Signal modulation and demodulation method, device and storage medium

Abstract

The application discloses a signal modulation and signal demodulation method, device and storage medium, and relates to the field of communication.The method comprises the following steps: a signal sending end extracts semantic features of source data, obtains a semantic feature vector, quantizes the semantic feature vector, obtains a bit sequence, modulates the bit sequence according to a pre-allocated feature domain, obtains a modulation signal, and sends the modulation signal to a signal receiving end through an interference channel, so that the signal receiving end demodulates the modulation signal; the signal receiving end demodulates the received modulation signal according to the pre-allocated feature domain, obtains a bit sequence, dequantizes the bit sequence, obtains a semantic feature vector, decodes the semantic feature vector, and obtains target semantic information, thereby effectively solving the problems that signals are easily interfered and counterfeited in the communication process, leading to poor signal transmission security and low transmission efficiency, and greatly improving the communication security and the communication transmission efficiency.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a signal modulation and demodulation method, device, and storage medium. Background Technology

[0002] In the field of communications, communication security is a critical issue. Currently, during communication, signal transmission may be subject to interference and forgery, leading to poor signal transmission security and low transmission efficiency.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this invention is to provide a signal modulation and demodulation method, device, and storage medium, aiming to solve the technical problems of poor transmission security and low transmission efficiency caused by the ease with which existing technologies can interfere with and forge signals during transmission.

[0005] To achieve the above objectives, the present invention provides a signal modulation method, which is applied at a signal transmitting end, and the method includes the following steps:

[0006] Semantic features are extracted from the source data to obtain semantic feature vectors;

[0007] The semantic feature vector is quantized to obtain a bit sequence;

[0008] The bit sequence is modulated according to the pre-allocated feature domain to obtain a modulated signal;

[0009] The modulated signal is transmitted to the signal receiving end through an interference channel so that the signal receiving end can demodulate the modulated signal.

[0010] Optionally, quantizing the semantic feature vector to obtain a bit sequence includes:

[0011] Determine the target codebook whose error meets the preset conditions;

[0012] Determine whether the error of the target codebook is less than a preset error threshold;

[0013] If the error of the target codebook is not less than the preset error threshold, the semantic feature vector is quantized to obtain a bit sequence.

[0014] Optionally,

[0015] Optionally, after determining whether the error of the target codebook is less than a preset error threshold, the method further includes:

[0016] If the error of the target codebook is less than the preset error threshold, then the semantic feature vector is subjected to codebook quantization to obtain the codebook index value;

[0017] The codebook index value is bit encoded to obtain a bit sequence.

[0018] Optionally, modulating the bit sequence according to a pre-allocated feature domain to obtain a modulated signal includes:

[0019] Map the bit sequence to a preset coordinate system to obtain at least one mapping group;

[0020] The bit sequence is converted into modulation symbols based on the mapping group;

[0021] The modulation symbols are modulated according to the pre-assigned feature domain to obtain a modulation signal.

[0022] Optionally, the step of extracting semantic features from the source data to obtain a semantic feature vector includes:

[0023] The feature domain is separated into multiple semantic feature subspaces;

[0024] Semantic features are extracted from the multiple semantic feature subspaces to obtain semantic feature vectors.

[0025] Furthermore, to achieve the above objectives, the present invention also proposes a signal demodulation method, wherein the signal demodulation is applied at a signal receiving end, and the signal demodulation method includes:

[0026] The received modulated signal is demodulated according to the pre-allocated feature domain to obtain the bit sequence;

[0027] The bit sequence is dequantized to obtain a semantic feature vector;

[0028] The semantic feature vector is decoded to obtain the target semantic information.

[0029] Optionally, the step of dequantizing the bit sequence to obtain a semantic feature vector includes:

[0030] Convert the bit sequence into a codebook index value;

[0031] The codebook index value is dequantized to obtain the semantic feature vector.

[0032] Furthermore, to achieve the above objectives, the present invention also proposes a communication device, the communication device comprising: a memory, a processor, and a signal modulation and / or signal demodulation program stored in the memory and executable on the processor, the signal modulation and / or signal demodulation program being configured to implement the steps of the signal modulation and / or signal demodulation method as described above.

[0033] Furthermore, to achieve the above objectives, the present invention also proposes a storage medium storing a signal modulation and / or signal demodulation program, which, when executed by a processor, implements the steps of the signal modulation and / or signal demodulation method as described above.

