Joint source channel coding voice wireless transmission method based on adjusting auto-encoder

By introducing the adjustment network and deregulation network of the signal-to-noise ratio adjustment module in deep learning combined source channel coding, the problem of degradation of deep learning method performance under channel conditions is solved, and the robustness and adaptability of voice transmission are achieved, and the The cliff effect and saturation effect in traditional methods are obtained.

CN119030662AActive Publication Date: 2024-11-26UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411077221.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-07
Publication Date
2024-11-26
Estimated Expiration
2044-08-07

AI Technical Summary

Technical Problem

The existing deep learning joint source channel encoding method has deteriorated performance when channel conditions do not match, and the image and voice data types are different, resulting in the inability to effectively adapt to different channel conditions in voice transmission tasks.

Method used

The adjustment network and deregulation network with a signal-to-noise ratio adjustment module are adopted to adjust the characteristics in the encoder and decoder through the channel signal-to-noise ratio information, and resist the noise of different channels and provide robust transmission performance.

Benefits of technology

Provides robust transmission performance in a larger channel signal-to-noise ratio range, overcomes cliff effect and saturation effect, and improves the adaptability and efficiency of voice wireless transmission.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119030662A_ABST
    Figure CN119030662A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of voice processing and transmission, and particularly provides a joint source channel coding voice wireless transmission method based on an adjusting auto-encoder, which is used for enabling a joint source channel model with a single depth to adapt to different channel conditions. The self-encoder comprises a transmitting end, a channel module and a receiving end, the transmitting end comprises a joint encoder and an adjusting network, the channel module comprises a signal average transmitting power limiter and a wireless channel, and the receiving end comprises a joint decoder and a deadjusting network, the adjusting network and the de-adjusting network only generate an adjusting factor from a signal-to-noise ratio value of a wireless channel, and the voice feature vectors in the joint encoder and the joint decoder are scaled to adjust the encoding and decoding processes, so that the capability of resisting channel noise with different intensities is obtained; on the basis of the method, the robust voice wireless transmission performance can be provided within the given channel signal-to-noise ratio range, and the method is superior to an existing deep joint information source channel method.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of speech processing and transmission, and specifically provides a joint source channel coding speech wireless transmission method based on adjusting an autoencoder. Background Art

[0002] Traditional communication systems originated from Shannon's separation theorem, which adopts a separate design in which source coding, channel coding, modulation and other modules are independently executed; although this separation design performs well and dominates in current communication networks, it still has two disadvantages: cliff effect and saturation effect; the cliff effect refers to the complete failure of transmission when the channel quality is lower than the channel coding error correction capability, and the saturation effect refers to the inability to further improve performance when the channel quality is better than the target channel quality.

[0003] In recent years, a joint source channel coding (JSCC) paradigm based on deep learning (DL) has emerged, in which the sender converts the input information into channel symbols and the receiver recovers the original information from the received information; by combining the JSCC encoder, channel, and JSCC decoder into an autoencoder and training it in an end-to-end distortion-loss manner, the encoder-decoder group can learn to extract important semantic features, overcome channel noise, and achieve the highest possible fidelity in the receiver. The goal of deep JSCC is to convey the basic semantics under a specified transmission task, which is closely related to the recent hot concept of semantic communication; many recent studies have shown that deep JSCC outperforms separate designs for different data type transmission tasks (such as images, videos, and text).

[0004] Although deep JSCC can learn the state that best suits the current channel conditions through end-to-end training, the mismatch between channel conditions in training and deployment will lead to severe performance degradation; deploying multiple networks suitable for multiple signal-to-noise ratios can alleviate this problem, but will greatly increase the complexity of the transceiver. In the image transmission task, the document "Xu J, Ai B, Chen W, et al. Wireless Image Transmission Using Deep Source Channel Coding With Attention Modules[J]. 2020. DOI: 10.48550 / arXiv.2012.00533." proposed an ADJSCC (Attention Deep Learning Based JointSource-Channel Coding) network structure to overcome this shortcoming. ADJSCC is a deep learning wireless image transmission JSCC method based on attention modules. This method can successfully operate in a wider signal-to-noise ratio range during transmission. A series of studies based on ADJSCC have verified its effectiveness in wireless image transmission; however, due to the different data types of images and speech, the ADJSCC network structure is not suitable for speech transmission tasks; therefore, the present invention provides a joint source channel coding speech wireless transmission method based on an adjusted autoencoder. Summary of the invention

