Continuous probing waveform design and probing method of low PAPR weak cross-correlation sequence

Abstract

The application discloses a kind of low PAPR weak correlation sequence's continuous probe waveform design and detection method, comprising: initialization generates a unit complex sequence set P, multiple iterations to each sequence p in P n Two-stage update is carried out, and the optimal sequence set P * ; Each sequence in P * Is modulated to the sending pulse and sequentially arranged with single frequency signal as interval, and is modulated into continuous probe waveform by up-conversion, and is sent out;After receiving end receives reflected waveform, baseband signal is obtained after down-conversion, and is input into initial matching filter set, and target object movement speed v1 and distance d1 at time t1 are preliminarily obtained;Based on the matching filter result of the n-1th sequence pulse signal, the target object movement speed v n At each time t n It is estimated in succession, and the target object distance d n Is obtained after CFO compensation and resampling and matching filtering of the nth sequence pulse signal.This application reduces the range-velocity ambiguity caused by echo aliasing, and realizes the continuous detection of underwater target.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of underwater detection, and particularly relates to a continuous detection waveform design and detection method of low PAPR weak cross-correlation sequence. BACKGROUND

[0002] In active detection of underwater targets, the traditional single pulse waveform (such as LFM, HFM) has a detection blind area due to the need to wait for the complete echo, and cannot continuously track the real-time distance-velocity changes of the target, especially the trajectory tracking of high-speed moving targets has the problem of insufficient accuracy. In order to realize continuous detection, if a fast repetitive pulse sequence is used, its strong cross-correlation will cause echo aliasing, causing distance-velocity ambiguity. Especially in underwater multipath channels, strong reverberation interference will further exacerbate signal distortion and reduce the detection probability of weak targets. If a set of orthogonal sequence signals are used for pulse modulation, although they have good weak cross-correlation, the time domain envelope fluctuates dramatically, and has a high peak-to-average power ratio (Peak-to-Average Power Ratio, hereinafter referred to as PAPR), which limits the average transmit power of the transducer and shortens the detection distance. Therefore, how to ensure the average transmit power of the pulse signal while designing a detection waveform with good continuous detection capability is a technical problem. SUMMARY

[0003] In view of the problem that the existing underwater target detection technology is difficult to realize continuous detection of underwater targets, the present application provides a continuous detection waveform design and detection method of low PAPR weak cross-correlation sequence, which realizes continuous detection and real-time tracking of distance-velocity of underwater targets through low PAPR weak cross-correlation sequence and multi-pulse continuous detection waveform design, overcomes the aliasing interference between pulses, improves the signal output efficiency of the transducer, and optimizes the noise immunity and sidelobe suppression performance, and significantly enhances the detection robustness in complex underwater acoustic environment.

[0004] The specific technical solutions are as follows:

[0005] A continuous detection waveform design and detection method of low PAPR weak cross-correlation sequence, comprising the following steps:

[0006] S1: initialize to generate a unit complex sequence set P containing N sequences with a sequence length of L, and give a PAPR constraint Γ PAPR , an initial target cross-correlation lower bound Γ mut , a scaling factor φ and a convergence threshold ε, a target cross-correlation lower bound update step γ, and an initial sequence update step τ q , τ PAPR , and update each sequence p n in the unit complex sequence set P in two stages each time to obtain an optimal sequence set P * ; pn Let n be the nth unit complex sequence in the set of unit complex sequences P, where n = 1, 2, 3, ..., N;

[0007] S2: The optimal sequence set P * Each sequence in the signal is modulated onto the transmitted pulse by a pulse shaping filter to form a pulse signal x. n (t); Arrange the modulated pulse signals in sequence, with a frequency of f. cw Duration is T cw The single-frequency signal is used as the interval between two adjacent pulse signals to form a baseband waveform x(t), which is then up-converted and modulated onto the carrier f. c Continuous detection waveform is formed on top And it is transmitted through the transmitter;

[0008] S3: The receiver receives the reflected waveform. After down-conversion, the baseband signal y(t) is obtained and input into the initial matched filter set to obtain a coarse estimate of the Doppler velocity α1, the target velocity v1, and the distance d1 at the moment the first sequence pulse signal is received. Based on the matched filtering result of the (n-1)th sequence pulse signal, the nth segment of single-frequency signal is extracted, and the moment t at the moment the nth sequence pulse signal is received is iteratively estimated. n Precise estimation of Doppler velocity α n The target object's velocity v n According to α n After CFO compensation and resampling of the nth sequence pulse signal, it is input into the nth matched filter, and the corresponding time t is obtained based on the peak value of the matched filter result. n The target distance d below n .

