KR20160148229A

KR20160148229A - Method for generating packet using a unequal error protection

Info

Publication number: KR20160148229A
Application number: KR1020150084940A
Authority: KR
Inventors: 정기웅
Original assignee: 주식회사 구버넷; 정기웅
Priority date: 2015-06-16
Filing date: 2015-06-16
Publication date: 2016-12-26
Also published as: KR101761812B1

Abstract

The present invention relates to a packet method of an AND-OR tree-based coded symbol, and more particularly, to a method and apparatus for encoding and decoding packets of a coded symbol transmitted from a transmitter in an unequal error protection (UEP) A stream is configured to transmit data according to priority, and a dependency between packets is given according to a weight (importance) of a packet to be transmitted, and a symbol (Common Symbol) having a common correlation with a source symbol is relatively A method of generating a packet of a symbol using an unequal error correction technique capable of preventing data loss due to packet loss occurring in a wireless network for a video streaming service by combining the packet with a high priority packet .

Description

[0001] METHOD FOR GENERATING PACKET USING A UNEQUAL ERROR PROTECTION [0002]

The present invention relates to a packet method of an AND-OR tree-based coded symbol, and more particularly, to a method and apparatus for encoding and decoding packets of a coded symbol transmitted from a transmitter in an unequal error protection (UEP) A stream is configured such that transmission of data according to priority is performed, and the encoded symbols are classified according to importance, and a symbol (Common Symbol) having a common correlation with the source symbol among the classified symbols is transmitted with a relatively high importance And more particularly, to a method of generating a symbol packet using an unequal error correction technique capable of reducing a packet transmission error by further combining and transmitting a packet (High Priority Packet).

According to the development of electronic and communication technologies, various information is transmitted and received through a network, and information transmitted through such a network is transmitted in the form of a truncated packet for easy transmission through the network.

These packets split the information to be transmitted, but reassemble it back to the original file where it is received. These packets consist of a header and a data terriler. The header contains the address and order to which data is to be transmitted, and the error information is recorded in the terrer.

Various techniques related to such a symbol packetization algorithm are disclosed in Patent Document 1 and various others.

Patent Document 1 discloses a method of generating at least one first signal including a first relative time value depending on an input signal and generating at least one second signal associated with at least one first signal depending on an input signal, And to generate at least one indicator associated with each of the at least one second signal, wherein each indicator is dependent on a first relative time value.

The data generated by the conventional packet generation method is a packet generation method that does not take account of the importance of data. When data is lost in one packet during data transmission, there is a problem that data can not be normally received. The Applicant has developed a method of packetizing a coded symbol and a device using this method in Patent Document 2 with respect to a symbol packetization algorithm.

The technique of Patent Document 2 relates to a method of packetizing a symbol stream that can improve the quality of a video streaming service. As shown in FIG. 3, a method of determining a first source symbol and an AND- When the first coded symbol that is not packetized is present among at least one first coded symbol which is a coded symbol of one source symbol, the non-packetized first coded symbol and the non-packetized first coded symbol are inserted A second source symbol based on at least one non-packetized first coded symbol is generated using a coded symbol and a target packet selection step for selecting a target packet and an AND-OR tree structure, Generating at least one second coded symbol based on the second source symbol using a tree structure, With the first coded symbol.

Here, the coded symbol may be a symbol coded using a luby transform. The method of packetizing a coded symbol may further include the steps of selecting the coded symbol and the target packet, and when the first coded symbol is not packetized in the packetizing step, The source symbols in the same AND-OR tree are set as the first source symbols, and the coded symbols and the target packet are selected and packetized.

When mapping the symbols of the LT code to the AND-OR tree, the source symbol is regarded as the Or- node and the coded symbol is regarded as the And-node, each of which associates with the parent and child nodes.

That is, packetization of coded symbols based on the And-Or tree is performed to help the video streaming service in the wireless network. The relationship between the encoded symbols can be analyzed by using the And-Or tree, and the effect of packets lost between the data transmissions using the method of encoding based on the analyzed result reduces the correlation between the packets Can be locally limited.

Accordingly, it is possible to minimize a packet loss occurring in a wireless network, and in particular, to provide a video streaming service that minimizes quality degradation of a video streaming service and reduces dependency among delivered packets to stably satisfy QoS.

However, in the technique of Patent Document 2, as described above, the first coded symbol is generated for the first source symbol using the AND-OR tree structure, and the first coded symbol is generated for the first coded symbol using the AND- After generating the second source symbol and again generating the second encoded symbol for the second source symbol, eventually encoding is performed for one source symbol, so that the other encoded symbols having connectivity between the source symbols So that the probability of errors in transmission of large-capacity media data increases.

