Digital down-conversion multichannel time delay alignment method

Abstract

The invention relates to the technical field of digital signal processing, and discloses a digital down-conversion multichannel time delay alignment method, which comprises the following steps of: performing down-conversion preprocessing on digital signals received by each channel, converting high-frequency signals into low-frequency signals, and simplifying the subsequent processing difficulty; performing dynamic threshold detection on the down-converted signal, and determining an effective starting point of each channel signal; estimating initial time delay inequality among the channels according to a dynamic threshold detection result; based on the initial time delay inequality, a self-adaptive PID controller is adopted to carry out time delay compensation on each channel signal; on the basis of initial time delay compensation, the time delay of each channel is further finely adjusted by using a sliding window technology; and after fine time delay alignment is completed, locking the time delay value of each channel. The digital down-conversion multichannel time delay alignment method aims at simplifying signal characteristics through down-conversion processing and combining dynamic threshold detection and a self-adaptive time delay compensation mechanism to achieve automatic alignment of multichannel time delay.

Description

Technical Field

[0001] This invention relates to the field of digital signal processing technology, specifically to a digital down-conversion multi-channel delay alignment method. Background Technology

[0002] In multi-channel digital receiving systems, due to differences in electronic components between channels, different signal transmission paths, and changes in environmental factors, there are often time delay differences between channels. These time delay differences can seriously affect the accuracy of signal processing and system performance.

[0003] Traditional time-delay alignment methods are severely limited in their universality when dealing with complex and ever-changing signal processing scenarios. They often rely heavily on specific signal formats, such as fixed frequencies, modulation methods, or coding rules. Once the signal characteristics change, such as frequency jumps or modulation method switching, the original alignment algorithm may fail and needs to be redesigned and adjusted, making it difficult to adapt to diverse practical application needs. On the other hand, from signal feature analysis and parameter setting to alignment effect evaluation, each step may rely on the experience and manual operation of professionals, which is not only time-consuming and labor-intensive, but also prone to errors introduced by human factors. Summary of the Invention

[0004] The purpose of this invention is to simplify signal characteristics through downconversion processing, and to achieve automatic alignment of multi-channel delays by combining dynamic threshold detection and adaptive delay compensation mechanisms. Therefore, a digital downconversion multi-channel delay alignment method is proposed.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] A method for aligning the delay of multiple channels in a digital down-conversion circuit includes the following steps:

[0007] S10: Perform down-conversion preprocessing on the digital signals received from each channel to convert high-frequency signals into low-frequency signals, simplifying the difficulty of subsequent processing;

[0008] S20: Perform dynamic threshold detection on the down-converted signal to determine the effective starting point of each channel signal;

[0009] S30: Estimate the initial time delay difference between each channel based on the dynamic threshold detection results;

[0010] S40: Based on the initial time delay difference, an adaptive PID controller is used to compensate for the time delay of each channel signal;

[0011] S50: Based on the initial delay compensation, the delay of each channel is further finely adjusted using sliding window technology;

[0012] S60: After completing fine-grained delay alignment, lock the delay values ​​of each channel and ensure the alignment effect through cross-validation.

[0013] Based on the above technical solution, the present invention can be further improved as follows.

[0014] Furthermore, the down-conversion preprocessing in S10 is achieved through the following formula:

[0015] y(t)=x(t)·cos(2πf c t+φ)

[0016] Where x(t) is the original signal, f c Where φ is the downconversion frequency, and φ is the phase offset;

[0017] The down-conversion frequency f c The frequency is dynamically adjusted based on the center frequency of the received signal.

[0018] Furthermore, the dynamic threshold detection in S20 dynamically calculates the threshold using the following formula:

[0019]

[0020] Where μ is the signal mean, σ is the amplitude adjustment factor, and ω is the angular frequency. This is the initial phase;

[0021] The amplitude adjustment factor σ is dynamically adjusted based on the signal strength and noise level.

[0022] Furthermore, the initial time delay difference estimation in step S30 includes:

[0023] Compare the time differences of the effective start points of signals in each channel to determine the preliminary time delay difference;

[0024] A weighted average algorithm is used to smooth the results of multiple detections, reducing random errors and improving the accuracy of the initial time delay difference estimation.

[0025] Furthermore, the adaptive PID controller in S40 achieves time delay compensation through the following formula:

[0026]

[0027] Where, e(n) = Δt i (n)-Δt ref (n) represents the time delay error, K p ,K i ,K d For PID parameters, Δt comp (n) represents the reference delay;

[0028] The PID parameters are dynamically adjusted according to the system response characteristics to achieve time delay compensation;

[0029] The dynamic adjustment strategy for the PID parameters includes:

[0030] A larger scaling factor K is used in the initial stage. p To quickly reduce time delay errors;

[0031] As the error decreases, gradually increase the integral coefficient K. i To eliminate static errors;

[0032] Introducing the differential coefficient K d This is to suppress overshoot and improve system stability.

