JP4226481B2 - Synchronous loss recovery digital communication system using forward erasure correction - Google Patents

Description

Detailed Description of the Invention

[Technical Field of the Invention]
The present invention relates generally to communication systems, and more particularly to synchronous loss recovery digital communication systems using forward erasure correction (FXC).
[Background of the invention]
In a wireless digital communication system, multiple communication paths and fading may occur, and synchronization may be lost. Data transmitted during the synchronization loss period is usually lost at the receiver. Therefore, the issue is how to design a wireless digital communication system that can recover the loss of synchronization due to multiple communication paths and fading while keeping the overhead rate as low as possible.

  The effects of multiple channels and fading on wireless digital communication links are widely known and their characteristics are recorded. For example, an expected distribution of fading periods is being studied in an Advanced Television Systems Committee (ATSC) 8-residual sideband (8VSB) transmission system for high definition television (HDTV) in the United States. When 8VSB is received by the portable device, the possibility of synchronization loss is further increased. Even after the synchronization is reacquired, valid data cannot be recovered in the ATSC8VSB system until the retraining of trellis coding is performed and the interleaver starts a new block.

  The ATSC8VSB system includes several types of channel coding to protect against noisy transmissions, including trellis coding, interleaving and Reed-Solomon (RS) Forward Error Correction (FEC). include. However, these channel coding methods do not contribute to data recovery when synchronization loss occurs.

  Repeating the transmission of data can improve resiliency against loss of synchronization, but requires high overhead. Allocating the available bandwidth to the data transmission iterations reduces the amount of initial data that can be transmitted. This means that the transmitted program is reduced or the quality of the program is lowered.

  Also, instead of repeating the initial data transmission, it has been proposed to repeatedly transmit a low bit rate version of the initial data to reduce the overhead rate. If the initial data is lost due to loss of synchronization, the receiver uses a version with reduced resolution. This allows for graceful degradation. That is, a lower quality version of the initial data is obtained rather than the initial data. However, when high resolution initial data is required, a reduced resolution version is not sufficient.

Therefore, it is desirable and effective to provide a synchronization loss recovery digital communication system that can solve the above-described problems of the prior art.
[Summary of Invention]
The above problems and related problems of the prior art are solved by the present invention relating to a synchronous loss recovery digital communication system using forward erasure correction (FXC).

  According to one aspect of the invention, an apparatus is provided that enables recovery of loss information in a digital communication system. The apparatus is for calculating a forward erasure correction (FXC) parity superpacket over the information superpacket for subsequent recovery of any one of the information superpackets at least partially compromised by synchronization loss. Has FXC encoder.

  According to a further aspect of the present invention, an apparatus for recovering loss information in a digital communication device is provided. The apparatus uses a forward erasure correction (FXC) parity superpacket previously calculated over the information superpacket to recover any one of the information superpackets at least partially compromised due to loss of synchronization. It has an FXC decoder for decoding.

  According to yet a further aspect of the invention, a method is provided that enables recovery of loss information in a digital communication system. The method includes calculating an FXC parity superpacket over the information superpacket for subsequent recovery of any one of the information superpackets at least partially compromised by synchronization loss.

  According to a still further aspect of the invention, a method for recovering loss information in a digital communication device is provided. The method includes the step of decoding the FXC parity superpacket previously calculated over the information superpacket to recover the whole of any one of the information superpackets at least partially compromised due to synchronization loss. Have.

The foregoing and aspects, features and advantages of the present invention will become apparent upon reading the following detailed description in conjunction with the accompanying drawings.
Detailed Description of the Invention
The present invention relates to a synchronization loss recovery digital communication system using forward erasure correction (FXC). The present invention treats the period of synchronization loss as packet loss and uses forward loss correction (FXC) to recover the synchronization loss. Additional parity data is transmitted upward compatible to recover from packet loss. This reduces the overhead rate compared to repeated transmission of data.

