A digital coding method and system for medium wave transmitter

Abstract

The present invention discloses a digital coding method and system for a medium wave transmitter, which relates to the field of digital coding technology, including: converting an input analog audio signal into a digital signal, extracting Mel-frequency cepstral coefficient features, instantaneous energy features and frequency domain features, and generating a composite feature vector; dividing the digital signal into a core layer, a secondary layer and a background layer, and performing layered differential coding and dynamic encryption; adopting a multi-antenna system in combination with a space-time coding scheme to perform spatial precoding, combining historical channel parameters and the current signal composite feature vector to predict the fading depth, switching the modulation mode and adjusting the transmission power when the depth threshold is exceeded; generating a signal copy for each layer respectively, and if the bit error rate exceeds the bit error rate threshold, borrowing resources from adjacent subcarriers, performing phase rotation on the signal copies of each layer and superimposing them to form a transmission signal, which is sent to the medium wave channel through the antenna, thereby significantly improving the accuracy, security, stability and anti-interference capability of signal transmission.

Description

Technical Field

[0001] The present invention relates to the field of digital coding technology, in particular to a digital coding method and system for a medium wave transmitter. Background Art

[0002] As the core equipment for broadcast signal transmission, medium wave transmitters play a vital role in the field of communications. With the rapid development of digital technology, medium wave transmitters have also undergone a transformation from analog modulation technology to digital modulation technology. Digital technology converts audio signals into digital signals for processing and transmission, which significantly improves the accuracy and stability of the signal and reduces distortion and interference.

[0003] However, although digital technology has provided strong support for the performance improvement of medium wave transmitters, it still has limitations in signal feature extraction. Traditional signal feature extraction methods can often only capture partial characteristics of the signal, such as time domain characteristics or frequency domain characteristics, and it is difficult to fully reflect the complexity and diversity of the signal. Although some feature extraction methods based on statistics or information theory can extract more signal features, they often have high computational complexity and poor real-time performance, which makes it difficult to meet the real-time and accuracy requirements of medium wave transmitters. This limitation may lead to the loss of key information during the signal processing process, thereby affecting the accuracy of subsequent analysis and processing.

[0004] In addition, the lack of adaptability of traditional coding methods to channels is also a major problem faced by existing technologies. Traditional coding methods often adopt a unified coding strategy and cannot perform differentiated processing based on the importance of different signals. As a result, during signal transmission, critical information may not be adequately protected, while non-critical information may occupy too many transmission resources. Traditional medium-wave transmitters often use fixed modulation methods and transmission powers, which cannot be dynamically adjusted according to channel quality. When the channel quality is poor, this fixed modulation method and transmission power may cause the reliability and efficiency of signal transmission to be seriously affected. Summary of the Invention

[0005] (1) Technical problems solved

[0006] In response to the shortcomings of the existing technology, the present invention provides a digital coding method and system for a medium wave transmitter. By digitally extracting Mel-frequency cepstral coefficient characteristics, instantaneous energy characteristics and frequency domain characteristics, integrating dimensionality reduction to generate a composite feature vector, signal layered differentiated coding and dynamic encryption are realized based on information entropy. The modulation mode and transmission power are dynamically adjusted by combining multi-antenna space-time coding and channel fading prediction, and redundant control and dynamic resource borrowing mechanisms are utilized to effectively solve the problems of incomplete signal feature extraction, rigid coding strategies that cannot adapt to differences in signal importance, and poor channel adaptability in traditional technologies, thereby significantly improving the accuracy, security, stability and anti-interference capability of signal transmission.

[0007] (2) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: A digital encoding method for a medium wave transmitter, comprising:

[0009] Convert the input analog audio signal into a digital signal, extract the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generate a composite feature vector after splicing, fusion, and dimensionality reduction.

[0010] The digital signal is divided into a core layer, a secondary layer, and a background layer according to the information entropy of the composite eigenvector. Huffman-arithmetic hybrid coding is performed on the core layer, Huffman coding is performed on the secondary layer, and wavelet compression coding is performed on the background layer. A key is generated and the coded data of each layer is encrypted separately. The encrypted data is grouped and traceability tags are added.

[0011] The coded data of each layer is interleaved at preset time intervals, and the available frequency band is divided into unequally spaced subcarrier groups. A multi-antenna system combined with a space-time coding scheme is used for spatial precoding. The fading depth is predicted by combining historical channel parameters and the current signal composite eigenvector. When the fading depth exceeds a depth threshold, the modulation mode is switched and the transmit power is adjusted.