[0033] Furthermore, the sliding window technology in S50 is implemented in the following way:

[0034] Set the sliding window length W and adjust it dynamically according to the signal characteristics;

[0035] Calculate the cross-correlation function of each channel signal within a sliding window to determine the optimal alignment position;

[0036] Adjusting the latency of each channel based on the peak value of the cross-correlation function achieves microsecond-level latency alignment.

[0037] The dynamic adjustment strategy for the sliding window length W includes:

[0038] The initial window length is determined based on the signal bandwidth and sampling rate;

[0039] During the alignment process, the window length is dynamically adjusted based on the sharpness of the peak value of the cross-correlation function to improve alignment accuracy;

[0040] The window length adjustment range is limited to a reasonable range to prevent excessive adjustment from causing system instability.

[0041] Furthermore, the cross-validation in S60 includes:

[0042] Cross-correlation function method: Calculate the time delay corresponding to the peak value of the cross-correlation function of each channel signal, and use it as an alignment standard;

[0043] Phase consistency check: Ensure that the phase difference of each channel signal is within the allowable error range after alignment;

[0044] Average of multiple verifications: The average of multiple verification results is taken to improve the reliability of the alignment effect.

[0045] Furthermore, the phase consistency check is performed through the following steps:

[0046] Step 1: Signal Phase Extraction

[0047] The Hilbert transform is applied to the aligned signals of each channel to extract the instantaneous phase information of the signal. For the signal of each channel, the corresponding analytic signal is obtained by the Hilbert transform, and then the real part and imaginary part are separated from the analytic signal. The instantaneous phase is calculated by the arctangent function using the ratio of the real part and the imaginary part.

[0048] Step 2: Phase Difference Calculation

[0049] Select one channel as the reference channel and calculate the phase difference between the other channels and the reference channel. To reduce the impact of noise on the phase difference calculation, the calculated phase difference is processed by moving average filtering. Specifically, a sliding window is set, which contains a certain number of sampling points. The average value of the phase difference of all sampling points in the window is calculated as the phase difference estimate at the current time.

[0050] Step 3: Phase Difference Consistency Determination

[0051] A maximum allowable phase difference threshold is preset. This threshold is set according to the system performance requirements and is usually a fraction of the signal period to ensure sufficient phase alignment accuracy. Within an observation time window, the phase difference between each channel and the reference channel is continuously monitored. If the absolute value of the estimated phase difference at all times within this observation time window does not exceed the preset maximum phase difference threshold, the phase consistency check is considered to have passed; otherwise, the phase consistency is considered to have failed the requirements, and the realignment process is triggered.

[0052] Step 4: Dynamic Threshold Adjustment

[0053] To adapt to different signal characteristics and environmental changes, a dynamic threshold adjustment mechanism can be introduced. Based on the statistical characteristics of historical phase difference data, including the mean and standard deviation, the maximum phase difference threshold can be dynamically updated. Specifically, a new threshold can be calculated by combining the historical phase difference mean and standard deviation in a certain proportion. The new threshold can be set as a certain multiple of the historical phase difference mean plus a certain multiple of the historical phase difference standard deviation. By adjusting these two multiples, the influence of the historical mean and standard deviation on the threshold can be balanced, ensuring that the threshold can both reflect the signal characteristics and adapt to environmental changes.

[0054] Furthermore, step S70 is also included:

[0055] Real-time monitoring of latency status of each channel; when abnormal latency changes are detected, triggering a realignment process.

[0056] The realignment process includes repeating steps S10 to S60 to ensure the system continues to operate stably.

[0057] Furthermore, the detection strategy for abnormal latency changes includes:

[0058] Set a threshold for latency changes; when a latency change in a channel is detected to exceed the threshold, it is considered abnormal.

[0059] By combining historical latency data, statistical methods are used to determine whether latency changes significantly deviate from the normal range;

[0060] Abnormal detection results are notified to maintenance personnel via an alarm mechanism for timely handling.