That is, although fuzzy codes such as LT codes and Raptor codes provide equal error protection for all data, in some application services such as multimedia streaming service, a part of data has higher reliability and priority than other parts However, in the symbol packet method described above, there is a problem that the reliability of data restoration is lowered due to packetizing of one source symbol.

1. Korean Patent Publication No. 10-2010-0096220 2. Korean Patent Registration No. 10-1153520

SUMMARY OF THE INVENTION The present invention has been developed in order to solve the problems of the related art as described above, and it is an object of the present invention to provide a method and apparatus for providing dependency between packets according to weight (importance) ) Is combined with a packet having a relatively high priority (High Priority Packet) to transmit the packet, thereby preventing the data loss due to the packet loss occurring in the wireless network for the video streaming service. And a method of generating a packet.

According to an aspect of the present invention, there is provided a method of generating a symbol packet using an unequal error correction technique, the method comprising: packetizing a coded symbol using an unequal error correction coding (UEP) (PEP) value of a predetermined packet after identifying the priority of the coded symbol based on the UEP of the packet, and inserting the symbol according to the importance of the packet into the packet space corresponding to the threshold value A common symbol is commonly combined with a relatively high priority packet (High Priority Packet) to form a symbol packet block, and then streaming transmission is performed.

Also, in the method of generating a symbol packet using the unequal error correction technique according to the present invention, a coded symbol is a symbol encoded using a luby transform, and a non-uniform error correction coding (UEP) The present invention is applied to block coding including LDPC according to characteristics, to fuzzy coding including LT, Rapto codes, or to coding using convolutional and turbo codes.

In addition, according to the present invention, there is provided a method of generating a symbol packet using an unequal error correction technique, wherein the unequal error correction coding (UEP) includes a binary symmetric channel (BSC), a binary cancel channel (BEC) Channel (AWGN). &Lt; / RTI >

A method of generating a symbol packet using an unequal error correction technique according to the present invention is characterized in that a dependency between packets is imparted according to a weight of a packet to be transmitted and a common symbol having a correlation with a source symbol is relatively (High Priority Packet) to a high priority packet, thereby transmitting the packet to the High Priority Packet. Thus, data loss due to packet loss occurring in the wireless network can be suppressed.

FIG. 1 is a block diagram for explaining generation of a LT coded symbol considering a UEP function according to the present invention.
FIG. 2 is a flow chart for explaining a process of generating a LT encoded symbol considering a UEP function.
FIG. 3 is a diagram for explaining LT coding symbol packetization using the conventional And-Or tree.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Like reference numerals are used for like elements in describing each drawing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

As shown in FIG. 1, a method of packetizing coded symbols using unequal error correcting coding (UEP) according to the present invention identifies a priority of a coded symbol based on unequal error correcting coding (UEP) A symbol corresponding to a priority of a packet is inserted into a packet space corresponding to a threshold value (Pth) of a predetermined packet, and a symbol (Common Symbol) having a correlation with a source symbol is relatively Among the High Priority Packet and the Low Priority Packet, a higher priority packet is further combined to form a symbol packet block, and then the stream packet is transmitted.

That is, as shown in FIG. 1, two common correlation symbols are combined into a high priority packet, and one common low priority packet is combined with a high priority packet. And combines the symbols having a correlation (common symbol) to generate a packet.

In FIG. 1, the importance is classified into two types. However, when the importance is divided into a plurality of classes, a symbol having a correlation in common from the packet having the highest importance to the packet having the least importance is gradually do.

The method of generating a symbol packet using the unequal error correction technique according to the present invention is performed by the present applicant in the same way as in the filing of Japanese Patent Application No. 1153520 filed by the present applicant. In other words, the priority of the coded symbol is identified based on the unequal error correction coding (UEP), and the priority of the packet is determined as the packet space corresponding to the threshold value (Pth) And inserting the symbol according to the symbol; B) inserting a symbol associated with each source symbol into a remaining packet space, and determining whether a remaining space of the packet exists; C) completing one packet when the remaining space of the packet does not exist in the step B), and inserting a symbol having no relevance to each source symbol when the remaining space of the packet exists; And D) forming a symbol packet block partitioned by a relatively high priority packet (a high priority packet) and a low priority packet (a low priority packet), and then transmitting the stream packet.

In the method of packetizing coded symbols according to the above process, in the process (A) of inserting symbols separately, a common symbol having a common correlation with the source symbol is referred to as a packet having a relatively high priority (High Priority Packets and Low Priority Packets.

