A Communication Transceiver Method and System with Adaptive Frequency Hopping Filtering

Abstract

This invention relates to the field of wireless communication anti-interference technology, specifically to an adaptive frequency hopping filtering communication transceiver method and system, comprising the following steps: S1: By continuously monitoring the interference level, signal-to-noise ratio, and spectral timing characteristics of the current communication channel, channel state information for channel quality assessment and interference behavior prediction is collected and stored. This invention achieves dynamic optimization generation of the frequency hopping pattern through interference trend prediction and weighted correction based on an attention mechanism-LSTM network. Combined with interference residual feature prediction and adaptive filtering parameter pre-configuration and real-time optimization using the least mean square algorithm, it effectively suppresses channel interference. Furthermore, by utilizing the characteristics of implicit pilots and chaotic frequency hopping sequences, high-precision blind synchronization is achieved through joint detection and maximum likelihood estimation. Ultimately, this achieves a synergistic improvement in the anti-interference capability, synchronization reliability, and data recovery accuracy of the communication link under complex electromagnetic environments.

Description

Technical Field

[0001] This invention relates to the field of wireless communication anti-interference technology, specifically to an adaptive frequency hopping filtering communication transceiver method and system. Background Technology

[0002] In the field of wireless communication, communication reliability under complex electromagnetic environments (such as industrial interference, multi-user frequency contention, and malicious electromagnetic interference) has always been a core technical challenge. Existing technologies often employ fixed-frequency transmission or simple frequency-hopping strategies. Parameters such as frequency-hopping patterns and rates are preset and cannot be dynamically adjusted according to real-time channel interference conditions, making communication links susceptible to dynamic interference such as sudden and time-varying interference. Simultaneously, receivers often rely on filters with fixed parameters to suppress interference, making it difficult to adapt to the random variations in interference, resulting in significant residual interference. Furthermore, traditional synchronization methods rely on explicit pilot signals, which are easily affected by interference, leading to low synchronization accuracy at both ends. This further reduces the effectiveness of data demodulation and error correction, ultimately causing problems such as poor data recovery reliability and unstable communication quality.

[0003] Based on this, the present invention provides an adaptive frequency hopping filter communication transceiver method and system to solve the above-mentioned technical problems. Summary of the Invention

[0004] The purpose of this invention is to provide a communication transceiver method and system based on adaptive frequency hopping filtering. This invention achieves dynamic optimization generation of frequency hopping patterns based on interference trend prediction and weighted correction using an attention mechanism-LSTM network. It effectively suppresses channel interference by combining interference residual feature prediction and adaptive filtering parameter pre-configuration and real-time optimization using the least mean square algorithm. Furthermore, it utilizes the characteristics of implicit pilots and chaotic frequency hopping sequences to achieve high-precision blind synchronization through joint detection and maximum likelihood estimation. Ultimately, it achieves a synergistic improvement in the anti-interference capability, synchronization reliability, and data recovery accuracy of communication links in complex electromagnetic environments.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] This invention provides an adaptive frequency hopping filtering communication transceiver method, comprising the following steps:

[0007] S1: By continuously monitoring the interference level, signal-to-noise ratio, and spectral timing characteristics of the current communication channel, collect and store channel state information for channel quality assessment and interference behavior prediction;

[0008] S2: Evaluate the current channel quality based on the acquired channel state information, and use the interference source behavior prediction model to predict the interference trend of future frequency bands, dynamically generating frequency hopping pattern sequences to avoid current and potential interference areas and available channels;

[0009] S3: Based on the selected frequency hopping pattern sequence, encode, encrypt, and frequency hopping modulate the transmitted data, and dynamically adjust the transmit power according to the channel conditions;

[0010] S4: At the receiving end, based on the interference residual characteristics in the historical frequency hopping cycle, the interference distribution of the current cycle is predicted through a lightweight timing model, the adaptive filter parameters are pre-configured, and the least mean square algorithm is combined for real-time optimization.

[0011] S5: Utilizing the deterministic characteristics of the implicit pilot signal embedded in the data stream and the chaotic frequency hopping sequence, blind synchronization of the transceiver is performed through joint detection and parameter estimation, and the filtered signal is demodulated, decrypted and error corrected.

[0012] Based on the above method, this invention also proposes an adaptive frequency hopping filtering communication transceiver system, including an intelligent spectrum sensing unit, a dynamic decision-making and waveform generation unit, an adaptive transmission unit, an intelligent anti-interference receiving unit, and a blind synchronization and signal processing unit, wherein:

[0013] The intelligent spectrum sensing unit is used to continuously monitor the interference level, signal-to-noise ratio, and spectrum timing characteristics of the communication channel, and to collect and store channel state information.

[0014] The dynamic decision-making and waveform generation unit: evaluates channel quality based on channel state information, predicts future interference trends using an interference source behavior prediction model, and dynamically generates frequency hopping pattern sequences;

[0015] The adaptive transmission unit is used to encode, encrypt, and frequency-hop modulate the transmitted data according to the frequency hopping pattern sequence, and dynamically adjust the transmission power in combination with the channel conditions.