[0061] Compared with the prior art, the technical solution of this application has the following beneficial technical effects:

[0062] This invention performs down-conversion preprocessing on the digital signals received from each channel, converting high-frequency signals into low-frequency signals. This prevents subsequent delay detection and compensation processes from being excessively interfered with by the complex characteristics of high-frequency signals. Dynamic threshold detection is used to determine the effective starting point of each channel signal. The dynamic threshold can be adjusted in real time according to the actual signal characteristics, eliminating dependence on fixed signal formats. Even if the signal undergoes changes in characteristics such as modulation mode switching, the effective starting point can still be accurately identified. Based on the dynamic threshold detection results, the initial delay difference is estimated, providing a precise initial basis for subsequent compensation and ensuring the directionality and accuracy of delay adjustment. An adaptive PID controller is introduced for delay compensation. Utilizing its adaptive characteristics, the compensation parameters can be automatically adjusted according to the real-time system status, enabling rapid response to delay fluctuations caused by changes in environmental factors and enhancing the system's dynamic adaptability. Based on the initial delay compensation, a sliding window technique is used for further fine-tuning, enabling precise correction of subtle delay differences and significantly improving the accuracy of delay alignment. The delay values ​​of each channel are locked, and cross-validation ensures the alignment effect. Simultaneously, locking the delay values ​​guarantees the stability of the delay of each channel in subsequent processing, avoiding performance fluctuations. Attached Figure Description

[0063] Figure 1 This is a structural block diagram of a digital downconversion multi-channel delay alignment method according to the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] The present invention provides a digital down-conversion multi-channel delay alignment method, comprising the following steps:

[0066] S10: Perform down-conversion preprocessing on the digital signals received from each channel to convert high-frequency signals into low-frequency signals, simplifying the difficulty of subsequent processing;

[0067] S20: Perform dynamic threshold detection on the down-converted signal to determine the effective starting point of each channel signal;

[0068] S30: Estimate the initial time delay difference between each channel based on the dynamic threshold detection results;

[0069] S40: Based on the initial time delay difference, an adaptive PID controller is used to compensate for the time delay of each channel signal;

[0070] S50: Based on the initial delay compensation, the delay of each channel is further finely adjusted using sliding window technology;

[0071] S60: After completing fine-grained delay alignment, lock the delay values ​​of each channel and ensure the alignment effect through cross-validation.

[0072] The down-conversion preprocessing in S10 is achieved through the following formula:

[0073] y(t)=x(t)·cos(2πf c t+φ)

[0074] Where x(t) is the original signal, f c Where φ is the downconversion frequency, and φ is the phase offset;

[0075] downconversion frequency f c The frequency is dynamically adjusted based on the center frequency of the received signal to ensure effective conversion to the low-frequency band. This formula multiplies the high-frequency signal x(t) by a cosine carrier wave, using the spectrum shifting principle to convert the high-frequency signal to a low-frequency signal, avoiding interference from the complex spectrum of the high-frequency signal on subsequent time delay detection. Simultaneously, the down-conversion frequency f... c It can dynamically adjust according to the center frequency of the received signal to ensure stable subsequent time delay processing even when the signal frequency changes.

[0076] In practical applications, the center frequency of the received signal can be detected by a real-time spectrum analysis module. When the center frequency is detected to jump from f0 to f1, the down-conversion frequency f1 is determined. c It will be synchronized from f c0 Adjust to f c1 (usually f) c1 =f1-f low , where f low The target low-frequency center frequency (e.g., 100MHz) is preset, and the phase offset φ can be calibrated in real time through the phase-locked loop circuit to ensure the phase stability of the low-frequency signal after down-conversion, avoid signal distortion caused by phase offset, and thus ensure the accuracy of subsequent effective starting point detection.

[0077] The dynamic threshold detection in S20 dynamically calculates the threshold using the following formula:

[0078]

[0079] Where μ is the signal mean, σ is the amplitude adjustment factor, and ω is the angular frequency. This is the initial phase;

[0080] The amplitude adjustment factor σ is dynamically adjusted according to the signal strength and noise level to improve detection robustness. The function of this formula is to adjust the threshold in real time with the dynamic changes of the signal. σ changes dynamically with the signal strength and noise level. Even if the signal characteristics change, such as the modulation mode changing from amplitude modulation to frequency modulation, the effective starting point can still be accurately identified, thus improving the adaptability to diverse signals.

[0081] The signal mean μ is calculated by averaging the signal values ​​within a sliding window, such as a window with a length of 1024 sampling points. The dynamic adjustment of the amplitude adjustment factor σ is as follows: when the signal strength (calculated from the peak value of the signal power spectrum) is higher than -30dBm and the noise level (calculated from the power mean outside the signal bandwidth) is lower than -60dBm, σ is set to 0.3 times the signal standard deviation to improve detection sensitivity; when the noise level is higher than -50dBm, σ is increased to 0.8 times the signal standard deviation to avoid noise being misjudged as a valid signal. The angular frequency ω is determined according to the symbol rate of the signal to ensure that the threshold change matches the signal period.