[0009] Furthermore, S1 is implemented through the following sub-steps:

[0010] S1.1: Initialize and generate a unit complex sequence set P containing N sequences, each of length L; set PAPR constraint Γ. PAPR Initial target cross-correlation lower bound Γ mut The scaling factor φ and the convergence threshold ε are used to initialize the sequence update step size τ. q and τ PAPR The target cross-correlation lower bound update step size γ, and the initialization unit complex sequence set overall update round a=1;

[0011] S1.2: Select the nth sequence p in P. n Calculate sequence p n With other sequences p in P m The minimum distance vector d between them n,m , m=1,2,3,…,N, and m≠n;

[0012] S1.3: Integrating sequence p into P n The minimum distance satisfies All p m The sequence yields the set P. mut Based on this, the formula used to represent sequence p is calculated. n The first collision vector u in the iteration direction during the first stage update n Update the step size τ with the first sequence. q In the iteration direction u n Above p n The first phase of the update was performed, and the results were obtained. and update the unit complex sequence set;

[0013] S1.4: Generate a set of unit norm vectors Q containing N sequences, each of length L, based on the updated set of unit complex sequences P; the nth sequence q in Q... n The standard unit modulus vector has non-zero values ​​in q. n The position in the sequence is consistent with the position of the element with the smallest modulus in the nth sequence of the unit complex sequence set P, and this value is the normalized result of the element with the smallest modulus in the nth sequence of P.

[0014] S1.5: Compare the nth sequence in the current set of unit complex sequences with the nth sequence q in the set of unit norm vectors Q. n Minimum distance vector between ;

[0015] S1.6: Integrate the current unit complex sequence set to satisfy the minimum distance Given all sequences, we obtain the set P. PAPR Based on this, calculations are performed to represent sequences. The second collision vector during the iteration direction of the second phase update And update the step size τ with the second sequence. PAPR In the direction of iteration Top Perform the second phase of the update to obtain... ;

[0016] S1.7: Determine the set P at this time. PAPR Is it empty? If so, then complete the processing of the current nth sequence p. n The two-stage update is denoted as . If not, execute S1.8; otherwise, proceed with... The value is updated to Then execute S1.4-S1.7 again;

[0017] S1.8: judge whether all sequences in the current unit complex sequence set have been updated completely, complete a round of update; if not, update the value of n, and return to execute S1.2-S1.7; if yes, calculate the unit complex sequence set P after the current round of update a , the peak cross-correlation coefficient μ a of all sequences in P a ;

[0018] S1.9: judge whether the peak cross-correlation coefficient exceeds the initial target cross-correlation lower bound, if μ mut ≤Γ mut , then tighten the target cross-correlation lower bound Γ mut , that is, subtract the target cross-correlation lower bound update step γ from Γ mut ; otherwise, relax the target cross-correlation lower bound Γ mut , that is, add the target cross-correlation lower bound update step γ to Γ

[0019] At the same time, compare the peak cross-correlation coefficient μ a-1 of the unit complex sequence set before the a-th round of update with the peak cross-correlation coefficient μ a of the unit complex sequence set after the a-th round of update; if the peak cross-correlation coefficient after the update is lower than that before the update, then enlarge the update step τ q , τ PAPR by the scaling factor φ; if the peak cross-correlation coefficient after the update is higher than that before the update, then reduce the update step τ q , τ PAPR by the scaling factor φ; if the peak cross-correlation coefficients before and after the update remain unchanged, then the current sequence update step remains unchanged;

[0020] S1.10: judge whether the absolute value of the difference between the peak cross-correlation coefficients of the unit complex sequence set before and after the a-th round of update is less than the convergence threshold ε, if not, update the value of a, and execute S1.2 to S1.9 again; if yes, output the current unit complex sequence set as the optimal sequence set P * .

[0021] Further, in S1.3, the first collision vector u n is obtained according to the difference between the minimum distance and of all sequences p mut in P n from the sequence p m , and is summed as the weight to the normalized minimum distance vector of all sequences in P mut from the sequence p n .

[0022] Further, in S1.6, the second collision vector is obtained according to the difference between the minimum distance and of all sequences p PAPR in P n from the sequence qn minimum distance and The difference, and use it as the weight for P PAPR The minimum distance vector between all sequences in the set and the nth sequence in the current unit complex sequence set after normalization is obtained.