[0005] The purpose of the present invention is to provide a joint source channel coding speech wireless transmission method based on adjusting the autoencoder, so that a single deep JSCC model can adapt to different channel conditions; specifically, the present invention proposes an adjustment network and a de-adjustment network with a signal-to-noise ratio adjustment module, uses the signal-to-noise ratio information of the channel to generate an adjustment factor, and adjusts the features in the encoder and decoder to control the importance of different channels, thereby obtaining the ability to resist channel noise of different intensities; based on this, the present invention can provide robust transmission performance within a larger channel signal-to-noise ratio range, which is better than the existing deep JSCC method.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is:

[0007] A joint source channel coding voice wireless transmission method based on an adjustment autoencoder, characterized in that the adjustment autoencoder comprises: a transmitting end, a channel module and a receiving end; the transmitting end comprises: a joint encoder and an adjustment network; the channel module comprises: a signal average transmission power limiter and a wireless channel; the receiving end comprises: a joint decoder and a demodulation network; the joint encoder is composed of four modules A and one module B in cascade; the joint decoder is composed of one module B, four modules A and another module B in cascade; the adjustment network and the demodulation network adopt the same structure, respectively including four signal-to-noise ratio adjustment modules, and the signal-to-noise ratio adjustment modules are connected to the module A in a one-to-one correspondence; the joint source channel coding voice wireless transmission method comprises the following steps:

[0008] Step S1, inputting the signal-to-noise ratio information of the wireless channel into the adjustment network and the de-adjustment network respectively, and the adjustment network and the de-adjustment network output the adjustment factor Q and the adjustment factor P, wherein Q = {q1, ..., q4}, P = {p1, ..., p4}, q1, ..., q4 respectively represent the adjustment factors of the four modules A in the joint encoder, and p1, ..., p4 respectively represent the adjustment factors of the four modules A in the joint decoder;

[0009] Step S2, at the transmitting end, the original speech signal X is input into the normalization layer to obtain a speech signal vector X1; the speech signal X1 is resized into a two-dimensional matrix to obtain a speech signal matrix X2; the speech signal matrix X2 is encoded by a joint encoder, and during the encoding process, the four modules A are adjusted by adjustment factors q1, q2, q3, and q4 respectively to obtain a speech feature vector X3 for channel transmission;

[0010] Step S3: In the channel module, the signal average transmission power limiter limits the average power of the speech feature vector X3 to obtain a transmission signal R, and transmits it to the receiving end through a wireless channel;

[0011] Step S4, at the receiving end, receiving the transmission signal from the channel module to obtain the speech feature vector Y1; the speech feature vector Y1 is decoded by the joint decoder, and during the decoding process, the four modules A are adjusted respectively by the adjustment factors p1, p2, p3, and p4 to obtain the reconstructed speech signal matrix Y2; the reconstructed speech signal matrix Y2 is resized into a one-dimensional vector to obtain the reconstructed speech signal vector Y3; the reconstructed speech signal Y3 is input into the denormalization layer to obtain the reconstructed speech signal Y.

[0012] Furthermore, the internal data processing flow of the module A is as follows: the input vector of module A is represented as J, the input vector J is respectively input into four parallel two-dimensional convolutional layers to obtain feature vectors K1, K2, K3 and K4, the feature vectors K1, K2, K3 and K4 are feature concatenated to obtain feature vector L, the feature vector L is passed through the cascaded two-dimensional convolutional layer and the normalization layer to obtain the feature vector Z, the feature vector Z is passed through the cascaded global pooling layer, the fully connected layer, the ReLU activation function layer, the fully connected layer, and the Sigmoid activation function layer to obtain the feature vector M, the feature vector M is dimensionally expanded according to the size of the feature vector Z and then multiplied with the feature vector Z, the multiplication result is then added to the input vector J, and the addition result is finally dot-multiplied with the adjustment factor output by the signal-to-noise ratio adjustment module to obtain the output of module A.

