A Joint Semantic Coding-Transmission Method for High-Mobility Scenarios

Abstract

This invention provides a joint semantic coding-transmission method for high-mobility scenarios, relating to the field of wireless communication. It includes: a transmitter extracting raw semantic features from raw data to obtain a raw data stream; mapping the raw data stream onto a time-delay-Doppler domain resource grid using an orthogonal time-frequency spatial modulator and inserting pilot sequences to obtain an orthogonal time-frequency spatial frame; modulating the orthogonal time-frequency spatial frame to obtain a time-domain transmission signal; the time-domain transmission signal reaching the receiver via a channel in a high-mobility scenario; the receiver demodulating the time-domain received signal using an orthogonal time-frequency spatial demodulator to obtain a time-delay-Doppler domain received signal; using the time-delay-Doppler domain received signal to obtain an estimated channel matrix; using the estimated channel matrix and the time-delay-Doppler domain received signal to obtain a reconstructed data stream; and using the reconstructed data stream to obtain reconstructed data, thereby improving the transmission performance of the semantic communication system in high-mobility scenarios.

Description

Technical Field

[0001] This invention relates to the field of wireless communication, and more particularly to a joint semantic coding-transmission method for high mobility scenarios. Background Technology

[0002] In highly mobile scenarios (such as high-speed rail, autonomous driving, drones, and aerospace), the communication channel environment exhibits high dynamism and complexity. Simultaneously, with the widespread application of high-speed mobile scenarios, users' demand for massive data transmission is increasing daily, and existing communication technologies struggle to meet the high-capacity, high-reliability communication requirements of complex high-speed mobile environments. Currently, to optimize semantic communication systems in multipath fading channel environments, Orthogonal Frequency Division Multiplexing (OFDM) technology, which has good resistance to multipath interference, has been introduced into semantic communication systems.

[0003] However, in high-speed, complex scenarios, the dynamic nature of transportation vehicles such as autonomous vehicles, high-speed trains, and aircraft leads to a highly time-varying communication environment. Channels exhibit significant time and frequency selectivity, which severely impacts the quality of semantic information transmission. Significant Doppler shift and severe signal attenuation pose serious challenges to semantic communication systems, resulting in a sharp decline in transmission performance for OFDM-based semantic communication systems. Therefore, improving the transmission performance of semantic communication systems in highly mobile scenarios is a crucial technical problem that this invention aims to solve. Summary of the Invention

[0004] Based on the above-mentioned technical problems, the present invention provides a joint semantic coding-transmission method for high mobility scenarios, which aims to overcome the above problems or at least partially solve them.

[0005] The first aspect of this invention provides a joint semantic coding-transmission method for high-mobility scenarios, the method comprising:

[0006] In a semantic communication system, the sending end extracts the original semantic features of the original data and obtains the original data stream based on the original semantic features.

[0007] The transmitting end maps the original data stream onto the time-delay-Doppler domain resource grid through an orthogonal time-frequency spatial modulator, inserts a pilot sequence to obtain an orthogonal time-frequency spatial frame, and modulates the orthogonal time-frequency spatial frame to obtain a time-domain transmission signal;

[0008] The transmitting end sends the time-domain transmission signal to the receiving end through a channel in a high-mobility scenario between itself and the receiving end in the semantic communication system;

[0009] The receiving end demodulates the received time-domain received signal through an orthogonal time-frequency spatial demodulator to obtain a time-delay-Doppler domain received signal;

[0010] The receiving end uses the time-delay-Doppler domain received signal to perform channel estimation through the channel estimation module, and obtains the estimated channel matrix;

[0011] The receiving end uses the estimated channel matrix and the time-delay-Doppler domain to receive the signal, obtains the reconstructed data stream, and obtains the reconstructed data based on the reconstructed data stream.

[0012] In the joint semantic coding-transmission method for high mobility scenarios proposed in this invention, orthogonal time-frequency space (OTFS) modulation technology is introduced into the semantic communication system by combining the channel characteristics of high mobility scenarios. By mapping the original data stream to the delay-Doppler (DD) domain for transmission, it has transmission advantages in the Doppler channel environment, effectively suppressing the Doppler effect, and providing a solution for semantic information transmission in high-speed and complex scenarios. Furthermore, to address the high complexity of DD domain channel estimation, this invention proposes a DD domain pilot pattern to reduce inter-signal interference: The original data stream obtained from the transmitter is mapped onto a time-delay-Doppler domain resource grid using an orthogonal time-frequency spatial modulator, and a pilot sequence is inserted to obtain an orthogonal time-frequency spatial frame. This frame is then modulated to obtain the time-domain transmission signal to be transmitted. Thus, when the receiver demodulates the received time-domain signal using an orthogonal time-frequency spatial demodulator to obtain the time-delay-Doppler domain received signal and performs channel estimation, the interference of data on the pilot is reduced, achieving low-overhead, low-complexity channel estimation. Through the coordinated optimization of semantic coding and transmission waveform, efficient and reliable transmission of semantic information in highly mobile scenarios is achieved, improving the transmission performance of semantic communication systems in highly mobile scenarios. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating the steps of a joint semantic coding-transmission method for high-mobility scenarios according to an embodiment of the present invention;

[0015] Figure 2This is a schematic diagram of a pilot pattern based on channel sparsity, as shown in an embodiment of the present invention;

[0016] Figure 3 This is a schematic diagram illustrating a specific example of a pilot pattern based on channel sparsity according to an embodiment of the present invention;

[0017] Figure 4 This is a schematic diagram illustrating an integrated semantic information transmission architecture based on orthogonal time-frequency spatial modulation for high-mobility scenarios, according to an embodiment of the present invention.

[0018] Figure 5 This is a structural block diagram of a joint semantic coding-transmission device for high-mobility scenarios provided in an embodiment of the present invention;

[0019] Figure 6 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of a joint semantic coding-transmission method for high-mobility scenarios, as shown in an embodiment of the present invention. Figure 1 As shown, the joint semantic coding-transmission method for high-mobility scenarios provided in this embodiment includes at least the following steps:

[0022] Step S11: The sending end in the semantic communication system extracts the original semantic features of the original data and obtains the original data stream based on the original semantic features.