[0023] Furthermore, in S2, the pulse shaping filter is a root-raised cosine pulse shaping filter with a symbol period of T and a roll-off coefficient of β.

[0024] Furthermore, S3 is specifically implemented through the following sub-steps:

[0025] S3.1: The receiving end receives the reflected waveform. Based on the optimal sequence set P * Given N sequences, design a matched filter that corresponds one-to-one with each sequence pulse. The matched filter f for the nth sequence pulse... n (t) is the pulse signal x obtained by sequence pulse modulation. n (t) conjugate flip; presuppose an initial Doppler rate set M represents the total number of Doppler velocities in the set, and the m-th Doppler velocity in the set is α. m Based on this set, the matched filter of the first sequence pulse is resampled to obtain the initial matched filter set under different Doppler rates. ;

[0026] S3.2: For the reflected waveform The baseband signal y(t) is obtained by down-conversion processing and then input into the initial matched filter set to obtain a set of M matched filtering results. τ is the time when the peak of the matched filter result of the current sequence pulse signal appears; let the index corresponding to the maximum value among the M matched filter results be the initial Doppler rate index. The coarse estimate of the Doppler velocity α1 corresponding to the first sequence of pulse signals is obtained as follows: At this time, the target object’s velocity is v1 = cα1, where c is the speed of sound wave propagation underwater.

[0027] Record number Matched filter results The time τ1 at which the maximum value occurs is used to obtain the target distance d1 at the time t1 when the current sequence pulse signal is received. ;

[0028] S3.3: Based on the matched filtering result of the (n-1)th sequence pulse signal, determine the time period of the nth segment of the cw signal; extract this segment of the cw signal, and define the frequency f of the cw signal under this segment of the cw signal. n loss function G n (fn ), calculate the first derivative of the loss function and second derivative ;

[0029] S3.4: Iteratively optimize the frequency of the nth segment of the cw signal:

[0030] In the first iteration, the frequency of the nth segment of the cw signal for Based on this, the first and second derivatives of the loss function at this frequency are calculated using method S3.3; during the k-th iteration, the frequency of the n-th segment of the cw signal is obtained. for k is an integer greater than or equal to 2;

[0031] After multiple iterations, the frequency difference of the cw signal before and after optimization was reduced to less than a set threshold. If the frequency of the nth segment of the cw signal is reached, then the iteration stops, and the frequency of the cw signal at this point is... Recorded as The precise estimated Doppler rate α at the moment when the nth sequence pulse signal is received is obtained. n for The target object's velocity v n for ;

[0032] S3.5: Based on the matched filtering result of the (n-1)th sequence pulse signal, extract the nth sequence pulse signal y from the baseband signal y(t) in the received waveform. n (t), according to the Doppler rate α n For y n (t) Perform CFO compensation and resampling to obtain ;

[0033] S3.6: The nth sequence pulse signal after CFO compensation and resampling Input to the nth matched filter f n In (t), the corresponding matched filter result is r. n (t), recording the matched filtering result r n (t) The time when the maximum value occurs τ n The time t when the nth sequence pulse signal is received is obtained. n The target distance d below n for ;

[0034] S3.7: Determine whether all sequence pulse signals of the received signal have been traversed. If not, return to S3.3 to calculate the target's speed and distance at the time the next sequence pulse signal is received. If yes, end the detection.

[0035] Further, in the S3.1, the mth match filter in the set of initial match filters is . .

[0036] Further, in the S3.2, the mth match filter result in the set of match filter results is . .

[0037] The beneficial effects of the present application are:

[0038] (1) The present application forms a multi-pulse waveform by weakly correlated sequences, effectively reduces the range-velocity ambiguity caused by echo aliasing, and realizes continuous detection of underwater targets.

[0039] (2) The present application designs low PAPR for weakly correlated sequences, realizes the improvement of transducer signal output efficiency, and enhances the anti-noise performance of the detection waveform. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 is a flow chart of the weakly correlated sequence set with low PAPR constraint generated in the embodiment of the present application.

[0041] Figure 2 is a structural schematic diagram of the transmitted waveform in the embodiment of the present application.

[0042] Figure 3 is a flow chart of the underwater target distance and velocity estimation implementation detection in the embodiment of the present application.

[0043] Figure 4 is the change of the peak cross-correlation coefficient in the optimal sequence set design process in the embodiment of the present application.

[0044] Figure 5 is a result schematic diagram of the autocorrelation and cross-correlation of the optimal sequence set changing with time in the embodiment of the present application.