[0013] Furthermore, the internal data processing flow of the module B is as follows: the input vector of the module B is represented as N, and the input vector N is passed through a cascaded two-dimensional convolutional layer and a normalization layer to obtain the output of the module B.

[0014] Furthermore, the internal data processing flow of the signal-to-noise ratio adjustment module is as follows: the input of the signal-to-noise ratio adjustment module is the signal-to-noise ratio information of the wireless channel, and the signal-to-noise ratio information is sequentially passed through a cascaded fully connected layer, a ReLU activation function layer, a fully connected layer, and a Sigmoid activation function layer to obtain a corresponding adjustment factor as the output of the signal-to-noise ratio adjustment module.

[0015] Furthermore, the regulated autoencoder is trained end-to-end, and the loss function of the training process is:

[0016]

[0017] Among them, L is the loss function, N is the number of speech signals in the training set, n is the number of sampling points of each speech signal in the training set, and s i,j represents the jth sampling point of the i-th original speech signal, s i ' ,j Represents the jth sampling point of the i-th reconstructed speech signal.

[0018] Based on the above technical solution, the beneficial effects of the present invention are:

[0019] The invention provides a joint source channel coding voice wireless transmission method based on an adjustment autoencoder, wherein the autoencoder comprises: a transmitting end, a channel module, and a receiving end, wherein the transmitting end comprises: an encoder and an adjustment network; the channel module comprises: a signal average transmission power limiter and a wireless channel; the receiving end comprises: a joint decoder and a demodulation network, and both the adjustment network and the demodulation network comprise a fully connected neural network; the invention adopts joint source channel coding, which has better effect than a traditional method of independently designing coding, and overcomes cliff effect and saturation effect; at the transmitting end, an original voice signal is extracted with a joint encoder, and the feature information enters the channel module, and after the transmission signal power is constrained by a signal average transmission power limiter, the feature information is transmitted to the receiving end through a wireless channel, and the receiving end reconstructs the original voice signal from the received signal through a joint decoder; and in the process of encoding and decoding, the signal-to-noise ratio information of the current channel is introduced, and under the adjustment of the adjustment network and the demodulation network, a more balanced resource allocation can be obtained between source coding and channel coding in voice transmission, so that the voice transmission task is more adapted to the current wireless channel conditions.

[0020] Furthermore, the autoencoder provided by the present invention adds a regulation network and a deregulation network, takes the signal-to-noise ratio information of the wireless channel as input, and adaptively generates a regulation factor, so that the model is more robust when facing wireless channels in different states; at the same time, the present invention uses a mean square error penalty function to end-to-end train the reconstruction and regulation network of the speech signal, effectively reducing the switching overhead of the signal-to-noise ratio scaling factor under different channel signal-to-noise ratio conditions, which is easier to deploy in practice.

[0021] In summary, the present invention is based on a deep learning method, and proposes a voice compression transmission method based on joint source channel coding of an autoencoder, which provides efficient voice transmission services; the signal-to-noise ratio information of the wireless channel is introduced in the encoding and decoding process, so that the transmission task is more adapted to the current channel state; the signal-to-noise ratio adjustment factor under the current channel state is adaptively generated by the adjustment network and the de-adjustment network, so as to further reduce the switching overhead of the adjustment factor when facing different channels, and achieve the effect of adaptive signal-to-noise ratio; in particular, the adaptive signal-to-noise ratio voice wireless transmission method of the present invention only needs to be trained once for the signal-to-noise ratio within a predetermined range, so the deployment cost is low, and the adaptive ability to face different channel transmission conditions is strong. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the structure of a joint source-channel coded voice wireless transmission system based on an adjusted autoencoder in Embodiment 1 of the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of the signal-to-noise ratio adjustment module of Example 1 of the present invention.

[0024] Figure 3 This is a schematic diagram of the structure of module A in Example 1 of the present invention.

[0025] Figure 4 Schematic diagram of the structure of the voice wireless transmission system based on the autoencoder in Comparative Example 1 of the present invention.