  It should be noted that the present invention may be implemented in the form of hardware, software, firmware, special purpose processors, or a combination thereof. Preferably, the present invention is implemented as a combination of hardware and software. Furthermore, it is preferable that such software is executed as an application program tangibly incorporated in the program storage device. This application program may be executed or uploaded to a device having any suitable architecture. Preferably, the apparatus is implemented on an electronic computer platform having one or more hardware such as a central processing unit (CPU), a random access memory (RAM), and an input / output (I / O) interface. The electronic computer platform also includes an OS (Operation System) and microinstruction code. Also, the various processes and functions described here may be part of microinstruction code or part of application program (or a combination thereof) executed via the OS. In addition, other peripheral devices such as additional data storage devices and printer devices may be coupled to the electronic computer platform.

  In addition, since the system elements and method steps shown in the accompanying drawings are preferably executed by software, the actual operation between system elements (or processing steps) depends on the manner in which the present invention is programmed. The connection relationship may be different. With this description, a person having ordinary knowledge in the technical field of the present invention can understand the present invention and methods and configurations similar to the present invention.

  A forward erasure correction (FXC) code is applied to transmission of packet data in order to prevent packet loss of data transmitted on a packet network using the Internet protocol (IP) or the like. Any kind of FXC code can be used, such as, but not limited to, Reed-Solomon (RS) code. The Reed-Solomon code is a systematic code. That is, the initial information byte is transmitted with an additional parity byte. If there is no loss of communication path, the receiver simply needs to use the initial information byte and does not need to perform FXC encoding. However, in a packet network, the entire packet data tends to be lost rather than the individual bytes of the packet being lost. When FXC is used to prevent packet loss, generally (but not necessarily) one byte from each packet is used to form an FXC codeword. For example, when 10 packets with a length of 1024 bytes are encoded using FXC of RS (15, 10) code, 10 information packets with a length of 1024 bytes and a length of 1024 bytes Five parity packets will be transmitted. Thus, using 1 byte from each packet, 1024 different RS codewords are formed.

  Therefore, to prevent data loss due to synchronization loss in digital communication systems by calculating parity data for the period corresponding to the expected length of synchronization loss period and adding a forward erasure correction (FXC) layer Can do. The period of data loss due to synchronization failure is considered to be packet loss in the FXC decoder. In this system, the term packet is generally used to mean a small time scale (eg, 188 byte MPEG-2 transmit packet), where the term “superpacket” is used in an FXC decoder. Used to mean a packet unit to be used or usable.

  The present invention does not depend on the type of data to be transmitted, but can be implemented for transmission of all types of data. Furthermore, the present invention is not limited to audio / video programs and can be implemented for all types of programs.

  According to the present invention, FXC is used in addition to the channel coding method of other digital communication systems for preventing impulse noise. For example, in an ATSC 8VSB system, FXC coding is in addition to system trellis coding, interleaving, and Reed-Solomon (RS) forward error correction (FEC) coding. FXC may be any systematic forward error correction code, such as, but not limited to, a Reed-Solomon (RS) code. If systematic codes are used, the present invention can be upward compatible. That is, existing decoders can ignore additional parity superpackets that are transmitted and can normally decode unmodified information superpackets. FXC encoders and decoders are different from existing RS / FEC codes used in existing digital communication systems. For example, the FXC codeword is calculated over a super packet (for example, 1 byte per super packet) to prevent loss (disappearance) of the entire super packet. In contrast, existing RS FEC protects random bit or byte errors with samples taken from close points in time, not erasures in any packet.

  In the preferred embodiment of the present invention, based on the desired loss prevention level and allowable delay, the FXC parameters n and k and the superpacket length s are selected. The parameter n represents the block length of the FXC codeword. The parameter k represents the number of information symbols in the FXC codeword. Here, the terms “FXC codeword” and “FXC parity superpacket” can be used interchangeably, and these are the data on which the FXC parity superpacket is calculated. Contrast with the element “information superpacket”.

The expected loss synchronization length should be: s * (n / k) * h (if h = n−k) or less.
The overhead rate of FXC encoding is h / k. Thus, for example, to protect against 500 msec fading in a system that broadcasts at 19.2 M bps (ie, 2.4 M bytes / sec), s * (n / k) * h ≦ 1.2 M bytes must be satisfied. For example, when n = 6 and k = 4 (and h = 2), the overhead rate is 50% and s = 400 Kbytes. The present invention may require additional storage of s * n bytes beyond that required by the standard ATSC8VSB system, in this example 2.4 Mbytes. When a plurality of programs share a communication bandwidth of 19.2 Mbps, the required memory can be reduced by performing FXC encoding only on the encoded program.