[0016] The intelligent anti-interference receiving unit: predicts the current interference distribution based on historical interference residual characteristics, pre-configures adaptive filter parameters, and performs real-time optimization using the least mean square algorithm to suppress interference and extract effective signals;

[0017] The blind synchronization and signal processing unit is used to perform blind synchronization by utilizing the characteristics of implicit pilot signals and chaotic frequency hopping sequences through joint detection and parameter estimation, and to perform demodulation, decryption and error correction processing on the signals.

[0018] The intelligent spectrum sensing unit includes a multi-dimensional signal acquisition module, a data preprocessing and storage module, and a feature extraction module, wherein:

[0019] The multi-dimensional signal acquisition module is used to acquire the original signal, including the interference level, signal-to-noise ratio, and spectral timing characteristics of the channel in real time.

[0020] The data preprocessing and storage module is used to perform noise reduction and normalization on the acquired raw signals and store them as structured channel state information.

[0021] The feature extraction module is used to extract key feature parameters from the preprocessed data for channel quality assessment and interference prediction.

[0022] The dynamic decision-making and waveform generation unit includes a channel quality assessment module, an interference trend prediction module, and a pattern generation module, wherein:

[0023] The channel quality assessment module: quantifies and assesses the communication capability of the current channel based on channel state information;

[0024] The interference trend prediction module is used to analyze historical data using an interference source behavior prediction model to predict the future interference distribution trend in frequency bands.

[0025] The pattern generation module is used to dynamically generate frequency hopping pattern sequences for interference avoidance and channel hopping by integrating the quality assessment and prediction results.

[0026] The interference trend prediction module uses an interference source behavior prediction model to analyze historical data and predict the future interference distribution trend in frequency bands. The specific operation is as follows:

[0027] A1: Normalize and align historical interference feature data with time series;

[0028] A2: Input the processed data into the interference source behavior prediction model and output the frequency band interference probability for the next N frequency hopping cycles;

[0029] A3: By combining the current channel quality assessment results, the predicted probability is weighted and corrected to obtain the final interference distribution trend. The weighted correction formula is as follows:

[0030] ,

[0031] in, The output of the interference source behavior prediction model is the "interference probability of the t-th frequency hopping cycle and frequency band j". This refers to the channel quality coefficients output by the current channel quality assessment module. Let "the actual interference probability of frequency band j in the (t-1)th frequency hopping cycle" be defined. The weighting coefficients are the historical measured values. This represents the final interference probability after weighted correction.

[0032] The adaptive transmission unit includes a source coding and encryption module, a frequency hopping modulation module, and a power adaptive adjustment module, wherein:

[0033] The source coding and encryption module is used to perform error correction coding and encryption processing on the transmitted data.

[0034] The frequency hopping modulation module is used to modulate the encoded data to the corresponding frequency band according to the frequency hopping pattern sequence.

[0035] The power adaptive adjustment module is used to dynamically adjust the transmission power based on real-time channel conditions, balancing communication quality and energy consumption.

[0036] The intelligent anti-interference receiving unit includes an interference feature memory module, an interference prediction and parameter pre-configuration module, and a filtering optimization module, wherein:

[0037] The interference feature memory module is used to store and update interference residual feature data in historical frequency hopping cycles.

[0038] The interference prediction and parameter pre-configuration module is used to predict the current periodic interference distribution using a lightweight time series model and pre-configure the initial parameters of the filter.

[0039] The filtering optimization module is used to run the least mean square algorithm to optimize filter parameters in real time, suppress interference, and extract effective signals.

[0040] The interference prediction and parameter pre-configuration module predicts the current period interference distribution and pre-configures the initial parameters of the filter using a lightweight time-series model. The specific operations are as follows:

[0041] B1: Retrieve the interference residual feature sequence of nearly M frequency hopping cycles from the interference feature memory module, and use Z-score normalization to eliminate dimensional differences;

[0042] B2: Input the standardized interference residual sequence into the compressed GRU model, and output the predicted interference power values ​​for each sub-band within the current frequency hopping cycle. The calculation formula is:

[0043] ,

[0044] in, The characteristics of the historical disturbance residuals after standardization. is the training parameter set for the lightweight model, and f is the sub-band frequency of the current period;

[0045] B3: Establish a "predicted interference power value - filter parameter mapping table", based on... Initial parameters of the matched filter:

[0046] ①If Pre-configured filter order Step size factor ;

[0047] ②If Pre-configuration ;

[0048] ③If Pre-configuration ;

[0049] in, The fundamental order, The base step size factor is k, and the order adjustment coefficient is k. The preset interference threshold;

[0050] B4: Send the pre-configured filter order and step size factor initial parameters to the filter optimization module.

[0051] The blind synchronization and signal processing unit includes an implicit pilot detection module, a blind synchronization module, and a demodulation and error correction module, wherein:

[0052] The implicit pilot detection module is used to extract the embedded implicit pilot signal from the received signal.

[0053] The blind synchronization module is used to synchronize the transmitting and receiving ends by utilizing the deterministic autocorrelation characteristics of implicit pilots and chaotic frequency hopping sequences through joint detection and parameter estimation.