[0026] Figure 5 It is a comparison chart of mean square error (MSE) between Example 1 of the present invention and comparative example 1.

[0027] Figure 6 It is a comparison diagram of the signal distortion ratio SDR between Example 1 of the present invention and Comparative Example 1.

[0028] Figure 7 It is a comparison chart of short-time objective intelligibility STOI between Example 1 of the present invention and Comparative Example 1.

[0029] Figure 8 It is a comparison chart of the mean square error (MSE) between Example 2 of the present invention and Comparative Example 2.

[0030] Fig. 9 It is a comparison diagram of the signal distortion ratio SDR of Example 2 of the present invention and Comparative Example 2.

[0031] Fig.10 It is a comparison chart of short-time objective intelligibility STOI between Example 2 of the present invention and Comparative Example 2. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and beneficial effects of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0033] Example 1

[0034] This embodiment provides a joint source channel coding voice wireless transmission method based on an adjustable autoencoder. The structure of the adjustable autoencoder is as follows: Figure 1 As shown, it includes: a transmitting end, a channel module and a receiving end; the transmitting end includes: a joint encoder and an adjustment network; the channel module includes: a signal average transmission power limiter and a wireless channel; the receiving end includes: a joint decoder and a demodulation network; the joint encoder is composed of four modules A and one module B in cascade; the joint decoder is composed of one module B, four modules A and another module B in cascade; the adjustment network and the demodulation network adopt the same structure, respectively including four signal-to-noise ratio adjustment modules, and the signal-to-noise ratio adjustment modules are connected to the module A in a one-to-one correspondence; more specifically:

[0035] The structure of module A is as follows Figure 2As shown in the figure, its internal data processing flow is as follows: the input vector of module A is represented as J, and the input vector J is respectively input into four parallel two-dimensional convolutional layers to obtain feature vectors K1, K2, K3 and K4, and the feature vectors K1, K2, K3 and K4 are feature concatenated to obtain feature vector L, and the feature vector L is passed through the cascaded two-dimensional convolutional layer and the normalization layer to obtain the feature vector Z, and the feature vector Z is passed through the cascaded global pooling layer, the fully connected layer, the ReLU activation function layer, the fully connected layer, and the Sigmoid activation function layer to obtain the feature vector M, and the feature vector M is multiplied with the feature vector Z after the dimension expansion operation according to the size of the feature vector Z, and the multiplication result is added to the input vector J, and the addition result is finally multiplied with the adjustment factor output by the signal-to-noise ratio adjustment module to obtain the output of module A;

[0036] The module B includes a cascaded two-dimensional convolution layer and a batch normalization layer, and its internal data processing flow is as follows: the input vector of the module B is represented as N, and the input vector N is passed through the cascaded two-dimensional convolution layer and the normalization layer to obtain the output of the module B;

[0037] The structure of the signal-to-noise ratio adjustment module is as follows: Figure 3 As shown in the figure, it includes a cascaded fully connected layer, a ReLU activation function layer, a fully connected layer and a Sigmoid activation function layer; its internal data processing flow is as follows: the input of the signal-to-noise ratio adjustment module is the signal-to-noise ratio information of the wireless channel, and the signal-to-noise ratio information sequentially passes through the cascaded fully connected layer, the ReLU activation function layer, the fully connected layer and the Sigmoid activation function layer to obtain the corresponding adjustment factor as the output of the signal-to-noise ratio adjustment module.

[0038] The regulated autoencoder is trained end-to-end. The training process uses the mean square error (MSE) between the original speech signal and the reconstructed speech signal as the penalty function L to update the parameters. The MSE calculation formula is:

[0039]

[0040] Among them, L is the loss function, N is the number of speech signals in the training set, n is the number of sampling points of each speech signal in the training set, and s i represents the i-th original speech signal, s i ′ represents the i-th reconstructed speech signal, s i,j represents the jth sampling point of the i-th original speech signal, s i ' ,j Represents the jth sampling point of the i-th reconstructed speech signal.