[0012] Redundant encoding is performed on each layer to generate a signal copy, and the subcarrier bit error rate is monitored. If the bit error rate exceeds the bit error rate threshold, resources are dynamically borrowed from adjacent subcarriers, and the signal copies of each layer are phase rotated and superimposed to form a transmission signal, which is sent to the medium wave channel through the antenna.

[0013] Furthermore, several Mel filter banks are set to calculate the Mel frequency cepstral coefficients, extract the Mel frequency cepstral coefficient features, and use the discrete Teager energy operator to operate on the digital signal to extract the instantaneous energy features;

[0014] The digital signal is modulated to the medium wave carrier frequency band through up-conversion, and the medium wave carrier frequency band is divided into multiple unequal width sub-bands. For each sub-band, the stopband attenuation parameters of the bandpass filter are set to perform interference filtering, and the filtered carrier signal is subjected to short-time Fourier transform to obtain frequency domain characteristics;

[0015] The Mel-frequency cepstral coefficient features, instantaneous energy features and frequency domain features are spliced ​​and fused to form an initial high-dimensional feature matrix. The principal component analysis is used to reduce the dimension of the initial high-dimensional feature matrix, and the linear discriminant analysis is used to further reduce the dimension to obtain a composite feature vector.

[0016] Furthermore, the division into multiple unequal width sub-bands includes:

[0017] Perform kernel density estimation on the historical interference intensity data within the frequency band to obtain the spatial distribution density function of the interference intensity;

[0018] The local minimum point of the density function is used as the candidate point of the sub-band boundary, and only the candidate points that meet the interference intensity less than the interference intensity threshold are retained as boundary points.

[0019] Furthermore, the information entropy of the composite feature vector is calculated, and the digital signal is divided into the core layer, the secondary layer and the background layer according to the size of the information entropy. , secondary layer , background layer , is the information entropy threshold, H represents information entropy;

[0020] The core layer is pre-coded using Huffman coding, and the pre-coded result is encoded into a real number in the interval [0,1) using arithmetic coding. The secondary layer is pre-coded using Huffman coding, and the background layer is pre-coded using Daubechies4 wavelet basis.

[0021] Furthermore, a true random number generator is used to generate two sets of keys. The long key is used for AES encryption of core layer signals, and the short key is used for AES encryption of secondary layer and background layer signals. The two sets of keys are regenerated after each transmission of a preset number of data packets of preset length.

[0022] A traceability tag based on a hash chain is added to each encrypted encoded data packet. The hash value of the previous packet is calculated using the SHA-256 hash algorithm. The tag contains the hash value of the previous packet, the current packet sequence number, and the key version number.

[0023] Furthermore, the coded data of the core layer, secondary layer, and background layer are interleaved at a time interval of 1:2:1. The available frequency band is divided into unequally spaced subcarrier groups using the golden section. The number of subcarriers is determined according to the available bandwidth and subcarrier frequency resolution: ,in, represents the number of subcarriers, B represents the available bandwidth, Indicates the subcarrier frequency resolution;

[0024] A multi-antenna system with a preset number of antennas is used, and spatial precoding is performed using the Alamouti space-time coding scheme. The coded data is divided into two groups and transmitted through the antennas in two consecutive symbol periods. m and antennas m +1 is emitted in the following mode: ,in, 、 Represents the coded data symbol, represents the complex conjugate.

[0025] Furthermore, a long short-term memory network neural network model is constructed, which includes an input layer, a hidden layer, and an output layer. The input layer receives historical channel parameters and the current signal composite feature vector. The historical channel parameters include signal-to-noise ratio, fading depth, and multipath delay. The hidden layer is set with two LSTM layers, each with 128 neurons. The activation function uses the ReLU function. The output layer predicts the channel fading depth within a preset time in the future.

[0026] Furthermore, if the subcarrier bit error rate exceeds the bit error rate threshold, the dynamic redundancy allocation mechanism is activated to select a subcarrier with a bit error rate lower than 1 / 10 of the bit error rate threshold from adjacent subcarriers, and borrow resources according to a preset ratio of the current transmission rate of the interfered subcarrier. The resource borrowing ratio formula is: ,in, α Indicates the preset ratio, BER represents the subcarrier bit error rate, Indicates the bit error rate threshold.

[0027] Furthermore, the signal replica generated by each coding layer is phase rotated, and the phase offset is: ,in, n is the subcarrier number, m is the antenna number, is the total number of subcarriers, M is the number of antennas, phase rotation is performed on different copies, and the rotated copy signals are superimposed through an adder to form the final coded signal for transmission, which is sent to the medium wave channel through the transmitting antenna.

