Calibration method for impulse radar transmitted pulse jitter

Abstract

The invention discloses a calibration method for impulse radar transmitted pulse jitter, and the method comprises the steps: capturing a transmitted pulse string sample of an impulse radar, generating a reference pulse waveform, and constructing a delay pulse sample set through random numbers; according to the reference pulse waveform and the delay pulse sample set, obtaining a measurement result by using a delay measurement mode based on multiple rules; according to the delay pulse sample set and the measurement result, constructing a delay measurement data set, carrying out fusion mapping relation training learning, and generating a fusion mapping relation function; a direct wave of the impulse radar is actually measured, and a measurement result is obtained through a delay measurement mode based on multiple rules; and according to the measurement result, obtaining a fusion value of the measurement result by using a fusion mapping relation function, and using the fusion value to delay the correction parameter.

Description

Technical Field

[0001] The present invention belongs to the technical field of signal calibration of impulse radar, and specifically relates to a neural network learning and training technology. Background Art

[0002] Compared with conventional radar, impulse radar has a smaller pulse width and a wider spectrum width, and has advantages and application potential in low intercept, anti-interference, and target recognition.

[0003] The characteristic of a small time width of the transmitted signal causes the electromagnetic wave energy radiated and received by the ultra-wideband transient detection system to be significantly reduced, limiting the detection power.

[0004] The fundamental means to improve the detection power is to radiate more electromagnetic energy and effectively accumulate the echo. In impulse radar, there is a contradiction between high radiation power and large number of pulse accumulations. The RF pulse source that can generate high peak power has unsatisfactory indexes in terms of pulse repetition frequency and its stability. There is a large cumulative gain loss in the multi-pulse accumulation of impulse radar echo. Summary of the Invention

[0005] In order to solve the technical problem of small energy of impulse radar, a technical solution of using a neural network learning mapping function is adopted, which produces the technical effect of accurately correcting jitter and provides accurate jitter compensation parameters.

[0006] Grab a series of transmitted pulse waveforms W of the impulse radar, perform iterative correction processing on multiple pulses to generate a single reference pulse waveform R with high signal-to-noise ratio. Let the nominal period of the transmitted pulse of the impulse radar be T; let t i be a random number uniformly distributed in the interval [−T / 2, T / 2), and use t i to randomly delay the waveform R to generate a randomly delayed signal x i , and construct a delayed pulse sample set X = {(x1, t1), (x2, t2)……(x i , t i )……}; use the rule m (k) to measure the delay t i of the delayed signal x i relative to the waveform R; t i (k) , and obtain data of multiple regularly measured random numbers t i , and construct a delayed measurement data set D = {(t1; t1 (1) , t1 (2) ……t1 (k) ……)……(t i ; t i (1) , t i (2) ……ti (k) ……)……}; Use a neural network model to train and learn the fusion mapping relationship function f based on set D; Use rule m (k) Measure the actual delay of the impulse radar pulse to obtain {s (1) ,s (2) ……s (k) ……}; Use the function f to calculate the fusion value of the actual pulse delay as f(s (1) ,s (2) ……s (k) ……); Input the fusion value into the information processing system of the impulse radar to correct the jitter of the transmitted pulse.

[0007] Perform iterative correction processing on multiple pulses: Use the period T to divide the waveform W into N single pulse waveforms, take the first one as the calibration reference waveform, and perform Fourier transforms on the N - 1 single pulse waveforms and the calibration reference waveform respectively; Calculate the spectral phase differences between the N - 1 single pulse waveforms and the calibration reference waveform to obtain N - 1 phase difference - frequency curves; Perform polynomial fitting on the N - 1 phase difference - frequency curves to obtain N - 1 slopes; Calculate the delays of the N - 1 single pulse waveforms according to the slopes, perform delay correction respectively, and calculate the arithmetic mean waveform S of the N - 1 single pulse waveforms in the time domain; Use S to replace the first single pulse waveform as the calibration reference waveform, repeat the above process until S converges, and use S to replace R. Description of the Drawings

[0008] Figure 1 is the calibration flow chart, Figure 2 is the flow chart for generating the reference pulse waveform, Figure 3 is the waveform diagram before calibration, Figure 4 is the waveform diagram after calibration. Detailed Implementation Manner

[0009] The technical solution of the present invention will be specifically described below with reference to the drawings.