  FIG. 1 is a schematic diagram illustrating a vestigial sideband (VSB) transmitter 100 according to an embodiment of the present invention. The transmitter configuration described herein and illustrated in FIG. 1 is exemplary, and the present invention is directed to other transmitter configurations capable of VSB symbol transmission while maintaining the spirit and scope of the present invention. May be applied.

  The transmitter 100 includes one source encoder 105, FXC encoder 110, mobile MUX 115, frame synchronizer 120, data randomizer 125, Reed-Solomon (RS) encoder 130, data interleaver 135, synchronous insertion module 140, pilot insertion. Module 145, 8-VSB modulator 150 and analog up-converter 155.

  The compressed bitstream from the source encoder 105 is processed by the FXC encoder 110 that generates parity superpackets and adds repeatability. Several inputs are combined into one composite stream by transport MUX 115 and sent to frame synchronizer 120. Data frames are generated and delineated by the frame synchronizer 120. Each data frame has two fields. Each data field has a 313 data segment in the ATSC standard, including a frame synchronization segment. Each data segment carries a 188 byte mobile packet and associated forward error correction (FEC) overhead. Data randomizer 125 is used for all input data to randomize only the data payload (not including overhead and synchronization elements). The Reed-Solomon (RS) encoder 130 is a block encoder that reserves 187 bytes and adds 20 parity bytes for external forward error correction at the receiver. The byte data is interleaved by a data interleaver 135 using a convolutional interleaver that calculates 52 data segments (approximately one sixth of the depth of the field). A trellis encoder 136 implements a two-thirds rate four-state trellis coding scheme with one pre-encoded bit encoded. The synchronization insertion module 140 inserts data segment synchronization into each data segment and inserts data field synchronization into each data field. These elements are not Reed-Solomon or trellis encoded. The pilot insertion module 145 introduces a pilot signal by adding a direct current (DC) component to each symbol. Thereafter, the 8VSB modulator 150 draws a symbol (generated at 10.76 M symbols / sec) on the set of 8VSB, and generates a root Nyquist pulse corresponding to each symbol. The analog up-converter 155 then converts the symbol to the desired carrier frequency for transmission.

  FXC encoder 110 is placed after source encoder 105, but before mobile MUX 115 and channel coding block (eg, RS encoder 130). k super packets and s of each length are input to FXC encoder 110. The FXC encoder 110 generates a super packet of parity of length s and h = n−k. The initial information super packet and the parity super packet are multiplexed by the mobile MUX 115. The ATSC8VSB system uses MPEG-2 mobile streams in mobile MUX that allow multiple programs to be multiplexed, each program being assigned a different process identifier (Process IDentifier, PID) and transmitted on the same channel To do. According to the present invention, the information super packet has not been changed in relation to a system that does not use FXC. According to the present invention, the additional parity superpacket to be transmitted is assigned a different PID by the mobile MUX 115 than the information superpacket. If multiple programs, each with a different PID, are sent, a parity superpacket can be calculated based on all the programs together or based on one or more individual programs. is there. All the data is sent to the remaining channel coding unit of ATSC8VSB that does not need to be changed to the standard system. To assist synchronization at the receiver, a special FXC synchronous mobile packet can be sent once for each superpacket n based on different PIDs. These special FXC synchronous mobile packets include super packet sequence number start position, MPEG-2 mobile packet PID, program clock reference (PCR), continuity counter field, super It shows the correspondence between the packet length s and the Reed-Solomon parameters (n, k).

  FIG. 2 is a schematic diagram illustrating a vestigial sideband (VSB) receiver 200 according to an embodiment of the present invention. The transmitter configurations described herein and illustrated in FIG. 2 are exemplary, and the present invention is directed to other transmitter configurations capable of VSB symbol transmission while maintaining the spirit and scope of the present invention. May be applied.