[0023] In this embodiment, the semantic communication system includes: a transmitter, a channel for high-mobility scenarios, and a receiver. High-mobility scenarios refer to communication or control environments on high-speed mobile carriers such as vehicles, airplanes, and high-speed trains, characterized by high dynamism and stringent real-time requirements, resulting in significant Doppler shift. In an optional embodiment, a high-mobility scenario can refer to a mobile scenario where the relative speed between the carrier and the ground base station exceeds a target threshold (e.g., 300 km / h, without specific limitation).

[0024] For the raw data, the sending end can first perform semantic encoding on the raw data to extract the raw semantic features, and then sequentially process the raw semantic features through channel coding and constellation modulation to obtain the raw data stream. The raw data stream includes multiple raw data symbols. In this embodiment, the raw data can be any type of data to be transmitted, such as image data, video data, text data, etc., without any limitation.

[0025] Step S12: The transmitting end maps the original data stream onto the time-delay-Doppler domain resource grid through an orthogonal time-frequency spatial modulator, inserts a pilot sequence to obtain an orthogonal time-frequency spatial frame, and modulates the orthogonal time-frequency spatial frame to obtain a time-domain transmission signal.

[0026] In this embodiment, orthogonal time-frequency spatial modulation technology is introduced into the semantic communication system. The transmitting end includes at least an orthogonal time-frequency spatial modulator. Based on the channel sparsity in the delay-Doppler domain (DD domain), this embodiment designs a DD domain pilot pattern: the original data stream is mapped onto the delay-Doppler domain resource grid through the orthogonal time-frequency spatial modulator, and a pilot sequence is inserted to obtain an orthogonal time-frequency spatial frame. After obtaining the orthogonal time-frequency spatial frame, the transmitting end can perform orthogonal time-frequency spatial modulation on the orthogonal time-frequency spatial frame through the orthogonal time-frequency spatial modulator to obtain a time-domain transmission signal. The pilot sequence includes multiple pilot symbols. In this embodiment, the orthogonal time-frequency spatial frame is a delay-Doppler domain transmission signal, including multiple delay-Doppler domain transmission symbols.

[0027] In an alternative example, the orthogonal time-frequency spatial modulator can first transform the orthogonal time-frequency spatial frame to the time-frequency domain through the inverse symplectic finite Fourier transform (ISFFT), and then generate the time-domain transmission signal through the inverse M-point discrete Fourier transform (IDFT) and pulse shaping.

[0028] Step S13: The transmitting end sends the time-domain transmission signal to the receiving end through a channel in a high-mobility scenario between itself and the receiving end in the semantic communication system.

[0029] In this embodiment, after receiving the time-domain transmission signal, the transmitting end can send the time-domain transmission signal to the receiving end through the channel of the high-mobility scenario between the receiving ends in the semantic communication system.

[0030] Step S14: The receiving end demodulates the received time-domain received signal through an orthogonal time-frequency spatial demodulator to obtain a time-delay-Doppler domain received signal.

[0031] In this embodiment, the receiving end can receive the time-domain transmitted signal through the channel in a high-mobility scenario between the sending and receiving ends in the semantic communication system, thus obtaining the received time-domain received signal. It can be understood that the relationship between the time-domain received signal and the time-domain transmitted signal is as follows: the time-domain received signal is obtained based on the time-domain channel matrix, the time-domain transmitted signal, and the noise of the channel.

[0032] The receiver includes at least an orthogonal time-frequency spatial demodulator. After receiving the time-domain received signal, the receiver can use the orthogonal time-frequency spatial demodulator to perform orthogonal time-frequency spatial demodulation on the received time-domain received signal to obtain the time-delay-Doppler domain received signal. The time-delay-Doppler domain received signal includes multiple time-delay-Doppler domain received symbols.

[0033] In an alternative example, the orthogonal time-frequency spatial modulator can sequentially pass the time-domain received signal through matched filtering, M-point Discrete Fourier Transform (DFT), and Symptotic Finite Fourier Transform (SFFT) to obtain the time-delay-Doppler domain received signal.

[0034] Step S15: The receiving end uses the time-delay-Doppler domain received signal to perform channel estimation through the channel estimation module, and obtains the estimated channel matrix.

[0035] In this embodiment, after receiving the time-delay-Doppler domain received signal, the receiver can use the channel estimation module in the receiver to perform channel estimation using the time-delay-Doppler domain received signal to obtain the estimated channel matrix for the detection of semantic information symbols.

[0036] Step S16: The receiving end uses the estimated channel matrix and the time delay-Doppler domain to receive the signal, obtains the reconstructed data stream, and obtains the reconstructed data based on the reconstructed data stream.

[0037] In this embodiment, after the receiver obtains the estimated channel matrix, it can process the estimated channel matrix and the obtained delay-Doppler domain received signal to obtain a reconstructed data stream. The reconstructed data stream is then processed to obtain reconstructed semantic features. The reconstructed semantic features are then semantically decoded to obtain the reconstructed data corresponding to the original data, thereby realizing the wireless transmission of the original data in high mobility scenarios.

[0038] In one optional embodiment, the receiver first performs signal detection based on the estimated channel matrix and the received signal in the delay-Doppler domain to obtain the reconstructed data stream, and then sequentially passes the reconstructed data stream through constellation demodulation and channel decoding to obtain the reconstructed semantic features.

[0039] In this embodiment, considering the channel characteristics of high-mobility scenarios, orthogonal time-frequency spatial modulation (ODSM) technology is introduced into the semantic communication system. By mapping the original data stream to the delay-Doppler domain for transmission, it has transmission advantages in the Doppler channel environment, effectively suppressing the Doppler effect and providing a solution for semantic information transmission in high-speed, complex scenarios. Furthermore, to address the high complexity of DD domain channel estimation, this invention proposes a DD domain pilot pattern to reduce inter-signal interference: the original data stream obtained from the transmitter is mapped onto the delay-Doppler domain resource grid using an orthogonal time-frequency spatial modulator, and a pilot sequence is inserted to obtain an orthogonal time-frequency spatial frame. This frame is then modulated to obtain the time-domain transmission signal to be sent. Thus, when the receiver demodulates the received time-domain signal using an orthogonal time-frequency spatial demodulator to obtain the delay-Doppler domain received signal and performs channel estimation, it can reduce data interference to the pilot, achieving low-overhead, low-complexity channel estimation. Therefore, through the coordinated optimization of semantic coding and transmission waveform, efficient and reliable transmission of semantic information in high-mobility scenarios is achieved.

