A method and device for joint estimation of angle and velocity in a broadband system

Abstract

The present application provides a method and device for joint angle-velocity estimation in a broadband system, which are used for estimating the target angle and velocity by a uniform linear array according to the received signal. The method includes: performing dechirp pulse compression processing on the received signal to obtain a pulse-compressed signal; performing time-domain sampling on the pulse-compressed signal to convert the analog signal into a discrete-time signal; arranging the discrete-time signal, designing a corresponding overcomplete dictionary, and constructing a corresponding multi-dictionary sparse recovery problem; using the multi-dictionary sparse Bayesian algorithm to solve the multi-dictionary sparse recovery problem; rearranging the solution result to obtain an angle-velocity spectrum, so as to realize joint angle-velocity estimation. The advantages of the present application are as follows: high-resolution and accurate joint angle-velocity estimation can be achieved under the condition of small samples; the solution speed is high and can meet the requirement of real-time performance; accurate joint angle-velocity estimation can be achieved in the presence of fast-slow time coupling and when the spatial domain steering vector changes with frequency.

Description

Technical Field

[0001] This application belongs to the technical field of sonar detection, and particularly relates to a method and device for joint angle-velocity estimation in a broadband system. Background Technique

[0002] The azimuth, distance, and velocity are important parameters of a moving target. Estimating them is a hot issue in the field of array signal processing and is widely applied in sonar systems. The sonar system estimates the angle, velocity, and distance of the target by estimating the direction of arrival (DOA), Doppler frequency shift, and time delay of the echo signal. Compared with other signals, the frequency modulated continuous wave (FMCW) signal has a larger time-bandwidth product and has advantages such as high measurement accuracy and low average transmit power, so it is widely applied in sonar systems. By performing operations such as dechirping and deramping on the echo signal, the received signal can be regarded as a linear superposition of multiple multi-dimensional complex sinusoidal signals, and its frequency is determined by the angle of the target echo, the velocity of the target, and the distance. Therefore, the problem of parameter estimation for a moving target can be transformed into a multi-dimensional spectral estimation problem.

[0003] In order to break through the Rayleigh limit, subspace-based methods represented by multiple signal classification (MUSIC) and estimation of signal parameters via rotational invariance techniques (ESPRIT) have been widely applied to spectral estimation problems. To generalize the subspace method to high dimensions, Hua et al. proposed a matrix enhancement and matrix pencil (MEMP) method to achieve two-dimensional spectral estimation. By introducing the idea of matrix pencil into the MUSIC algorithm, they further proposed the two-dimensional MUSIC algorithm. Subsequently, Haardt et al. proposed the two-dimensional unitary ESPRIT algorithm, extending the ESPRIT algorithm to two dimensions.

[0004] In recent years, to meet the growing demand for high-precision range resolution, the use of broadband signals has become increasingly common. As the bandwidth increases, the distance that a high-speed target moves within a coherent processing interval (CPI) may exceed the range cell, making the coupling term between fast time and slow time non-negligible. Currently, the commonly used methods of Keystone transform and matched filtering are used to eliminate the above coupling terms. In addition, the steering vector of the array changes with frequency and cannot be ignored due to the increase in bandwidth, making it difficult to directly apply narrowband DOA estimation methods. To solve this problem, the incoherent signal-subspace method (ISSM) and the coherent signal-subspace method (CSSM) have been proposed successively.

[0005] Subspace algorithms are all based on the covariance matrix and require sufficient samples to ensure performance. Compared with subspace methods, methods based on compressive sensing and sparse recovery can accurately estimate parameters under the condition of a limited number of samples, and even achieve good results in the case of a single snapshot. To avoid solving the NP-hard l0 norm minimization problem, sparse algorithms based on l1 norm minimization such as the basis pursuit algorithm are widely used in the problem of moving target parameter estimation. However, when the problem scale is large, the computational complexity of the basis pursuit algorithm cannot meet the application requirements. Different from the basis pursuit algorithm, the sparse Bayesian learning (SBL) algorithm regards the coefficient vector with sparse characteristics as a random vector with unknown parameters, and obtains the estimation of the unknown parameters by maximizing its evidence. Since the estimation of the unknown parameters satisfies sparsity, sparse coefficient estimation can also be obtained. Compared with the basis pursuit algorithm, the SBL algorithm has a significant improvement in computational speed and has thus received increasing attention.