[0041] On the basis of the above-mentioned adjustment of the autoencoder, the joint source channel coding voice wireless transmission method based on the adjustment of the autoencoder comprises the following steps:

[0042] Step S1, inputting the signal-to-noise ratio information of the wireless channel into the adjustment network of the transmitting end and the demodulation network of the receiving end respectively, and outputting the adjustment factor Q and the adjustment factor P containing the signal-to-noise ratio information through each signal-to-noise ratio adjustment module in the adjustment network and the demodulation network, wherein Q={q1, ..., q4}, P={p1, ..., p4}, q1, ..., q4 respectively represent the adjustment factors of the four modules A in the joint encoder of the transmitting end, and p1, ..., p4 respectively represent the adjustment factors of the four modules A in the joint decoder of the receiving end;

[0043] Step S2, at the transmitting end, the original speech signal X is input into the normalization layer to limit the speech signal within the range of [0,1], and a speech signal vector X1 is obtained; the speech signal X1 is resized to transform the one-dimensional vector into a two-dimensional matrix, and a speech signal matrix X2 is obtained; the speech signal matrix X2 is encoded by a joint encoder, and the four modules A are adjusted by adjustment factors q1, q2, q3, and q4 respectively during the encoding process, and a speech feature vector X3 for channel transmission is obtained;

[0044] Step S3: In the channel module, the signal average transmission power limiter limits the average power of the speech feature vector X3 to obtain a transmission signal R, and transmits it to the receiving end through a wireless channel (additive white Gaussian noise wireless channel);

[0045] Step S4, at the receiving end, receiving the transmission signal from the channel module, and obtaining the speech feature vector Y1; the speech feature vector Y1 is decoded by the joint decoder, and the four modules A are adjusted respectively by the adjustment factors p1, p2, p3, and p4 during the decoding process to obtain the reconstructed speech signal matrix Y2; the reconstructed speech signal matrix Y2 is resized to convert the two-dimensional matrix into a one-dimensional vector, and the reconstructed speech signal vector Y3 is obtained; the reconstructed speech signal Y3 is input into the denormalization layer to obtain the reconstructed speech signal Y.

[0046] Example 2

[0047] This embodiment provides a joint source channel coding voice wireless transmission method based on adjusting the autoencoder, which is different from the embodiment 1 in that the wireless channel of the channel module adopts a Rayleigh fading channel.

[0048] Comparative Example 1

[0049] This comparative example provides a joint source channel coding voice wireless transmission method based on an autoencoder, such as Figure 4 As shown, the difference between this embodiment and embodiment 1 is that the transmitting end and the receiving end do not include a regulation network and a deregulation network.

[0050] Comparative Example 2

[0051] This comparative example provides a joint source-channel coded voice wireless transmission method based on an autoencoder, which differs from comparative example 1 in that the wireless channel of the channel module adopts a Rayleigh fading channel.

[0052] For the wireless voice transmission method, the present invention uses a public voice data set from the database of the University of Edinburgh, which is used to store research data generated by researchers of the university related to existing or upcoming publications. The voice data set provides more than 10,000 clear and noisy voice training files. The present invention pre-processes the data set and downsamples it to 8kHz. The batch size of all embodiments and comparative examples in the present invention is 32, the target size of the voice data size conversion operation at the transmitting end is 128*128, the mean square error between the reconstructed voice data and the original voice data is used as a penalty function to update the model parameters, and the RMSprop optimizer with a learning rate of 0.001 is used.

[0053] The present invention uses two common speech quality evaluation indicators, Signal Distortion Ratio (SDR) and Short-Time Objective Intelligibility (STOI), to measure the quality of the reconstructed speech signal at the receiving end; SDR can be calculated by the following formula:

[0054]

[0055] Where s represents the original speech signal, s′ represents the reconstructed speech signal, and the higher the SDR value is, the smaller the difference between the reconstructed speech signal and the original speech signal is, that is, the better the performance of the speech wireless transmission system is.