[0040] In conjunction with the above embodiments, in one implementation, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios, wherein, in addition to the above steps, the method may further include steps S21 and S22:

[0041] Step S21: The receiver feeds back the estimated channel matrix to the semantic encoder in the transmitter.

[0042] In this embodiment, the transmitting end also includes a semantic encoder to semantically encode the original data and obtain the original semantic features. After obtaining the estimated channel matrix, the receiving end can also feed the estimated channel matrix back to the semantic encoder in the transmitting end as channel state information (CSI), serving as one of the inputs to the semantic encoder.

[0043] Step S22: The transmitting end uses the estimated channel matrix as a priori and performs semantic extraction on the next raw data through the semantic encoder to obtain the original semantic features of the next raw data.

[0044] In this embodiment, after receiving the estimated channel matrix, the transmitting end can use the estimated channel matrix as a priori to perform semantic extraction on the next raw data through a semantic encoder to obtain the original semantic features of the next raw data. Specifically, for the transmission of the next raw data, the input of the semantic encoder includes at least the estimated channel matrix and the next raw data. The semantic encoder obtains the original semantic features of the next raw data by performing semantic encoding on the estimated channel matrix and the next raw data.

[0045] In another optional embodiment, the input to the semantic encoder, in addition to the next raw data and the estimated channel matrix, also includes the channel conditions of the high-mobility scenario (such as using the signal-to-noise ratio (SNR) of the channel in the high-mobility scenario as the channel conditions). The transmitting end can input the next raw data, the estimated channel matrix, and the channel conditions together into the semantic encoder for semantic encoding to extract semantic features and obtain the raw semantic features of the next raw data. It should be noted that the next raw data can be any raw data. When the raw data is the first raw data, the estimated channel matrix and the channel conditions of the high-mobility scenario can be empty, or can be empirical values, historical values, etc., without restriction.

[0046] In this embodiment, the channel matrix estimated by the receiver is fed back to the transmitter as input to the semantic encoder, enabling the semantic encoder to adaptively adjust according to the estimated channel matrix, thereby achieving CSI-aware semantic coding, i.e., channel-adaptive semantic coding, which improves coding and transmission performance.

[0047] In conjunction with any of the above embodiments, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios. In this method, the step S12 above, "the transmitting end maps the original data stream onto the time-delay-Doppler domain resource grid through an orthogonal time-frequency spatial modulator and inserts a pilot sequence to obtain an orthogonal time-frequency spatial frame," specifically includes the following steps S31 to S32:

[0048] Step S31: Generate a pilot pattern with the goal of making a portion of the time-delay-Doppler domain received symbols in the time-delay-Doppler domain received signal only related to the pilot sequence;

[0049] Step S32: According to the pilot pattern, map the original data stream onto the time-delay-Doppler domain resource grid and insert the pilot sequence to obtain an orthogonal time-frequency spatial frame.

[0050] This embodiment takes into account that for orthogonal time-frequency spatial modulation, let and Let the number of symbols represent the time delay and the Doppler dimension, respectively. Then the total number of DD domain resources is: .make and Let represent the subcarrier spacing and symbol period, respectively. Then the resolutions of the time delay and Doppler dimension are respectively... and Therefore, the equivalent channel matrix in the DD domain can be expressed by the following equation (1):

[0051] (1);

[0052] in, Indicates a total of One path; Indicates the first Complex gain of the path, Indicates the first The orthogonal basis corresponding to each path. express Fourier matrix, express 3D identity matrix express The conjugate transpose of . Indicates the Kronecker product. and These represent the time delay and the size of the Doppler tap, respectively. It is a forward cyclic shift matrix. It is a diagonal matrix, where .

[0053] Furthermore, considering integer delay and Doppler, and These represent the time delay and the size of the Doppler tap, respectively. and These represent the resource grid sizes for the time delay and Doppler dimensions, respectively. The resource grid sizes for the time delay domain and Doppler domain are defined as follows: and and order Indicating the time-delay-Doppler domain resource grid index The corresponding complex gain; then, the DD-domain equivalent channel matrix in equation (1) can be re-expressed using the following equation (2):

[0054] (2);

[0055] in, Indicating the time-delay-Doppler domain resource grid index The corresponding orthogonal basis.

[0056] This embodiment takes into account that, as can be seen from equations (1) and (2) above, although the number of potential elements constituting the channel ( ) is very large, but only a few ( Non-zero elements of ) satisfy The DD domain channel exhibits sparsity. Therefore, this embodiment transforms OTFS channel estimation into a sparse signal recovery problem within compressed sensing.

[0057] In this embodiment, the specific modeling method for the sparse signal recovery problem is as follows:

[0058] The first embodiment q The orthogonal time-frequency space frame is composed of the first q pilot sequence and the q One raw data stream composition, q ∈[1, Q ],and q and Q If the integer is , then the first... q Time-delay-Doppler domain received signal It can be expressed by the following formula (3):

[0059] (3)

[0060] in, For the first q Equivalent channel matrix in DD domain Indicates the first q DD domain receives noise.

[0061] Substituting equation (2) into equation (3), we get:

[0062] (4)

[0063] in, Equation (4) above can be written in matrix form as follows:

[0064] (5)

[0065] in, For the sensing matrix, each column consists of... composition. For sparse vectors, only The non-zero element is the nth non-zero element, where the non-zero element is the nth non-zero element. q Equivalent channel matrix in DD domain Complex gain If there is no data interference, equation (5) above can be equivalent to the sparse signal recovery problem in compressed sensing, that is, according to the observation vector (the first... q (Time Delay-Doppler Domain Received Signal) and sensor matrix Solving using algorithms from compressed sensing This leads to the channel matrix. The sensing matrix is ​​constructed based on the pilot sequence, which is known to both the receiver and transmitter; therefore, the sensing matrix is ​​also known.

