RU2613021C1 - Method for coding and decoding messages - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: method for coding and decoding messages, comprising receiving information bits, after which at least two messages are formed, wherein each contains above information bits coded by the error-correcting ZigZag code, to which a group of check bits is added, and then at least, two coded messages are transmitted, wherein each message is transmitted at different frequencies; thereinafter at least one above coded message is received, and finally at least one coded message is decoded, each message being decoded separately, and in case of failure, all undecoded messages are decoded together.

EFFECT: increasing the information transmission interference immunity.

11 cl, 3 dwg

Description

FIELD OF TECHNOLOGY

[0001] This technical solution relates to a communication system with the transmission of data packets, and more particularly to methods of noiseless coding and decoding of information.

BACKGROUND

[0002] At the current level of communications development in radio communications, various error-correcting coding schemes are widely used. Codes are known and theoretically investigated that reach or approach arbitrarily close to the Shannon limit for the data rate on a channel with Gaussian noise. Many classes of error-correcting codes are also known that allow practical implementation at the modern level of electronics development - algebraic codes, convolutional codes, codes with a low density of parity checks, various types of product codes, etc.

[0003] Codes with a low density of parity checks are of great interest both theoretically and from a practical point of view, various versions of LDPC codes are used in such systems and radio standards as DVB-S2, 10GBase-T, WiFi 802.11, etc.

[0004] A class of low density codes called ZigZag codes is known in the art. This class of codes is of interest because of the relatively simple decoding algorithm, while maintaining a fairly good quality of error correction under various conditions (within 1.5 dB from the Shannon limit).

[0005] Patent EP 1,432,129 A2, "Iterative Decoding of parallel concatenated Zigzag codes", patent holder: Nokia Corporation, published: 06/23/2004, is known in the art. An iterative decoder using parallel cascading ZigZag code containing a data circuit for using external information during iterative decoding, in which verification information for parity bits is updated during iterations.

[0006] However, there are no known systems of the prior art that practically use similar encoding and decoding methods with the separation of test bits into several messages that can be decoded both separately and together in systems with the simultaneous reception of a large number of narrow-band messages with frequency division and high probability collisions.

SUMMARY OF THE INVENTION

[0007] The technical task is to develop a method for encoding and decoding messages, in which the verification information is divided into several messages, and the information contained in each message should be sufficient to decode it with sufficient efficiency (with a loss of not more than 1-2 dB relative to the circuit with independent encoding of each message with the same code rate).

[0008] The technical result of this technical solution is to increase the noise immunity of information transmission.

[0009] This technical result is achieved by a method of encoding and decoding messages, in which information bits are received, after which at least two messages are generated, each of which contains the above information bits encoded by the noise-resistant ZigZag code, to which a group of check bits is added, then at least two encoded messages are transmitted, where each is transmitted at different frequencies, then at least one of the above encoded messages is received, and as a result, I decode t, at least one encoded message, and each message is decoded separately, and in case of failure, decode all undecoded messages together.

[00010] In some embodiments, the error-correcting code is a ZigZag code with a code rate of 1/3.

[00011] In some embodiments of the technical solution, when the verification bits are divided into groups, in the case of two groups, the odd bits fall into the first group and the even bits into the second.

.[00012] In some embodiments of the technical solution for transmitting encoded messages, each of the transmitted messages is a code code word with perforated code rate

.

[00013] In some embodiments, when transmitting coded messages, messages are transmitted sequentially at different frequencies.

[00014] In some embodiments, when decoding encoded messages, the initial metrics for the check bits that were not transmitted in the above decoded message are assumed to be zero at the start of decoding.

[00015] In some embodiments of the technical solution, when decoding encoded messages from metrics fails for all bits of the original messages, new metrics are compiled to decode the verification code.

[00016] In some embodiments of the technical solution, when receiving messages in case the first message is received with errors, errors are corrected using the check bits transmitted in the first message.

[00017] In some embodiments of the technical solution for joint decoding, the metrics corresponding to all bits in the original messages are added.

[00018] In some embodiments of the technical solution, the transmission of messages at different frequencies occurs with a time interval.

BRIEF DESCRIPTION OF THE DRAWINGS

[00019] The features and advantages of this technical solution will become apparent from the following detailed description of the invention and the accompanying drawings, in which:

[00020] FIG. 1 - probability of a bit error for coding by noise-resistant ZigZag with code rates 1/2 and 1/3 in a channel with normal noise;

[00021] In FIG. 2 - an example of the implementation of the distribution scheme of information and check bits for messages;

[00022] FIG. 3 is a flowchart of a method for encoding and decoding messages.