[0028] A medium wave transmitter digital coding system, comprising:

[0029] The signal preprocessing module converts the input analog audio signal into a digital signal, extracts the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generates a composite feature vector after splicing, fusion, and dimensionality reduction.

[0030] The layered coding module divides the digital signal into a core layer, a secondary layer, and a background layer based on the information entropy of the composite eigenvector. It performs Huffman-arithmetic hybrid coding on the core layer, Huffman coding on the secondary layer, and wavelet compression coding on the background layer. It generates a key and encrypts the coded data of each layer separately. It groups the encrypted data and adds traceability tags.

[0031] The signal modulation module interleaves the coded data of each layer at preset time intervals, divides the available frequency band into unequally spaced subcarrier groups, uses a multi-antenna system combined with a space-time coding scheme for spatial precoding, combines historical channel parameters and the current signal composite eigenvector to predict the fading depth, and switches the modulation mode and adjusts the transmit power when the fading depth exceeds the depth threshold;

[0032] The redundancy control module performs redundant encoding on each layer to generate signal copies and monitors the subcarrier bit error rate. If the bit error rate exceeds the bit error rate threshold, it dynamically borrows resources from adjacent subcarriers, performs phase rotation on the signal copies of each layer, and superimposes them to form a transmission signal, which is sent to the medium wave channel through the antenna.

[0033] (3) Beneficial effects

[0034] The present invention provides a digital coding method and system for a medium wave transmitter, which has the following beneficial effects:

[0035] (1) By digitizing the analog audio signal and extracting multi-dimensional features, a composite feature vector is generated through splicing, fusion and dimensionality reduction. This can not only fully capture the key characteristics of the signal, such as the Mel frequency characteristics that are consistent with human hearing, but also filter out interference and reduce data dimensions, laying an efficient and accurate feature foundation for subsequent layered coding, modulation and other processing.

[0036] (2) The signal levels are divided according to the information entropy of the composite eigenvector, and hybrid coding, Huffman coding and wavelet compression coding are used for the core layer, secondary layer and background layer respectively. Combined with dynamic key encryption and hash chain traceability tags, it not only achieves differentiated and efficient compression of signals of different importance, but also improves the security and traceability of data transmission through encryption and traceability mechanisms.

[0037] (3) By interleaving the coded data of each layer in proportion and dividing it into unequally spaced subcarrier groups, combined with multi-antenna space-time coding and channel fading prediction, the modulation mode and transmission power can be dynamically adjusted in complex channel environments, effectively improving spectrum utilization, enhancing signal anti-fading capabilities, and ensuring transmission stability and reliability.

[0038] (4) By redundantly encoding the signals of each layer to generate copies, combining the subcarrier bit error rate monitoring to dynamically borrow adjacent subcarrier resources, and then performing phase rotation and superposition processing, resources can be adaptively allocated when channel interference occurs, effectively reducing the bit error rate, improving the signal's anti-interference ability, and ensuring the stability and integrity of the medium wave channel transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a schematic diagram of the steps of the digital encoding method for a medium wave transmitter of the present invention;

[0040] Figure 2 Schematic diagram of the coded signal preprocessing flow chart of the present invention;

[0041] Figure 3 This is a structural diagram of the digital coding system of the medium wave transmitter of the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] See also Figure 1 and Figure 2 The present invention provides a digital coding method for a medium wave transmitter, comprising the following steps:

[0044] Step 1: Convert the input analog audio signal into a digital signal, extract the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generate a composite feature vector after splicing, fusion, and dimensionality reduction.

[0045] The step 1 includes the following contents:

[0046] Step 101: Convert the input analog audio signal into a digital signal at a preset sampling rate and quantization accuracy that meet broadcast-grade audio quality requirements. The preset sampling rate is no less than twice the highest frequency of the audio signal, and the quantization accuracy is no less than 12 bits to ensure the signal dynamic range. For example, a 48kHz sampling rate (corresponding to a 20kHz audio bandwidth) and 16-bit quantization accuracy (dynamic range 96dB) can be used to achieve digital conversion.