[0056] STOI aims to measure the correlation between clarity and intelligibility. It is a quality assessment method for speech signals. It evaluates the quality of speech signals by comparing the spectral correlation between the original speech and the reconstructed speech. The calculation steps are as follows: First, the original speech signal and the reconstructed speech signal are divided into short time segments, and short-time Fourier transform (STFT) is performed to convert the signal segments into spectral segments; then the spectral similarity between the original speech signal and the reconstructed speech signal is calculated; finally, the similarity of all spectral segments is averaged to obtain the STOI value between the two speech signals. The STOI value is between 0 and 1. The closer the value is to 1, the better the signal quality.

[0057] Further, such as Figure 5 , Figure 6 and Figure 7The figures show the MSE, SDR and STOI comparison diagrams of Example 1 and Comparative Example 1 when the above-mentioned public data set is used in an additive white Gaussian noise wireless channel, respectively. The signal-to-noise ratio of different additive white Gaussian noise wireless channel conditions is used as the horizontal coordinate, and the evaluation index between the reconstructed speech signal and the original speech signal is used as the vertical coordinate. It can be seen from the figure that the MSE of Example 1 is smaller in the entire test signal-to-noise ratio range, and the SDR and STOI of Example 1 are larger in the entire test signal-to-noise ratio range. The performance of Comparative Example 1 is poor when the training signal-to-noise ratio does not match the test signal-to-noise ratio, and the performance is still worse than that of Example 1 even at the training signal-to-noise ratio in the dynamic range. This shows that under the action of the adjustment module, the voice wireless transmission system is more robust when facing additive white Gaussian noise wireless channels in different states, which proves the effectiveness of the method proposed in the present invention.

[0058] like Figure 8 , Fig. 9 and Fig.10 The figures show the MSE, SDR and STOI comparison diagrams of Example 2 and Comparative Example 2 when using the above-mentioned public data set under Rayleigh fading channel, respectively. The signal-to-noise ratio of different Rayleigh fading channel conditions is used as the horizontal coordinate, and the evaluation index between the reconstructed speech signal and the original speech signal is used as the vertical coordinate. It can be seen from the figure that the conclusion is the same as that under the additive white Gaussian noise wireless channel. Under the Rayleigh fading channel, the adjustment module proposed in the present invention can also enhance the robustness of the system, further verifying the effectiveness of the proposed method.

[0059] In summary, the present invention provides a voice wireless transmission method based on joint source channel coding of the regulated autoencoder, providing efficient voice transmission services; introducing the signal-to-noise ratio information of the wireless channel during the encoding and decoding process, and adjusting the voice feature vectors in the joint encoder and the joint decoder to make the voice transmission task more adaptable to the current channel conditions; using the regulation network and the deregulation network to adaptively generate the signal-to-noise ratio adjustment factor under the current channel state, further reducing the switching overhead of the signal-to-noise ratio adjustment factor when facing different channels, and achieving the effect of adaptive signal-to-noise ratio. In particular, the adaptive signal-to-noise ratio voice wireless transmission method of the present invention only needs to be trained once for the signal-to-noise ratio within a predetermined range, has low deployment costs, and has strong adaptability to different channel transmission conditions.

[0060] The above description is only a specific implementation mode of the present invention. Any feature disclosed in this specification, unless otherwise stated, can be replaced by other alternative features that are equivalent or have similar purposes; all the disclosed features, or all the steps in the methods or processes, except for mutually exclusive features and / or steps, can be combined in any way.