[0034] This invention extracts semantic features from source data at the signal transmitting end to obtain a semantic feature vector, quantizes the semantic feature vector to obtain a bit sequence, modulates the bit sequence according to a pre-allocated feature domain to obtain a modulated signal, and transmits the modulated signal to the signal receiving end through an interference channel so that the signal receiving end can demodulate the modulated signal; the signal receiving end demodulates the received modulated signal according to the pre-allocated feature domain to obtain a bit sequence, dequantizes the bit sequence to obtain a semantic feature vector, and performs feature decoding on the semantic feature vector to obtain the target semantic information. This effectively solves the problem that signals are easily interfered with and forged during communication, resulting in poor signal transmission security and low transmission efficiency, and significantly improves communication security and communication transmission efficiency. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the structure of a communication device in the hardware operating environment involved in the embodiments of the present invention;

[0036] Figure 2 This is a flowchart illustrating the first embodiment of the signal modulation method of the present invention;

[0037] Figure 3 This is a schematic diagram of multi-user information feature domain separation in the first embodiment of the signal modulation method of the present invention;

[0038] Figure 4 This is a schematic diagram of signal transmission and reception in the first embodiment of the signal modulation method of the present invention;

[0039] Figure 5 This is a flowchart illustrating the second embodiment of the signal modulation method of the present invention;

[0040] Figure 6 This is a flowchart illustrating the third embodiment of the signal modulation method of the present invention;

[0041] Figure 7 This is a flowchart illustrating a first embodiment of a signal demodulation method according to the present invention;

[0042] Figure 8 This is a schematic diagram of the semantic coding network of a first embodiment of the signal demodulation method of the present invention;

[0043] Figure 9 This is a schematic diagram of a semantic decoding network for a first embodiment of a signal demodulation method according to the present invention.

[0044] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0045] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0046] Reference Figure 1 , Figure 1 This is a schematic diagram of the communication device structure of the hardware operating environment involved in the embodiments of the present invention.

[0047] like Figure 1 As shown, the communication device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0048] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the communication device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0049] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and signal modulation and / or signal demodulation programs.

[0050] exist Figure 1In the communication device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the communication device of the present invention can be set in the communication device, and the communication device calls the signal modulation and / or signal demodulation program stored in the memory 1005 through the processor 1001 and executes the signal modulation and / or signal demodulation method provided in the embodiment of the present invention.

[0051] This invention provides a signal modulation method, referring to... Figure 2 , Figure 2 This is a flowchart illustrating a first embodiment of a signal modulation method according to the present invention.

[0052] In this embodiment, the signal modulation method is applied at the signal transmitting end, and the signal modulation method includes the following steps:

[0053] Step S10: Extract semantic features from the source data to obtain semantic feature vectors.

[0054] It should be understood that the execution subject of the method in this embodiment can be a communication device with data processing, network communication and program running functions, such as a computer, or other devices or equipment that can achieve the same or similar functions. The above-mentioned signal transmitting end is used as an example for illustration here.

[0055] It should be noted that this embodiment uses source data... Perform interpretable decoupling encoding to extract semantic features and encode them into feature vectors. Refer to Formula 1 below, where, This represents the i-th interpretable decoupled semantic encoder. This represents the trainable parameters of an interpretable, decoupled semantic encoder.

[0056]

[0057] Furthermore, in order to effectively extract semantic features, step S10 above may include:

[0058] The feature domain is separated into multiple semantic feature subspaces;

[0059] Semantic features are extracted from the multiple semantic feature subspaces to obtain semantic feature vectors.

[0060] In the implementation, assume there are N sending users and M receiving users, and each sending user needs to send M source data to each receiving user. M source data The implicit semantic information is

[0061]

[0062] Understandably, referring to Figure 3 , Figure 3 This is a schematic diagram of feature domain separation for multi-user information. This embodiment utilizes deep learning to analyze source data. semantic feature representation And combined with CSI Multi-user joint source-channel coding is used to obtain The feature domain is separated into multiple semantic feature subspaces. To achieve multi-user semantic information encoded and transmitted in a separate feature subspace. i,j ∈T i,j ,and This enables multi-user interference feature-based multiple access and multi-user semantic interference feature-based multiple access model training. Users access the network based on their assigned feature domains and jointly train their respective anti-jamming channel encoding and decoding models. IB As the loss function, refer to Formula 2 below, where λ>0 is an adjustable weight factor.