[0047] Step 102: Mel-frequency cepstral coefficient (MFCC) extraction is performed in parallel. Several Mel-frequency cepstral coefficients (e.g., 24) are set to calculate MFCC coefficients (e.g., 13th order) to capture MFCC features. The digital signal is operated using a discrete Teager energy operator: ,in,E [ x ( t )] is a digital signal x ( t ) at the time t The instantaneous energy, x ( t ) is the digital signal at time t The sampling value of is used to extract the instantaneous energy characteristics of the signal;

[0048] It should be noted that the Mel filter bank is designed based on the human auditory characteristics. It converts linear frequencies into Mel frequencies, making the filter denser in the low-frequency band and sparser in the high-frequency band, which is more in line with human auditory perception. The 13th-order MFCC coefficient can capture the spectral envelope characteristics of the audio signal and reflect information such as the timbre and pitch of the sound.

[0049] The Teager energy operator is a nonlinear signal processing tool used to capture instantaneous energy changes in signals. It is particularly adept at analyzing non-stationary signals (such as speech, biomedical signals, and mechanical vibrations). By calculating the product of a signal and its derivative, it can extract the instantaneous energy characteristics of the signal. It is sensitive to sudden changes in signals such as plosives and fricatives in speech, and can effectively capture the dynamic changes and energy distribution of the signal.

[0050] Step 103: Up-convert the digital signal to the medium wave carrier frequency band (526.5kHz-1606.5kHz). The medium wave carrier frequency band is divided into multiple unequal width sub-bands using a density peak-based clustering algorithm based on the probability density distribution data of historical interference intensity. Specifically, the following steps are performed:

[0051] Perform kernel density estimation on the historical interference intensity data within the frequency band to obtain the spatial distribution density function of the interference intensity;

[0052] The local minimum point of the density function is used as the candidate point of the sub-band boundary, and only the candidate points that meet the interference intensity less than the interference intensity threshold (such as 0.5) are retained as boundary points;

[0053] For example, the medium wave carrier frequency band is divided into 8 sub-bands by the above method: S_1: 526.5-650kHz, corresponding to the density valley point of the high-incidence area of ​​industrial electromagnetic interference, S_2: 650-780kHz, corresponding to the boundary point of the broadcast intermodulation interference band, S_3: 780-900kHz, based on the inflection point of the atmospheric noise spectrum distribution, S_4: 900-1050kHz, matching the regional electromagnetic interference source band, S_5: 1050-1200kHz, corresponding to the power system harmonic characteristic frequency band, S_6: 1200-1350kHz, based on the boundary of the sky wave and ground wave interference mixed area, S_7: 1350-1450kHz, corresponding to the ionospheric interference characteristic frequency point, S_8: 1450-1606.5kHz, covering the main distribution area of ​​high-frequency background noise;

[0054] For each sub-band, according to the maximum historical interference intensity, set the bandpass filter stopband attenuation parameters for interference filtering: ,in, Indicates the maximum historical interference intensity, represents the interference intensity threshold, is the stopband attenuation parameter of the bandpass filter. This parameter is used to dynamically adjust the filter to specifically filter out common interference signals in each sub-band, improve the purity of the signal, and obtain the filtered carrier signal. The filtered carrier signal is subjected to short-time Fourier transform (STFT) to obtain its frequency domain characteristics.

[0055] Step 104: The MFCC features, instantaneous energy features, and frequency domain features are concatenated and fused to form an initial high-dimensional feature matrix. Principal component analysis (PCA) is used to reduce the dimensionality of the feature matrix. The specific steps are as follows:

[0056] Perform standardization preprocessing on the initial high-dimensional feature matrix to eliminate the dimension effect;

[0057] Calculate the covariance matrix of the initial high-dimensional features, perform eigendecomposition on the covariance matrix, obtain the eigenvalues ​​and corresponding eigenvectors, and sort them from large to small according to the eigenvalues;

[0058] Before selection k The eigenvectors form a transformation matrix so that the cumulative variance contribution rate reaches a preset contribution rate threshold (for example, 90%, which can be determined experimentally or adjusted according to signal characteristics), and the original features are projected onto the selected eigenvectors. ,in, represents the features after dimensionality reduction, X represents the initial high-dimensional feature matrix, Represents the transformation matrix, and obtains the features after PCA dimensionality reduction;

[0059] Step 105: Use linear discriminant analysis (LDA) to further reduce the dimension of the feature vector after PCA dimension reduction. Suppose the number of sample categories is c (Determined by the signal feature classification requirements, such as speech / music / noise, etc.) The specific steps are as follows:

[0060] Calculate the sample mean vector for each category, calculate the intra-class scatter matrix and the inter-class scatter matrix, and solve the generalized eigenvalue problem: ,in, is the between-class scatter matrix, is the intra-class scatter matrix, and the eigenvalues ​​are obtained λ and the corresponding eigenvector w , and sort by eigenvalue from large to small, select the first m The eigenvectors form the transformation matrix , m To maximize the ratio of inter-class divergence to intra-class divergence, the features after PCA dimensionality reduction are projected onto the selected feature vector to obtain the final composite feature vector: ,in, represents a composite eigenvector.