[0010] The calibration process is as Figure 1 shown. Grab the transmitted pulse train samples of the impulse radar, generate the reference pulse waveform, and construct a delay pulse sample set with random numbers. According to the reference pulse waveform and the delay pulse sample set, use a delay measurement method based on multiple rules to obtain the measurement results. According to the delay pulse sample set and the measurement results, construct a delay measurement data set, perform fusion mapping relationship training and learning, and generate a fusion mapping relationship function. Measure the direct wave of the actual impulse radar, use a delay measurement method based on multiple rules to obtain the measurement results. According to the measurement results, use the fusion mapping relationship function to obtain the fusion value of the measurement results for the delay correction parameters.

[0011] The process of generating the reference pulse waveform is as Figure 2As shown, the pulse is segmented and the calibrated reference waveform is taken. The segmented pulse is successively subjected to Fourier transform, phase difference - frequency curve calculation, curve polynomial fitting, delay correction, and average waveform calculation using the calibrated reference waveform. If the waveform converges at this time, this waveform is used as the reference pulse waveform. Otherwise, this waveform is used to replace the original calibrated reference waveform, and the above - mentioned process is repeated for pulse segmentation until convergence.

[0012] Rule m (1) For cross - correlation weighted average: The cross - correlation operation is performed between the pulse sequence to be measured and the reference pulse sequence to obtain the peak value. M points are taken on each side of the peak value, and the 2M + 1 values are weighted - averaged to improve the accuracy.

[0013] Rule m (2) For the sum of squares or absolute values of the difference signal: The difference operation is performed between the pulse waveform to be measured and waveform R to obtain the difference waveform, and the square of the difference waveform is calculated, or the sum of the absolute values of the difference waveform is calculated.

[0014] Rule m (3) For the zero - crossing point of the main rising edge: The two points closest to the zero - crossing point are taken and weighted - averaged to improve the accuracy.

[0015] Rule m (4) For Fourier transform: The spectra of the pulse sequence to be measured and the reference pulse sequence are calculated respectively, the difference operation is performed to obtain the phase difference - frequency curve, and the slope is obtained by fitting.

[0016] The waveform before calibration is as Figure 3 shown. After adopting the technical solution of the present invention, the delay measurement accuracy can be improved by one order of magnitude compared with the sampling interval of digital reception, more accurate jitter correction can be realized, the radiated signals are fully utilized for learning and calibration, the consistency with the actual measurement scenario is high, and it can adapt to the characteristics of waveform distortion in the impulse radar transceiver link. The improvement effect is as Figure 4 shown.

[0017] The above are examples of the present invention and do not limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are all included in the protection scope of the present invention.

Claims

1. A calibration method for impulse radar transmitted pulse jitter, characterized in that, Including: Grabbing the transmitted pulse train samples of the impulse radar, generating a reference pulse waveform, and constructing a set of delayed pulse samples with random numbers; According to the reference pulse waveform and the set of delayed pulse samples, using a delay measurement method based on multiple rules to obtain a measurement result; According to the set of delayed pulse samples and the measurement result, constructing a set of delay measurement data, performing training and learning on the fusion mapping relationship, and generating a fusion mapping relationship function; measuring the direct wave of the impulse radar, using a delay measurement method based on multiple rules to obtain a measurement result; according to the measurement result, using the fusion mapping relationship function to obtain the fusion value of the measurement result for the delay correction parameter.

2. The calibration method for impulse radar transmitted pulse jitter according to claim 1, characterized in that, The generating of the reference pulse waveform includes: grabbing a series of transmitted pulse waveforms W of the impulse radar, performing iterative correction processing on multiple pulses to generate a single reference pulse waveform R with high signal-to-noise ratio, and the nominal period of the transmitted pulse of the impulse radar is T.