[0054] The demodulation and error correction module is used to demodulate and decrypt the synchronized signal, and to recover the original data through an error correction algorithm.

[0055] The blind synchronization module utilizes the determinism of implicit pilot signals and chaotic frequency hopping sequences to perform transceiver synchronization through joint detection and parameter estimation. The specific operation is as follows:

[0056] C1: Extract the embedded implicit pilot sequence from the received signal;

[0057] C2: Utilizing the deterministic autocorrelation characteristics of chaotic frequency hopping sequences, coarse synchronization of time and frequency is achieved through matched filtering;

[0058] C3: Based on the coarse synchronization results, maximum likelihood estimation is used to jointly estimate the frequency offset and time delay parameters to complete the synchronization of the transmitting and receiving ends.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] This invention achieves dynamic optimization generation of frequency hopping patterns through interference trend prediction and weighted correction based on attention mechanism-LSTM network. It effectively suppresses channel interference by combining interference residual feature prediction and adaptive filtering parameter pre-configuration and real-time optimization of least mean square algorithm. Furthermore, it utilizes the characteristics of implicit pilot and chaotic frequency hopping sequence to achieve high-precision blind synchronization through joint detection and maximum likelihood estimation. Ultimately, it achieves a synergistic improvement in the anti-interference ability, synchronization reliability and data recovery accuracy of communication links in complex electromagnetic environments. Attached Figure Description

[0061] Figure 1 This is a system diagram of an adaptive frequency hopping filter communication transceiver system according to the present invention.

[0062] Figure 2 This is a flowchart of an adaptive frequency hopping filtering communication transceiver method according to the present invention.

[0063] Explanation of icon numbers:

[0064] 1. Intelligent spectrum sensing unit; 11. Multi-dimensional signal acquisition module; 12. Data preprocessing and storage module; 13. Feature extraction module; 2. Dynamic decision-making and waveform generation unit; 21. Channel quality assessment module; 22. Interference trend prediction module; 23. Pattern generation module; 3. Adaptive transmission unit; 31. Source coding and encryption module; 32. Frequency hopping modulation module; 33. Power adaptive adjustment module; 4. Intelligent anti-interference receiving unit; 41. Interference feature memory module; 42. Interference prediction and parameter pre-configuration module; 43. Filter optimization module; 5. Blind synchronization and signal processing unit; 51. Implicit pilot detection module; 52. Blind synchronization module; 53. Demodulation and error correction module. Detailed Implementation

[0065] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0066] Example 1:

[0067] like Figure 1As shown, this embodiment provides a communication transceiver system with adaptive frequency hopping filtering, including an intelligent spectrum sensing unit 1, a dynamic decision-making and waveform generation unit 2, an adaptive transmission unit 3, an intelligent anti-interference receiving unit 4, and a blind synchronization and signal processing unit 5. Specifically: the intelligent spectrum sensing unit 1 continuously monitors the interference level, signal-to-noise ratio, and spectral timing characteristics of the communication channel, collects and stores channel state information; the dynamic decision-making and waveform generation unit 2 evaluates channel quality based on channel state information, predicts future interference trends using an interference source behavior prediction model, and dynamically generates a frequency hopping pattern sequence; the adaptive transmission unit 3 encodes, encrypts, and modulates the transmitted data according to the frequency hopping pattern sequence, and dynamically adjusts the transmission power based on channel conditions; the intelligent anti-interference receiving unit 4 predicts the current interference distribution based on historical interference residual characteristics, pre-configures adaptive filter parameters, and performs real-time optimization using a least mean square algorithm to suppress interference and extract effective signals; and the blind synchronization and signal processing unit 5 utilizes the characteristics of implicit pilot signals and chaotic frequency hopping sequences to perform blind synchronization through joint detection and parameter estimation, and demodulates, decrypts, and corrects the signal.

[0068] It should be noted that the intelligent spectrum sensing unit 1 monitors the channel environment in real time and transmits the status information to the dynamic decision-making and waveform generation unit 2 to generate an anti-interference frequency hopping sequence. This sequence drives the adaptive transmission unit 3 to complete the signal encoding, modulation and power controllable transmission. At the receiving end, the intelligent anti-interference receiving unit 4 predicts and filters out interference based on historical interference characteristics. Finally, the blind synchronization and signal processing unit 5 uses implicit pilot and chaotic characteristics to achieve blind synchronization and signal recovery.

[0069] In this embodiment, it should also be noted that the intelligent spectrum sensing unit 1 includes a multi-dimensional signal acquisition module 11, a data preprocessing and storage module 12, and a feature extraction module 13, wherein: the multi-dimensional signal acquisition module 11 is used to acquire the original signal of the channel's interference level, signal-to-noise ratio, and spectral timing characteristics in real time; the data preprocessing and storage module 12 is used to perform noise reduction and normalization processing on the acquired original signal and store it as structured channel state information; the feature extraction module 13 is used to extract key feature parameters for channel quality assessment and interference prediction from the preprocessed data.

[0070] It should be noted that the multi-dimensional signal acquisition module 11 is responsible for acquiring the original channel signal. After the data preprocessing and storage module 12 performs noise reduction, normalization and structured storage, the feature extraction module 13 extracts the key feature parameters used for channel quality assessment and interference prediction.