[0061] When using, combine the contents of steps 101 to 105:

[0062] By digitizing analog audio signals and extracting multi-dimensional features, and then generating composite feature vectors through splicing, fusion and dimensionality reduction, we can not only fully capture the key characteristics of the signal, such as the Mel frequency characteristics that are consistent with human hearing, but also filter out interference and reduce data dimensions, laying an efficient and accurate feature foundation for subsequent layered coding, modulation and other processing.

[0063] Step 2: Divide the digital signal into a core layer, a secondary layer, and a background layer based on the information entropy of the composite eigenvector. Perform Huffman-arithmetic hybrid coding on the core layer, Huffman coding on the secondary layer, and wavelet compression coding on the background layer. Generate a key and encrypt the encoded data of each layer separately. Group the encrypted data and add traceability tags.

[0064] The second step includes the following contents:

[0065] Step 201: Calculate the information entropy of the composite feature vector: ,in, H represents information entropy, Indicates the i The probability that an eigenvalue appears in the eigenvector, N Indicates the number of different eigenvalues ​​in the eigenvector. According to the size of information entropy, the digital signal is divided into three levels, including the core layer, the secondary layer and the background layer. , The information entropy threshold is determined through a large number of experiments and is generally 1.2 times the average information entropy. This layer contains key information of the digital signal, such as the fundamental frequency of speech and the melody of music, which plays a decisive role in the quality of the digital signal. , contains secondary information of digital signals, such as the formant of speech, harmony of music, etc., which has an important impact on the clarity and richness of the signal; background layer , including background noise and ambient sound of the signal, which has little impact on the overall perception of the signal;

[0066] Step 202: Pre-encode the core layer using Huffman coding, and use arithmetic coding to encode the pre-encoding result into a real number in the interval [0, 1). Huffman coding is used for the secondary layer, and lossy wavelet compression coding with 3-layer decomposition using Daubechies4 (db4) wavelet basis is used for the background layer.

[0067] It should be noted that the wavelet transform can decompose the signal into subbands of different frequencies. The db4 wavelet basis has good compact support and symmetry. If a function is always 0 outside a finite interval, it is said to have compact support, which is suitable for audio signal processing. The three-layer decomposition decomposes the signal into low-frequency approximate components and high-frequency detail components. The high-frequency components are threshold quantized and entropy coded to achieve lossy compression. Because the background layer signal has little impact on perception, lossy compression can significantly reduce the amount of data while ensuring a certain sound quality.

[0068] Step 203: Generate two sets of keys using a true random number generator. The long key (e.g., 256 bits) is used for AES encryption of core layer signals, and the short key (e.g., 128 bits) is used for AES encryption of secondary layer and background layer signals. The key update mechanism is as follows: after each transmission of a preset number of data packets of a preset length, the two sets of keys are regenerated according to a preset pseudo-random algorithm based on a chaotic map, and the keys are replaced through a key update protocol. For example, a key update can be triggered every time 1024 data packets of 512 bytes are transmitted.

[0069] It should be noted that chaotic mapping is extremely sensitive to initial conditions and parameters, and the generated sequence has good randomness and unpredictability. Key updates are completed through a key update protocol, which ensures the secure handover of old and new keys and prevents key leakage and man-in-the-middle attacks. For example, when updating the key, the sender first sends the hash value of the new key. The receiver verifies that the hash value is correct before receiving the new key, ensuring the security of key transmission.

[0070] Step 204: Add a traceability tag based on a hash chain to each encrypted encoded data packet, and use the SHA-256 hash algorithm to calculate the hash value of the previous packet. The tag contains information such as the hash value of the previous packet, the current packet sequence number, and the key version number. The tag length is fixed to a preset number of bytes (such as 64 bytes).

[0071] When using, combine the contents of steps 201 to 204:

[0072] The signal levels are divided according to the information entropy of the composite eigenvector, and hybrid coding, Huffman coding and wavelet compression coding are used for the core layer, secondary layer and background layer respectively. Combined with dynamic key encryption and hash chain traceability tags, it not only achieves differentiated and efficient compression of signals of different importance, but also improves the security and traceability of data transmission through encryption and traceability mechanisms.

