CN105072588B - The multi-medium data method of multicasting that full linear is protected without error correction - Google Patents

Abstract

The invention discloses a completely linear multicast method AnalogCast without error correction protection. The method includes: at the sending end, performing decorrelation transformation on multimedia data, L-shape-based block and data stretching, whitening transformation and group-of-pictures (GoP)-based interleaving, through linear modulation without error correction protection. communication channel for transmission. At the receiving end, de-interleaving, de-whitening, stretching factor estimation based on received data, model correction based on fitting correction factor, de-stretching and de-correlation transformation are performed on the received data. The present invention can obtain the following advantages: (1) superior fairness. (2) There is no "cliff" effect in metadata transmission, and it has better robustness and transmission distance. (3) There is no digital sidewalk, and the computational complexity and memory overhead are reduced. (4) There is no need for metadata transmission, reducing bandwidth occupation. (5) In lines with high packet loss rate, better image quality is obtained.

Description

Fully Linear Multimedia Multicast Method Without Error Correction Protection

technical field

The invention relates to the technical field of wireless multicasting, in particular to a method for multimedia data multicasting with full linearity and no error correction protection.

Background technique

Wireless multicast is a video broadcast service applied to different channel situations. The existing wireless multicast system has a short-board effect, that is, the video or picture quality received by all receivers is the same as that received by the receiver with the worst channel. This short board effect makes the existing wireless multicast system unable to be "fair" to all receivers. At the same time, because the existing multicast system adopts a redundant picture compression method to be transmitted, the traditional wireless multicast system cannot tolerate the occurrence of packet loss during wireless transmission, which makes the traditional wireless multicast system suffer from channel changes. In severe cases, it shows poor robustness.

Also in the field of satellite remote sensing pictures, because the transmission bandwidth of a given picture is relatively narrow, it is necessary to compress the picture with a higher compression rate (using JPEG or JPEG2000), and the high compression rate will cause distortion of the picture, that is, the peak signal of the picture The noise ratio drops because the loss has been introduced in the source coding process. Then, in the case of no packet loss in channel transmission, even if the channel condition improves, the transmission quality will not improve, thus wasting the improvement brought about by channel optimization.

At the same time, for the existing network system, due to the increase of users' demand for network services, the network will be congested and packet loss will occur. In the traditional image transmission method, the data source first performs nonlinear data compression to compress the larger data source, and then transmits it through a channel encoder with a forward error correction code. Once data loss occurs, it will cause packet loss. , will cause a large drop in image quality, resulting in unreadable or unusable images. Therefore, in a system with high packet loss, how to ensure the robustness of the system becomes the key to affecting user experience and becomes the first consideration for system designers. question.

Existing research mainly focuses on proposing some solutions or improving algorithms around the problems existing in practical application requirements. These technologies mainly use the method of removing entropy coding at the source to break the transferability of transmission errors caused by source compression in traditional wireless picture transmission methods. This enables the data source to tolerate a certain amount of data loss and bit errors, and greatly improves the system stability in high packet loss and harsh channel environments such as low signal-to-noise ratio environments and strong multipath environments. These technologies use a combination of analog (linear) modulation and digital (non-linear) modulation at the channel to double the transmission rate. The transmission method of video or picture data in a kind of wireless video multicast system in the prior art is: Softcast (soft transmission scheme), and Softcast uses linear coding method (2D-DCT transformation, energy allocation, Hadamard Transformation) to form data and metadata, use analog modulation communication system (except protection error correction and interleaving, etc.) to transmit data at the channel, and use QAM (Quadrature Amplitude Modulation, quadrature amplitude modulation) modulation communication system to transmit elements at the side channel data.

But the shortcoming of the Softcast scheme in the above-mentioned prior art is:

1. There is a digital side channel in the Softcast solution. When the channel quality is poor to a certain extent, a cliff effect will appear, and the image quality will suddenly drop.

2. The channel has a digital side channel and an analog main channel, and the calculation overhead of the digital side channel is relatively large.

3. Metadata transmission takes up a certain amount of bandwidth.

4. In the channel environment with high packet loss, the image quality needs to be improved.

Contents of the invention

Embodiments of the present invention provide a fully linear multicast method for multimedia data without error correction protection, so as to realize effective full linear transmission of multimedia data without error correction protection.

In order to achieve the above object, the present invention adopts the following technical solutions.

A fully linear multimedia data multicast method without error correction protection, comprising:

At the sending end, decorrelation transformation is performed on the multimedia data:

Data stretching based on block partitioning is performed on the data after decorrelation transformation;

Perform whitening transformation processing on the stretched data blocks, perform intra-frame interleaving processing on the data after the whitening transformation processing, and form the data packets after the intra-frame interleaving processing;

After analog modulation is performed on the data packet, it is transmitted.

Preferably, at the sending end, performing decorrelation transformation on the multimedia data includes:

Separate each frame of multimedia data to be transmitted independently, and then split each frame of image into a large block X of MN×MN, and translate the large block X of each frame of image by 2 ^b-1 , where b is the image sample Depth, the entire frame decorrelation transformation is performed on the translated large block X:

y=Tr(x-2 ^b-1 )

Y is the large block after the decorrelation transformation, Tr is the decorrelation transformation, and x is the pixel matrix of the large block.