[0082] The initial time delay difference estimation in step S30 includes:

[0083] Compare the time differences of the effective start points of signals in each channel to determine the preliminary time delay difference;

[0084] A weighted average algorithm is used to smooth multiple detection results, reducing random errors and improving the accuracy of initial delay difference estimation. The initial delay difference is determined by comparing the time difference of the effective starting points of signals in each channel, and the weighted average algorithm is used to smooth multiple detection results. The function of this algorithm is to assign greater weight to detection results with high signal-to-noise ratio (e.g., the weight of detection results with a signal-to-noise ratio higher than 20dB is 0.8, and the weight of those with a signal-to-noise ratio lower than 10dB is 0.2), which reduces the impact of random noise on the initial delay difference estimation, provides a more accurate initial basis for subsequent delay compensation, and improves the accuracy of delay adjustment.

[0085] In practice, 5-8 valid starting point detections are performed consecutively, with each detection interval being twice the signal period (e.g., 20 μs interval if the signal period is 10 μs) to cover the complete characteristic period of the signal. The weights of the weighted average are determined based on the signal-to-noise ratio (SNR) of each detection result. For example, when the SNR of a detection is 25 dB, the weight is set to 0.2; when the SNR is 30 dB, the weight is set to 0.3, ensuring that detection results with higher SNR have a greater impact on the final initial time delay difference. During smoothing, the delay difference obtained from each detection is first calculated, then each delay difference is multiplied by its corresponding weight and summed, and finally divided by the total weight to obtain the smoothed initial delay difference. For example, if the delay differences of 5 detections are 12μs, 11μs, 13μs, 12μs, and 14μs, and the corresponding weights are 0.15, 0.2, 0.25, 0.2, and 0.2, then the smoothed result is (12×0.15+11×0.2+13×0.25+12×0.2+14×0.2)÷(0.15+0.2+0.25+0.2+0.2)=12.35μs, which reduces the impact of random fluctuations in a single detection.

[0086] The adaptive PID controller in S40 achieves time delay compensation through the following formula:

[0087]

[0088] Where, e(n) = Δt i (n)-Δt ref (n) represents the time delay error, K p ,K i ,K d For PID parameters, Δt comp (n) represents the reference delay;

[0089] The PID parameters are dynamically adjusted according to the system response characteristics to achieve time delay compensation;

[0090] Dynamic adjustment strategies for PID parameters include:

[0091] A larger scaling factor K is used in the initial stage. p To quickly reduce time delay errors;

[0092] As the error decreases, gradually increase the integral coefficient K. i To eliminate static errors;

[0093] Introducing the differential coefficient K d To suppress overshoot and improve system stability, the adaptive PID controller dynamically adjusts the proportional coefficient K. p Integral coefficient K i Differential coefficient K d To achieve delay compensation, a relatively large K is used in the initial stage. pIt can quickly reduce time delay error; as the error decreases, K increases. i It can eliminate static error and introduce K d It can suppress overshoot, and the synergistic effect of the three factors enables time delay compensation to automatically adapt to changes in system state without manual intervention. Whether it is environmental temperature fluctuations or changes in signal characteristics, it can maintain the timeliness and accuracy of compensation, greatly reducing the cost and error of manual operation.

[0094] When the initial delay error is greater than 80μs, the scaling factor K p When the value is 6-8, the system adjusts the delay relatively quickly. For example, when the error is 100μs, a single compensation can reduce the error to within 40μs; when the error decreases to 20-80μs, the integral coefficient K... i By gradually increasing K from 0.1 to 0.6, cumulative error compensation is used to eliminate long-standing small deviations, such as when a static error of 5 μs persists. i After increasing the value, the error can be reduced to below 0.5 μs within 3 adjustment cycles; when the error is less than 20 μs, the differential coefficient K... d The value is set between 0.3 and 0.5 to prevent over-adjustment from causing the time delay to exceed the reference value, such as avoiding a sudden reversal from 10μs error to -5μs error. Parameter adjustment is triggered by the error change rate monitored in real time by the system. When the error change rate is less than 5μs / cycle for two consecutive cycles, it automatically increases K. p Phase switch to increase K i During this phase, ensure that parameter adjustments are precisely matched with error conditions.

[0095] The sliding window technology in S50 is implemented in the following way:

[0096] Set the sliding window length W and adjust it dynamically according to the signal characteristics;

[0097] Calculate the cross-correlation function of each channel signal within a sliding window to determine the optimal alignment position;

[0098] Adjusting the latency of each channel based on the peak value of the cross-correlation function achieves microsecond-level latency alignment.

[0099] The dynamic adjustment strategies for the sliding window length W include:

[0100] The initial window length is determined based on the signal bandwidth and sampling rate;

[0101] During the alignment process, the window length is dynamically adjusted based on the sharpness of the peak value of the cross-correlation function to improve alignment accuracy;

[0102] The window length adjustment range is limited to a reasonable range to prevent over-adjustment from causing system instability. The sliding window technology dynamically adjusts the window length and calculates the cross-correlation function of each channel signal within the window to determine the optimal alignment position. The window length is dynamically adjusted according to the signal bandwidth (the wider the bandwidth, the shorter the window can be), the sampling rate (the higher the sampling rate, the more feature points the window can contain), and the sharpness of the cross-correlation peak (the sharper the peak, the shorter the window can be to improve accuracy). It can accurately capture subtle signal features and achieve microsecond-level time delay alignment. Even if the signal undergoes characteristic changes such as modulation mode switching, it can still maintain high alignment accuracy and improve the system's adaptability to diverse signal scenarios.