[0071] Furthermore, it should be noted that the multi-dimensional signal acquisition module 11 employs a wideband RF front-end, operating in the 20MHz-6GHz frequency band, a 12-bit high-speed ADC (sampling rate ≥100MSPS), and a signal-to-noise ratio detector to acquire data. The interference level is measured in real-time using a power meter (measurement range -120dBm to 0dBm, accuracy ±1dB), the signal-to-noise ratio is calculated using "signal power / noise power" (measurement resolution 0.1dB), and the spectral timing characteristics are obtained through FFT transformation (1024 FFT points, frequency resolution 97.66kHz) to obtain the power spectral density timing curve. Key feature parameters in the feature extraction module 13 include: channel fading rate, interference signal center frequency and bandwidth, and spectral hole duration.

[0072] In this embodiment, it should also be noted that the dynamic decision-making and waveform generation unit 2 includes a channel quality assessment module 21, an interference trend prediction module 22, and a pattern generation module 23, wherein: the channel quality assessment module 21: quantitatively assesses the communication capability of the current channel based on channel state information; the interference trend prediction module 22: uses an interference source behavior prediction model to analyze historical data and predict the interference distribution trend of future frequency bands; the specific operations are as follows: A1: normalize and align the historical interference feature data with the time series; A2: input the processed data into the interference source behavior prediction model and output the frequency band interference probability for the next N frequency hopping cycles; A3: combine the current channel quality assessment results to perform weighted correction on the predicted probability to obtain the final interference distribution trend, and the weighted correction formula is as follows:

[0073]

[0074] in, The output of the interference source behavior prediction model is the "interference probability of the t-th frequency hopping cycle and frequency band j". The channel quality coefficients output by the current channel quality assessment module 21. Let "the actual interference probability of frequency band j in the (t-1)th frequency hopping cycle" be defined. The weighting coefficients are the historical measured values. This represents the final interference probability after weighted correction. Pattern generation module 23: Used to integrate quality assessment and prediction results to dynamically generate frequency hopping pattern sequences for interference avoidance and channel mitigation.

[0075] It should be noted that the channel quality assessment module 21 quantifies the current channel communication capability, the interference trend prediction module 22 predicts the future interference distribution trend based on historical data and performs weighted correction on the probability, and the pattern generation module 23 dynamically generates an anti-interference frequency hopping sequence by combining the assessment and prediction results.

[0076] Furthermore, it should be noted that the channel quality assessment module 21 uses a "weighted scoring method" to quantify channel quality:

[0077] ,

[0078] Wherein, SNR is the normalized signal-to-noise ratio, representing the relative strength of the ratio of signal power to noise power. Its original value is obtained in real time by the signal-to-noise ratio detector in the intelligent spectrum sensing unit 1, and the unit is dB. Then, it is linearly mapped to the [0,1] interval for normalization processing. The specific formula is as follows:

[0079]

[0080] in, The minimum acceptable signal-to-noise ratio (e.g., 5 dB). I represents the maximum signal-to-noise ratio supported by the system (e.g., 40 dB), thus ensuring the comparability of signal-to-noise ratios under different channel conditions; I represents the current interference power. For the maximum tolerable interference power, This is the normalized value for the duration of the spectral hole. A channel with Q ≥ 80 is considered a high-quality channel, a channel with Q ≤ 80 is considered a usable channel, and a channel with Q < 60 is considered an unusable channel.

[0081] In A2, the interference source behavior prediction model employs an attention mechanism-LSTM network. The network structure is: 'Input layer (20-dimensional features) → Attention layer (based on Bahdanau attention mechanism) → LSTM layer (2 layers, 64 neurons each) → Fully connected layer (output layer, dimension is "future N cycles × number of sub-bands")'. The model is trained using the Adam optimizer with a learning rate of 0.001. The training dataset consists of 100,000 pairs of historical interference features and actual interference distribution samples. The prediction period N can be adaptively adjusted within 5-20 frequency hopping cycles. In the weighted correction formula in A3, the channel quality coefficient... (Q represents the channel quality score), historical measured weighting coefficients Frequency hopping pattern generation follows three main rules: ① Prioritize patterns with Q≥70 and... (Low interference probability) frequency band; ② Frequency hopping interval ≥ 2MHz to avoid adjacent channel interference; ③ Pseudo-random frequency hopping sequence generated using 'm-sequence + interference avoidance constraint', the primitive polynomial of the m-sequence is: This ensures the anti-interception capability of the frequency hopping pattern; the frequency hopping rate is dynamically adjusted according to the channel fading rate: 100 hops / second for slow fading channels and 500 hops / second for fast fading channels.

[0082] In this embodiment, it should also be noted that the adaptive transmission unit 3 includes a source coding and encryption module 31, a frequency hopping modulation module 32, and a power adaptive adjustment module 33, wherein: the source coding and encryption module 31 is used to perform error correction coding and encryption processing on the transmitted data; the frequency hopping modulation module 32 is used to modulate the encoded data to the corresponding frequency band according to the frequency hopping pattern sequence; and the power adaptive adjustment module 33 is used to dynamically adjust the transmission power in combination with the real-time channel conditions to balance communication quality and energy consumption.