[0063]

[0064] Step S20: Quantize the semantic feature vector to obtain a bit sequence.

[0065] It should be noted that the bit sequence can be a transmitted bit stream. In this embodiment, the semantic feature vector is quantized to convert the semantic information to be transmitted into binary to obtain a bit stream. A bit stream is a sequence of numbers consisting of a series of 0s and 1s, with each bit representing a binary digit.

[0066] It is understood that this embodiment can be implemented by codebook quantization of the semantic feature vector or by direct quantization of the semantic feature vector. When a feature vector is input, codebook quantization is performed first. The codebook with the largest inner product and the smallest error is found in the codebook. An error threshold is set. If the error is less than the threshold, the codebook index is output. If the error is greater than the threshold, the feature vector is directly quantized. At the same time, the reconstructed feature vector is fed into the codebook and updated to establish a learnable codebook.

[0067] It should be understood that the correlation between the dimensions of the feature vector is very low. Codebook quantization effectively utilizes the various interrelationships between the components of the feature vector, thus exhibiting high compressibility. Direct quantization, on the other hand, simplifies the process. Employing both quantization methods can minimize distortion and efficiently quantize the feature vector into a bitstream.

[0068] It's important to note that semantic feature vectors typically contain a large number of dimensions and continuous values, incurring significant overhead during transmission and storage. Vector quantization maps high-dimensional continuous vectors to a low-dimensional discrete codeword set, thereby compressing the data. This saves bandwidth and storage space, and improves system efficiency and performance. By quantizing vectors into discrete codewords, the details and specific values ​​of the original data can be hidden. This helps protect data privacy, preventing unauthorized access to sensitive information in the original data.

[0069] Step S30: Modulate the bit sequence according to the pre-allocated feature domain to obtain a modulated signal.

[0070] It should be noted that the semantic feature vector after codebook quantization becomes the index value of the codebook, which takes the value of a certain integer. Then, each index value is bit-encoded, thus obtaining the transmitted bitstream. The feature vector after direct quantization directly becomes the transmitted bitstream. By modulating the transmitted bitstream, a modulated signal is obtained.

[0071] Step S40: The modulated signal is transmitted to the signal receiving end through the interference channel so that the signal receiving end can demodulate the modulated signal.

[0072] It should be noted that this embodiment proposes an implicit protocol networking mechanism. The signal transmitter uses source-content-destination modulation to transmit a unified waveform, while the signal receiver uses artificial intelligence decoding to obtain the source and destination, and demodulates the corresponding content to achieve implicit protocol networking. Implicit protocol networking transmits signals in parallel across time, frequency, space, and feature domains (a new dimension of signal transmission), and distinguishes users in the feature domain. Based on AI-driven joint feature encoding / decoding of source-content-destination, it enables intelligent and rapid network communication between people.

[0073] It should be understood that, with reference Figure 4 , Figure 4 The diagram illustrates signal transmission and reception. The signal transmitter extracts semantic features from the source data based on a feature knowledge base, performs source-content-destination waveform modulation based on a waveform library, and transmits the modulated signal to the signal receiver through a physical channel. The signal receiver performs high-speed, high-precision sampling, performs waveform reconstruction calculations based on the waveform library, reconstructs feature information based on the feature knowledge base, obtains the target semantics, and achieves source-content-destination waveform demodulation.

[0074] This embodiment extracts semantic features from source data at the signal transmitting end to obtain a semantic feature vector, quantizes the semantic feature vector to obtain a bit sequence, modulates the bit sequence according to a pre-allocated feature domain to obtain a modulated signal, and sends the modulated signal to the signal receiving end through an interference channel so that the signal receiving end can demodulate the modulated signal. This effectively solves the problem that signals are easily interfered with and forged during communication, resulting in poor signal transmission security and low transmission efficiency, and greatly improves communication security and communication transmission efficiency.

[0075] refer to Figure 5 , Figure 5 This is a flowchart illustrating a second embodiment of a signal modulation method according to the present invention.

[0076] Based on the first embodiment described above, in this embodiment, step S20 includes:

[0077] Step S21: Determine the target codebook whose error meets the preset conditions;

[0078] Step S22: Determine whether the error of the target codebook is less than a preset error threshold;

[0079] Step S23: If the error of the target codebook is not less than the preset error threshold, then the semantic feature vector is quantized to obtain a bit sequence.