Claims

1. A joint source channel coding speech wireless transmission method based on adjusting the autoencoder, It is characterized in that The regulating autoencoder comprises: a transmitting end, a channel module and a receiving end; the transmitting end comprises: a joint encoder and a regulating network; the channel module comprises: a signal average transmission power limiter and a wireless channel; the receiving end comprises: a joint decoder and a demodulation network; the joint encoder is composed of four modules A and one module B in cascade; the joint decoder is composed of one module B, four modules A and another module B in cascade; the regulating network and the demodulation network adopt the same structure, respectively comprising four signal-to-noise ratio regulating modules, and the signal-to-noise ratio regulating modules are connected to the modules A in a one-to-one correspondence; the joint source channel coding voice wireless transmission method comprises the following steps: Step S1: Input the signal-to-noise ratio information of the wireless channel to the adjustment network and the de-adjustment network respectively, and the adjustment network and the de-adjustment network output the adjustment factor Q and the adjustment factor P, where Q={q 1 ,…,q 4 }, P = {p 1 ,…,p 4 }, q 1 ,…,q 4 They represent the adjustment factors of the four modules A in the joint encoder, p 1 ,…,p 4 They represent the adjustment factors of the four modules A in the joint decoder respectively; Step S2: At the transmitting end, the original speech signal X is input into the normalization layer to obtain the speech signal vector X 1 ; For speech signal X 1 Transform the size into a two-dimensional matrix and obtain the speech signal matrix X 2 ; Speech signal matrix X 2 After the joint encoder is used for encoding, the adjustment factor q 1 ,q 2 ,q 3 ,q 4 Adjust the four modules A respectively to obtain the speech feature vector X for channel transmission 3 ; Step S3: In the channel module, the signal average transmission power limiter performs a power control on the speech feature vector X 3 The average power is limited to obtain the transmission signal R, and transmitted to the receiving end through the wireless channel; Step S4: At the receiving end, receive the transmission signal from the channel module and obtain the speech feature vector Y 1 ; Speech feature vector Y 1 After decoding by the joint decoder, the adjustment factor p 1 、p 2 、p 3 、p 4 Adjust the four modules A respectively to obtain the reconstructed speech signal matrix Y 2 ; Reconstruct the speech signal matrix Y 2 Transform the size into a one-dimensional vector to obtain the reconstructed speech signal vector Y 3 ; Reconstruct the speech signal Y 3 Input to the denormalization layer to obtain the reconstructed speech signal Y.

2. According to claim 1, the joint source channel coding voice wireless transmission method based on adjusting the autoencoder, It is characterized in that The internal data processing flow of module A is as follows: the input vector of module A is represented as J, and the input vector J is input into four parallel two-dimensional convolutional layers to obtain the feature vector K 1 , K 2 , K 3 With K 4 , the feature vector K 1 , K 2 , K 3 With K 4 Feature concatenation is performed to obtain feature vector L. Feature vector L is passed through a cascade of two-dimensional convolutional layers and normalization layers to obtain feature vector Z. Feature vector Z is passed through a cascade of global pooling layers, fully connected layers, ReLU activation function layers, fully connected layers, and Sigmoid activation function layers to obtain feature vector M. Feature vector M is dimensionally expanded according to the size of feature vector Z and then multiplied by feature vector Z. The multiplication result is then added to the input vector J. The addition result is finally dot-multiplied with the adjustment factor output by the signal-to-noise ratio adjustment module to obtain the output of module A.

3. According to the joint source channel coding voice wireless transmission method based on adjusting the autoencoder according to claim 1, It is characterized in that The internal data processing flow of the module B is as follows: the input vector of the module B is represented as N, and the input vector N is passed through a cascaded two-dimensional convolutional layer and a normalization layer to obtain the output of the module B.

4. According to claim 1, the joint source channel coding voice wireless transmission method based on adjusting the autoencoder, It is characterized in that The internal data processing flow of the signal-to-noise ratio adjustment module is as follows: the input of the signal-to-noise ratio adjustment module is the signal-to-noise ratio information of the wireless channel, and the signal-to-noise ratio information is sequentially passed through a cascaded fully connected layer, a ReLU activation function layer, a fully connected layer, and a Sigmoid activation function layer to obtain a corresponding adjustment factor as the output of the signal-to-noise ratio adjustment module.

5. According to claim 1, the joint source channel coding voice wireless transmission method based on adjusting the autoencoder, It is characterized in that The regulated autoencoder is trained end-to-end, and the loss function of the training process is: Where L is the loss function, N is the number of speech signals in the training set, n is the number of sampling points of each speech signal in the training set, and s i,j represents the jth sampling point of the i-th original speech signal, s i ' ,j Represents the jth sampling point of the i-th reconstructed speech signal.

Citation Information

Patent Citations

  • Communication method based on deep learning and joint source channel coding optimization

    CN117939150A

  • Joint source channel coding image wireless transmission method based on adjusting auto-encoder

    CN118611824A