[0083] It should be noted that the source coding and encryption module 31 performs error correction coding and encryption processing on the transmitted data, the frequency hopping modulation module 32 modulates the encoded data to the corresponding frequency band according to the frequency hopping pattern sequence, and the power adaptive adjustment module 33 dynamically adjusts the transmission power according to the real-time channel status to balance communication quality and energy consumption.

[0084] Furthermore, it should be noted that the transmit power adjustment is performed based on the channel quality score Q: ① If Q≥80 (high-quality channel), the transmit power is set to 10dBm (optimal energy consumption); ② If 60≤Q<80 (available channel), the transmit power is set to 20dBm (balancing quality and energy consumption); ③ If Q<60 (unavailable channel), the transmit power is set to 30dBm (maximum coverage), and the dynamic decision-making and waveform generation unit 2 is triggered to regenerate the frequency hopping pattern; the response delay of the power adjustment is ≤1ms, adapting to real-time changes in the channel.

[0085] In this embodiment, it should also be noted that the intelligent anti-interference receiving unit 4 includes an interference feature memory module 41, an interference prediction and parameter pre-configuration module 42, and a filter optimization module 43, wherein: the interference feature memory module 41 is used to store and update the interference residual feature data in the historical frequency hopping cycle; the interference prediction and parameter pre-configuration module 42 is used to predict the interference distribution of the current cycle through a lightweight time series model and pre-configure the initial parameters of the filter; the specific operation is as follows: B1: call the interference residual feature sequence of the past M frequency hopping cycles from the interference feature memory module 41, and use Z-score normalization to eliminate the difference in dimensions; B2: input the normalized interference residual sequence into the compressed GRU model, and output the predicted interference power value of each sub-band in the current frequency hopping cycle. The calculation formula is:

[0086]

[0087] in, The characteristics of the historical disturbance residuals after standardization. This is the training parameter set for the lightweight model, where f is the sub-band frequency of the current cycle; GRU is a variant of recurrent neural network (RNN), mainly used for processing sequence data, and can effectively capture long-term dependencies in the sequence. In this embodiment, the compressed GRU model is used to process historical interference residual feature sequences. By learning the patterns in these historical sequences, the interference power of each sub-band in the current frequency hopping cycle is predicted; B3: Establish an "Interference Power Prediction Value - Filter Parameter Mapping Table", based on... Matched filter initial parameters: ① If Pre-configured filter order Step size factor ; ②If Pre-configuration ③If Pre-configuration ;in, The fundamental order, The base step size factor is k, and the order adjustment coefficient is k. B4: Sends the pre-configured filter order and step size factor initial parameters to the filter optimization module 43. Filter optimization module 43: Used to run the least mean square algorithm to optimize filter parameters in real time, suppress interference and extract effective signals.

[0088] It should be noted that the interference feature memory module 41 stores and updates historical interference feature data, the interference prediction and parameter pre-configuration module 42 predicts the interference distribution and pre-configures filter parameters based on the data, and the filter optimization module 43 optimizes the filter in real time and extracts the effective signal based on the obtained parameters through an adaptive algorithm.

[0089] Furthermore, it should be noted that the core formula of the least mean square algorithm is: ① Filter output: ② Error calculation: ③ Weight update: ;in, Let n be the filter coefficient vector at time n. The input signal vector, For the desired signal, For error signals, The step size factor (the initial value output by the interference prediction and parameter pre-configuration module 42, and based on the error) Variance dynamic fine-tuning: If If >0.1, then μ increases by 10%; if If μ < 0.01, then μ decreases by 10%.

[0090] In this embodiment, it should also be noted that the blind synchronization and signal processing unit 5 includes an implicit pilot detection module 51, a blind synchronization module 52, and a demodulation and error correction module 53. Specifically: the implicit pilot detection module 51 extracts the embedded implicit pilot signal from the received signal; the blind synchronization module 52 utilizes the deterministic autocorrelation characteristics of the implicit pilot and the chaotic frequency hopping sequence to perform transceiver synchronization through joint detection and parameter estimation; the specific operations are as follows: C1: Extract the embedded implicit pilot sequence from the received signal; C2: Utilize the deterministic autocorrelation characteristics of the chaotic frequency hopping sequence to perform coarse time and frequency synchronization through matched filtering; C3: Based on the coarse synchronization result, use maximum likelihood estimation to jointly estimate the frequency offset and time delay parameters to complete the transceiver synchronization. The demodulation and error correction module 53 demodulates and decrypts the synchronized signal and recovers the original data through an error correction algorithm.

[0091] It should be noted that the implicit pilot detection module 51 extracts the implicit pilot signal from the received signal, the blind synchronization module 52 uses the pilot signal and the chaotic sequence characteristics to complete the synchronization of the transmitting and receiving ends through joint parameter estimation, and the demodulation and error correction module 53 demodulates, decrypts and corrects the synchronized signal to recover the original data.