[0080] It's important to note that semantic feature vectors typically contain a large number of dimensions and continuous values, incurring significant overhead during transmission and storage. Vector quantization maps high-dimensional continuous vectors to a low-dimensional discrete codeword set, thereby compressing the data. This saves bandwidth and storage space, and improves system efficiency and performance. By quantizing vectors into discrete codewords, the details and specific values ​​of the original data can be hidden. This helps protect data privacy, preventing unauthorized access to sensitive information in the original data.

[0081] It is understood that this embodiment can either quantize the semantic feature vector using codebook quantization or directly quantize the semantic feature vector. If the error exceeds a threshold, the feature vector is directly quantized. Simultaneously, the reconstructed feature vector is fed into the codebook and updated to establish a learnable codebook.

[0082] It should be understood that the correlation between the dimensions of the feature vector is very low. Codebook quantization effectively utilizes the various interrelationships between the components of the feature vector, thus exhibiting high compressibility. Direct quantization, on the other hand, simplifies the process. Employing both quantization methods can minimize distortion and efficiently quantize the feature vector into a bitstream.

[0083] Furthermore, to improve compression efficiency, after step S22, the following steps are also included:

[0084] If the error of the target codebook is less than the preset error threshold, then the semantic feature vector is subjected to codebook quantization to obtain the codebook index value;

[0085] The codebook index value is bit encoded to obtain a bit sequence.

[0086] It's important to note that in certain scenarios, processing high-dimensional continuous vectors can lead to complex computations and operational costs. Vector quantization can convert the original vector into a discrete symbolic representation, thereby simplifying computation and reducing computational complexity. This is crucial for real-time applications or resource-constrained environments.

[0087] It should be understood that this embodiment maps continuous semantic feature vectors to a predefined codebook. The codebook is a set of discrete codewords, each codeword representing a cluster center or representative sample. In codebook quantization, the distance between the input vector and the codewords in the codebook is calculated, and the best-matching codeword is selected as the quantized representation. Therefore, codebook quantization converts semantic feature vectors into discrete codewords, achieving vector compression and discretization.

[0088] Understandably, when the input semantic feature vector is given, codebook quantization is performed first. The codebook with the largest inner product and smallest error is found within the codebook. An error threshold is set; if the error is less than the threshold, the codebook index value is output. The feature vector after codebook quantization becomes the codebook index value, which takes the value of a certain integer. Then, each index value is bit-encoded, thus obtaining the transmitted bit stream.

[0089] This embodiment determines a target codebook whose error meets a preset condition, and then judges whether the error of the target codebook is less than a preset error threshold. If the error of the target codebook is not less than the preset error threshold, the semantic feature vector is quantized to obtain a bit sequence, thereby effectively reducing the dimension of the semantic feature vector, greatly improving the compression rate and simplification rate, minimizing the distortion rate, and efficiently quantizing the feature vector into a bit stream.

[0090] refer to Figure 6 , Figure 6 This is a flowchart illustrating a third embodiment of a signal modulation method according to the present invention.

[0091] Based on the first embodiment described above, in this embodiment, step S30 includes:

[0092] Step S301: Map the bit sequence to a preset coordinate system to obtain at least one mapping group;

[0093] Step S302: Convert the bit sequence into modulation symbols based on the mapping group;

[0094] Step S303: Modulate the modulation symbol according to the pre-allocated feature domain to obtain a modulation signal.

[0095] It should be noted that the preset coordinate system can be a constellation diagram, and the mapping group can be a mapping relationship group consisting of each bit in the bit sequence and the corresponding point on the preset coordinate system.

[0096] Understandably, in this embodiment, the information to be transmitted is first converted into binary form to obtain a bitstream, which is then modulated using M-PSK modulation. In M-PSK modulation, each discrete phase state corresponds to a point in a constellation diagram. A constellation diagram is a two-dimensional coordinate system where points represent different phase states. For example, for 4-PSK, a square constellation diagram can be used, with each vertex representing a phase state; for 8-PSK, the constellation diagram can be an octagon. Based on each bit of the bitstream, a mapping table or mapping algorithm is used to convert a continuous bit sequence into a modulation symbol corresponding to a point in the constellation diagram.