[0103] The initial window length is determined based on the signal bandwidth and sampling rate. For example, with a signal bandwidth of 2MHz and a sampling rate of 20MHz, the initial window length is set to 200 sampling points (corresponding to a duration of 10μs). When the ratio of the peak value to the sidelobe of the cross-correlation function is greater than 8 (high sharpness), the window length is shortened to 150 sampling points to focus on finer signal features. When the ratio is less than 3 (low sharpness), the window length is extended to 250 sampling points to avoid losing key features. The window length adjustment range is limited to 100-300 sampling points to prevent insufficient features due to excessive length or increased computational load due to excessive length. When calculating the cross-correlation function, one channel signal is moved point by point within the window, and the similarity with the other channel signal is calculated. The position with the highest similarity (maximum cross-correlation value) is the optimal alignment point. This method can control the time delay alignment error within 0.3μs, achieving microsecond-level accuracy.

[0104] Cross-validation in S60 includes:

[0105] Cross-correlation function method: Calculate the time delay corresponding to the peak value of the cross-correlation function of each channel signal, and use it as an alignment standard;

[0106] Phase consistency check: Ensure that the phase difference of each channel signal is within the allowable error range after alignment;

[0107] Multiple verifications and averaging: The average value of multiple verification results is taken to improve the reliability of the alignment effect. Cross-verification is a combination of cross-correlation function method, phase consistency check and multiple verifications and averaging. Cross-correlation function method verifies the time delay alignment accuracy from the perspective of signal similarity, phase consistency check ensures signal phase synchronization, and multiple verifications and averaging reduces the random error of a single verification. The three work together to form a multi-dimensional evaluation system, making the judgment of the alignment effect more objective and reliable.

[0108] In the cross-correlation function method, the time delay error corresponding to the peak cross-correlation value between each channel and the reference channel must be less than 0.5 μs; otherwise, the alignment is deemed unqualified. The phase consistency check has a preset maximum phase difference threshold of 30 degrees, and the phase difference between each channel and the reference channel must consistently remain within this range. When averaging multiple verifications, six consecutive verifications are performed. After removing one maximum and one minimum value, the average of the remaining four results must be less than 0.3 μs to be considered aligned successfully. If any verification fails (e.g., cross-correlation error reaches 0.6 μs, phase difference reaches 35 degrees, or average value reaches 0.4 μs), the system will automatically trigger readjustment until all indicators meet the requirements.

[0109] Phase consistency checks are performed through the following steps:

[0110] Step 1: Signal Phase Extraction

[0111] The Hilbert transform is applied to the aligned signals of each channel to extract the instantaneous phase information of the signal. For the signal of each channel, the corresponding analytic signal is obtained by the Hilbert transform, and then the real part and imaginary part are separated from the analytic signal. The instantaneous phase is calculated by the arctangent function using the ratio of the real part and the imaginary part.

[0112] Step 2: Phase Difference Calculation

[0113] Select one channel (usually the main channel or the best performing channel) as the reference channel, calculate the phase difference between the other channels and the reference channel. To reduce the impact of noise on the phase difference calculation, perform a moving average filtering on the calculated phase difference. Specifically, set a sliding window containing a certain number of sampling points, calculate the average of the phase differences of all sampling points in the window, and use it as the phase difference estimate at the current time.

[0114] Step 3: Phase Difference Consistency Determination

[0115] A maximum allowable phase difference threshold is preset. This threshold is set according to the system performance requirements and is usually a fraction of the signal period to ensure sufficient phase alignment accuracy. Within an observation time window, the phase difference between each channel and the reference channel is continuously monitored. If the absolute value of the estimated phase difference at all times within this observation time window does not exceed the preset maximum phase difference threshold, the phase consistency check is considered to have passed; otherwise, the phase consistency is considered to have failed the requirements, and the realignment process is triggered.