[0045] Figure 6 is a result diagram of the continuous estimation of the underwater target distance in the embodiment of the present application.

[0046] Figure 7 is a result diagram of the continuous estimation of the underwater target motion velocity in the embodiment of the present application. DETAILED DESCRIPTION

[0047] The present application will be described in detail below with reference to the accompanying drawings and preferred embodiments, the objects and effects of the present application will become more apparent. It should be understood that the specific embodiments described herein are merely intended for the purpose of illustration of the present application and are not intended to limit the present application.

[0048] A low-PAPR weak cross-correlation sequence continuous probe waveform design and a probe method, specifically comprising the following steps:

[0049] S1: initialize to generate a unit complex sequence set P, and give a low PAPR constraint Γ PAPR (value range 1~2), an initial target cross-correlation lower bound Γ mut and algorithm parameters, update each sequence p n in the unit complex sequence set P and other parameters in each iteration, and finally obtain the optimal sequence set P with low PAPR constraint weak cross-correlation * . As shown in S1, the following sub-steps are implemented: Figure 1

[0050] S1.1: initialize to generate a unit complex sequence set P containing N unit complex sequences (hereinafter referred to as sequences), and each sequence has a length of L , wherein , N represents the total number of sequences in the unit complex sequence set, . Given a low PAPR constraint Γ PAPR , a target cross-correlation lower bound update step γ, a scaling factor φ and a convergence threshold ε, set an initial target cross-correlation lower bound Γ mut , an initial sequence update step τ q and τ PAPR , initialize the sequence set overall update round (i.e. iteration number) a=1.

[0051] S1.2: select the nth sequence p n (n initial value is 1) in P, calculate the minimum distance vector d n between the sequence p m and other sequences p n,m in P, and calculate the expression as follows:

[0052]

[0053] In the formula, m=1,2,3,…,N, and m≠n.

[0054] S1.3: for the sequence p n , screen out the sequences p m in P that do not satisfy , and integrate all the remaining sequences p m to obtain the set ​Based on this, the first collision vector u is calculated. n And according to the following rule, p n Perform the first phase of the update, and update the unit complex sequence set accordingly:

[0055]

[0056]

[0057] In the formula, u n Let p be the first collision vector. n The direction of iteration during updates.

[0058] S1.4: Generate a set of identity norm vectors from the updated set of identity complex sequences P. For the nth sequence Its generation rules are as follows (if the corresponding The first phase of the update has been completed. Replace with If the corresponding The second phase of the update has been completed. Replace with ):

[0059]

[0060] S1.5: Calculate the nth sequence in the current set of unit complex sequences and the nth sequence q in the current set Q. n Minimum distance vector between .

[0061] S1.6: Screen out those that do not meet the requirements in P. The sequence of conditions, integrating all remaining sequences (i.e., those that satisfy the conditions) The sequence of elements yields the set P. PAPR The second collision vector is calculated based on this. and according to the following rules The second phase of the update will be carried out.

[0062]

[0063]

[0064] S1.7: Determine the set P at this time. PAPR Is it empty? If so, then complete the processing of the current nth sequence p. n The two-stage update is denoted as . If not, execute S1.8; otherwise, proceed with... The value is updated to Then execute S1.4-S1.7 again.

[0065] S1.8: Determine whether all sequences in the current unit complex sequence set P have been completely updated (i.e., one round of update is completed). If not, update the value of n (e.g., increment n by 1) and return to execute S1.2-S1.7; if yes, calculate the unit complex sequence set P after the current round of update. a Peak cross-correlation coefficient , where P a Let P be the sequence set after the a-th round of updates.

[0066] S1.9: Determine whether the peak cross-correlation coefficient exceeds the initial target cross-correlation lower bound. If μ a ≤Γ mut Then tighten the lower bound of the target cross-correlation Γ mut ,Right now If μ a >Γ mut Then relax the lower bound of the cross-correlation of the targets Γ mut ,Right now .

[0067] Simultaneously, compare the peak cross-correlation coefficient μ of the unit complex sequence set before the a-th round update. a-1 The peak cross-correlation coefficient μ of the unit complex sequence set after the a-th round update a The step size τ is updated based on the cross-correlation coefficient between the two peak values. q τ PAPR The expression is as follows:

[0068]

[0069] S1.10: Determine whether the peak cross-correlation coefficient of the current unit complex sequence set satisfies the convergence condition. If not, update the value of 'a' (increment 'a' by 1) and execute S1.2 to S1.9 again for the next round of updates; if yes, output the current unit complex sequence set as the optimal sequence set P with weak cross-correlation under low PAPR constraints. * .