[0066] According to equation (5) above, the receiving end only knows the sensing matrix. Then, it consists of unknown raw data streams. This is considered interference. Since the original data stream has a high power (compared to noise), the resulting interference is not negligible and will affect the performance of channel estimation. Therefore, this embodiment proposes a solution: a specially designed pilot pattern to eliminate interference from the received data to the pilot and improve channel estimation performance.

[0067] Specifically, this embodiment takes into account the equivalent channel matrix in the DD domain based on the above equation (1). Each line has only The nth non-zero value. Therefore, the nth q Time-delay-Doppler domain received signal Each symbol in the text is only associated with... The pilot pattern is related to the transmitted symbols in the time-delay-Doppler domain. Therefore, the specific method for designing the pilot pattern in this embodiment is as follows: A pilot pattern is generated with the goal of ensuring that a portion of the received symbols in the time-delay-Doppler domain are only related to the pilot sequence. Based on this, channel estimation is performed only on a portion of the received symbols in the time-delay-Doppler domain that are related to the pilot sequence, which can effectively eliminate interference from unknown data streams and improve channel estimation performance. Then, according to the generated pilot pattern, the original data stream is mapped onto the time-delay-Doppler domain resource grid, and the pilot sequence is inserted to obtain an orthogonal time-frequency spatial frame.

[0068] In conjunction with any of the above embodiments, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios, in which the number of generated orthogonal time-frequency spatial frames is... The above step S32 may specifically include the following step S41:

[0069] Step S41: According to the pilot pattern, from the first position to the second position... The pilot sequence is inserted at the nth position, and in the remaining nth position... Position 1 to 2 Insert the original data stream at the nth position to obtain the nth position. q indivual Dimensional orthogonal time-frequency spatial frames.

[0070] In this embodiment, the generated pilot pattern can be used in the first... q indivual The first position to the second position of the orthogonal time-frequency space frame of dimension Insert pilot sequences at position n (the pilot sequences are agreed upon by both the transmitter and receiver), and at position n... q indivual The remaining third dimension in the orthogonal time-frequency space frame Position 1 to 2 Insert the original data stream at the nth position to generate the nth... q indivual An orthogonal time-frequency spatial frame of dimension . Indicates the maximum Doppler tap. and These represent the number of symbols for time delay and Doppler dimension, respectively.

[0071] Thus, by generating orthogonal time-frequency spatial frames in this embodiment, it is possible to achieve the following in the first... q Time-delay-Doppler domain received signal In the middle, due to the sparsity of the channel matrix, the first Each delayed-Doppler domain received symbol is only related to the pilot sequence. Furthermore, because... much smaller In this embodiment, only a small number of pilot symbols are placed in the orthogonal time-frequency space frame, which will significantly reduce pilot overhead.

[0072] In conjunction with any of the above embodiments, in one implementation, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios. In this method, step S15 may specifically include step S51:

[0073] Step S51: The receiving end uses the channel estimation module to perform channel estimation using a portion of the delay-Doppler domain received symbols in the delay-Doppler domain received signal that are only related to the pilot sequence, and the pilot sequence inserted by the transmitting end, to obtain the estimated channel matrix.

[0074] In this embodiment, the receiver can use the channel estimation module to perform channel estimation only on a portion of the delayed-Doppler domain received symbols that are only related to the pilot sequence in the delayed-Doppler domain received signal, as well as the pilot sequence inserted by the transmitter, to obtain the estimated channel matrix. This can effectively eliminate interference from unknown data streams and improve channel estimation performance.

[0075] In conjunction with any of the above embodiments, in one implementation, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios. In this method, the number of delay-Doppler domain received signals is... Q ;No. q The received signal in the time-delay-Doppler domain corresponds to the first... q indivual An orthogonal time-frequency spatial frame of dimension; and, step S51 above may specifically include the following step S61:

[0076] Step S61: The receiving end, through the channel estimation module, utilizes the... q The first time-delay-Doppler domain received signal that is only related to the pilot sequence Position 1 to 2 The time-delay-Doppler domain received symbols at each location, and the transmitting end at the [number]th location. q indivual The pilot sequence inserted into the orthogonal time-frequency space frame of dimension 1 yields the 1st... q An estimated channel matrix.

[0077] Based on the first example generated in this embodiment q indivual orthogonal time-frequency space frame of dimension Due to the sparsity of the channel matrix, it can be seen that the first... q Time-delay-Doppler domain received signal In the middle, the first Position 1 to 2 The time-delay-Doppler domain received symbols at each location are only related to the pilot sequence.

[0078] Thus, in this embodiment, the receiver can utilize the channel estimation module to... q The first time-delay-Doppler domain received signal that is only related to the pilot sequence Position 1 to 2 The delay-Doppler domain received symbol at position n, and the transmitter at position n. q indivual The pilot sequence inserted in the orthogonal time-frequency space frame of dimension (known to the receiver) yields the th... q The estimated channel matrix is ​​thus obtained, thereby avoiding data interference. In an alternative implementation, the first estimated channel matrix can be obtained by the following method. q indivual The channel matrix corresponding to the orthogonal time-frequency space frame of dimension :

[0079] After eliminating data interference, equation (3) becomes:

[0080] (6)

[0081] in, That is, the selected portion of the received signal, i.e., the first... q Time-delay-Doppler domain received signal The first one, which is only related to the pilot sequence Position 1 to 2 Delay-Doppler domain received symbols at each location; For the first q Equivalent channel matrix in DD domain The to Line, 1 to A matrix composed of the elements of each column; For the sending end at the first q indivual The pilot sequence inserted in the orthogonal time-frequency space frame of dimension, i.e. the th q indivual orthogonal time-frequency space frame of dimension The first position to the first Delay-Doppler domain transmission of symbols at each location; For the first q DD domain received noise The The elements up to the first A vector consisting of 10 elements.