Preferably, the data stretching based on block division for the data after decorrelation transformation includes:

At the sending end, data stretching based on L-shaped block division is performed on the multimedia data after decorrelation transformation, and the large block is divided into L-shaped small blocks according to the set block division method. The length of each L block is equal to The width is a fixed value, and the average energy λ _i of each small block

Among them, λ _i is the average energy of each small block, i represents the sequence of small blocks, chunknum is the number of blocks that is N, y _{i, k} is the kth data of the i-th small block in the large block Y;

Calculate the stretching factor g _i by the average energy of each patch:

Where g _i is the stretching factor of each small block, P is the energy stretching factor, i represents the sequence of small blocks, and then multiply the data in each small block by the respective stretching factor to obtain the stretched data block u _{i ,j} :

u _i,j =g _i .y _i,j

Preferably, performing whitening transformation processing on the stretched data blocks, performing intra-frame interleaving processing on the data after whitening transformation processing, and forming data packets after intra-frame interleaving processing, including:

At the sending end, perform whitening transformation processing on the stretched data block u, let H be the whitening matrix, and the whitening transformation is as follows:

v= ^HuHT

Where v is the data processed by whitening transformation;

Perform intra-frame interleaving on the data after whitening transformation, classify the data in V according to their respective positions, divide the data in the frame into four types according to their positions: A pixel, B pixel, C pixel and D pixel, Randomly rearrange each type of pixel, and reorganize the randomly rearranged pixels into data blocks;

For each frame of the multimedia data, the decorrelation transformation, data stretching, whitening transformation processing and intra-frame interleaving processing are respectively performed, and the data blocks after the intra-frame interleaving processing of all frames are completed to form a GoP, and then A layer of GoP is encapsulated into a packet.

Preferably, the described data in the frame is divided into four types of pixels according to positions: A pixel, B pixel, C pixel and D pixel, including:

Determine the pixel at the first position of the first row of the large block after the whitening transformation process as an A pixel, the next pixel as a B pixel, the next pixel as a C pixel, and the next pixel as a D pixel, and so on. A row of pixel definitions;

In the second row, the first pixel starts with D pixel, the next pixel is A pixel, the next pixel is B pixel, and the next pixel is C pixel, and so on to complete the pixel definition of the second row;

In the third row, the first pixel starts with a C pixel, the next pixel is a D pixel, the next pixel is an A pixel, and the next pixel is a B pixel, and so on to complete the pixel definition of the three rows;

In the fourth row, the first pixel starts with B pixel, the next pixel is C pixel, the next pixel is D pixel, and the next pixel is A pixel, and so on to complete the definition of four rows of pixels;

Henceforth, the division of the pixel positions of the data in the frame of every four lines is the same as the division of the pixel positions of the first line, the second line, the third line and the fourth line.

Preferably, said transmitting the data packet after performing analog modulation includes:

The data in the data packet is encoded into a data frame through the physical layer, and the data frame is analog modulated, a real number in the data frame is mapped into an I vector, and another real number is mapped into a Q vector, and the analog The modulated data frame is transmitted.

Preferably, the method also includes:

At the receiving end, all received data packets are formed into GoP, and then GoP is decomposed into data blocks corresponding to 1 data frame;

The data block corresponding to each data frame includes various types of pixels after rearrangement. According to the preset pixel rearrangement rules, restore the pixels after various types of disorder, and then reset their respective positions to obtain The data after whitening transformation corresponding to each data frame;

Performing de-whitening transformation processing on the data after whitening transformation corresponding to each data frame, and performing stretch coefficient estimation on the data after de-whitening transformation processing;

Perform de-data stretching processing on the data after the de-whitening transformation process according to the estimated stretching coefficient, and perform inverse decorrelation transformation on the data after the de-data stretching process, to obtain the multimedia data sent by the sending end.

Preferably, performing de-whitening transformation processing on the data after whitening transformation corresponding to each data frame, and performing stretch coefficient estimation on the data after de-whitening transformation processing include:

At the receiving end, after de-whitening transformation processing is performed on the multimedia data, the stretching coefficient g _i is estimated, and the average energy in the L-shaped small block of the received data is calculated:

in is the average energy of the L-shaped small block of received data, and i is the sequence number of the small block; is the received data, i is the serial number of the small block, k is the kth data in the small block, and chunknum is the number of elements in the small block;

because F:

in is the average energy of the small block data after destretching, i is the small block sequence, is the data after destretching, k is the kth data in the small block, chunknum is the number of elements in the small block, and the following relationship is obtained:

Stretch factor The estimation formula for is as follows:

in is the stretch factor and i is the sequence of small blocks.

Preferably, the de-data stretching process is performed on the data after the de-whitening transformation according to the estimated stretching coefficient, including:

At the receiving end, after the multimedia data is estimated by the stretching coefficient g _i , the data aggregation coefficient mean _L is calculated, and the average amplitude of each small block is calculated

The data aggregation coefficient mean _L is the average amplitude of the first 1-M small blocks divided by the average amplitude of the previous M+1-2M small blocks. The calculation formula of the data aggregation coefficient mean _L is as follows:

Estimate the channel SNR and calculate the system packet loss rate PLR at the physical layer level, and calculate the correction factor through the channel SNR, packet loss rate PLR and data aggregation coefficient mean _L and Used to correct the estimation bias caused by noise, Used to correct the estimation bias caused by packet loss;

The following formula is to calculate the correction factor and The revised model:

in is the noise repair factor that mainly repairs the estimation error of the stretch factor caused by noise, is the packet loss repair factor which mainly repairs the estimation error of the stretch factor caused by packet loss. i is the serial number of the small block, MCHUNK is the number of small blocks, round(MCHUNK)<NUM<MCHUNK. Among them, C1, C2, C3, C4, C5, C6, C7, C8, C9, the parameters are obtained through a large number of experiments.

get correction factor and Afterwards, reverse stretching is carried out in combination with the stretching factor, the formula is as follows:

in is the estimated data after decorrelation change, i is the small block sequence, and j is the jth data in the small block; is the data after de-whitening, i is the small block sequence, j is the jth data in the small block; P is the energy parameter known locally by the receiver.