[0070] S2: The optimal sequence set P will be obtained. * sequence in The pulses are modulated onto the transmitted pulses through a root-raised cosine pulse shaping filter g(t) to form pulses such as... Figure 2 The pulse signal shown has the following expression:

[0071]

[0072]

[0073] In the formula, T sThe sampling period is P represents * The l-th value in the n-th sequence; β is the roll-off factor, and T is the symbol period.

[0074] The modulated N pulse signals are arranged sequentially, with a frequency of f. cw Duration is T cw single-frequency signal As the interval between two adjacent pulse signals, it forms the baseband waveform. An LFM signal is inserted at the frame header, and finally up-converted and modulated to a frequency of f. c Continuous detection waveforms are formed on the carrier frequency band. And it is transmitted through the transmitter.

[0075] S3: The receiver receives the reflected waveform. After down-conversion, the baseband signal y(t) is obtained. This signal is then input into the initial matched filter set to obtain the first sequence pulse signal at the corresponding time t. n The Doppler velocity α1 and the target distance d1 are roughly estimated. Based on the matched filtering result of the (n-1)th sequence pulse signal, the nth segment of the cw signal is extracted, and the corresponding time t is iteratively estimated. n Precise estimation of Doppler velocity α n The target object's velocity v n According to α n After CFO compensation and resampling of the nth sequence pulse signal, it is input into the nth matched filter, and the corresponding time t is obtained based on the peak value of the matched filter result. n The target distance d below n The exploration is complete. Figure 3 As shown, S3 is implemented through the following sub-steps:

[0076] S3.1: The receiving end receives the reflected waveform. Based on the optimal sequence set P * Given N sequences, design a matched filter that corresponds one-to-one with each sequence pulse. The matched filter for the nth sequence pulse is the pulse signal x obtained by modulating the sequence pulses. n The conjugate flip of (t), i.e. Pre-determine an initial Doppler velocity set. M represents the total number of Doppler rates in the set. Then, based on this set, the matched filter of the first sequence pulse is resampled to obtain the initial matched filter set under different Doppler rates. Where m represents the index in the Doppler velocity set α, Let α represent the m-th Doppler velocity in the Doppler velocity set α, where m = 1, 2, 3, ..., M.

[0077] S3.2: For the reflected waveform The baseband signal is obtained by performing down-conversion processing. The baseband signal is input to the initial matched filter set in real time. In this process, the set of M matched filtering results is obtained as follows: Where τ is the time when the peak of the matched filter result of the sequence pulse signal appears, and the initial Doppler rate index is... This is the index corresponding to the maximum value of the matched filter result, i.e. The coarse estimate of the Doppler velocity corresponding to the current first sequence of pulse signals is: The peak time of the matched filter result of the first sequence pulse signal is... Then, the velocity of the target object at the time t1 corresponding to the moment when the current sequence pulse is received (hereinafter referred to as the corresponding time) is: Distance is , where c is the speed of sound propagation underwater.

[0078] S3.3: Based on the matched filtering result of the (n-1)th (n=2,…,N)th sequence pulse signal, determine the time interval of the next adjacent cw signal (corresponding to the nth cw segment) of the nth sequence pulse signal as follows. , where N cw This indicates the number of samples in the cw signal segment; the cw signal segment is truncated. And define the frequency f of the cw signal under this segment of the cw signal. n loss function And calculate the first derivative of the loss function. and second derivative The expression is as follows:

[0079]

[0080]

[0081] In the formula, f n This represents the frequency of the nth segment of the cw signal.

[0082] S3.4: Iteratively optimize the frequency of the nth segment of the cw signal. In the initial iteration, the frequency of the nth segment of the cw signal (i.e., the initial value) is: Based on this, the method in S3.3 was used to calculate... and .

[0083] When performing the k-th iteration (where k is an integer greater than or equal to 2), the frequency of the n-th segment of the cw signal is:

[0084]

[0085] After several iterations until , stop iteration, is a threshold value set artificially; the frequency of the nth segment of the cw signal at this time is recorded as , and the time t n at which the nth sequence pulse signal is received is calculated, and the fine estimation of the Doppler rate at this time is , and the target motion speed is .