[0092] Furthermore, it should be noted that the formula for coarse synchronization of time and frequency using matched filtering in C2 is as follows: Where c(t) is the received chaotic sequence, For time delay; when When the maximum value is reached, the corresponding The coarse synchronization delay is defined as ±1μs. Coarse frequency synchronization is achieved through 'FFT frequency offset estimation', with an estimation range of ±500Hz and an error of ±100Hz. Based on the coarse synchronization results, a maximum likelihood function is constructed in C3. Where r(n) is the received signal and s(n) is the local reference signal. For frequency offset, For time delay, The sampling interval is given; the solution is obtained using the gradient descent method. The minimum value is obtained. Accuracy ±1Hz and With an accuracy of ±0.1μs, precise synchronization is achieved.

[0093] Example 2:

[0094] like Figure 2 As shown in this embodiment, an adaptive frequency hopping filtering communication transceiver method specifically includes the following steps:

[0095] S1. Spectrum Sensing and Channel State Information Acquisition

[0096] S1.1: Multi-dimensional signal acquisition:

[0097] The raw channel signal is acquired in real time using a wideband RF front-end (operating frequency band 20MHz-6GHz), a 12-bit high-speed ADC (sampling rate ≥100 MSPS) and a signal-to-noise ratio detector;

[0098] Interference level: Real-time measurement by power meter (range -120dBm to 0dBm, accuracy ±1dB).

[0099] Signal-to-noise ratio: Calculated as the ratio of signal power to noise power (resolution 0.1dB);

[0100] Spectral time series characteristics: Power spectral density time series curves were obtained by FFT transformation (1024 points, frequency resolution 97.66kHz);

[0101] S1.2: Data Preprocessing and Storage

[0102] The original signal is denoised and normalized, and then stored as structured channel state information;

[0103] S1.3: Feature Extraction:

[0104] Key feature parameters are extracted from the preprocessed data: channel fading rate, center frequency and bandwidth of the interference signal, and duration of spectral holes.

[0105] S2. Dynamic Decision-Making and Frequency Hopping Pattern Generation

[0106] S2.1: Channel quality assessment:

[0107] Channel quality is quantified using a weighted scoring method:

[0108]

[0109] in: Here, I is the normalized signal-to-noise ratio value, and I is the current interference power. For the maximum tolerable interference power, Let Q be the normalized value of the duration of the spectrum hole, where Q∈[0,100]. Q≥80 indicates a high-quality channel, 60≤Q<80 indicates a usable channel, and Q<60 indicates an unusable channel.

[0110] S2.2: Interference Trend Prediction:

[0111] ① Data preprocessing: Normalize historical interference feature data and align time series;

[0112] ② Model prediction: The processed data is input into an attention mechanism-LSTM network (structure: 20-dimensional input layer → attention layer → LSTM layer (2×64 neurons) → output layer), which outputs the interference probability of each frequency band in the next N cycles (N=5-20). ;

[0113] Model training: Adam optimizer (learning rate 0.001), 100,000 historical data samples;

[0114] ③ Weighted correction: Correct the predicted probability by combining the current channel quality and historical measured values.

[0115]

[0116] in, The output of the interference source behavior prediction model is the "interference probability of the t-th frequency hopping cycle and frequency band j". The channel quality coefficients output by the current channel quality assessment module 21. Let "the actual interference probability of frequency band j in the (t-1)th frequency hopping cycle" be defined. The weighting coefficients are the historical measured values. , This is the final interference probability after weighted correction;

[0117] S2.3: Frequency hopping pattern generation:

[0118] Based on quality assessment (Q value) and interference prediction Dynamically generate frequency hopping sequences:

[0119] ① Prioritize selection with Q≥70 and The frequency band;

[0120] ② Frequency hopping interval ≥ 2MHz (to avoid adjacent channel interference);

[0121] ③ Using m-sequences (primitive polynomials) Add interference avoidance constraints to generate pseudo-random sequences).

[0122] ④ Adaptive frequency hopping rate: 100 hops / second for slow fading channels, 500 hops / second for fast fading channels;

[0123] S3. Adaptive Emission Processing

[0124] S3.1: Source Coding and Encryption:

[0125] The transmitted data is then subjected to forward error correction coding (such as Turbo coding) and encryption;

[0126] S3.2: Frequency hopping modulation:

[0127] Based on the frequency hopping pattern sequence generated by S2, the coded data is modulated to the corresponding frequency band (such as FSK or QPSK modulation).

[0128] S3.3: Power Adaptive Adjustment: Dynamically adjusts transmit power based on channel quality score Q.