  The receiver 200 includes a tuner 205, an intermediate frequency (IF) filter and synchronization detector 210, a synchronization and timing module 215, an equalizer 220, a phase tracker 225, a trellis decoder 230, a data interleaver 235, a Reed-Solomon ( RS) decoder 240, data de-randomizer 245, mobile DEMUX 250, FXC decoder 260, and source decoder 265.

  In the 8VSB transmission signal, the digital information is transmitted exclusively with the amplitude of the RF envelope, not the phase. Eight levels of the transmitted signal are recovered by sampling only the I channel or in-phase information. Since the dependence on the Q channel is eliminated, the receiver 200 need only process the I channel. Therefore, the number of digital signal processing circuits required at different stages of the receiver 200 can be reduced to half. This clearly makes it possible to simplify the formation of the receiver and reduce costs.

  The symbols are demodulated at the receiver 200 by applying the opposite principle applied to the transmitter 100 of FIG. That is, incoming VSB symbols are received, downconverted, filtered, and detected. Then, segment synchronization and frame synchronization are recovered. The incoming signal is converted to an intermediate frequency by tuner 205. Communication path selection is performed by the IF filter and synchronization detector 210. Timing recovery and coarse carrier recovery are then achieved by the synchronization and timing module 215. Then, the signal is equalized by the equalizer 220 to remove all the multiplexed channel components. Since the equalizer performs calculation using the I channel or actual information, as one advantage of the VSB system, complicated equalization is not required. The output of equalizer 220 is applied to phase tracking 225 to remove residual phase jitter. The output of the phase tracker 225 is applied to a trellis decoder 230 corresponding to the inner code of the concatenated forward error correction (FEC) system. The output of the trellis decoder 230 is applied to a data deinterleaver 235 that spreads bursts of errors due to trellis decoder induction. The de-interleaved data corresponds to the outer code of the concatenated coding system and is applied to a block-based Reed-Solomon (RS) decoder 240. The output of the RS decoder 240, that is, the channel encoded data, is processed by the data de-randomizer 245 corresponding to the inversion processing at the transmitter. The de-randomized data is then sent to the mobile de-multiplexer 250 for audio / source decoder, video source decoder, and data source decoder (collectively represented as source decoder 260). The combined stream is separated into component streams.

FXC decoder 255 is placed after other channel decoding blocks (eg, RS decoder 240 ) and mobile DEMUX 250, and before source decoder 260 . By using the FXC synchronous mobile packet, it is possible to determine and determine the number and position of the super packet sequence. By using the error indication signal from one of the preceding channel coding blocks such as the RS decoder 240, the FXC decoder 255 can use the erasure position. For use in the MPEG-2 transmission stream system, the movement error indication field of the movement packet can be used to indicate the position of the error.

If no synchronization loss occurs, FXC decoding is unnecessary, and the FXC decoder 255 only passes the data of the information superpacket to the source decoder 260 . If synchronization loss occurs and is detected, superpackets with loss or bad data are marked with loss. If k or more superpackets are successfully received, the FXC decoder 255 obtains 1 byte from each superpacket to form a codeword, whether it is an information packet or a parity packet. The sRS (n, k) codeword is decrypted to completely reconstruct the lost information packet. Of course, more than one byte can be obtained from each superpacket to form a codeword. That is, the description of the present invention allows one skilled in the art to anticipate other data units used to form codewords from the above and superpackets.

  In the FXCRS (6, 4) system, an example will be given in which six super packets having a length of 400 Kbytes are transmitted. Super packets 0-3 are information super packets, and super packets 4 and 5 are parity super packets. Due to the loss of synchronization, superpackets 3 and 4 become bad at the receiver. FXC encoding is performed on 400,000 codewords, each codeword gets the first through fourth bytes of superpackets 0, 1, 2, and 5 and takes the third and fourth positions. Is formed by marking as erasure. Superpacket 3 is completely reconstructed by FXC decoder 260 and superpackets 0-3 are sent to source decoder 265. An example of this is shown in FIG. 3, which shows an example of a super packet loss pattern according to an embodiment of the present invention.

  Although the present invention has been described for an Advanced Television Systems Committee (ATSC) 8VSB digital communication system, the present invention can be applied to any packet-based digital communication system.