3. The calibration method for impulse radar transmitted pulse jitter according to claim 1, characterized in that, It also includes: Let t i be a random number uniformly distributed in the interval [﹣T / 2, T / 2), and use t i to randomly delay the waveform R to generate a randomly delayed signal x i , and construct a delayed pulse sample set X = {(x1, t1), (x2, t2)……(x i , t i )……}; Use the rule m (k) to measure the delay t i of the delayed signal x relative to the waveform R i ; t i (k) , and obtain data of multiple regularly measured random numbers t i , and construct a delayed measurement data set D = {(t1; t1 (1) , t1 (2) ……t1 (k) ……)……(t i ; t i (1) , t i (2) ……t i (k) ……)……}; Use a neural network model to train and learn a fusion mapping relationship function f according to the set D; Use the rule m (k) to measure the actual delay of the impulse radar pulse and obtain {s (1) , s (2) ……s (k) ……}; Use the function f to calculate the fusion value of the actual pulse delay as f(s (1) , s (2) ……s (k) ……); Input the fusion value into the information processing system of the impulse radar to correct the jitter of the transmitted pulse.

4. The calibration method for impulse radar transmitted pulse jitter according to claim 2, characterized in that, The generating of the reference pulse waveform includes: splitting the pulse, selecting a calibrated reference waveform; successively performing Fourier transform, phase difference-frequency curve calculation, curve polynomial fitting, delay correction, and average waveform calculation on the pulse split by the calibrated reference waveform; if the waveform converges, using this waveform as the reference pulse waveform; otherwise, replacing the original calibrated reference waveform with this waveform and repeating the above process for the pulse split until the waveform converges.

5. The calibration method for impulse radar transmitted pulse jitter according to claim 4, characterized in that, The constructing of the set of delay measurement data includes: splitting the waveform W into N single pulse waveforms with the period T, taking the first one as the calibrated reference waveform, and performing Fourier transform on N - 1 single pulse waveforms and the calibrated reference waveform respectively; calculating the spectral phase differences between N - 1 single pulse waveforms and the calibrated reference waveform to obtain N - 1 phase difference-frequency curves; performing polynomial fitting on N - 1 phase difference-frequency curves to obtain N - 1 slopes; calculating the delays of N - 1 single pulse waveforms according to the slopes, performing delay correction respectively, and calculating the arithmetic average waveform S of N - 1 single pulse waveforms in the time domain; using S to replace the first single pulse waveform as the calibrated reference waveform and repeating the above process until S converges, and using S to replace R.

6. The calibration method for impulse radar transmitted pulse jitter according to claim 3, characterized in that, The rule m (k) , including: the rule m (1) is a cross-correlation weighted average. Perform a cross-correlation operation on the pulse sequence to be measured and the reference pulse sequence to obtain a peak value. Take M points on each side of the peak value and perform a weighted average on the 2M + 1 values.

7. The calibration method for impulse radar transmitted pulse jitter according to claim 3, characterized in that, The rule m (k) , including: the rule m (2) is the sum of the square or absolute value of the difference signal. The difference operation is performed between the pulse waveform to be measured and the waveform R to obtain a differential waveform, and then the square of the differential waveform is calculated, or the absolute value of the differential waveform is summed up.

8. The calibration method for impulse radar transmitted pulse jitter according to claim 3, characterized in that, The rule m (k) , including: the rule m (3) is the zero crossing point of the main rising edge: Take the two points closest to the zero crossing point position and perform weighted averaging.

9. The calibration method for impulse radar transmitted pulse jitter according to claim 3, characterized in that, The rule m (k) , including: the rule m (4) is the Fourier transform: calculate the spectra of the pulse sequence to be measured and the reference pulse sequence respectively, perform a difference operation to obtain a phase difference - frequency curve, and fit to obtain the slope.

Images