[0116] Step 4: Dynamic Threshold Adjustment

[0117] To adapt to different signal characteristics and environmental changes, a dynamic threshold adjustment mechanism can be introduced. Based on the statistical characteristics of historical phase difference data, including the mean and standard deviation, the maximum phase difference threshold is dynamically updated. Specifically, a new threshold can be calculated by combining the historical phase difference mean and standard deviation in a certain proportion. The new threshold can be set as a certain multiple of the historical phase difference mean plus a certain multiple of the historical phase difference standard deviation. By adjusting these two multiples (i.e., weighting coefficients), the influence of the historical mean and standard deviation on the threshold is balanced, ensuring that the threshold can reflect signal characteristics and adapt to environmental changes. The phase consistency check involves steps such as extracting the instantaneous phase through Hilbert transform, calculating the phase difference, and dynamic threshold determination. Hilbert transform can accurately extract the instantaneous phase of the signal, ensuring the integrity of phase information. The phase difference calculation uses a sliding window averaging to reduce noise interference, making the phase difference assessment more stable. Dynamic threshold adjustment allows the phase difference determination to adapt to signal characteristics (such as the phase change patterns of different modulation methods) and environmental noise changes, ensuring that after alignment, the signals of each channel are not only time-synchronized but also have consistent phases, improving the signal-to-noise ratio and system processing accuracy after multi-channel signal synthesis.

[0118] During signal phase extraction, a Hilbert transform is performed on the real signal to obtain an analytic signal (containing real and imaginary parts). The instantaneous phase is calculated by the ratio of the imaginary part to the real part. For example, when the real part of a channel signal is 2 and the imaginary part is 3, the instantaneous phase is arctan(3 / 2)≈56.3 degrees. The phase difference calculation uses a sliding window of 50 sampling points. For example, at a sampling rate of 10MHz, the window duration is 5μs. The average phase difference between each channel and the reference channel within the window is calculated to reduce the impact of phase jumps caused by sudden noise. During dynamic threshold adjustment, the mean and standard deviation are calculated using the phase difference data of the past 50 sliding windows. The new threshold is 0.6 times the mean plus 0.4 times the standard deviation. When the signal switches from a Gaussian channel to a Rayleigh fading channel (increased noise), the standard deviation increases, and the threshold is automatically increased by 10%-20% to avoid phase fluctuations caused by noise being misjudged as inconsistencies.

[0119] It also includes step S70:

[0120] Real-time monitoring of latency status of each channel; when abnormal latency changes are detected, triggering a realignment process.

[0121] The realignment process involves repeating steps S10 to S60 to ensure continuous and stable system operation. It monitors the delay status of each channel in real time and triggers the realignment process in case of anomalies. This solves the problem of long-term delay offset caused by environmental factors (such as temperature changes causing component delay drift) or sudden interference (such as strong electromagnetic interference causing instantaneous delay abrupt changes) that traditional methods cannot handle. Real-time monitoring ensures timely detection of delay anomalies, and the realignment process quickly restores the alignment state by repeating the previous processing steps. This ensures that the system maintains stable delay synchronization throughout long-term operation, preventing signal processing accuracy from decreasing due to accumulated delay differences and improving the continuous reliability of the system.

[0122] Real-time monitoring collects latency data for each channel every 8ms (accuracy up to 0.1μs), and calculates the change by comparing it with the data from the previous moment. The anomaly detection condition is that the latency change in 4 consecutive collections exceeds 3μs, or the single change exceeds 15μs (such as instantaneous jumps caused by sudden interference). During realignment, the system will call historical data such as the downconversion frequency and PID initial parameters from the previous alignment as the basis, which shortens the realignment time from 120ms in the first alignment to 60ms, reducing the impact on normal system operation and ensuring stable operation in scenarios with high real-time requirements such as radar and communication.

[0123] Strategies for detecting abnormal latency changes include:

[0124] Set a threshold for latency changes; when a latency change in a channel is detected to exceed the threshold, it is considered abnormal.

[0125] By combining historical latency data, statistical methods are used to determine whether latency changes significantly deviate from the normal range;

[0126] Anomaly detection results are notified to maintenance personnel via an alarm mechanism for timely handling. The detection strategy for abnormal latency changes combines threshold judgment, statistical methods, historical data, and alarm mechanisms. Threshold judgment can quickly identify significant anomalies (such as sudden large latency changes), statistical methods (combined with historical data) can discover subtle but continuous abnormal trends (such as slow drift), and the alarm mechanism promptly notifies maintenance personnel. This achieves early detection and early handling of anomalies, reduces system performance degradation or downtime caused by abnormal latency, and improves system operation and maintenance efficiency.

[0127] The latency variation threshold is set according to the system type, for example, 4μs for civilian communication systems and 1μs for military radar systems. The statistical method adopts the 3σ principle, calculating the mean and standard deviation of the past 1000 latency data. When the latency value of a certain channel exceeds the mean ± 3 times the standard deviation (e.g., when the mean is 10μs and the standard deviation is 1μs, exceeding the range of [7μs, 13μs] is considered abnormal), the alarm mechanism includes local red indicator light flashing + buzzer beeping intermittently (once every 2 seconds), and remotely sending information to the monitoring center containing the abnormal channel number (e.g., channel 3), the current latency value (e.g., 25μs), and the historical mean (e.g., 10μs). Maintenance personnel can quickly locate whether it is a channel hardware failure or environmental interference problem through the information.