[0082] Accordingly, equation (5) is transformed into:

[0083] (7)

[0084] in, For sensing matrix The to A matrix composed of row elements. dimensional vector Only by It consists of 3 non-zero elements, and the non-zero elements are the channel coefficients. ,and ,therefore Given a sparse vector, the channel estimation problem is transformed into a sparse signal recovery problem in compressed sensing, which can be solved using algorithms from compressed sensing. Noise effects are not considered when performing channel estimation.

[0085] In an alternative example, equation (7) can be solved using a low-complexity orthogonal matching pursuit algorithm. This gives the value and index of a non-zero element; the value is then obtained. The size can be obtained through the position index. and .get Then, substituting into equation (2) yields the first... q indivual The channel matrix corresponding to the orthogonal time-frequency space frame of dimension The estimated value is the first... q Given an estimated channel matrix, we can obtain... Q An estimated channel matrix.

[0086] In one embodiment, such as Figure 2 As shown, Figure 2 This is a schematic diagram of a pilot pattern based on channel sparsity, as shown in an embodiment of the present invention. Figure 2 In the middle, the formula highlighted by the double-dotted line at the top represents the above formula (6), in the first... q indivual orthogonal time-frequency space frame of dimension middle, express The to One element, n ∈[1, N ]; in the q Time-delay-Doppler domain received signal middle, express The to One element, Indicates the first q indivual The channel matrix corresponding to the orthogonal time-frequency space frame of dimension The i Line number j Column block matrix.

[0087] So, in the first q indivual orthogonal time-frequency space frame of dimension In the middle, if the front Locations ( , , …, Place pilot signals (pilot sequence), other positions ( to If the data (raw data stream) is released, then in the first... q Time-delay-Doppler domain received signal In the middle, due to the sparsity of the channel matrix, the first A vector composed of symbols It is only related to the pilot sequence (the data portion is all 0). Furthermore, due to... much smaller Only a small number of pilot symbols are placed in the transmitted frame (orthogonal time-frequency space frame), which can significantly reduce pilot overhead.

[0088] In combination with the above Figure 2 In one embodiment, a specific example is given: such as Figure 3 As shown, Figure 3 This is a schematic diagram illustrating a specific example of a pilot pattern based on channel sparsity, according to an embodiment of the present invention. Figure 3 In the middle, the largest Doppler tap In the q indivual orthogonal time-frequency space frame of dimension In the middle, if the front Locations ( , and ) Place pilot signals, other locations ( to ) Put the data, then in the first... q Time-delay-Doppler domain received signal In the middle, due to the sparsity of the channel matrix, the first A vector composed of symbols It is only related to the pilot symbol.

[0089] In conjunction with any of the above embodiments, in one implementation, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios. In this method, S21 specifically includes steps S71 and S72:

[0090] Step S71: The receiving end constructs a low-dimensional channel matrix based on the estimated channel matrix.

[0091] In this embodiment, the channel matrix estimated by direct feedback (i.e., the high-dimensional channel matrix) is taken into consideration. This would cause significant feedback overhead. Therefore, in this embodiment, the channel matrix is ​​determined based on the above equation (1). The key variables include only: p Delay tap size of the path , No. p Doppler tap size of the path and the p Complex gain of a path Therefore, this embodiment only needs to provide a small amount of feedback information: In this embodiment, a low-dimensional channel matrix is ​​constructed based on the obtained estimated channel matrix. In one alternative embodiment, the constructed low-dimensional channel matrix... line, number The column elements are the magnitudes of the complex gain for the corresponding path, and the path delay tap size is... Doppler tap size is .

[0092] In an alternative example, the low-dimensional channel matrix can be obtained using the following equation (8):

[0093] (8);

[0094] in, For the first p Complex gain of the path; The channel matrix is ​​of low dimension. The line, number Column elements; Indicates a total of One path; other representations: .

[0095] Step S72: Feed back the low-dimensional channel matrix to the semantic encoder.

[0096] In this embodiment, only the low-dimensional channel matrix can be fed back to the semantic encoder as prior information for semantic coding, thereby reducing the feedback overhead without affecting the semantic coding adaptation.

[0097] In conjunction with any of the above embodiments, in one implementation, the present invention also provides a joint semantic coding-transmission method for high-mobility scenarios. In this method, the joint semantic coding-transmission method for high-mobility scenarios is implemented through a joint semantic coding-transmission model for high-mobility scenarios; the training process of the joint semantic coding-transmission model for high-mobility scenarios may include steps S81 to S85:

[0098] Step S81: Construct the joint semantic coding-transmission model to be trained based on the semantic encoder to be trained, the semantic decoder to be trained, the channel encoder to be trained, the channel decoder to be trained, the constellation modulator to be trained, the constellation demodulator to be trained, the orthogonal time-frequency space modulator, and the orthogonal time-frequency space demodulator.

[0099] In this embodiment, a joint semantic coding-transmission model to be trained can be constructed based on the semantic encoder to be trained, the semantic decoder to be trained, the channel encoder to be trained, the channel decoder to be trained, the constellation modulator to be trained, the constellation demodulator to be trained, the orthogonal time-frequency space modulator, and the orthogonal time-frequency space demodulator.

[0100] Step S82: The transmitting end uses the original sample data to obtain the time-domain sample transmission signal through the semantic encoder to be trained, the channel encoder to be trained, the constellation modulator to be trained, and the orthogonal time-frequency spatial modulator.

[0101] In this embodiment, the transmitting end can extract semantics from the original sample data using a semantic encoder to be trained, thereby obtaining the original semantic features of the sample; then, the original semantic features of the sample are processed sequentially by a channel encoder to be trained and a constellation modulator to be trained, thereby obtaining the original sample data stream; finally, the original sample data stream is mapped onto the time-delay-Doppler domain resource grid by an orthogonal time-frequency spatial modulator, and a sample pilot sequence is inserted to obtain the sample orthogonal time-frequency spatial frame, and the sample orthogonal time-frequency spatial frame is modulated to obtain the time-domain sample transmission signal.

[0102] Step S83: The transmitting end sends the time-domain sample transmission signal to the receiving end through a channel in a high-mobility scenario between itself and the receiving end.

[0103] In this embodiment, the transmitting end can send the obtained time-domain sample transmission signal to the receiving end through the channel in a high-mobility scenario between the transmitting end and the receiving end.