[0073] Step 3: Interleave the coded data of each layer at a preset time interval, divide the available frequency band into unequally spaced subcarrier groups, and use a multi-antenna system combined with a space-time coding scheme for spatial precoding. Combined with historical channel parameters and the current signal composite eigenvector, the fading depth is predicted. When the fading depth exceeds the depth threshold, the modulation mode is switched and the transmit power is adjusted.

[0074] The step three includes the following contents:

[0075] Step 301: The core layer, secondary layer, and background layer coded data are interleaved at a time interval of 1:2:1. For example, in every four time units, the first unit transmits the core layer data, the second and third units transmit the secondary layer data, and the fourth unit transmits the background layer data.

[0076] Step 302: Divide the available frequency band into unequally spaced subcarrier groups using the golden section (0.618). The number of subcarriers is determined based on the available bandwidth and subcarrier frequency resolution: ,in, represents the number of subcarriers, B represents the available bandwidth, Indicates the subcarrier frequency resolution;

[0077] Step 303: Use a multi-antenna system equipped with a preset number of antennas (e.g., 4) and perform spatial precoding using the Alamouti space-time coding scheme to divide the coded data into two groups. m and antennas m +1 is emitted in the following mode: ,in, 、 Represents the coded data symbol, represents complex conjugate;

[0078] Step 304: Construct a long short-term memory (LSTM) neural network model comprising an input layer, a hidden layer, and an output layer. The input layer receives historical channel parameters and a composite feature vector of the current signal. The historical channel parameters include signal-to-noise ratio, fading depth, and multipath delay. The hidden layer is configured with two LSTM layers (e.g., 2 layers), each with 128 neurons. The activation function uses the ReLU function. The output layer predicts the channel fading depth within a preset time period (e.g., 500 ms) in the future.

[0079] Step 305: When it is predicted that the channel fading depth exceeds a preset depth threshold (e.g., 20 dB), the following operations are performed: the modulation mode of the corresponding subcarrier is switched to π / 4-DQPSK or other anti-fading modulation modes; the transmit power is increased by a preset gain (e.g., 6 dB) through the power control module; otherwise, the current configuration is maintained.

[0080] When using, combine the contents of steps 301 to 305:

[0081] By interleaving the coded data of each layer in proportion, dividing it into unequally spaced subcarrier groups, and combining multi-antenna space-time coding with channel fading prediction, the modulation mode and transmission power can be dynamically adjusted in complex channel environments, effectively improving spectrum utilization, enhancing signal anti-fading capabilities, and ensuring transmission stability and reliability.

[0082] Step 4: Redundantly encode each layer to generate a signal replica, monitor the subcarrier bit error rate, and if the bit error rate exceeds the bit error rate threshold, dynamically borrow resources from adjacent subcarriers, perform phase rotation on the signal replicas of each layer, and superimpose them to form a transmission signal, which is sent to the medium wave channel through the antenna.

[0083] The fourth step includes the following contents:

[0084] Step 401: The core layer uses low-density parity check code and Reed-Solomon code, the secondary layer uses LDPC code, and the background layer uses fountain code. Each coding layer (i.e., core layer, secondary layer, and background layer) generates a corresponding coding copy, monitors the real-time bit error rate of each subcarrier at a preset period (e.g., 10ms), and sets a bit error rate threshold (e.g., );

[0085] Step 402: If the subcarrier bit error rate exceeds the bit error rate threshold, the dynamic redundancy allocation mechanism is activated to select a subcarrier with a bit error rate lower than 1 / 10 of the bit error rate threshold from adjacent subcarriers, and borrow resources according to a preset ratio of the current transmission rate of the interfered subcarrier. The resource borrowing ratio formula is: ,in, α Indicates the preset ratio, BER indicates the subcarrier bit error rate, Indicates the bit error rate threshold. Otherwise, keep the current configuration.

[0086] Step 403: Perform phase rotation on the signal replica generated by each coding layer. The phase offset is: , where n is the subcarrier number and m is the antenna number. is the total number of subcarriers, M is the number of antennas, phase rotation is performed on different replicas, and the rotated replica signals are superimposed through an adder to form the final coded signal for transmission, which is sent to the medium wave channel through the transmitting antenna.

[0087] When using, combine the contents of step 401 to step 403:

[0088] By redundantly encoding the signals at each layer to generate copies, dynamically borrowing adjacent subcarrier resources based on subcarrier bit error rate monitoring, and then performing phase rotation and superposition processing, resources can be adaptively allocated when channel interference occurs, effectively reducing the bit error rate, improving the signal's anti-interference capability, and ensuring the stability and integrity of medium-wave channel transmission.