[0128] First, the system performs down-conversion preprocessing on the high-frequency digital signals received by each channel. By multiplying the high-frequency signal with the dynamically adjusted cosine carrier down-conversion frequency in real time according to the center frequency of the received signal, the high-frequency signal is converted into a low-frequency signal by spectrum shifting, which simplifies the complexity of subsequent processing. Even if the signal frequency jumps, the down-conversion frequency can be adjusted synchronously through the real-time spectrum analysis module to ensure stable output of the low-frequency signal.

[0129] Next, dynamic threshold detection is used to determine the effective starting point of each channel signal. The threshold is calculated by the signal mean, an amplitude adjustment factor that is dynamically adjusted according to the signal strength and noise level (such as taking a small value to improve sensitivity when the signal-to-noise ratio is high and taking a large value to avoid misjudgment when the noise is high), and an angular frequency that matches the signal period. This eliminates the dependence on fixed modulation methods and coding rules and accurately identifies the starting point of different characteristic signals.

[0130] Subsequently, the initial time delay difference is estimated based on the time difference of the effective starting point. The result is smoothed by performing 5-8 consecutive detections and using a weighted average algorithm. The result with the higher signal-to-noise ratio has a higher weight, which reduces random errors and provides an accurate initial basis for subsequent compensation.

[0131] Subsequently, an adaptive PID controller is introduced for time delay compensation. In the initial stage, a large proportional coefficient is used to quickly reduce the time delay error. When the error decreases, the integral coefficient is increased to eliminate the static error. When the error is small, the derivative coefficient is used to suppress overshoot. The parameters automatically switch with the error change rate, and can adapt to environmental and signal changes without manual intervention.

[0132] After compensation, the window length is further finely adjusted using sliding window technology. The window length can be shortened according to the signal bandwidth, the sampling rate can include more features, and the sharpness of the cross-correlation peak can be shortened. The window is dynamically adjusted, and the cross-correlation function is calculated within the window to determine the optimal alignment position, achieving microsecond-level accuracy.

[0133] After fine-tuning, cross-validation is used to ensure alignment, including checking the peak delay error of the cross-correlation function to be less than 0.5μs, setting a 30-degree threshold for phase consistency, extracting instantaneous phase through Hilbert transform, averaging the phase difference using a sliding window and dynamically adjusting the threshold to adapt to signal and environmental changes, and ensuring that the average value after removing extreme values ​​from multiple verifications is less than 0.3μs. Multi-dimensional verification replaces manual evaluation, improving objectivity. At the same time, the system adds a real-time monitoring link S70, which collects delay data for each channel every 8ms. Threshold judgment, such as a 1μs threshold for military systems, and the 3σ statistical method combined with historical data are used to identify subtle drifts and detect abnormal delay changes. Once multiple consecutive out-of-tolerance errors or a single abrupt change are detected, the realignment process is immediately triggered, reusing historical parameters to shorten the time to 60ms. Maintenance personnel are notified through local audible and visual alarms and remote information push. Ultimately, the system achieves fully automated, high-precision delay alignment of multi-channel signals from preprocessing to continuous stable operation, effectively overcoming the shortcomings of traditional methods such as dependence on specific signal formats, human operation errors, and poor environmental adaptability.

[0134] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0135] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital down-conversion multi-channel time delay alignment method, characterized in that, Includes the following steps: S10: Perform down-conversion preprocessing on the digital signals received from each channel to convert high-frequency signals into low-frequency signals, simplifying the difficulty of subsequent processing; S20: Perform dynamic threshold detection on the down-converted signal to determine the effective starting point of each channel signal; S30: Estimate the initial time delay difference between each channel based on the dynamic threshold detection results; S40: Based on the initial time delay difference, an adaptive PID controller is used to compensate for the time delay of each channel signal; S50: Based on the initial delay compensation, the delay of each channel is further finely adjusted using sliding window technology; S60: After completing fine-grained delay alignment, lock the delay values ​​of each channel and ensure the alignment effect through cross-validation.

2. The digital down-conversion multi-channel delay alignment method according to claim 1, characterized in that, The down-conversion preprocessing in S10 is achieved through the following formula: y(t)=x(t)·cos(2πf c t+φ) Where x(t) is the original signal, f c Where φ is the downconversion frequency, and φ is the phase offset; The down-conversion frequency f c The frequency is dynamically adjusted based on the center frequency of the received signal.

3. The digital down-conversion multi-channel delay alignment method according to claim 1, characterized in that, The dynamic threshold detection in S20 dynamically calculates the threshold using the following formula: Where μ is the signal mean, σ is the amplitude adjustment factor, and ω is the angular frequency. This is the initial phase; The amplitude adjustment factor σ is dynamically adjusted based on the signal strength and noise level.