[0129] ①Q≥80: Power is set to 10dBm (optimal energy consumption);

[0130] ②60≤Q<80: Power set to 20dBm (balanced mode);

[0131] ③Q<60: Power is set to 30dBm (maximum coverage), and frequency hopping pattern is regenerated;

[0132] Power adjustment response delay ≤1ms;

[0133] S4. Intelligent anti-interference receiver processing

[0134] S4.1: Interference Feature Memory:

[0135] Store and update the interference residual characteristic data of historical frequency hopping cycles;

[0136] S4.2: Interference Prediction and Filter Pre-configuration:

[0137] ① Data standardization: Use the residual sequences of interference from nearly M periods and perform Z-score standardization;

[0138] ② Interference Prediction: Input a compressed GRU model, output the predicted interference power for each sub-band in the current period. :

[0139]

[0140] in, The characteristics of the historical disturbance residuals after standardization. is the training parameter set for the lightweight model, and f is the sub-band frequency of the current period;

[0141] ③ Parameter mapping: Pre-configure LMS filter parameters by looking up the predicted values ​​in a table:

[0142] Ⅰ. If Order Step size factor ;

[0143] II. If : , ;

[0144] III. If : , ;

[0145] S4.3: Filtering Optimization:

[0146] Run the LMS algorithm to optimize filtering in real time:

[0147] Filter output: ;

[0148] Error calculation: ;

[0149] Weight update: ;

[0150] Step length Dynamic fine-tuning: If >0.1, Increase by 10%; if <0.01, Reduce by 10%;

[0151] S5. Blind Synchronization and Signal Processing

[0152] S5.1: Implicit Pilot Detection

[0153] Extract the embedded implicit pilot sequence from the received signal;

[0154] S5.2: Blind Synchronization:

[0155] ① Coarse synchronization:

[0156] Time delay estimation: through autocorrelation matching of chaotic sequences

[0157] ,when When the maximum value is reached, the corresponding This is the coarse synchronization delay (error ±1μs);

[0158] Frequency offset estimation: FFT frequency offset estimation (range ±500Hz, error ±100Hz);

[0159] ②Precise synchronization:

[0160] Construct the maximum likelihood function:

[0161] ,

[0162] Jointly optimize latency using gradient descent method and frequency offset With an accuracy of ±0.1μs and ±1Hz;

[0163] S5.3: Demodulation and Error Correction

[0164] Demodulate (e.g., FSK / QPSK demodulation), decrypt (AES decryption) the synchronization signal, and recover the original data through error correction decoding (e.g., Viterbi decoding).

[0165] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0166] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A communication transceiver method with adaptive frequency hopping filtering, characterized in that, Includes the following steps: S1: By continuously monitoring the interference level, signal-to-noise ratio, and spectral timing characteristics of the current communication channel, collect and store channel state information for channel quality assessment and interference behavior prediction; S2: Evaluate the current channel quality based on the acquired channel state information, and use the interference source behavior prediction model to predict the interference trend of future frequency bands, dynamically generating frequency hopping pattern sequences to avoid current and potential interference areas and available channels; S3: Based on the selected frequency hopping pattern sequence, encode, encrypt, and frequency hopping modulate the transmitted data, and dynamically adjust the transmit power according to the channel conditions; S4: At the receiving end, based on the interference residual characteristics in the historical frequency hopping cycle, the interference distribution of the current cycle is predicted through a lightweight timing model, the adaptive filter parameters are pre-configured, and the least mean square algorithm is combined for real-time optimization. S5: Utilizing the deterministic characteristics of the implicit pilot signal embedded in the data stream and the chaotic frequency hopping sequence, blind synchronization of the transceiver is performed through joint detection and parameter estimation, and the filtered signal is demodulated, decrypted and error corrected. In S2, historical data is analyzed using an interference source behavior prediction model to predict future interference distribution trends in frequency bands. The specific operation is as follows: A1: Normalize and align historical interference feature data with time series; A2: Input the processed data into the interference source behavior prediction model and output the frequency band interference probability for the next N frequency hopping cycles; A3: The predicted probability is weighted and corrected based on the current channel quality assessment results to obtain the final interference distribution trend; In S4, the current period interference distribution is predicted using a lightweight time series model, and adaptive filter parameters are pre-configured. The specific operation is as follows: B1: Call the interference residual feature sequence of nearly M frequency hopping cycles, and use Z-score normalization to eliminate dimensional differences; B2: Input the standardized interference residual sequence into the compressed GRU model and output the predicted interference power of each sub-band in the current frequency hopping cycle; B3: Establish an "Interference Power Prediction Value-Filter Parameter Mapping Table" and match the initial filter parameters according to the interference power prediction value; B4: Outputs the initial parameters of the pre-configured filter order and step size factor; S5 utilizes the deterministic characteristics of implicit pilot signals embedded in the data stream and chaotic frequency hopping sequences to perform blind synchronization at the transmitting and receiving ends through joint detection and parameter estimation. The specific operation is as follows: C1: Extract the embedded implicit pilot sequence from the received signal; C2: Utilizing the deterministic autocorrelation characteristics of chaotic frequency hopping sequences, coarse synchronization of time and frequency is achieved through matched filtering; C3: Based on the coarse synchronization results, maximum likelihood estimation is used to jointly estimate the frequency offset and time delay parameters to complete the synchronization of the transmitting and receiving ends.