Preferably, performing inverse decorrelation transformation on the data after data stretching processing to obtain the multimedia data sent by the sending end includes:

At the receiving end, perform anti-correlation transformation on the entire frame of the data matrix Y after data de-stretching, and then perform inverse translation of b-depth on the data matrix after inverse de-correlation transformation to obtain a large block of data The formula is as follows:

where x is the estimated pixel matrix.

The above-mentioned processing procedure is repeatedly executed, and all the large-block pixel matrices X obtained from one M×M large-block pixel matrix X are spliced to obtain the multimedia data transmitted by the sending end.

As can be seen from the technical solutions provided by the above-mentioned embodiments of the present invention, the embodiments of the present invention propose a transmission method for fully linearly transmitting multimedia data such as video/pictures without error correction protection, and obtain the following advantages: (1) superior fairness. (2) There is no "cliff" effect in metadata transmission, and it has better robustness and transmission distance. (3) There is no digital sidewalk, and the computational complexity and memory overhead are reduced. (4) There is no need for metadata transmission, reducing bandwidth occupation. (5) In lines with high packet loss rate, better image quality is obtained.

Additional aspects and advantages of the invention will be set forth in part in the description which follows, and will become apparent from the description, or may be learned by practice of the invention.

Description of drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For Those of ordinary skill in the art can also obtain other drawings based on these drawings without any creative effort.

1 is a schematic diagram of the processing flow of the encoding and transmitting part in the method for performing a new full-linear non-error-correction-protected transmission of multimedia data according to an embodiment of the present invention;

2 is a schematic diagram of the processing flow of the receiving and decoding part in the method for performing a new full-linear non-error-correction-protected transmission of multimedia data according to an embodiment of the present invention;

Fig. 3 is a system block diagram of an embodiment of the present invention;

FIG. 4 is a schematic diagram of analog modulation according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of GoP interleaving according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of an L-shaped block based on an embodiment of the present invention;

Fig. 7 is a schematic diagram of the hardware system structure of the encoding and transmitting part of the embodiment of the present invention;

FIG. 8 is a schematic diagram of a hardware system structure of a receiving and decoding part according to an embodiment of the present invention;

Fig. 9 is a software flow chart of the code transmitting part of the embodiment of the present invention;

Fig. 10 is a software flowchart of the receiving and decoding part of the embodiment of the present invention;

FIG. 11 is a structural diagram of a software radio architecture according to an embodiment of the present invention;

FIG. 12 is a flowchart of ASIC implementation of the AnalogCast baseband system according to the embodiment of the present invention.

Detailed ways

Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

Those skilled in the art will understand that unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the word "comprising" used in the description of the present invention refers to the presence of said features, integers, steps, operations, elements and/or components, but does not exclude the presence or addition of one or more other features, Integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Additionally, "connected" or "coupled" as used herein may include wirelessly connected or coupled. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

Those skilled in the art can understand that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It should also be understood that terms such as those defined in commonly used dictionaries should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless defined as herein explain.

In order to facilitate the understanding of the embodiments of the present invention, several specific embodiments will be taken as examples for further explanation below in conjunction with the accompanying drawings, and each embodiment does not constitute a limitation to the embodiments of the present invention.

Embodiment one

This embodiment provides a processing flow of a fully linear multimedia data multicast method without error correction protection, as shown in Figure 1, including the following processing steps:

Step 1. Separate each frame of multimedia data to be transmitted independently, and then split each frame of image into a large block X of MN×MN.

Step 2. Translate the large block X of each frame image by 2 ^b-1 , where b is the image sampling depth, and then perform the whole frame decorrelation transformation on the large block after translation:

y=Tr(x-2 ^b-1 )

Y is the large block after the decorrelation transformation, Tr is the decorrelation transformation, and x is the pixel matrix of the large block.

The above decorrelation transformation can concentrate a large amount of energy (or information) of the image in a small area, and the decorrelation transformation can use K-L transformation, DCT transformation, DST transformation, DWT transformation, etc.

Step 3. In order to obtain the best error protection capability, data stretching must be performed. For the large block Y after transformation, block division and data stretching based on L shape are performed. Divide the large block into N L-shaped small blocks, as shown in Figure 6. The length and width of each L-shape are the same, and the division method of small blocks is pre-set locally at the receiving end.

Step 4. After determining the division of small blocks, calculate the average energy λ _i of each small block

Among them, λ _i is the average energy of each small block, i represents the sequence of small blocks, chunknum is the number of blocks that is N, y _{i, k} is the i-th small block in the large block Y, and the k-th data.

The contour information of the image is concentrated in the low frequency part, and the detail information of the image is concentrated in the high frequency part, so that the image has a certain ability to protect against errors.

Step 5. Determine the energy parameter P (locally known by the receiver) of each small block, and calculate the stretching factor g _i parameter:

Among them, g _i is the stretching factor of each small block, and i represents the sequence of small blocks. Then multiply the data in each small block by the respective stretching factor to obtain the stretched data u _i,j :

u _i,j =g _i .y _i,j

Data stretching can improve the noise immunity of the system.

Step 6. At the sending end, after data stretching is performed on the multimedia data, anti-packet loss processing is performed. The first is to make the information in the block have the same importance, which requires whitening processing on the data block, such as anti-correlation transformation, Walsh-Hadamard transformation and so on. Let H be the whitening matrix, and the whitening transformation is as follows:

v= ^HuHT

Among them, V is the data after whitening transformation, which enhances the anti-packet loss capability of the system.