[0089] See also Figure 3 The present invention also provides a medium wave transmitter digital coding system, comprising: a signal preprocessing module, a layered coding module, a signal modulation module and a redundancy control module, wherein:

[0090] The signal preprocessing module converts the input analog audio signal into a digital signal, extracts the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generates a composite feature vector after splicing, fusion, and dimensionality reduction.

[0091] The layered coding module divides the digital signal into a core layer, a secondary layer, and a background layer based on the information entropy of the composite eigenvector. It performs Huffman-arithmetic hybrid coding on the core layer, Huffman coding on the secondary layer, and wavelet compression coding on the background layer. It generates a key and encrypts the coded data of each layer separately. It groups the encrypted data and adds traceability tags.

[0092] The signal modulation module interleaves the coded data of each layer at preset time intervals, divides the available frequency band into unequally spaced subcarrier groups, uses a multi-antenna system combined with a space-time coding scheme for spatial precoding, combines historical channel parameters and the current signal composite eigenvector to predict the fading depth, and switches the modulation mode and adjusts the transmit power when the fading depth exceeds the depth threshold;

[0093] The redundancy control module performs redundant encoding on each layer to generate signal copies and monitors the subcarrier bit error rate. If the bit error rate exceeds the bit error rate threshold, it dynamically borrows resources from adjacent subcarriers, performs phase rotation on the signal copies of each layer, and superimposes them to form a transmission signal, which is sent to the medium wave channel through the antenna.

[0094] In the application, the several formulas involved are all calculated by taking their numerical values ​​after removing the dimensions, and the formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the latest real situation. The coefficients in the formula are set by technical personnel in this field according to actual conditions.

[0095] The above embodiments can be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0096] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0097] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A digital coding method for a medium wave transmitter, characterized in that: include: Convert the input analog audio signal into a digital signal, extract the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generate a composite feature vector after splicing, fusion, and dimensionality reduction. The digital signal is divided into a core layer, a secondary layer, and a background layer according to the information entropy of the composite eigenvector. Huffman-arithmetic hybrid coding is performed on the core layer, Huffman coding is performed on the secondary layer, and wavelet compression coding is performed on the background layer. A key is generated and the coded data of each layer is encrypted separately. The encrypted data is grouped and traceability tags are added. The coded data of each layer is interleaved at preset time intervals, and the available frequency band is divided into unequally spaced subcarrier groups. A multi-antenna system combined with a space-time coding scheme is used for spatial precoding. The fading depth is predicted by combining historical channel parameters and the current signal composite eigenvector. When the fading depth exceeds a depth threshold, the modulation mode is switched and the transmit power is adjusted. Redundant encoding is performed on each layer to generate a signal copy, and the subcarrier bit error rate is monitored. If the bit error rate exceeds the bit error rate threshold, resources are dynamically borrowed from adjacent subcarriers, and the signal copies of each layer are phase rotated and superimposed to form a transmission signal, which is sent to the medium wave channel through the antenna.

2. A medium wave transmitter digital encoding method according to claim 1, characterized in that: Set up several Mel filter banks, calculate the Mel frequency cepstral coefficients, extract the Mel frequency cepstral coefficient features, use the discrete Teager energy operator to operate on the digital signal, and extract the instantaneous energy features; The digital signal is modulated to the medium wave carrier frequency band through up-conversion, and the medium wave carrier frequency band is divided into multiple unequal width sub-bands. For each sub-band, the stopband attenuation parameters of the bandpass filter are set to perform interference filtering, and the filtered carrier signal is subjected to short-time Fourier transform to obtain frequency domain characteristics; The Mel-frequency cepstral coefficient features, instantaneous energy features and frequency domain features are spliced ​​and fused to form an initial high-dimensional feature matrix. The principal component analysis is used to reduce the dimension of the initial high-dimensional feature matrix, and the linear discriminant analysis is used to further reduce the dimension to obtain a composite feature vector.

3. A digital encoding method for a medium wave transmitter according to claim 2, characterized in that: The division into multiple unequal width sub-bands includes: Perform kernel density estimation on the historical interference intensity data within the frequency band to obtain the spatial distribution density function of the interference intensity; The local minimum point of the density function is used as the candidate point of the sub-band boundary, and only the candidate points that meet the interference intensity less than the interference intensity threshold are retained as boundary points.