[0086] S3.5: According to the matched filtering result of the (n-1)th sequence pulse signal, the nth sequence pulse signal y n (t) in the received waveform is cut from the baseband signal y(t), CFO compensation and resampling are performed on the nth sequence pulse signal y n (t) at the corresponding time t n according to the fine estimation of the Doppler rate a n of the nth sequence pulse signal at this time, to obtain the resampled nth sequence pulse signal .

[0087] S3.6: The CFO-compensated and resampled nth sequence pulse signal is input into the nth matched filter f n (t), and the corresponding matched filtering result is output as , and the time at which the peak value of the matched filtering result appears is recorded as , and the target distance at the time t n at which the nth sequence pulse signal is received is .

[0088] S3.7: It is judged whether all sequence pulse signals of the received signal have been traversed, if not, returning to execute S3.3 to calculate the target distance and speed at the corresponding time of the next sequence pulse signal; if yes, ending the detection.

[0089] To verify the performance of the method of the application, corresponding performance simulation experiments are performed, which are specifically described below through two embodiments.

[0090] Embodiment one: To verify the effectiveness of the optimal sequence set method with weak cross-correlation under low PAPR constraint in the method of the application, the parameter settings of this embodiment are shown in Table 1.

[0091] Table 1 Specific parameters of the optimal sequence set design method

[0092]

[0093] Under the simulation parameter condition, 256 unit complex sequences with length of 450 are randomly initialized. The optimal sequence set P is successfully obtained by using the method provided in the application * . The iteration process is shown in Figure 4 . After multiple rounds of iteration, the peak cross-correlation coefficient is reduced from 0.21 to 0.16. The autocorrelation and cross-correlation results are shown in * Figure 5 . It can be found that the peak of the autocorrelation result is 20 dB higher than that of the cross-correlation result, indicating that the optimal sequence set designed by the method has good weak cross-correlation.

[0094] In order to verify the reliability of the continuous detection waveform design and detection method of the low PAPR weak cross-correlation sequence provided in the application for continuous detection of underwater targets, the parameter settings of the embodiment are as follows: the waveform center frequency is 4 kHz, the signal bandwidth is 2 kHz, and the sampling frequency is 24 kHz. The received signal is affected by the ocean environmental noise simulated by Gaussian white noise, and the signal-to-noise ratio is -5 dB. The initial distance of the underwater target is 3000 m, and the initial velocity is 0 m / s, and the acceleration is 0.5 m / s 2 .

[0095] The specific parameters of the simulation signal are shown in Table 2.

[0096] Table 2 Specific parameters of simulation signal

[0097]

[0098] Under the simulation condition, first, the 0.25s cw signal is inserted between the 0.25s single sequence pulse signals as an interval to design the continuous detection waveform. By using the detection method described in the application, the target distance and velocity at the corresponding time of each sequence pulse signal are estimated. The estimation result of the target distance is shown in Figure 6 , and the estimation result of the target velocity is shown in Figure 7 . As can be seen from the figure, the method can estimate the distance and velocity of the underwater target every 0.5s, and the estimation error is controlled within a very small range, indicating that the method of the application can effectively realize the continuous detection of the distance and velocity of the target.

[0099] The PAPR and cross-correlation of the randomly generated sequence set are alternately optimized by the method of the application, and the parameters are updated at the same time. After multiple rounds of iteration, the optimal sequence set P with weak cross-correlation under the constraint of low PAPR is obtained * ​The obtained optimal sequence set is sequentially arranged on the base band after pulse shaping, and then modulated to the carrier band to form a waveform capable of continuous detection. After the waveform reflected by the target is down-converted and synchronized at the receiving end, each sequence pulse signal and the adjacent cw signal are processed to realize continuous tracking and detection of the distance and speed of the underwater target at each time. In summary, the method can realize weak correlation optimization of any number and length of sequences under low PAPR constraint, and realize continuous detection of the distance and speed of the underwater target.