Step 7. After the whitening transformation, the positional correlation of the data still remains to a certain extent. Next, intra-frame interleaving is performed on the data after the whitening process, and the data in V are classified according to their respective positions. The data in the frame is divided into: A pixel, B pixel, C pixel, and D pixel according to the position. As shown in Figure 5(a),

Determine the pixel at the first position of the first row of the large block after the whitening transformation process as an A pixel, the next pixel as a B pixel, the next pixel as a C pixel, and the next pixel as a D pixel, and so on. Pixel definition for a row.

In the second row, the first pixel starts with D pixel, the next pixel is A pixel, the next pixel is B pixel, and the next pixel is C pixel, and so on to complete the pixel definition of the second row.

In the third row, the first pixel starts with a C pixel, the next pixel is a D pixel, the next pixel is an A pixel, and the next pixel is a B pixel, and so on to complete the pixel definitions of the three rows.

In the fourth row, the first pixel starts with B pixel, the next pixel is C pixel, the next pixel is D pixel, and the next pixel is A pixel, and so on to complete the definition of four rows of pixels.

Henceforth, the division of the pixel positions of the data in the frame of every four lines is the same as the division of the pixel positions of the above-mentioned first line, second line, third line and fourth line. That is, the division of the pixel positions of the fifth row is the same as that of the first row, the division of the pixel positions of the sixth row is the same as that of the second row, the division of the pixel positions of the seventh row is the same as that of the third row, and the division of the pixel positions of the eighth row is the same as that of the third row. The division of pixel positions is the same as that of the fourth row. So on and so forth,…..

Then each type of pixel is randomly rearranged as shown in Fig. 5(b), and the random rearrangement method is known locally by the receiver. Eliminate the location dependence of the data. The various rearranged pixels are reorganized into large data blocks according to Fig. 5(c), and intra-frame interleaving is completed.

Step 8. Repeat steps 2 to 7 for one time, and for each frame of the multimedia data, respectively perform the decorrelation transformation, data stretching, whitening transformation processing and intra-frame interleaving processing.

Sometimes packet loss occurs suddenly, that is, burst packet loss. In a period of time, the packet loss rate will become very high. In order to smooth the burst packet loss rate, the present invention adopts the method of interleaving between frames to resist, and the data blocks after completing the interleaving process in the frame of all frames are formed into GoP (picture group), that is, large blocks of data interleaved within several frames are arranged to form a three-dimensional matrix, see Figure 5(d).

Then, a layer of GoP is encapsulated into a packet. As shown in Figure 5(d).

Step 9. At the sending end, after the multimedia data is packaged, the data package is encoded into a data frame through the physical layer, and the data frame is subjected to analog modulation. As shown in Figure 4(a), the two real numbers in the data frame are mapped to I and Q respectively. The feature of this mapping is that when the transmitter and receiver are close, the distance between the transmitted complex vector and the received complex vector is very small. close (as shown in Figure 4(b)), when the transmitter and receiver are far apart, the distance between the transmitted complex vector and the received complex vector is very far away (as shown in Figure 4(c)). During the analog demodulation, it is only necessary to map the I vector into a real number, and map the Q vector into another real number. Compared with QPSK/QAM modulation, encoding and decoding are simpler, and compared with digital modulation, it saves computing resources.

Step 10, decoding the received data frame and remapping it into a data packet.

Step 11, receive all the data packets, form the GoP with the received data packets, and then decompose the GoP into data blocks corresponding to 1 data frame.

Step 12. The data block corresponding to each data frame includes various types of pixels after rearrangement, according to the preset pixel rearrangement rules, restore the pixels after various types of disorder, and then reset their respective positions On, the data v after whitening transformation corresponding to each data frame is obtained.

Step 13, for the received data after whitening transformation We have:

Perform demodulation to obtain the dewhitened data of each small data block

data affected by channel noise.

Step 14. At the receiving end, perform de-whitening on the multimedia data, and then estimate the stretching coefficient g _i . First calculate the average energy μ _i in the L-shaped small data block of the received data, the calculation formula is as follows:

in is the average energy of the L-shaped small block of received data, and i is the sequence number of the small block; is the received data, i is the serial number of the small block, and k is the kth data in the small block. chunknum is the number of elements within the chunk.

Step 15, because At the same time we have:

in is the average energy of the small block data after destretching, and i is the small block sequence. is the data after destretching, i is the small block sequence, k is the kth data in the small block, and chunknum is the number of elements in the small block. Then, we can get the following relation:

From this we can get the stretch factor estimate:

in is the stretch factor and i is the sequence of small blocks.

Step 16. Estimating the receiving channel SNR (Signal Noise Ratio, signal-to-noise ratio) through the channel physical layer.

Step 17. At the receiving end, calculate the data aggregation coefficient mean _L after estimating the stretching coefficient g _i of the multimedia data. First calculate the average magnitude of each patch

Where i is the serial number of the small block, and chunknum is the number of data in the small block.

Step 18. Then, the present invention defines the data aggregation coefficient mean _L as the average amplitude of the first 1-M small blocks divided by the average amplitude of the previous M+1-2M small blocks. That is, as shown in the following formula. The larger the data aggregation coefficient mean _L , the higher the energy aggregation degree of the decorrelation transformation, and the better the anti-noise ability of the block.

by average magnitude Calculate the data aggregation coefficient mean _L , divide the average amplitude of the first 1 to M small blocks by the average amplitude of the previous M+1 to 2M small blocks:

The larger the mean _L data, the stronger the compression capability of the decorrelation transformation.

Step 19. Estimate channel SNR (Signal Noise Ratio, signal-to-noise ratio) and calculation system PLR (Packet Loss Ratio, packet loss rate) at the physical layer level, where 0<PLR<1.