[0104] Step S84: The receiving end processes the received time-domain sample received signal through the orthogonal time-frequency spatial demodulator, the constellation demodulator to be trained, the channel decoder to be trained, and the semantic decoder to be trained to obtain sample reconstruction data.

[0105] In this embodiment, after the receiving end receives the time-domain sample received signal through the channel in a high-mobility scenario, it can demodulate the received time-domain sample received signal using the orthogonal time-frequency spatial demodulator to obtain the sample delay-Doppler domain received signal. Then, using the orthogonal time-frequency spatial demodulator, channel estimation is performed using the sample delay-Doppler domain received signal to obtain the estimated sample channel matrix. Signal detection is then performed using the estimated sample channel matrix and the sample delay-Doppler domain received signal to obtain the sample reconstruction data stream. This sample reconstruction data stream is then sequentially passed through a constellation demodulator to be trained and a channel decoder to be trained to obtain the sample reconstruction semantic features. Finally, the sample reconstruction semantic features are processed by the semantic decoder to be trained to obtain the sample reconstruction data.

[0106] It should be noted that the implementation details of steps S82 to S84 in the model training process are the same as or similar to the implementation details of the model application process in any of the foregoing embodiments, and will not be repeated here.

[0107] Step S85: Based on the difference between the reconstructed sample data and the original sample data, update the model parameters of the joint semantic coding-transmission model to be trained, and obtain the joint semantic coding-transmission model for high mobility scenarios, which consists of a trained semantic encoder, a trained semantic decoder, a trained channel encoder, a trained channel decoder, a trained constellation modulator, a trained constellation demodulator, the orthogonal time-frequency space modulator, and the orthogonal time-frequency space demodulator.

[0108] In this embodiment, the difference between the reconstructed sample data and the original sample data can be determined. Based on the difference between the reconstructed sample data and the original sample data, with the goal of minimizing the difference between the reconstructed sample data and the original sample data, the model parameters of the joint semantic coding-transmission model to be trained are updated. That is, the model parameters of the semantic encoder, semantic decoder, channel encoder, channel decoder, constellation modulator, constellation demodulator, orthogonal time-frequency space modulator, and orthogonal time-frequency space demodulator to be trained are updated to obtain a joint semantic coding-transmission model for high mobility scenarios. The joint semantic coding-transmission model for high mobility scenarios includes: a trained semantic encoder, a trained semantic decoder, a trained channel encoder, a trained channel decoder, a trained constellation modulator, a trained constellation demodulator, an orthogonal time-frequency space modulator, and an orthogonal time-frequency space demodulator.

[0109] In one embodiment, such as Figure 4 As shown, Figure 4 This is a schematic diagram illustrating an integrated semantic information transmission architecture based on orthogonal time-frequency spatial modulation for high-mobility scenarios, as shown in one embodiment of the present invention. In this embodiment, considering the significant Doppler shift and other challenges faced by high-mobility scenarios, which severely affect the reliability and stability of semantic information transmission, an integrated semantic information transmission architecture based on orthogonal time-frequency spatial modulation is designed, taking into account the channel characteristics of high-mobility scenarios. Specifically, based on the relevant semantic communication system framework, orthogonal time-frequency spatial modulation technology is introduced to effectively suppress the Doppler effect. Through the coordinated optimization of semantic coding and transmission waveform, efficient and reliable transmission of semantic information in high-mobility scenarios is achieved.

[0110] exist Figure 4 The specific functions and modules of the proposed semantic information integrated transmission architecture include: (1) Semantic encoding and decoding: At the sending end, a neural network is used to extract the meaning of the original data (i.e., the source S), and it is encoded to obtain semantic features A, reducing data redundancy, retaining core semantic features, and improving transmission efficiency. At the receiving end, the obtained reconstructed semantic features are processed... Decode and recover the original data to obtain the destination. (2) Semantic transmission waveform (i.e., including semantic transmission waveform and semantic reception waveform): including channel coding, channel decoding, constellation modulation, constellation demodulation, orthogonal time-frequency space modulation, orthogonal time-frequency space demodulation, etc. (3) Pilot design and channel estimation: the transmitter designs pilot patterns, and the receiver recovers channel state information from the received signal for the detection of semantic information symbols. (4) Channel feedback: the channel state information estimated by the receiver is fed back to the transmitter as the input of the semantic encoder, so that the semantic encoder can make adaptive adjustments according to the channel state information and realize channel state-aware semantic coding.

[0111] In summary, this invention addresses the Doppler effect in high-mobility scenarios by designing an adapted OTFS-based semantic information integrated transmission architecture (i.e., the aforementioned semantic communication system), proposing a channel-adaptive semantic coding method, and further proposing a joint optimization method for semantic coding, channel coding, constellation modulation, and OTFS modulation. Compared with related technologies, this invention has the following technical advantages:

[0112] (1) Currently, related semantic communication systems based on separate coding and transmission architecture lack environmental adaptability and are difficult to cope with dynamic and complex transmission environments. Compared with the related semantic communication systems based on separate coding and transmission architecture, the semantic information integrated transmission architecture and waveform proposed in this invention fully consider the dynamic and varied channel characteristics in high mobility scenarios, reveal the transmission evolution law of semantic features in complex environments, and jointly optimize semantic coding and transmission waveforms, providing a brand-new solution for efficient and reliable transmission of massive data in high mobility scenarios.

[0113] (2) Currently, high-quality transmission of semantic information depends on the accurate detection of semantic features by the receiver, which requires accurate channel state information. However, there is an inherent contradiction between the accuracy and computational complexity of traditional channel estimation algorithms, making it difficult to achieve both high accuracy and low complexity at the same time. Compared with related channel estimation schemes, the channel estimation method proposed in this invention can reduce the complexity of channel estimation, reduce pilot overhead, and ensure the accuracy of channel estimation and signal detection, thus guaranteeing the recovery performance of semantic features.