4. A digital encoding method for a medium wave transmitter according to claim 1, characterized in that: Calculate the information entropy of the composite feature vector and divide the digital signal into core layer, secondary layer and background layer according to the size of the information entropy. , secondary layer , background layer , is the information entropy threshold, H represents information entropy; The core layer is pre-coded using Huffman coding, and the pre-coded result is encoded into a real number in the interval [0,1) using arithmetic coding. The secondary layer is pre-coded using Huffman coding, and the background layer is pre-coded using Daubechies4 wavelet basis.

5. A digital encoding method for a medium wave transmitter according to claim 4, characterized in that: Two sets of keys are generated using a true random number generator. The long key is used for AES encryption of core layer signals, and the short key is used for AES encryption of secondary layer and background layer signals. The two sets of keys are regenerated after each transmission of a preset number of data packets of preset length. A traceability tag based on a hash chain is added to each encrypted encoded data packet. The hash value of the previous packet is calculated using the SHA-256 hash algorithm. The tag contains the hash value of the previous packet, the current packet sequence number, and the key version number.

6. A medium wave transmitter digital encoding method according to claim 1, characterized in that: The coded data of the core layer, secondary layer, and background layer are interleaved at a time interval of 1:2:

1. The available frequency band is divided into unequally spaced subcarrier groups using the golden section. The number of subcarriers is determined by the available bandwidth and subcarrier frequency resolution: ,in, represents the number of subcarriers, B represents the available bandwidth, Indicates the subcarrier frequency resolution; A multi-antenna system with a preset number of antennas is used, and spatial precoding is performed using the Alamouti space-time coding scheme. The coded data is divided into two groups and transmitted through the antennas in two consecutive symbol periods. m and antennas m +1 is emitted in the following mode: ,in, 、 Represents the coded data symbol, represents the complex conjugate.

7. A digital encoding method for a medium wave transmitter according to claim 6, characterized in that: A long short-term memory neural network model is constructed, which includes an input layer, a hidden layer, and an output layer. The input layer receives historical channel parameters and the composite feature vector of the current signal. The historical channel parameters include signal-to-noise ratio, fading depth, and multipath delay. The hidden layer is set with two LSTM layers, each with 128 neurons. The activation function uses the ReLU function. The output layer predicts the channel fading depth within a preset time in the future.

8. The digital encoding method for a medium wave transmitter according to claim 1, characterized in that: If the subcarrier bit error rate exceeds the bit error rate threshold, the dynamic redundancy allocation mechanism is activated. A subcarrier with a bit error rate lower than 1 / 10 of the bit error rate threshold is selected from adjacent subcarriers. Resources are borrowed according to the preset ratio of the current transmission rate of the interfered subcarrier. The resource borrowing ratio formula is: ,in, α Indicates the preset ratio, BER represents the subcarrier bit error rate, Indicates the bit error rate threshold.

9. A medium wave transmitter digital encoding method according to claim 8, characterized in that: The signal replica generated by each coding layer is phase rotated by the phase offset: ,in, n is the subcarrier number, m is the antenna number, is the total number of subcarriers, M is the number of antennas, phase rotation is performed on different copies, and the rotated copy signals are superimposed through an adder to form the final coded signal for transmission, which is sent to the medium wave channel through the transmitting antenna.

10. A medium wave transmitter digital coding system, used to implement the method according to any one of claims 1 to 9, characterized in that: include: The signal preprocessing module converts the input analog audio signal into a digital signal, extracts the Mel-frequency cepstral coefficient features, instantaneous energy features, and frequency domain features, and generates a composite feature vector after splicing, fusion, and dimensionality reduction. The layered coding module divides the digital signal into a core layer, a secondary layer, and a background layer based on the information entropy of the composite eigenvector. It performs Huffman-arithmetic hybrid coding on the core layer, Huffman coding on the secondary layer, and wavelet compression coding on the background layer. It generates a key and encrypts the coded data of each layer separately. It groups the encrypted data and adds traceability tags. The signal modulation module interleaves the coded data of each layer at preset time intervals, divides the available frequency band into unequally spaced subcarrier groups, uses a multi-antenna system combined with a space-time coding scheme for spatial precoding, combines historical channel parameters and the current signal composite eigenvector to predict the fading depth, and switches the modulation mode and adjusts the transmit power when the fading depth exceeds the depth threshold; The redundancy control module performs redundant encoding on each layer to generate signal copies and monitors the subcarrier bit error rate. If the bit error rate exceeds the bit error rate threshold, it dynamically borrows resources from adjacent subcarriers, performs phase rotation on the signal copies of each layer, and superimposes them to form a transmission signal, which is sent to the medium wave channel through the antenna.

Images