Step 20, at the receiving end, after calculating the data aggregation coefficient mean _L for the multimedia data, estimate the correction factor. Estimate the channel SNR and calculate the system packet loss rate PLR at the physical layer level. The correction factor is calculated by channel SNR, packet loss rate PLR and data aggregation coefficient mean _L. Due to the influence of channel noise, the average energy estimate of the large block will be too large, resulting in a small estimate of the stretch factor. At the same time, due to the impact of packet loss, some energy will be lost, resulting in a smaller average energy estimate of the large block, resulting in a larger stretch factor estimate.

The embodiment of the present invention introduces a correction factor to correct the estimation deviation caused by noise and packet loss. Used to correct for estimation bias caused by noise. Used to correct estimation bias caused by packet loss. There are options for different system estimates: no correction is used for systems with only low noise and low packet loss; only correction is used for systems with high noise Fix; for systems with high packet loss only use correction; use both corrections for systems with high noise and high packet loss at the same time.

Calculation of correction factors by fit-based correction model and Among them, C1, C2, C3, C4, C5, C6, C7, C8, C9, the parameters are obtained through a large number of experiments.

in is the noise repair factor, which mainly repairs the estimation error of the stretch factor caused by noise, Is the packet loss repair factor, which mainly repairs the estimation error of the stretch factor caused by packet loss. i is the serial number of the small block, MCHUNK is the number of small blocks, round(MCHUNK)<NUM<MCHUNK.

Step 21. After obtaining the correction factor, combine the stretch factor to perform data back-stretching. The formula is as follows:

in is the estimated data matrix after destretching, i is the small block sequence, and j is the jth data in the small block; is the data after de-whitening, i is the small block sequence, j is the jth data in the small block; P is the energy parameter known locally by the receiver.

Step 22. Perform anti-correlation transformation on the entire frame of the data matrix Y after the data de-stretching, and then perform de-translation of the b-depth on the data matrix after the inverse de-correlation transformation to obtain a large block of data The formula is as follows:

in is the estimated pixel matrix, ITr is the decorrelation transformation, is the data matrix after destretching the estimated data.

Step 23. Repeat steps 10-20. The obtained I large block pixel matrix X of M×M is spliced next to the estimated large block pixel matrix X to obtain the restored image and the multimedia data transmitted by the sending end.

Embodiment two

FPGA (Field-Programmable Gate Array, Field Programmable Gate Array) + DSP (digital signal processing, digital signal processing) implementation method: Due to the need for real-time video encoding and decoding (especially the receiving and decoding part), the FPGA+DSP implementation method becomes very necessary. The implementation of this scheme is divided into two parts: encoding and transmitting and receiving and decoding. The system block diagram is shown in Figure 8, which are respectively implemented on two Xilinx KC705FPGA development boards.

In this implementation scheme, the maximum module size is , and MN=1024 at the same time. The communication system adopts the OFDM system. The energy parameter P is determined with reference to the frame structure and the rated transmission power. The random sorting can be customized by using seeds. The decorrelation transformation is 2D-DCT transformation. The number of small blocks is N=1024. Data aggregation calculation is in progress M=20, the whitening transformation adopts the Walsh-Hadamard transformation, and the revised model parameters are as follows C1=1, C2=0.9, C3=0.8, C4=10, C5=1.36, C6=31.2, C6=0.115, C7=4.66, MCHUNK=1024, NUM=900. The specific implementation plan is as follows:

1.1) The schematic diagram of the hardware system structure of the encoding transmitting part of the embodiment of the present invention is as shown in Figure 7, the hardware part of the encoding transmitting end: by brand-new AnalogCast encoding transmitting baseband module, microprocessor, double rate DDR (DynamicRandom Access Memory, synchronous Dynamic Random Access Memory), UART (Universal Asynchronous Receiver/Transmitter, Universal Asynchronous Receiver Transmitter), IIC (Inter-Integrated Circuit, integrated circuit bus), RF transceiver chip. Among them, except the RF transceiver chip, other modules are integrated in the FPGA development board in the form of IP, and are mounted on the system bus through the AXI interface protocol (see Figure 7 for the hardware system diagram). The microprocessor is responsible for controlling the baseband module and peripherals; the new AnalogCast encoding and transmitting baseband module is responsible for source channel encoding of the data incoming from the AXI bus; the IIC is responsible for the communication and control of the FPGA and the RF transceiver chip; DDR is responsible for reading and writing data; The RF transceiver chip is responsible for the medium and high frequency processing of the data obtained by the new AnalogCast encoding and transmitting baseband processing and transmitting through the antenna; the UART is responsible for communicating with the PC.

The new AnalogCast encoding and transmitting baseband module: it is divided into three parts:

New AnalogCast source coding terminal: the main function is the new AnalogCast source coding. Use the HLS software that comes with the Xilinx Vivado FPGA design software to write fixed-point C codes according to the coding flow chart (see the source coding part in Figure 1), and then synthesize the required verilog codes through the HLS software.

Data Converter: The main function is to serve as the interface between the new AnalogCast source coder and the linear OFDM transmitter. The transmission matrix output by the new AnalogCast source coder is input into the linear OFDM transmitter according to the data transmission format of the linear OFDM transmitter. This module is written in verilog language.

Linear OFDM Transmitter: Hardware implementation of linear OFDM channel coding. Features include: analog modulation, adding pilots, IFFT, PAPR reduction, windowing, and frame shaping. Implemented using verilog coding.

RF radio frequency chip: use ADI's fmcomms1 chip, a high-speed transceiver chip, FPGA controls and communicates with the RF transceiver chip through IIC (configure radio frequency parameters, sampling rate, clock frequency of each module, etc.), and the new AnalogCast code transmits the baseband through FMC The interface provides data and reference clock for radio transmission. The fmcomms1 chip provides a 400Mhz to 4Ghz radio frequency range. The module is customizable to a wide range of frequencies via software without any hardware changes, provides synchronization for GPS or IEEE1588, and has MIMO configuration options.