DETAILED DESCRIPTION OF THE INVENTION

[00023] Below will be described the concepts and definitions of terminology necessary for the detailed disclosure of the invention.

[00024] Interference-proof coding is the coding of messages in which the elements are connected by a certain dependency, which, if violated, indicates errors and restores information that reduces redundancy in transmitted messages.

[00025] An information bit is a bit generated by a data source and intended to transmit user data [from clause 57 of GOST 24402-88].

[00026] Check bit - a bit designed to detect errors in a data sign or data block [from clause 57 of GOST 24402-88].

[00027] Code rate is the ratio of the length of the code block at the input to the length of the converted code block at the output of the encoder.

[00028] A code with a low density of parity checks is a code used in transmitting information, a special case of a block linear code with parity. A feature is the low density of significant elements of the verification matrix, due to which the relative simplicity of the implementation of the encoding means is achieved.

[00029] ZigZag code - in coding theory, this is a linear error-correcting code in which the input data is divided into segments of a fixed size and a sequence of check bits is added, where each check bit is exclusive in the segment.

[00030] The Shannon limit is the limit (maximum) information transfer rate (bandwidth) in a noisy communication channel.

[00031] A transmitter is a device that receives information bits from their source, for example, from a counter, encodes them, generates data packets, adds overhead information and broadcasts several times at various frequencies with DBPSK modulation.

[00032] Perforation is a method of obtaining from a code word a code with one codeword speed at a different speed by puncturing (in Russian-language literature, usually “perforating”) parts of check or information bits. In the claimed solution, in one message only even bits are transmitted, then it can be said that the odd bits were perforated, they were not transmitted, and at the reception they are replaced with zeros. Messages received in this way are called code words of the perforated code, and the method of receiving them is called perforated code.

[00033] Metric is a numerical indicator of the probability that the value of bits "1" is accepted. As a rule, a negative metric value corresponds to the accepted value of “1”, and a positive value is “0”. The larger the metric modulo, the greater the confidence of the algorithm in the accepted value.

[00034] According to a variant of the claimed technical solution (Fig. 3): information bits 101 are preliminarily received (Fig. 3), then at least two messages are generated, each of which contains the above information bits encoded by the noise-resistant ZigZag code, to which a group is added check bits 102 (Fig. 3) as follows.

[00035] The information bits on the transmitter are encoded by an error-correcting composite ZigZag code, after which the test bits are evenly distributed into groups - for the case of two messages, even test bits of each of the four component codes are selected for the first message, and odd for the second. Moreover, in each message all information bits and one of the parts of the verification bits are transmitted, i.e. one group, and at different frequencies 103 (Fig. 3). Coding methods for low-density codes and ZigZag codes, in particular, are widely known from the literature, for example, these codes are described in [2]. In general, for systematic codes, which include ZigZag coding, this is an addition to the original information bits, which remain unchanged, of an additional number of verification bits using a special algorithm that will subsequently correct incorrectly received information bits using verification bits. One skilled in the art will appreciate that the number of groups may be more than two, for example three or more groups. In this case, three or more messages will be transmitted, in each of which there will be all information bits and a group of test bits. The rules for splitting into groups are fundamentally important for the high-quality operation of the method. The used even and odd partitioning is both simple and optimal for the selected noise-resistant ZigZag code and for two messages. When using three messages, a similar splitting will be applied - in the first message, bits 1, 4, 7 ... will be transmitted, in the second 2,5,8, ..., in the third 3,6.9 ... and so on, without limitation.

[00036] Then, each of the transmitted messages will be a code word of a perforated ZigZag code. The source of information [1] shows that the specified method of perforation ensures the quality of decoding of the codes obtained with its help, which is in good agreement with the quality of decoding of other noise-resistant codes with a similar speed. When a message is received at the receiver 104 (Fig. 3), the detected messages are decoded separately 105 (Fig. 3), while the initial metrics for the check bits that were not transmitted in the decoded message are taken equal to zero at the beginning of decoding.

[00037] In the event that a single message could not be decoded and if messages were found that also could not be decoded separately, they are jointly decoded, for which new metrics are compiled from the metrics for all bits of the original messages to decode the verification code. For this, the metrics corresponding to all bits in the original messages are added up. Then the metrics for the check bits will be zero in all messages except one, and the metrics for the data bits will add up, which will correspond to the generally accepted methods of receiving the code with repetition. This is the easiest way to get metrics together from several received messages. In fact, this is just the addition of metrics for all received messages. The maximal-ratio combining method is known in the literature, simple addition is its simplest option.