[0114] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0115] Based on the same inventive concept, one embodiment of the present invention provides a joint semantic coding-transmission apparatus for high-mobility scenarios. (Reference) Figure 5 , Figure 5 This is a structural block diagram of a joint semantic coding-transmission device for high-mobility scenarios provided in an embodiment of the present invention. Figure 5 As shown, the device includes:

[0116] The first processing module, deployed at the sending end of the semantic communication system, is used to extract the original semantic features of the original data and obtain the original data stream based on the original semantic features;

[0117] The second processing module, deployed at the transmitting end, is used to map the original data stream onto the time-delay-Doppler domain resource grid through an orthogonal time-frequency spatial modulator, insert pilot sequences to obtain orthogonal time-frequency spatial frames, and modulate the orthogonal time-frequency spatial frames to obtain a time-domain transmission signal.

[0118] The first transmitting module, deployed at the transmitting end, is used to transmit the time-domain transmitting signal to the receiving end through a channel in a high-mobility scenario between the transmitting end and the receiving end in the semantic communication system.

[0119] The third processing module, deployed at the receiving end, is used to demodulate the received time-domain received signal through an orthogonal time-frequency spatial demodulator to obtain a time-delay-Doppler domain received signal.

[0120] A channel estimation module, deployed at the receiving end, is used to perform channel estimation using the time-delay-Doppler domain received signal to obtain the estimated channel matrix.

[0121] The first reconstruction module, deployed at the receiving end, is used to obtain a reconstructed data stream using the estimated channel matrix and the time-delay-Doppler domain received signal, and to obtain reconstructed data based on the reconstructed data stream.

[0122] Optionally, the device further includes:

[0123] A channel feedback module is used by the receiving end to feed back the estimated channel matrix to the semantic encoder in the transmitting end;

[0124] The semantic extraction module is used by the transmitting end to extract the semantics of the next raw data through the semantic encoder, based on the estimated channel matrix as a priori, to obtain the original semantic features of the next raw data.

[0125] Optionally, the second processing module includes:

[0126] The first generation module is used to generate a pilot pattern with the goal of making a portion of the time-delay-Doppler domain received symbols in the time-delay-Doppler domain received signal only related to the pilot sequence.

[0127] The second generation module is used to map the original data stream onto the time-delay-Doppler domain resource grid according to the pilot pattern, and insert pilot sequences to obtain orthogonal time-frequency spatial frames.

[0128] Optionally, the second generation module includes:

[0129] The third generation module is used to generate signals from the first position to the second position according to the pilot pattern. The pilot sequence is inserted at the nth position, and in the remaining nth position... Position 1 to 2 Insert the original data stream at the nth position to obtain the nth position. q indivual orthogonal time-frequency space frames of dimension; Indicates the maximum Doppler tap. and These represent the number of symbols for time delay and Doppler dimension, respectively.

[0130] Optionally, the channel estimation module includes:

[0131] The channel estimation processing module uses a portion of the delay-Doppler domain received symbols in the delay-Doppler domain received signal that are only related to the pilot sequence, as well as the pilot sequence inserted by the transmitting end, to perform channel estimation and obtain the estimated channel matrix.

[0132] Optionally, the number of time-delay-Doppler domain received signals is Q ;No. q The received signal in the time-delay-Doppler domain corresponds to the first... q indivual An orthogonal time-frequency spatial frame of dimension 1; the channel estimation processing module includes:

[0133] The fourth generation module is used to, through the channel estimation module, utilize the... q The first time-delay-Doppler domain received signal that is only related to the pilot sequence Position 1 to 2 The time-delay-Doppler domain received symbols at each location, and the transmitting end at the [number]th location. q indivual The pilot sequence inserted into the orthogonal time-frequency space frame of dimension 1 yields the 1st... q An estimated channel matrix.

[0134] Optionally, the channel feedback module includes:

[0135] A construction module is used by the receiving end to construct a low-dimensional channel matrix based on the estimated channel matrix;

[0136] The information feedback module is used to feed back the low-dimensional channel matrix to the semantic encoder.

[0137] Optionally, the first of the low-dimensional channel matrix line, number The column elements are the magnitudes of the complex gain for the corresponding path, and the path delay tap size is... Doppler tap size is .

[0138] Optionally, the joint semantic coding-transfer device for high-mobility scenarios is implemented using a joint semantic coding-transfer model for high-mobility scenarios; the device further includes: a training module for executing the training process of the joint semantic coding-transfer model for high-mobility scenarios; the training module includes:

[0139] The model determination module is used to construct the joint semantic coding-transmission model to be trained based on the semantic encoder, semantic decoder, channel encoder, channel decoder, constellation modulator, constellation demodulator, orthogonal time-frequency space modulator, and orthogonal time-frequency space demodulator to be trained.

[0140] The first signal generation module, deployed at the transmitting end, is used to obtain a time-domain sample transmission signal using the original sample data through the semantic encoder to be trained, the channel encoder to be trained, the constellation modulator to be trained, and the orthogonal time-frequency spatial modulator.

[0141] The second transmitting module, deployed at the transmitting end, is used to transmit the time-domain sample transmission signal to the receiving end through a channel in a high-mobility scenario between the transmitting end and the receiving end.

[0142] The second reconstruction module, deployed at the receiving end, is used to process the received time-domain sample received signal through the orthogonal time-frequency spatial demodulator, the constellation demodulator to be trained, the channel decoder to be trained, and the semantic decoder to be trained, to obtain sample reconstruction data.

[0143] The parameter update module is used to update the model parameters of the joint semantic coding-transmission model to be trained based on the difference between the sample reconstructed data and the sample original data, so as to obtain the joint semantic coding-transmission model for high mobility scenarios, which consists of the trained semantic encoder, trained semantic decoder, trained channel encoder, trained channel decoder, trained constellation modulator, trained constellation demodulator, the orthogonal time-frequency space modulator and the orthogonal time-frequency space demodulator.

[0144] Based on the same inventive concept, another embodiment of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the joint semantic coding-transmission method for high-mobility scenarios as described in any of the above embodiments of the present invention.

[0145] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, such as... Figure 6 As shown, Figure 6This is a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When executed by the processor, the program implements the steps of the joint semantic coding-transmission method for high-mobility scenarios described in any of the above embodiments of the present invention.

[0146] Based on the same inventive concept, another embodiment of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the steps in the joint semantic coding-transmission method for high-mobility scenarios described in any of the above embodiments of the present invention.