Other IPs: Use Xilinx's own IP.

1.2) The software part burns the .elf file into the microprocessor, and controls the work of the baseband module and peripherals through the microprocessor, and performs real-time control through UART-PC. The software flow chart of the code transmitting part of the embodiment of the present invention is as shown in accompanying drawing 9:

Step 1: Drive the RF communication chip through the IIC bus to set the radio frequency, sampling frequency, clock of each sub-module, and make the antenna work.

Step 2: Drive the new AnalogCast baseband module through the AXI bus, and tell the new AnalogCast baseband module to start working.

Step 3: Read data from the AXI bus DDR and write it into the new AnalogCast baseband module.

Step 4: Repeat steps 2 to 3.

Step 5: If the UART inputs a termination command to MicroBlaze, the system will stop working.

1.3) The schematic diagram of the hardware system structure of the receiving decoding part of the embodiment of the present invention is as shown in Figure 8, the hardware part of receiving the decoding end: receiving and decoding baseband module by brand-new AnalogCast, microprocessor, double rate synchronous dynamic random access memory (DDR) , Universal Asynchronous Transceiver Transmitter (UART), IIC, and RF transceiver chip, among which other modules except the RF transceiver chip are integrated in the FPGA in IP mode, and mounted on the system bus through the AXI interface protocol (see the hardware system diagram Figure 8). The microprocessor is responsible for controlling the baseband module and peripherals; the new AnalogCast receiving and decoding baseband module is responsible for source channel decoding of the data sent by the RF chip; the IIC is responsible for the communication and control between the FPGA and the RF receiving chip; DDR is responsible for reading and writing data ; The RF receiving chip receives the source signal, and sends the obtained antenna signal to a series of medium and high frequency processing such as digital-to-analog conversion and carrier modulation, and sends it to the new AnalogCast receiving and decoding baseband module; UART is responsible for communicating with the PC.

The new AnalogCast receiving and decoding baseband module: divided into three parts:

1. Linear OFDM receiver: hardware implementation of linear OFDM channel decoding. Functions include: packet detection, carrier compensation, symbol synchronization, FFT, channel estimation, residual phase compensation, sample compensation, pilot removal, analog modulation. Implemented using verilog coding.

2. Data converter: The main function is to serve as the interface between the new AnalogCast source decoder and the linear OFDM receiver. The interface receives data packets from the linear OFDM receiver and forms a receiving matrix, and sends the receiving matrix to the new AnalogCast source decoder for data processing according to the data format requirements of the new AnalogCast source decoder. This module is written in verilog language.

3. The new AnalogCast source decoding end: the same implementation method as the new AnalogCast source encoding end, use the HLS software that comes with the Xilinx Vivado FPGA design software to write the fixed-point C Code, and then synthesized by HLS software to obtain the required verilog code.

RF radio frequency chip: use the same chip as the transmitter.

Other IPs: Use Xilinx's own IP.

1.4) The software part burns the .elf file into the microprocessor, controls the work of the baseband module and peripherals through the microprocessor, and performs real-time control through UART-PC. The software flowchart of the receiving and decoding part of the embodiment of the present invention is shown in Figure 10, including:

Step 1: Drive the RF communication chip through the IIC bus to set the radio frequency, sampling frequency, clock of each sub-module, and make the antenna work.

Step 2: Drive the new AnalogCast baseband module through the AXI bus and tell the baseband to start working.

Step 3: Observe whether the new AnalogCast baseband system works through the working signal of the AXI bus. If it does not work, it means that the data stored in the DDR is valid. And re-drive the new AnalogCast baseband module through the AXI bus, tell the baseband to start working, and repeat step 3.

Step 4: Stop working if the UART informs that there is a termination signal.

Embodiment Three

2): Implementation of software radio:

In this implementation scheme, the maximum module size is , and MN=1024 at the same time. The communication system adopts the OFDM system. The energy parameter P is determined with reference to the frame structure and the rated transmission power. The random sorting can be customized by using seeds. The decorrelation transformation is 2D-DCT transformation. The number of small blocks is N=1024. Data aggregation calculation is in progress M=20, the whitening transformation adopts the Walsh-Hadamard transformation, and the revised model parameters are as follows C1=1, C2=0.9, C3=0.8, C4=10, C5=1.36, C6=31.2, C6=0.115, C7=4.66, MCHUNK=1024, NUM=900. The system block diagram is shown in Figure 3, and the specific implementation plan is as follows:

Source channel coding and decoding is completed on the PC by software Matlab, and high-speed digital signal processing and RF transmission and reception are completed on the USRP N210 motherboard. The structure diagram of the software radio architecture of the embodiment of the present invention is shown in Figure 11, and the hardware part is composed of a PC, a software radio development board, and an RF radio frequency chip.

2.1) PC: the new AnalogCast source codec end and linear OFDM channel codec end through the Matlab software architecture.

2.2) Software radio development board: use a software radio platform called GUNRadio, USRP N210 motherboard, responsible for high-speed digital signal processing (digital-to-analog/analog-to-digital conversion, digital up-down conversion, etc.).

2.3) RF radio frequency chip: RFX2400 chip provides 2.4Ghz carrier modulation and demodulation function.

Software part: The software part is mainly the baseband encoding and decoding of the source channel, which is mainly realized by writing codes on the PC through Matlab software, including the encoding and decoding of the new AnalogCast system, channel encoding of encoded data, synchronization and decoding of received data. Among them, the new AnalogCast system encoding and data channel encoding belong to the encoding and transmitting process, and the synchronization of the new AnalogCast system decoding and receiving data and decoding belong to the encoding and transmitting process. See attached drawing 1 for the flow chart of encoding and transmitting, and see attached drawing 2 for the flow chart of receiving and decoding.