[00038] As already mentioned, due to the fact that this encoding method involves the transfer of all useful data, ie information bits, in each message, successful reception is possible even if all messages except one ... are lost due to collisions or interference in their band.

[00039] On the other hand, if several messages are jointly decoded, the coding gain will be greater than the repetition of identical messages in accordance with the theory (a code with a more complex structure than repetition is decoded more efficiently).

[00040] The technical result is also achieved due to the fact that the transmission of messages with the same information at several different frequencies occurs with a time interval. This is a fundamental point, since transmitting devices do not track the occupancy of the frequencies at which they transmit, and some frequencies may not be suitable for transmitting messages. The transmission of all information at each of the selected frequencies reduces the likelihood of falling into an unsuitable frequency for transmission. Similarly, a pause between the transmission of messages allows you to wait for a more favorable situation on the air.

[00041] Secondly, according to the claimed method of error-correcting coding and decoding of messages, in each message only a part of the check bits is transmitted, and even this part allows to correct errors on reception.

[00042] According to an embodiment of the technical solution, when coding 240 information bits with a noise-resistant code 1/3, 480 check bits are obtained. The check bits are divided into groups of even and odd bits, the even bits being transmitted in the first message, the odd bits in the second.

[00043] It is possible that the first or second messages will be received, or both of these messages. The selected coding scheme allows you to decode all these cases in exactly the same way, writing zeros at the positions of those check bits that were not received at all. It is important that a message containing only half the check bits can also be decoded, using the same methods as when all the bits were received. Moreover, this message (in the text it is called a perforated code) allows you to fix as many errors as if we used a special code with a speed of 1/2, developed separately.

[00044] That is, the intermediate result is the receipt of such a 1/2 speed code (from the existing 1/3 code) that, in terms of error correction efficiency, does not lose to other 1/2 speed codes obtained separately and independently. At the same time, when receiving several messages, you can decode the source code 1/3, which allows you to fix even more errors. By itself, this process of transmitting only part of the test bits (called perforation) is known from the literature.

[00045] Thus, it becomes necessary to transmit the same message several times at different frequencies, since some of the frequencies may not be suitable for transmission. In this case, all information bits are transmitted, and the verification bits are generated and decoded in such a way as to further increase the number of correctable errors. This is possible thanks to the use of error-correcting code (for example, ZigZag), as well as the division of bits into even / odd.

[00046] The claimed technical solution can be used in electrical engineering and digital communications, as well as, for example, in satellite communications and in other fields where it is necessary to achieve maximum reliability and energy efficiency of data transmission over a communication channel with noise in a limited frequency band.

[00047] The present detailed description is made with a variety of non-limiting and exhaustive embodiments. At the same time, for specialists having an average level of competence in the considered field of technology, it is obvious that various substitutions, modifications or combinations of any of the embodiments disclosed herein (including partially) can be reproduced within the scope of this technical solution. Thus, it is understood and understood that the present description includes additional embodiments, the essence of which is not set forth here in an explicit form. Such embodiments may be obtained, for example, by combining, modifying, or transforming any actions, components, elements, properties, aspects, characteristics, limitations, etc., related to the embodiments presented herein and not being restrictive.

EXAMPLES OF IMPLEMENTATION

[00048] Consider an example implementation of a technical solution in which it is necessary to transmit 240 bits of data (information bits). For coding, the noise-resistant ZigZag code with a code rate of 1/3 is used. After the coding process is completed, the initial 240 bits of data (information bits) and 480 (twice the number of information) verification bits, 720 bits in total, will be ready for transmission. The check bits are divided into two equal groups of 240 bits, after which the first message is composed of 240 information data bits and the first 240 check bits, a total of 480 bits. Similarly, the second message consists of the same 240 data bits and the remaining 240 verification bits. The check bits are divided into two groups and are transmitted in separate messages as follows - 1,3,5,7, ..., 479 are transmitted in the first message, and 2,4,6,8, ..., 480 are transmitted in the second message. Information bits are transmitted in each message as a whole without modification (Fig. 2).

[00049] Each of the transmitted messages is a codeword with perforated code rate 1/2, the two messages being decoded together represent the original codeword speed 1/3.

[00050] Thus formed messages are transmitted sequentially at different frequencies. After detecting the first message, a check is made to correctly receive information bits. Under good transmission conditions, when the noise in the channel is small enough and does not lead to errors in reception, the information bits are immediately received correctly. If the first message is received with errors, the error correction procedure is started using the check bits transmitted in the first message. For decoding, the code word of the error-correcting ZigZag code is restored, while the metrics of the information bits and the verification bits 1,3, ..., 479 are calculated based on the received data, and the metrics of the verification bits 2,4, ..., 480 are taken equal to zero.