[0147] As the apparatus embodiment is basically similar to the method embodiment, it is described in a relatively simple manner. For relevant details, please refer to the description of the method embodiment.

[0148] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0149] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0150] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0151] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0152] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0153] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0154] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device 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 terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0155] The foregoing has provided a detailed description of a joint semantic coding-transmission method for high-mobility scenarios. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A joint semantic coding-transmission method for high-mobility scenarios, characterized in that, The method includes: In a semantic communication system, the sending end extracts the original semantic features of the original data and obtains the original data stream based on the original semantic features. The transmitting end maps the original data stream onto the time-delay-Doppler domain resource grid through an orthogonal time-frequency spatial modulator, inserts a pilot sequence to obtain an orthogonal time-frequency spatial frame, and modulates the orthogonal time-frequency spatial frame to obtain a time-domain transmission signal; The transmitting end sends the time-domain transmission signal to the receiving end through a channel in a high-mobility scenario between itself and the receiving end in the semantic communication system; The receiving end demodulates the received time-domain received signal through an orthogonal time-frequency spatial demodulator to obtain a time-delay-Doppler domain received signal; The receiving end uses the time-delay-Doppler domain received signal to perform channel estimation through the channel estimation module, and obtains the estimated channel matrix; The receiving end uses the estimated channel matrix and the time-delay-Doppler domain to receive the signal, obtains the reconstructed data stream, and obtains the reconstructed data based on the reconstructed data stream; The number of orthogonal time-frequency spatial frames is The transmitting end maps the original data stream onto a time-delay-Doppler domain resource grid using an orthogonal time-frequency spatial modulator, and inserts pilot sequences to obtain an orthogonal time-frequency spatial frame, including: The pilot pattern is generated with the goal of making a portion of the time-delay-Doppler domain received symbols in the time-delay-Doppler domain received signal related only to the pilot sequence; According to the pilot pattern, from the first position to the second position... The pilot sequence is inserted at the nth position, and in the remaining nth position... Position 1 to 2 Insert the original data stream at the nth position to obtain the nth q indivual orthogonal time-frequency space frames of dimension; Indicates the maximum Doppler tap. and These represent the number of symbols for time delay and Doppler dimension, respectively. q ∈[1, Q ],and q and Q It is an integer.

2. The joint semantic coding-transmission method for high-mobility scenarios according to claim 1, characterized in that, The method further includes: The receiver feeds back the estimated channel matrix to the semantic encoder in the transmitter; The transmitting end uses the estimated channel matrix as a priori and extracts the semantics of the next raw data through the semantic encoder to obtain the original semantic features of the next raw data.

3. The joint semantic coding-transmission method for high-mobility scenarios according to claim 1, characterized in that, The receiving end uses the time-delay-Doppler domain received signal to perform channel estimation through the channel estimation module, obtaining the estimated channel matrix, including: The receiving end uses the channel estimation module to perform channel estimation using a portion of the delay-Doppler domain received symbols in the delay-Doppler domain received signal that are only related to the pilot sequence, and the pilot sequence inserted by the transmitting end, to obtain the estimated channel matrix.

4. The joint semantic coding-transmission method for high-mobility scenarios according to claim 3, characterized in that, The number of received signals in the time-delay-Doppler domain is Q ;No. q The received signal in the time-delay-Doppler domain corresponds to the first... q indivual The receiver, through the channel estimation module, uses a portion of the time-delay-Doppler domain received symbols in the time-delay-Doppler domain received signal that are only related to the pilot sequence, and the pilot sequence inserted by the transmitter, to perform channel estimation and obtain the estimated channel matrix, including: The receiving end, through the channel estimation module, utilizes the first... q The first time-delay-Doppler domain received signal that is only related to the pilot sequence Position 1 to 2 The time-delay-Doppler domain received symbols at each location, and the transmitting end at the [number]th location. q indivual The pilot sequence inserted into the orthogonal time-frequency space frame of dimension 1 yields the 1st... q An estimated channel matrix.

5. The joint semantic coding-transmission method for high-mobility scenarios according to claim 2, characterized in that, The receiver feeds back the estimated channel matrix to the semantic encoder in the transmitter, including: The receiving end constructs a low-dimensional channel matrix based on the estimated channel matrix; The low-dimensional channel matrix is ​​fed back to the semantic encoder.

6. The joint semantic coding-transmission method for high-mobility scenarios according to claim 5, characterized in that, The low-dimensional channel matrix of the first line, number The column elements are the magnitudes of the complex gain for the corresponding path, and the path delay tap size is... Doppler tap size is .

7. The joint semantic coding-transmission method for high-mobility scenarios according to any one of claims 1 to 6, characterized in that, The joint semantic coding-transfer method for high-mobility scenarios is implemented through a joint semantic coding-transfer model for high-mobility scenarios; the training process of the joint semantic coding-transfer model for high-mobility scenarios includes: Based on the semantic encoder, semantic decoder, channel encoder, channel decoder, constellation modulator, constellation demodulator, orthogonal time-frequency space modulator, and orthogonal time-frequency space demodulator to be trained, a joint semantic coding-transmission model to be trained is constructed. The transmitting end uses the original sample data to obtain the time-domain sample transmission signal through the semantic encoder to be trained, the channel encoder to be trained, the constellation modulator to be trained, and the orthogonal time-frequency spatial modulator. The transmitting end transmits the time-domain sample transmission signal to the receiving end through a channel in a high-mobility scenario between itself and the receiving end; The receiving end processes the received time-domain sample signal through the orthogonal time-frequency spatial demodulator, the constellation demodulator to be trained, the channel decoder to be trained, and the semantic decoder to be trained to obtain sample reconstruction data. Based on the difference between the reconstructed sample data and the original sample data, the model parameters of the joint semantic coding-transmission model to be trained are updated to obtain the joint semantic coding-transmission model for high mobility scenarios, which consists of a trained semantic encoder, a trained semantic decoder, a trained channel encoder, a trained channel decoder, a trained constellation modulator, a trained constellation demodulator, the orthogonal time-frequency space modulator, and the orthogonal time-frequency space demodulator.

Images