Embodiment Four

3): ASIC implementation method:

In this implementation scheme, the maximum module size is , and MN=1024 at the same time. The communication system adopts the OFDM system. The energy parameter P is determined with reference to the frame structure and the rated transmission power. The random sorting can be customized by using seeds. The decorrelation transformation is 2D-DCT transformation. The number of small blocks is N=1024. Data aggregation calculation is in progress M=20, the whitening transformation adopts the Walsh-Hadamard transformation, and the revised model parameters are as follows C1=1, C2=0.9, C3=0.8, C4=10, C5=1.36, C6=31.2, C6=0.115, C7=4.66, MCHUNK=1024, NUM=900. The system block diagram is shown in Figure 3, and the specific implementation plan is as follows:

The ASIC implementation flow chart of the new AnalogCast baseband system in the embodiment of the present invention is shown in Figure 12. The ASIC implementation is mainly aimed at the new AnalogCast receiving and decoding baseband module and the new AnalogCast encoding and transmitting baseband module. Model, and finally get the IP physical model (GDSII), test model, power consumption model, timing model. The construction of the RTL model is the same as the FPGA+DSP solution. The part of the new AnalogCast source codec uses HLS to produce verilog, and the rest is handwritten.

In summary, the embodiment of the present invention proposes a method for fully linearly transmitting multimedia data such as video/picture without error correction protection, which has the following advantages compared with the prior art:

1) Superior fairness.

2) Due to the fully linear transmission, there will be no "cliff effect" for metadata, and it has better mobility and longer transmission distance.

3) Due to the use of full linear modulation, there is no need to transmit metadata, which saves bandwidth.

4) Compared with the traditional multicast system and semi-linear system (Softcast, G-cast, HAD-cast, D-cast, etc.), digital side information is omitted, which greatly reduces the complexity of channel coding (eliminating Source compression and decompression, source protection and error correction, interleaving and deinterleaving, and channel protection and error correction calculations), greatly reducing the computational complexity.

5) In the channel environment with high packet loss, the image quality is greatly improved.

Those skilled in the art can understand that the accompanying drawing is only a schematic diagram of an embodiment, and the modules or processes in the accompanying drawing are not necessarily necessary for implementing the present invention.

It can be seen from the above description of the implementation manners that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in storage media, such as ROM/RAM, disk , CD, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments of the present invention.

Each embodiment in this specification is described in a progressive manner, the same and similar parts of each embodiment can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for relevant parts, refer to part of the description of the method embodiments. The device and system embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, It can be located in one place, or it can be distributed to multiple network elements. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. It can be understood and implemented by those skilled in the art without creative effort.

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present invention can easily think of changes or Replacement should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (7)

1. the multi-medium data method of multicasting that a kind of full linear is protected without error correction, which is characterized in that including：

In transmitting terminal, decorrelation transformation is carried out to multi-medium data：

The data that data after being converted to decorrelation divided based on block stretch；

Data block after being stretched to data carries out whitening transformation processing, to by whitening transformation treated data hand in frame Processing is knitted, by intra-frame interleaving treated data chunk at data packet；

After carrying out analog-modulated to the data packet, launch；

The data for the data after decorrelation transformation divided based on block are stretched, including：

In transmitting terminal, the data that the multi-medium data after being converted to decorrelation carries out the division of the block based on L shape stretch, and press Bulk is divided into the fritter of L shape according to the block dividing mode of setting, the length of each L blocks and it is wide be definite value, each fritter Average energy λ_i

Wherein λ_iIt is the average energy of each fritter, i represents the sequence of fritter, and chunknum is block number i.e. N, y_{I, k}It is in bulk Y K-th of data of i-th of fritter；

Stretching factor g is calculated by the average energy of each fritter_i：

Wherein g_iIt is the stretching factor of each fritter, P is energy stretching factor, and i represents the sequence of fritter, then will be in each fritter Data be multiplied by respective stretching factor, the data block u after being stretched_{I, j}:

u_{I, j}=g_i·y_{I, j}

It is described data are stretched after data block carry out whitening transformation processing, to by whitening transformation, treated that data carry out Intra-frame interleaving processing, by intra-frame interleaving treated data chunk at data packet, including：

In transmitting terminal, whitening transformation processing is carried out to the data block u after stretching, if H is whitening matrix, whitening transformation is as follows：

V=H.u.H^T

Wherein v is by whitening transformation treated data；

To carrying out intra-frame interleaving by whitening transformation treated data, the data in V are divided according to respective positions Data in frame are divided by class according to position:Four kinds of A pixels, B pixels, C pixels and D pixels pixels, each pixel is carried out Random rearrangement, by the various pixel reorganizations after random rearrangement at data block；

For each frame of the multi-medium data, the decorrelation transformation is executed respectively, data stretch, whitening transformation processing With intra-frame interleaving processing, by the completion intra-frame interleaving of all frames treated data chunk at GoP, then by a layer of GoP It is packaged into a data packet；Described is divided into the data in frame according to position:Four kinds of A pixels, B pixels, C pixels and D pixels Pixel, including：

The pixel of first position of the first row of whitening transformation treated bulk is determined as A pixels, next pixel is B pictures Element, next one pixel be C pixels, next one pixel be D pixels, and so on complete a line pixel definition；

It is started with D pixels in first pixel of the second row, next pixel is A pixels, next one pixel is B pixels, then under One pixel is C pixels, and so on complete the pixel definitions of two rows；

It is started with C pixels in first pixel of the third line, next pixel is D pixels, next one pixel is A pixels, then under One pixel is B pixels, and so on complete the pixel definitions of three rows；

It is started with B pixels in first pixel of fourth line, next pixel is C pixels, next one pixel is D pixels, then under One pixel is A pixels, and so on complete the pixel definitions of four rows；

After, the dividing condition and above-mentioned the first row of the location of pixels of the data in the frame of every four row, the second row, the third line and The dividing condition of the location of pixels of fourth line is identical.