Unequal Error Protection (UEP) used in the present invention is a bit wise UEP that divides a message bit by unit so that a bit set at a specific position has an error correcting capability different from that of the remaining bit sets It will be applied separately to the message wise UEP.

In the present invention, the UEP is used to increase the dependency between packets in a packet process of a coded symbol. UEP using a block code including an LDPC or a fuzzy code such as LT or Rapto code is used according to the characteristics of a subject code UEP may be applied, or UEP using convolutional or turbo coding may be applied.

Furthermore, it is desirable to apply the unequal error correction technique (UEP) in various scenarios on various channels such as a binary symmetric channel (BSC), a binary clear channel (BEC), a Gaussian noise channel (AWGN)

FIG. 1 is a block diagram illustrating a method of generating a coded symbol using the UEP technique according to the present invention. Referring to FIG.

1, a coded symbol generated by using the first source symbol and the first source symbol using the And-Or tree is referred to as a first coded symbol, a first coded symbol is referred to as an And-Or tree, And a coded symbol generated by using an And-Or tree as a second source symbol is defined as a second coded symbol.

That is, when the depth of the tree increases, a source symbol located at the top of the generalized tree structure is used as an n-th source symbol, a coded symbol generated by using an And-Or tree as an n-th source symbol is referred to as an n-th coded symbol, An encoded symbol generated by using an And-Or tree as an (n + 1) -th source symbol and an (n + 1) th source symbol as a source symbol generated by using an And-Or tree is defined as an (n + 1) .

As shown in the figure, the LT coded symbol packet block first determines a first source symbol to be coded, and selects a first source symbol, which is a coded symbol of a first source symbol generated using an And-Or tree structure, If there is one coded symbol, the non-packetized first coded symbol is selected.

Then, a target packet to be inserted with the first coded symbol not packetized is determined, a first coded symbol is inserted into the target packet, and a coded symbol based on the first coded symbol generated using the And- 2 source symbols, generates a second coded symbol based on the second source symbol using the And-Or tree structure, and at least one of the second coded symbols is packetized with the first coded symbol in the target packet do.

However, if the depth of the tree is equal to or greater than 3, the n-th source symbol to be coded is determined and the n-th coded symbol, which is the coded symbol of the n-th source symbol generated using the And- The non-packetized n-th coded symbol is selected.

Then, the target packet to be inserted with the non-packetized n-th coded symbol is determined, and the first coded symbol is inserted into the target packet. Further, an (n + 1) -th source symbol based on the generated n-th coded symbol is generated using the And-Or tree structure, and an n + 1 encoded symbols, and packetizes at least one of the (n + 1) -th coded symbols together with the n-th encoded symbols in the target packet.

The LT encoded symbol packet thus configured includes a plurality of source symbols (Root 1, Root 2, ...), a common symbol having a common correlation with each source symbol, and a correlation with each source symbol (Non Symbol).

In addition, the symbols have different priorities in unequal error correction coding (UEP), and are classified into a plurality of levels of MIB (More Important Bits) and a plurality of levels of LIB (Less Important Bits). This distinguishes the importance of symbols in UEP-LDPC coding and UEP-LT coding. For the sake of convenience of explanation, it is a matter of course that the importance of symbols will be classified into MIBs and LIBs and packetized, and packets can be packetized for each level of importance.

Therefore, according to the present invention, bit blending is performed for a symbol having a high priority (High Priority) and a symbol having a low priority (Low Priority), so that a message bit is divided into units, Correcting ability.

As described above, the bit mixture is associated with each of the source symbols (Root1 and Root2), with one source symbol Root1 having a higher priority and another source symbol Root2 having a lower priority, A symbol for each source symbol, and a symbol for which there is no probability, and so forth to form a coded block.

Then, UEP packets are classified into low-priority packets and high-priority packets, which are generated by bit-wise UEPs, that is, information sets requiring a high communication success rate, that is, Most-Significant Bits ), And an information set that is less important but provides improved quality upon reception, i.e., LSB (Least-Significant Bits).

That is, when UEP is applied as in the present invention, a user who is close to a base station or able to access a plurality of base stations can receive a relatively large amount of data, and a service can be performed. However, Data is received. UEP packetization corresponding to cell coverage according to UEP application should be made.

For example, in the case of a video stream, a higher error correction rate is given to the SHD-class image and a lower error correction rate is given to the HD-class image, thereby changing the service quality flexibly according to the environment without changing the transmission policy according to the user environment It will be possible.

FIG. 2 is a flowchart for explaining UEP packet timing according to the present invention.