[0100] Those skilled in the art can understand that the above description is only preferred examples of the application and is not used to limit the application, although the application is described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions recorded in the foregoing examples or make equivalent replacement for part of the technical features. Any modification, equivalent replacement, etc. within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A low PAPR weak cross-correlation sequence continuous probe waveform design and probe method, characterized in that, Comprising the following steps: S1: initialize to generate a set of N sequences, sequence length L unit complex sequence set P, given PAPR constraint Γ PAPR , initial target cross-correlation lower bound Γ mut , scaling factor φ and convergence threshold ε, target cross-correlation lower bound update step γ, initial sequence update step τ q , τ PAPR , each iteration of each sequence p n in the unit complex sequence set P two-stage update, get the optimal sequence set P * ; p n Pn= nth unit complex sequence in the unit complex sequence set P, n = 1, 2, 3, …, N; S2: modulate each sequence in the optimal sequence set P * through a pulse-shaping filter to form a pulse signal x n (t); The modulated pulse signals are arranged in sequence, taking a single-frequency signal with a frequency of f cw and a time length of T cw as the interval between two adjacent pulse signals, to form a baseband waveform x(t), which is up-converted to a carrier f c to form a continuous detection waveform , which is emitted by the transmitting end; S3: receiving end receives the reflected waveform , after down-conversion processing, a baseband signal y(t) is obtained, which is input into an initial matched filter set, to obtain a coarse estimated Doppler rate α1, target motion velocity v1 and distance d1 at the time when the first sequence pulse signal is received; based on the matched filtering result of the n-1th sequence pulse signal, the n-th single frequency signal segment is intercepted backward, to iteratively estimate a fine estimated Doppler rate αn, target motion velocity vn and distance dn at the time t n when the n-th sequence pulse signal is received. n n ; after CFO compensation and resampling of the n-th sequence pulse signal according to α n , the signal is input into the n-th matched filter, and according to the peak value of the matched filtering result, the target distance dn at the time t n is obtained. n ​​ 2.The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 1, wherein, In the S1, each iteration is performed for each sequence p in the set of unit complex sequences P n The two-stage update is performed by the following sub-steps: S1.1: initialize the generation of a unit complex sequence set P and initialize the setting sequence update step, target cross-correlation lower limit update step, unit complex sequence set overall update round a = 1; S1.2: select the nth sequence p in P n , calculate the minimum distance vector d n between p and other sequences p m in P n,m , integrate all p in P that satisfy m to obtain the set P mut ; according to which, calculate the first collision vector u n , which is used to represent the iteration direction when the sequence p n is updated in the first stage, and the first sequence update step size τ q updates the sequence p n in the first stage in the iteration direction to obtain , and updates the unit complex sequence set; S1.3: generating a set Q of N unit norm vectors, each of length L, from the updated set of unit complex sequences; the nth vector q n is a standard unit modulus vector whose non-zero value is at the position corresponding to the position of the smallest modulus element in the nth sequence in the set of unit complex sequences, and the value is the normalized result of the smallest modulus element in the nth sequence in the set of unit complex sequences. n is a standard unit modulus vector whose non-zero value is at the position corresponding to the position of the smallest modulus element in the nth sequence in the set of unit complex sequences, and the value is the normalized result of the smallest modulus element in the nth sequence in the set of unit complex sequences. S1.4: Calculate the minimum distance vector between the nth sequence in the current unit complex sequence set and the nth sequence q in the unit norm vector set Q n ; integrate all sequences in the current unit complex sequence set satisfying to obtain set P PAPR ; according to which, calculate the second collision vector , which is used to represent the iteration direction when the sequence performs the second stage update, and the second sequence update step size τ PAPR Update the sequence in the iteration direction to obtain ;​ S1.5: Determine the set P at this time. PAPR Is it empty? If so, then complete the processing of the current nth sequence p. n The two-stage update is denoted as . If not, execute S1.6; otherwise, proceed with... The value is updated to Then execute S1.3-S1.5 again; S1.6: judge whether all sequences in the current unit complex sequence set have been updated completely, if yes, complete one round of update; if not, update the value of n, and return to execute S1.2-S1.

5. 3.The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 2, wherein, In the S1, the unit complex sequence set P is updated for multiple rounds of iteration, which is specifically implemented through the following sub-steps: S1.7: Calculate the updated unit complex sequence set P after the current round a the peak cross-correlation coefficient μ a ; determine whether the peak cross-correlation coefficient exceeds the initial target cross-correlation lower bound, if μ a ≤ Γ mut , then tighten the target cross-correlation lower bound Γ mut , i.e. subtract the target cross-correlation lower bound update step γ from Γ mut ; otherwise, relax the target cross-correlation lower bound Γ mut , i.e. add the target cross-correlation lower bound update step γ to Γ mut ; Meanwhile, compare the peak cross-correlation coefficient μ of the unit complex sequence set before the a-th round of update a-1 and the peak cross-correlation coefficient μ of the unit complex sequence set after the a-th round of update a ; if the peak cross-correlation coefficient after the update is lower than that before the update, then the update step τ is enlarged by a scaling factor φ q , τ PAPR ; if the peak cross-correlation coefficient after the update is higher than that before the update, then the update step τ is reduced by a scaling factor φ q , τ PAPR ; if the peak cross-correlation coefficient before and after the update remains unchanged, then the current sequence update step remains unchanged; S1.8: judge whether the absolute value of the difference between the peak cross-correlation coefficients of the unit complex sequence sets before and after the a-th round of updating is less than the convergence threshold ε, if not, update the value of a, and execute S1.2 to S1.7 again; if yes, output the current unit complex sequence set as the optimal sequence set P * .