[0097] In the specific implementation, the transmitter sends the modulated signal to N users through an interference channel, and performs M-PSK modulation on the modulation symbols passed through the interference channel to obtain the damaged vector. Formula 3 is shown below.

[0098]

[0099] Where h(.) represents modulation, d(.) represents demodulation, and g ii ,i∈{1,..,N} is the channel gain from sending user i to receiving user i, g ij Let i,j∈{1,..,N} be the channel gain from sending user j to receiving user i. i Representing channel noise, additive channel noise is sampled from a zero-mean Gaussian distribution, with... The noise variance, i.e.

[0100]

[0101] This embodiment maps the bit sequence to a preset coordinate system to obtain at least one mapping group. Based on the mapping group, the bit sequence is converted into a modulation symbol. The modulation symbol is then modulated according to a pre-allocated feature domain to obtain a modulated signal. This effectively improves noise immunity, reduces the complexity of modulation and demodulation, and improves signal transmission efficiency.

[0102] refer to Figure 7 , Figure 7 This is a flowchart illustrating a first embodiment of a signal demodulation method according to the present invention.

[0103] In this embodiment, the signal demodulation method is applied at the signal receiving end, and the signal demodulation method includes:

[0104] Step S50: Demodulate the received modulated signal according to the pre-allocated feature domain to obtain the bit sequence;

[0105] Step S60: Dequantize the bit sequence to obtain a semantic feature vector;

[0106] Step S70: Decode the semantic feature vector to obtain the target semantic information.

[0107] It should be understood that the execution subject of the method in this embodiment can be a communication device with data processing, network communication and program running functions, such as a computer, or other devices or equipment that can achieve the same or similar functions. The above-mentioned signal receiving end is used as an example for illustration here.

[0108] It should be noted that this embodiment proposes an implicit protocol networking mechanism. The signal transmitter uses source-content-destination modulation to transmit a unified waveform, while the signal receiver uses artificial intelligence decoding to obtain the source and destination, and demodulates the corresponding content to achieve implicit protocol networking. Implicit protocol networking transmits signals in parallel across time, frequency, space, and feature domains (a new dimension of signal transmission), and distinguishes users in the feature domain. Based on AI-driven joint feature encoding / decoding of source-content-destination, it enables intelligent and rapid network communication between people.

[0109] It should be understood that the signal transmitting end extracts semantic features from the source data based on the feature knowledge base, performs source-content-sink waveform modulation based on the waveform library, and sends the modulated signal obtained after modulation to the signal receiving end through the physical channel; the signal receiving end performs high-speed and high-precision sampling, performs waveform reconstruction calculation based on the waveform library, and reconstructs feature information based on the feature knowledge base to obtain the target semantics, thereby realizing source-content-sink waveform demodulation.

[0110] It is understandable that this embodiment converts the damaged vector (bit stream) into a feature vector by performing vector dequantization on the received modulated signal. In the case of direct quantization, the bit stream is directly dequantized into a feature vector. Then semantic feature recovery is performed, and the feature vector is... Input semantic decoder to recover semantics, thereby obtaining estimated semantics. Refer to Formula 4 below, where Let θ represent the i-th semantic decoder. i The trainable parameters of the semantic decoder are represented by [reference]. Figure 8 and Figure 9 , Figure 8This is a schematic diagram of a semantic coding network. Figure 9 This is a schematic diagram of a semantic decoding network, such as... Figure 9 Semantic decoding network, for user i's received signal y i Upsampling involves performing a deconvolution operation that halves the number of feature channels. Each upsampling operation is performed with... Figure 8 Semantic coding networks with the same number of channels are fused together. This allows us to obtain the estimated semantics of the original source data.

[0111]

[0112] Furthermore, to improve semantic recovery efficiency, step S60 above may include:

[0113] Convert the bit sequence into a codebook index value;

[0114] The codebook index value is dequantized to obtain the semantic feature vector.

[0115] It should be noted that when using codebook quantization, the damaged vector (bit stream) must first be converted into an integer to obtain the codebook index. Then, the codebook index is used to find the corresponding codebook basis vector in the codebook, thus forming the feature vector. When using direct quantization, the bitstream is directly dequantized into feature vectors.