As shown in the figure, in step S301, the length P of the entire packet is set. The length of the packet may be determined according to the protocol. Thereafter, the flow advances to step S303 to determine the threshold value (Pth) of the packet, which can be set to 0 Pth. If the threshold value (Pth) is '0', it can be assumed that the streaming is performed without using the UEP (Non-uniform Error Correction) technique.

When the threshold value Pth is determined, the priority of each source symbol of the LT encoded symbol packet is recognized as in step S305. The significance of the symbols can be recognized in the UEP-LDPC encoding or the UEP-LT encoding process, and each level value for importance is provided. Each level value is identified by a plurality of steps. In the present invention, symbols having a high level of importance, that is, MIB (More Important Bits) and LIB (Less Important Bits), are separately described.

As described above, the importance is applied to MSB (Most Significant Bits), which is an important information set, and LSB (Least Significant Bits), which is a less important information set. Can be distinguished.

Accordingly, in step S305, the MIB and LIB for the symbol are identified to form a level-specific packet for the symbol. In FIG. 2, low priority packets and high priority packets are identified through UEP packet timing.

Thereafter, the flow advances to step S307 and a symbol according to importance is introduced into a packet space defined by a predetermined threshold value (Pth). For example, when the importance of the source symbol corresponding to Root 1 is high and the importance of the source symbol corresponding to Root 2 is low, symbols with low importance are collected and received as one packet, symbols having high importance are collected, As shown in Fig. The configuration of such a packet repeatedly extracts and pulls in accordance with a predetermined threshold value (Pth).

That is, the MIB symbol and the LIB symbol are inserted into the packet space corresponding to the threshold value (Pth), and it is determined whether or not each symbol is inserted into the corresponding packet space. If inserting of the corresponding symbol is not completed in this process, the process returns to step S305 to insert a continuous symbol, and if all the symbols are inserted into the corresponding packet space, step S309 is entered.

In this step, the common symbol is inserted into the remaining space excluding the packet space corresponding to the threshold value (Pth) from the total packet length (P). A common symbol is inserted into the remaining packet space as a symbol correlating to each source symbol Root. If necessary, the number of inserted common symbols may be limited. In this case, the threshold value Pcth for the common symbol may be defined. Also, as a preferred embodiment, the number of insertions of the common symbols may be determined in proportion to the importance of each source symbol to which the common symbol is related.

Accordingly, if the insertion of the common symbol is completed as in step S311, the process proceeds to step S313 to determine whether there is a remaining packet space. If there is no remaining packet space, packet streaming is formed in step S317. If there is a remaining packet space, step S315 is performed to identify and insert the non-random symbol into the remaining packet space. Non-plausible symbols are logically generated symbols regardless of source symbols (Root1, Root2, ...) and are defined as symbols with very low importance.

When the UEP symbol packet block is completed as described above, packet streaming is performed by attaching an identifier or a header file preset in the first and / or the end of each packet. Accordingly, when an arbitrary packet is lost on the wireless network, the reliability of the packet restoration is improved by restoring the lost packet data based on the association of each packet.

root: the target source symbol
Low Priority Packet:
High Priority Packet: High Priority Packet
Pth: Threshold value of the packet
P: length of the entire packet

Claims

A method of packetizing coded symbols using non-uniform error correction coding (UEP), the method comprising:
The priority of the coded symbol is identified based on the unequal error correction coding (UEP), the symbol corresponding to the importance of the packet is classified into the packet space corresponding to the threshold value (Pth) value of the predetermined packet, However,
A common symbol having a common correlation with the source symbol is combined with a relatively high priority packet in a relatively high priority packet and a low priority packet, And forming a packet block and streaming the packet. The method of generating a packet of a symbol using an unequal error correction method.

The method according to claim 1,
Wherein the common symbol is divided into a number proportional to the importance of the source symbol, and inserted as a packet of high importance and a packet of low importance, and a method of generating a packet of a symbol using the unequal error correction method.

The method according to claim 1,
Wherein the first coded symbol among the coded symbols is a coded symbol using a luby transform. 10. The method of claim 1, wherein the first coded symbol is a coded symbol using a luby transform.

3. The method according to claim 1 or 2,
The unequal error correction coding (UEP) may be applied to block coding including LDPC according to the characteristics of a subject code, to fuzzy coding including LT and Rapto codes, or to coding using convolutional and turbo codes And generating a symbol packet using the unequal error correction technique.

3. The method according to claim 1 or 2,
Wherein the unequal error correcting coding (UEP) is applied to one of a binary symmetric channel (BSC), a binary clear channel (BEC) and a Gaussian noise channel (AWGN) according to a communication environment. A Packet Generation Method of Symbols Using Correction Technique.