4. The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 2, wherein, In S1.2, the first collision vector u n According to P mut In all sequences p n The minimum distance of sequence p m The difference of the minimum distance of sequence p As a weight to P mut In all sequences p n The normalized minimum distance vector is summed up; In S1.4, the second collision vector According to P PAPR The minimum distance of all sequences in P n and the difference value is taken as the weight of P The minimum distance vector of all sequences in P PAPR is summed up with the n-th sequence in the current unit complex sequence set after normalization.

5. The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 1, wherein, In the S3, the coarse estimated Doppler rate α1, target motion speed v1 and distance d1 at the moment when the first sequence pulse signal is received are obtained, which is specifically implemented through the following sub-steps: S3.1: The receiving end receives the reflected waveform , based on the N sequences in the optimal sequence set P * , a matched filter corresponding to each sequence pulse is designed, and the matched filter f n (t) of the nth sequence pulse is the conjugate inversion of the pulse signal x n (t) obtained by modulating the sequence pulse. presetting an initial Doppler rate set , M represents the total number of Doppler rates in the set, the mth Doppler rate in the set is m ; resampling the matched filter of the first sequence pulse based on the set to obtain an initial matched filter set under different Doppler rates ; S3.2: to the reflected waveform The baseband signal y(t) is obtained by down-conversion processing, which is input to the initial matched filter set, and a set of M matched filtering results is obtained τ is the time when the peak value of the matched filtering result of the current sequence pulse signal appears. The index corresponding to the maximum value in the M matched filtering results is an initial Doppler velocity index The coarse estimated Doppler velocity a1 corresponding to the first sequence pulse signal is obtained At this time, the target motion speed v1 is c a1, wherein c is the propagation speed of sound waves in water; Record the first matching filter result The time τ1 at which the maximum value occurs, and the distance d1 of the target object at the time t1 at which the current sequence pulse signal is received is obtained as .

6. The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 5, wherein, In the S3, the target object distance d n at the corresponding time t n is obtained, specifically through the following sub-steps: S3.3: according to the matched filtering result of the n-1th sequence pulse signal, determine the time period of the nth segment of cw signal; intercept the segment of cw signal, and define the loss function G n (f n ) of the cw signal frequency f n under the segment of cw signal, calculate the first derivative and the second derivative of the loss function; S3.4: iteratively optimize the frequency of the nth segment of cw signal: the frequency of the nth segment of cw signal in the first iteration is , the first derivative and the second derivative of the loss function of the frequency are calculated by the method of S3.3, and the frequency of the nth segment of cw signal obtained in the kth iteration is is , and k is an integer greater than or equal to 2. If the frequency difference of the CW signal before and after optimization is less than the set threshold If the frequency of the nth segment of the cw signal is reached, then the iteration stops, and the frequency of the cw signal at this point is... Recorded as The precise estimated Doppler rate α at the moment when the nth sequence pulse signal is received is obtained. n for The target object's velocity v n for ; S3.5: According to the matched filtering result of the n-1th sequence pulse signal, the n th sequence pulse signal y n (t) is cut from the baseband signal y(t), and the Doppler rate a n is calculated according to the Doppler rate a n (t) is CFO compensated and resampled, and y is obtained. S3.6: CFO compensation and resample the nth sequence pulse signal Input to the nth matched filter f n (t), output the corresponding matched filtering result as r n (t), record the matched filtering result r n (t) the moment τ when the maximum value appears n , get the moment t when the nth sequence pulse signal is received n The target distance d under the moment t n is ; S3.7: judge whether all sequence pulse signals of the received signal have been traversed, if not, return to execute S3.3 to calculate the target motion speed and distance at the moment when the next sequence pulse signal is received; if yes, end the detection.

7. The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 6, wherein, In S3.1, the mth matched filter in the set of initial matched filters is .​ 8.The low PAPR weak cross-correlation sequence continuous probe waveform design and probe method of claim 7, wherein, In the S3.2, the mth matched filter result in the set of matched filter results is .​

Images