2. the multi-medium data method of multicasting that full linear according to claim 1 is protected without error correction, which is characterized in that described In transmitting terminal, decorrelation transformation is carried out to multi-medium data, including：

Each frame of multi-medium data to be transmitted is independently separated, each frame image is then isolated into the big of MN × MN Block X, the bulk X translations 2 to each frame image^b-1, wherein b is image sampling depth, and carrying out whole frame to the bulk X after translation goes Correlation converts：

Y=Tr (x-2^b-1)

Y is the bulk after being converted for decorrelation, and Tr is decorrelation transformation, and x is the picture element matrix of bulk.

3. the multi-medium data method of multicasting that full linear according to claim 1 is protected without error correction, which is characterized in that described To the data packet carry out analog-modulated after, launch, including：

Data in the data packet are encoded into data frame by physical layer, and analog-modulated is carried out to the data frame, it will A real number in the data frame is mapped to I vectors, another real number is mapped to Q vector, by the data frame after analog-modulated Launch.

4. the multi-medium data method of multicasting that full linear according to claim 3 is protected without error correction, which is characterized in that described Method further include：

In receiving terminal, the entire packet received is formed into GoP, then GoP is resolved into the corresponding data block of I data frame；

Each corresponding data block of data frame includes various pixels after resetting, and is reset according to preset pixel Rule, the pixel after various types are upset is restored, then resets in respective positions, obtains the corresponding albefaction of each data frame and becomes Data after changing；

Whitening transformation is carried out to the data after the corresponding whitening transformation of each data frame to handle, to going whitening transformation to handle Data afterwards carry out drawing coefficient estimation；

Whitening transformation is removed to described treated that data carry out data stretch processing, according to the drawing coefficient estimated to going to count Anti- decorrelation transformation is carried out according to the data after stretch processing, obtains the multi-medium data that the transmitting terminal is sent.

5. the multi-medium data method of multicasting that full linear according to claim 3 is protected without error correction, which is characterized in that described Whitening transformation carried out to the data after the corresponding whitening transformation of each data frame handle, after going whitening transformation to handle Data carry out drawing coefficient estimation include：

In receiving terminal drawing coefficient g is carried out after carrying out whitening transformation processing to multi-medium data_iEstimation calculates and receives number According to the small average energy in the block of L-type：

WhereinIt is the average energy for receiving data L-type fritter, i is fritter sequence number；It is the data received, i is fritter sequence Row number, k are k-th of data in fritter, and chunknum-s is first prime number in fritter；

BecauseSo having：

WhereinIt is the average energy of small block data after reverse-drawing, i is fritter sequence,It is the data after reverse-drawing, k is small K-th of data in block, chunknum-s are first prime numbers in fritter, obtain following relationship：

Drawing coefficientEstimation formulas it is as follows：

WhereinIt is drawing coefficient, i is fritter sequence.

6. the multi-medium data method of multicasting that full linear according to claim 5 is protected without error correction, which is characterized in that described The drawing coefficient that estimates of basis remove whitening transformation treated that data carry out data stretch processing to described, including：

In receiving terminal, multi-medium data is carried out in drawing coefficient g_iAfter estimation, carry out to data convergence factor mean_LIt is counted It calculates, calculates the average amplitude of each fritter

The data convergence factor mean_LThe average width of former M+1~2M fritter is removed for the average amplitude of preceding 1~M fritter Degree, data convergence factor mean_LCalculation formula it is as follows：

In physical layer level estimation channel SNR and computing system packet loss PLR, by channel SNR, packet loss PLR is poly- with data Collect Coefficient m ean_LCalculate modifying factorWith Calculating error caused by for correcting noise,Cause for correcting packet loss Calculating error；

Following formula is to calculate modifying factorWithCorrection model:

WhereinIt is the estimation mistake that the noise repair factor mainly repairs the stretching factor caused by noise,Packet loss recovery because The sub main estimation mistake for repairing the stretching factor caused by packet loss, i is fritter sequence number, and MCHUNK is small block number, round (MCHUNK)<NUM<MCHUNK, wherein C1, C2, C3, C4, C5, C6, C7, C8, C9, parameter are obtained by experiment；

Obtain modifying factorWithLater, reverse-drawing is carried out in conjunction with stretching factor, formula is as follows：

WhereinIt is the data after the decorrelation of estimation changes, i is fritter sequence, and j is small j-th of data in the block；It is It is fritter sequence to go the data after albefaction, i, and j is small j-th of data in the block；P is that energy parameter receiver is locally known.

7. the multi-medium data method of multicasting that full linear according to claim 6 is protected without error correction, which is characterized in that described To going the data after data stretch processing to carry out anti-decorrelation transformation, obtain the multi-medium data that the transmitting terminal is sent, Including：

In receiving terminal, anti-decorrelation transformation is carried out to the whole frames of data matrix Y after data reverse-drawing, then anti-decorrelation is become Data matrix after changing carries out the anti-translation of b depth, obtains chunk dataFormula is as follows：

Wherein x is estimation picture element matrix,

Above-mentioned processing procedure is repeated, by the bulk picture element matrix X of I obtained M × M, by all bulk picture element matrix X Spliced, obtains the multi-medium data of transmitting terminal transmission.