S21, a first means ML1, Step S22, a second one Means ML2 the step S23, a third means ML3 the step 24, a fourth means ML4 implements and executes the step S25. The Device V may be executable in hardware, in a computing unit Software or implemented in a combination of hardware and software be.

The inventive method is performed in a terminal EG by means of the device V. In 5 Such a terminal EG is shown in the form of a portable device. This portable device operates, for example, according to a mobile radio standard, in particular according to the GSM standard (GSM Global System for Mobile Communications), UMTS standard (UMTS Universal Mobile Telecommunications System), DAB standard (DAB Digital Audio Broadcast) or DVB Standard (DVB Digital Video Broadcast).

bibliography

• [1] Prof. C. Preston, "Ein Script for Linear Algebra I and II ", 2003/04, "http://dhcp24-159.mathematik.unibielefeld.de/~preston/files/teaching/linalg/skripten/linalg.pdf"
• [2] Prof. W. Ebeling, "Linear Algebra ", winter semester 2001/2002, "http://www-ifm.math.unihannover.de/~ebeling/LA-A/LAA.pdf"

Claims (6)

Method for transmitting at least one error correction symbol (TCS), wherein code symbols (CC) coded based on source symbols (Q) are generated using a generator matrix (G) of a systematic block code (BC) and the coded code symbols (CC) are transmitted by a transmitting device (SV ) are transmitted to a receiving device (EV), wherein transmission errors occur in received code symbols (C), wherein the at least one error correction symbol (FKS) to be transmitted is generated according to the following steps: - a first matrix (M1) is generated on the basis of the generator matrix (G ) are formed such that those lines of the generator matrix (G) are copied which generate the code symbols (C) which have been received without errors, - each column of the first matrix (M1) is marked with a respective column index (SI), - the first matrix (M1) is exchanged by elementary line transformations and / or column exchange, whereby the respective column indices (SI) are exchanged in a column exchange a second matrix (M2) is transformed such that - RG independent rows (ZU) are formed, - a square third matrix (M3) with a dimension RG × RG is formed in a row-step form by RG columns of the RG independent rows (ZU) a fourth matrix (M4) is generated by the columns of the RG independent lines (ZU) lying outside the third matrix (M3), the code indices (SI) of the fourth matrix (M4) encode those coded code symbols (CC) which are requested in the form of error correction symbols (FKS) by the receiving device (EV) for error-free reconstruction of the source symbols (Q) and transmitted by the transmitting device (SV). Method according to one of the preceding claims, characterized characterized in that the generator matrix (G) by linking several Generator matrices (G1, G2) of systematic block codes (BC1, BC2) is produced. Method according to one of the preceding claims, characterized characterized in that the source symbols (Q), coded code symbols (CC) and the code symbols (C) are binary values or values of a Galois field (GF). Method according to one of the preceding claims, characterized characterized in that systematic raptor codes, LDGM codes or LPDC codes are used as block code (BC). Device (V) for carrying out the method of transmission of at least one error correction symbol (FKS), based on code symbols (CC) coded by source symbols (Q) using a Generator matrix (G) of a systematic block code (BC) generated and the coded code symbols (CC) from a transmitting device (SV) to a receiving device (EV), wherein transmission errors occur in received code symbols (C), in particular after one of the preceding claims, marked by, - one first means (ML1) for forming a first matrix (M1) is opened Basis of the generator matrix (G), wherein those lines (ZX) of the Generator matrix (G) are copied, those code symbols (C) generate that have been received without errors, - a second Means (ML2) for marking each column of the first matrix (M1) with a respective column index (SI), - a third resource (ML3) for transforming the first matrix (M1) by elementary line transformations and / or Column exchange, whereby with a column exchange also the respective Column indices (SI) are exchanged into a second matrix (M2), in which - RG independent Lines (TO) arise, - by RG columns of the RG independent Rows (ZU) a square third matrix (M3) with a dimension RG × RG with a line-step shape is formed, - through the outside the third matrix (M3) lying columns of RG independent lines (TO) a fourth matrix (M4) is generated, - a fourth Means (ML4) for identifying those coded code symbols (CC) due to the column indices (SI) of the fourth matrix (M4), the in the form of error correction symbols (FKS) by the receiving device (EV) for error-free reconstruction of all source symbols (Q) requested and transmitted from the transmitting device (SV). Device (V) according to claim 5, characterized in that that the device (V) in a terminal (EG), in particular a portable device according to a mobile radio standard, in particular according to GSM standard, UMTS standard or DVB standard, is integrated.