2. A communication transceiver system with adaptive frequency hopping filtering, wherein the communication transceiver method with adaptive frequency hopping filtering according to claim 1 is characterized in that, It includes an intelligent spectrum sensing unit (1), a dynamic decision-making and waveform generation unit (2), an adaptive transmission unit (3), an intelligent anti-interference receiving unit (4), and a blind synchronization and signal processing unit (5), wherein: The intelligent spectrum sensing unit (1) is used to continuously monitor the interference level, signal-to-noise ratio and spectrum timing characteristics of the communication channel, and to collect and store channel state information. The dynamic decision-making and waveform generation unit (2) evaluates the channel quality based on channel state information and predicts future interference trends using an interference source behavior prediction model, and dynamically generates frequency hopping pattern sequences. The adaptive transmission unit (3) is used to encode, encrypt, and frequency-hop modulate the transmitted data according to the frequency hopping pattern sequence, and dynamically adjust the transmission power in combination with the channel conditions; The intelligent anti-interference receiving unit (4) predicts the current interference distribution based on the historical interference residual characteristics, pre-configures the adaptive filter parameters, and combines the least mean square algorithm for real-time optimization to suppress interference and extract effective signals. The blind synchronization and signal processing unit (5) is used to perform blind synchronization by using the characteristics of implicit pilot signals and chaotic frequency hopping sequences through joint detection and parameter estimation, and to demodulate, decrypt and correct the signals.

3. The adaptive frequency hopping filtering communication transceiver system according to claim 2, characterized in that, The intelligent spectrum sensing unit (1) includes a multi-dimensional signal acquisition module (11), a data preprocessing and storage module (12), and a feature extraction module (13), wherein: The multi-dimensional signal acquisition module (11) is used to acquire the original signal of the channel's interference level, signal-to-noise ratio and spectral timing characteristics in real time. The data preprocessing and storage module (12) is used to perform noise reduction and normalization on the acquired raw signals and store them as structured channel state information. The feature extraction module (13) is used to extract key feature parameters from the preprocessed data for channel quality assessment and interference prediction.

4. The adaptive frequency hopping filtering communication transceiver system according to claim 2, characterized in that, The dynamic decision-making and waveform generation unit (2) includes a channel quality assessment module (21), an interference trend prediction module (22), and a pattern generation module (23), wherein: The channel quality assessment module (21) quantifies and assesses the communication capability of the current channel based on channel state information; The interference trend prediction module (22) is used to analyze historical data using the interference source behavior prediction model and predict the interference distribution trend of future frequency bands. The pattern generation module (23) is used to dynamically generate frequency hopping pattern sequences to avoid interference and channel based on the comprehensive quality assessment and prediction results.

5. The adaptive frequency hopping filtering communication transceiver system according to claim 4, characterized in that, The weighted correction formula is as follows: ， in, The output of the interference source behavior prediction model is "the probability of interference in the t-th frequency hopping cycle and frequency band j in the future". The channel quality coefficients output by the current channel quality assessment module (21) are the channel quality coefficients. Let "the actual interference probability of frequency band j in the (t-1)th frequency hopping cycle" be defined. The weighting coefficients are the historical measured values. This represents the final interference probability after weighted correction.

6. The adaptive frequency hopping filtering communication transceiver system according to claim 2, characterized in that, The adaptive transmission unit (3) includes a source coding and encryption module (31), a frequency hopping modulation module (32), and a power adaptive adjustment module (33), wherein: The source coding and encryption module (31) is used to perform error correction coding and encryption processing on the transmitted data; The frequency hopping modulation module (32) is used to modulate the encoded data to the corresponding frequency band according to the frequency hopping pattern sequence; The power adaptive adjustment module (33) is used to dynamically adjust the transmission power based on real-time channel conditions to balance communication quality and energy consumption.

7. The adaptive frequency hopping filtering communication transceiver system according to claim 2, characterized in that, The intelligent anti-interference receiving unit (4) includes an interference feature memory module (41), an interference prediction and parameter pre-configuration module (42), and a filtering optimization module (43), wherein: The interference feature memory module (41) is used to store and update interference residual feature data in historical frequency hopping cycles; The interference prediction and parameter pre-configuration module (42) is used to predict the current period interference distribution through a lightweight time series model and pre-configure the initial parameters of the filter. The filtering optimization module (43) is used to run the least mean square algorithm to optimize the filter parameters in real time, suppress interference and extract effective signals.

8. The adaptive frequency hopping filtering communication transceiver system according to claim 7, characterized in that, The predicted value of interference power The calculation formula is: ， in, The characteristics of the historical disturbance residuals after standardization. is the training parameter set for the lightweight model, and f is the sub-band frequency of the current period; according to The initial parameters of the matched filter include the following: ①If Pre-configured filter order Step size factor ; ②If Pre-configuration ; ③If Pre-configuration ; in, The fundamental order, The base step size factor is k, and the order adjustment coefficient is k. This is a preset interference threshold.

9. The adaptive frequency hopping filtering communication transceiver system according to claim 2, characterized in that, The blind synchronization and signal processing unit (5) includes an implicit pilot detection module (51), a blind synchronization module (52), and a demodulation and error correction module (53), wherein: The implicit pilot detection module (51) is used to extract the embedded implicit pilot signal from the received signal; The blind synchronization module (52) is used to synchronize the transmitting and receiving ends by means of joint detection and parameter estimation, utilizing the deterministic autocorrelation characteristics of implicit pilots and chaotic frequency hopping sequences. The demodulation and error correction module (53) is used to demodulate and decrypt the synchronized signal and recover the original data through the error correction algorithm.

Images