4. The digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, The initial time delay difference estimation in step S30 includes: Compare the time differences of the effective start points of signals in each channel to determine the preliminary time delay difference; A weighted average algorithm is used to smooth the results of multiple detections, reducing random errors and improving the accuracy of the initial time delay difference estimation.

5. The digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, The adaptive PID controller in S40 achieves time delay compensation through the following formula: Where, e(n) = Δt i (n)-Δt ref (n) represents the time delay error, K p ,K i ,K d For PID parameters, Δt comp (n) represents the reference delay; The PID parameters are dynamically adjusted according to the system response characteristics to achieve time delay compensation; The dynamic adjustment strategy for the PID parameters includes: A larger scaling factor K is used in the initial stage. p To quickly reduce time delay errors; As the error decreases, gradually increase the integral coefficient K. i To eliminate static errors; Introducing the differential coefficient K d This is to suppress overshoot and improve system stability.

6. The digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, The sliding window technology in S50 is implemented in the following way: Set the sliding window length W and adjust it dynamically according to the signal characteristics; Calculate the cross-correlation function of each channel signal within a sliding window to determine the optimal alignment position; Adjusting the latency of each channel based on the peak value of the cross-correlation function achieves microsecond-level latency alignment. The dynamic adjustment strategy for the sliding window length W includes: The initial window length is determined based on the signal bandwidth and sampling rate; During the alignment process, the window length is dynamically adjusted based on the sharpness of the peak value of the cross-correlation function to improve alignment accuracy; The window length adjustment range is limited to a reasonable range to prevent excessive adjustment from causing system instability.

7. The digital down-conversion multi-channel delay alignment method according to claim 1, characterized in that, The cross-validation in S60 includes: Cross-correlation function method: Calculate the time delay corresponding to the peak value of the cross-correlation function of each channel signal, and use it as an alignment standard; Phase consistency check: Ensure that the phase difference of each channel signal is within the allowable error range after alignment; Average of multiple verifications: The average of multiple verification results is taken to improve the reliability of the alignment effect.

8. The digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, The phase consistency check is performed through the following steps: Step 1: Signal Phase Extraction The Hilbert transform is applied to the aligned signals of each channel to extract the instantaneous phase information of the signal. For the signal of each channel, the corresponding analytic signal is obtained by the Hilbert transform, and then the real part and imaginary part are separated from the analytic signal. The instantaneous phase is calculated by the arctangent function using the ratio of the real part and the imaginary part. Step 2: Phase Difference Calculation Select one channel as the reference channel and calculate the phase difference between the other channels and the reference channel. To reduce the impact of noise on the phase difference calculation, the calculated phase difference is processed by moving average filtering. Specifically, a sliding window is set, which contains a certain number of sampling points. The average value of the phase difference of all sampling points in the window is calculated as the phase difference estimate at the current time. Step 3: Phase Difference Consistency Determination A maximum allowable phase difference threshold is preset. This threshold is set according to the system performance requirements and is usually a fraction of the signal period to ensure sufficient phase alignment accuracy. Within an observation time window, the phase difference between each channel and the reference channel is continuously monitored. If the absolute value of the estimated phase difference at all times within this observation time window does not exceed the preset maximum phase difference threshold, the phase consistency check is considered to have passed; otherwise, the phase consistency is considered to have failed the requirements, and the realignment process is triggered. Step 4: Dynamic Threshold Adjustment To adapt to different signal characteristics and environmental changes, a dynamic threshold adjustment mechanism can be introduced. Based on the statistical characteristics of historical phase difference data, including the mean and standard deviation, the maximum phase difference threshold can be dynamically updated. Specifically, a new threshold can be calculated by combining the historical phase difference mean and standard deviation in a certain proportion. The new threshold can be set as a certain multiple of the historical phase difference mean plus a certain multiple of the historical phase difference standard deviation. By adjusting these two multiples, the influence of the historical mean and standard deviation on the threshold can be balanced, ensuring that the threshold can both reflect the signal characteristics and adapt to environmental changes.

9. A digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, It also includes step S70: Real-time monitoring of latency status of each channel; when abnormal latency changes are detected, triggering a realignment process. The realignment process includes repeating steps S10 to S60 to ensure the system continues to operate stably.

10. A digital down-conversion multi-channel time delay alignment method according to claim 1, characterized in that, The detection strategy for abnormal delay changes includes: Set a threshold for latency changes; when a latency change in a channel is detected to exceed the threshold, it is considered abnormal. By combining historical latency data, statistical methods are used to determine whether latency changes significantly deviate from the normal range; Abnormal detection results are notified to maintenance personnel via an alarm mechanism for timely handling.

Images