[0116] This embodiment demodulates the received modulated signal according to the pre-allocated feature domain at the signal receiving end to obtain a bit sequence, dequantizes the bit sequence to obtain a semantic feature vector, and performs feature decoding on the semantic feature vector to obtain the target semantic information. This effectively solves the problem that signals are easily interfered with and forged during communication, resulting in poor signal transmission security and low transmission efficiency, and greatly improves communication security and communication transmission efficiency.

[0117] Furthermore, embodiments of the present invention also propose a storage medium storing a signal modulation and / or signal demodulation program, wherein when the signal modulation and / or signal demodulation program is executed by a processor, it implements the steps of the signal modulation and / or signal demodulation method as described above.

[0118] Since this storage medium adopts all the technical solutions of all the above embodiments, it has at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be repeated here.

[0119] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0120] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0121] In addition, for technical details not described in detail in this embodiment, please refer to the signal modulation and / or signal demodulation methods provided in any embodiment of the present invention, which will not be repeated here.

[0122] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0123] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0124] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0125] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A signal modulation method, characterized in that, The signal modulation method is applied at the signal transmitting end, and the signal modulation method includes: Semantic features are extracted from the source data to obtain semantic feature vectors; The semantic feature vector is quantized to obtain a bit sequence, which is a transmitted bit stream; The bit sequence is modulated according to the pre-allocated feature domain to obtain a modulated signal; The modulated signal is transmitted to the signal receiving end through an interference channel so that the signal receiving end can demodulate the modulated signal. The step of modulating the bit sequence according to the pre-allocated feature domain to obtain the modulated signal includes: Map the bit sequence to a preset coordinate system to obtain at least one mapping group; The bit sequence is converted into modulation symbols based on the mapping group; The modulation symbols are modulated according to the pre-assigned feature domain to obtain a modulation signal.

2. The signal modulation method as described in claim 1, characterized in that, The step of quantizing the semantic feature vector to obtain a bit sequence includes: Determine the target codebook whose error meets the preset conditions; Determine whether the error of the target codebook is less than a preset error threshold; If the error of the target codebook is not less than the preset error threshold, the semantic feature vector is quantized to obtain a bit sequence.

3. The signal modulation method as described in claim 2, characterized in that, After determining whether the error of the target codebook is less than a preset error threshold, the method further includes: If the error of the target codebook is less than the preset error threshold, then the semantic feature vector is subjected to codebook quantization to obtain the codebook index value; The codebook index value is bit encoded to obtain a bit sequence.

4. The signal modulation method as described in claim 1, characterized in that, The step of extracting semantic features from the source data to obtain a semantic feature vector includes: The feature domain is separated into multiple semantic feature subspaces; Semantic features are extracted from the multiple semantic feature subspaces to obtain semantic feature vectors.

5. A signal demodulation method, characterized in that, The signal demodulation is applied at the signal receiving end, and the signal demodulation method includes: The received modulated signal is demodulated according to the pre-allocated feature domain to obtain the bit sequence; The bit sequence is dequantized to obtain a semantic feature vector; The semantic feature vector is decoded to obtain the target semantic information; The modulation signal is obtained by modulating the bit sequence according to a pre-allocated feature domain, wherein the modulation of the bit sequence according to the pre-allocated feature domain includes: Map the bit sequence to a preset coordinate system to obtain at least one mapping group; The bit sequence is converted into modulation symbols based on the mapping group; The modulation symbols are modulated according to the pre-assigned feature domain to obtain a modulation signal.

6. The signal demodulation method as described in claim 5, characterized in that, The step of dequantizing the bit sequence to obtain a semantic feature vector includes: Convert the bit sequence into a codebook index value; The codebook index value is dequantized to obtain the semantic feature vector.

7. The signal demodulation method as described in claim 5, characterized in that, The step of decoding the semantic feature vector to obtain the target semantic information includes: Obtain the decoding model corresponding to the pre-allocated feature domain; The semantic feature vector is input into the decoding model for feature decoding to obtain the target semantic information.

8. A communication device, characterized in that, The communication device includes: a memory, a processor, and a signal modulation and / or demodulation program stored in the memory and executable on the processor, the signal modulation and / or demodulation program being configured to implement the signal modulation and / or demodulation method as described in any one of claims 1 to 7.

9. A storage medium, characterized in that, The storage medium stores a signal modulation and / or signal demodulation program, which, when executed by a processor, implements the signal modulation and / or signal demodulation method as described in any one of claims 1 to 7.

Images