[00051] Thus, from the codeword with perforated code rate 1/2, the codeword of the source code of speed 1/3 is restored with the replacement of part of the metrics by zeros. The resulting codeword is decoded.

[00052] Similarly, after the second message is received, the correctness of the information bits transmitted in it is checked, and if necessary, the error correction procedure is started using the verification bits transmitted in the second message. If the first message with the same data was received, all 480 check bits are used for recovery. To do this, the code word is restored, and the metrics of the check bits 2,4, ..., 480 are calculated according to the received data, and the metrics of the bits 1,3, ..., 479 are replaced by zeros, or, if the first message was received and not decoded, the metrics of these bits calculated according to the first message.

[00053] If any of the messages was received without errors, or if one of the messages was successfully decoded individually, joint decoding is not performed. Also, it is not performed when only one message has been received. Thus, messages are jointly decoded when they were both received, and the former was not independently decoded.

[00054] The simulation results of the noise-resistant ZigZag code under normal Gaussian noise channel conditions for rates 1/3 and 1/2 are shown in FIG. 1. It can be seen that using the 1/3 speed code gives a gain in the required Еb / N0 ratio of about 0.5 dB relative to the 1/2 speed. This means that when using coding with a speed of 1/3, the same level of reception errors is achieved at a lower ratio Eb / N0. That is, to transmit data with such encoding, less energy is required for the transmitted information bit. That is, despite the fact that when encoding with a speed of 1/3, it is necessary to transmit more bits than with 1/2, for example, 480 at 1/2 and 720 at 1/3, the transmitter power when working with 1/3 can be significantly lower.

[00055] The proposed encoding-decoding scheme can be implemented in practice using affordable and high-performance solutions, and there is no need to use specialized integrated circuits or programmable logic integrated circuits. The applied coding scheme using the noise-resistant ZigZag code has the following properties:

[00056] firstly, the decoding of the ZigZag code is performed by a linear complexity algorithm with respect to the length of the code and mainly contains only addition / subtraction and comparison operations. This implies that for decoding low-density codes, it is not necessary to perform the operations of multiplication, division, and to calculate trigonometric and other special functions. Due to this, low-density codes are widely used, including, it is possible to use them in the proposed technical solution. In many ways, this is why the ZigZag code was chosen, and the entire proposed coding scheme was developed taking into account that it should be practicable;

[00057] secondly, the scheme for uniformly perforating the check bits of the code for generating individual messages is easy to implement in comparison with other error-correcting codes.

INFORMATION SOURCES

1. Construction and analysis of rate-compatible punctured concatenated zigzag codes, Information Theory, 2005. ISIT 2005. Proceedings. International Symposium on, IEEE.

2. Ping L., Huang X., Phamdo N. Zigzag codes and concatenated zigzag codes // Information Theory, IEEE Transactions on. - 2001. - T. 47. - No. 2. - C. 800-807.

Claims (18)

1. A method of encoding and decoding messages, in which: receive information bits; at least two messages are generated, where each contains the above information bits encoded by the noise-resistant ZigZag code, to which a group of check bits is added; transmit at least two encoded messages, where each is transmitted at different frequencies; receive at least one of the above encoded message; decode at least one encoded message, and each message is decoded separately; if unsuccessful, decode all undecoded messages together. 2. The method according to p. 1, characterized in that the error-correcting code is a ZigZag code with a code rate of 1/3. 3. The method according to p. 1, characterized in that each of the transmitted messages is a code word of a composite ZigZag code with perforation. 4. The method according to claim 1, characterized in that when distributing the test bits into groups, in the case of two groups, the odd bits fall into the first group and the even bits into the second. 5. The method according to p. 1, characterized in that when transmitting encoded messages, each of the transmitted messages is a code word code with perforated code rate

. 6. The method according to p. 1, characterized in that when transmitting encoded messages, messages are transmitted sequentially at different frequencies. 7. The method according to claim 1, characterized in that when decoding the encoded messages, the initial metrics for the check bits that were not transmitted in the above decoded message are assumed to be zero at the beginning of decoding. 8. The method according to claim 1, characterized in that if the decoding of encoded messages from the metrics for all bits of the original messages is unsuccessful, new metrics are compiled to decode the verification code. 9. The method according to p. 8, characterized in that the metrics corresponding to all bits in the source messages are added. 10. The method according to claim 1, characterized in that when receiving messages in case the first message is received with errors, errors are corrected using the verification bits transmitted in the first message. 11. The method according to p. 1, characterized in that the transmission of messages at different frequencies occurs with an interval of time.

Images