    Although the embodiments of the present invention have been described with reference to the accompanying drawings, the present invention is not limited to the above-described embodiments, and may be modified and changed by those skilled in the art without departing from the spirit and technical scope of the present invention. Improvements can be made. These changes and modifications are intended to be included within the scope of the claims.

FIG. 1 is a schematic diagram illustrating a vestigial sideband (VSB) transmitter according to an embodiment of the present invention. FIG. 2 is a schematic diagram illustrating a vestigial sideband (VSB) receiver according to an embodiment of the present invention. FIG. 6 is a block diagram illustrating the use of virtual repetition of high priority packets. FIG. 3 is a diagram illustrating an example of a super packet loss pattern according to an embodiment of the present invention.

Claims (14)

A device that enables recovery of lost information in a digital communication system, which:
FXC encoder for calculating a forward erasure correction (FXC) parity superpacket over the information superpacket for subsequent recovery of any one of the information superpackets at least partially compromised by synchronization loss ; and
Said information before performing any of the transmission of the super packet and the FXC parity superpackets, these multiplexer server for multiplexed; consists,
The apparatus, wherein the FXC parity superpacket is calculated for a period corresponding to an expected length of at least one synchronization loss period .   The apparatus of claim 1, wherein the FXC encoder calculates the FXC parity superpacket over the information superpacket with one byte for each of the information superpackets.   2. The apparatus according to claim 1, wherein when the information packet is k and each length is s, the FXC encoder calculates the FXC parity superpacket h having the length s. Where h = nk, n = block length of each FXC parity superpacket, and k = number of information symbols in each FXC parity superpacket.   The apparatus of claim 1, wherein the multiplexer assigns a process identifier (PID) different from the information superpacket to the FXC parity superpacket.   The apparatus of claim 1, wherein the FXC encoder calculates the FXC parity superpacket using a Reed-Solomon (RS) code.   The apparatus according to claim 1, further comprising a multiplexer for generating an FXC synchronous mobile packet indicating a correspondence with a super packet sequence number start position. A device for recovering lost information in a digital communication device, comprising:
To decode a forward erasure correction (FXC) parity superpacket previously calculated over the information superpacket to recover the whole of any one of the information superpackets at least partially compromised due to synchronization loss FXC decoder;
Consisting of
The FXC decoder further decodes the FXC synchronous mobile packet to determine a super packet sequence number and super packet position for both the FXC parity super packet and the information super packet ;
The apparatus, wherein the FXC decoder is adapted to receive an error signal indicating an erasure location corresponding to the information superpacket at least partially compromised by the synchronization loss . A method that enables recovery of lost information in a digital communication system comprising:
Calculating an FXC parity superpacket over the information superpacket for subsequent recovery of any one of the information superpackets at least partially compromised by synchronization loss; and
Wherein before performing any transmission of information superpackets and the FXC parity superpackets, these steps of the multiplex; consists,
The method wherein the calculating step calculates for a period corresponding to an expected length of at least one synchronization loss period . 9. The method of claim 8 , wherein the calculating step includes calculating the FXC parity superpacket over the information superpacket with one byte for each of the information superpackets. 9. The method of claim 8 , wherein the information packet is k and each length is s, the calculating step includes calculating an FXC parity superpacket h of the length s. Where h = nk, n = block length of each FXC parity superpacket, and k = number of information symbols in each FXC parity superpacket. 9. The method of claim 8 , wherein the multiplexing step comprises assigning a process identifier (PID) different from the information superpacket to the FXC parity superpacket. 9. The method of claim 8 , wherein the calculating step calculates the FXC parity superpacket using a Reed-Solomon (RS) code. 9. The method of claim 8 , further comprising the step of generating an FXC synchronous mobile packet indicating a correspondence with a super packet sequence number starting position. A method for recovering lost information in a digital communication device comprising:
Decoding the FXC parity superpacket previously calculated over the information superpacket to recover any one of the information superpackets at least partially compromised by synchronization loss; and the FXC parity Decoding the FXC synchronous mobile packet to determine the superpacket sequence number and superpacket position for both the superpacket and the information superpacket;
A method consisting of:

Images