[0006] Although Keystone transform and matched filtering can eliminate the coupling term, they require a large amount of data to complete interpolation and thus cannot achieve super-resolution estimation. In addition, to achieve a large bandwidth, a longer pulse repetition interval is usually required, resulting in velocity ambiguity. The ISSM and CSSM methods applicable to broadband signals are both based on the covariance matrix and cannot adapt to the scenario of small samples. Summary of the Invention

[0007] The purpose of this application is to overcome the defects of the prior art that super-resolution estimation cannot be achieved, a longer pulse repetition interval is required, resulting in velocity ambiguity, or it cannot adapt to the scenario of small samples.

[0008] To achieve the above object, the present application proposes a method for joint angle-velocity estimation in a broadband system, which is used for estimating the target angle and velocity according to the received signal by a uniform linear array, including:

[0009] Step 1: Perform dechirp pulse compression processing on the received signal to obtain a pulse-compressed signal;

[0010] Step 2: Perform time-domain sampling on the pulse-compressed signal to convert the analog signal into a discrete-time signal;

[0011] Step 3: Arrange the discrete-time signal, design a corresponding overcomplete dictionary, and construct a corresponding multi-dictionary sparse recovery problem;

[0012] Step 4: Use the multi-dictionary sparse Bayesian algorithm to solve the multi-dictionary sparse recovery problem;

[0013] Step 5: Rearrange the solution result to obtain an angle-velocity spectrum, and realize joint angle-velocity estimation.

[0014] As an improvement of the above method, the said Step 1 includes:

[0015] The pulse-compressed signal obtained by performing dechirp pulse compression on the signal received by the l-th array element is

[0016]

[0017] where N Tar represents the number of targets to be measured; represents noise; The transmitting array element transmits a string of Chirp signals in a coherent processing interval, T c represents the pulse duration of the signal, T p represents the pulse repetition interval of the signal, N p represents the number of pulses of the signal, B represents the pulse width of the signal, f c represents the starting frequency of the signal; t m = mT p , m ∈ {1, 2,..., N p};

[0018] represents the difference-frequency output obtained by correlating the reflected signal received by the l-th array element from the i-th target with the reference signal s:

[0019]

[0020] where, represents the loss coefficient after the transmitted signal propagates and is reflected; R irepresents the radial distance of the \(i\)-th target from the sonar; \(\theta\) i represents the azimuth angle of the \(i\)-th target from the sonar; \(v\) i represents the radial velocity of the \(i\)-th target from the sonar; \(c\) represents the speed of sound in water; \(\mu = B / T\) c ;

[0021] As an improvement to the above method, step 2 includes:

[0022] Perform time-domain sampling on the pulse compression signal at a sampling rate of \(f\) s . The discrete-time signal after sampling is :

[0023]

[0024] where \(k\) represents the fast-time sampling index.

[0025] As an improvement to the above method, step 3 includes:

[0026] Let There is:

[0027]

[0028] where \(N\) m represents the number of array elements; \(T\) represents the transpose;

[0029]

[0030] Vectorize \(Y\) k to obtain:

[0031]

[0032] where represents the Kronecker product,

[0033] Expand \(A\) k into an over-complete dictionary, and the parameters to be estimated are obtained by the compressive sensing algorithm; specifically, first divide the value ranges of the radial velocity \(v\) and the azimuth angle \(\theta\) into fine grids, and respectively form sets and set \(\Theta\); let

[0034] \(|\cdot|\) represents the cardinality of the set; the over-complete dictionary obtained by expanding \(A\) k is represented by . There is:

[0035]

[0036] Construct the corresponding multi-dictionary sparse recovery problem:

[0037]

[0038] where, is a sparse vector, and its support set is obtained by the compressive sensing algorithm, and the estimates of the parameters v and θ to be estimated are obtained therefrom; N t represents the total number of fast-time samplings; ∈ k represents the noise threshold; ||·||2 represents the 2-norm of the matrix; ||·|| 2,1 represents the 2,1-norm of the matrix.

[0039] As an improvement of the above method, step 4 includes:

[0040] All have the same covariance matrix, which is Λ = diag(λ), where

[0041] The estimate of λ is:

[0042]

[0043] where, H represents the conjugate matrix;

[0044] is solved by an iterative algorithm to obtain:

[0045]

[0046] where, represents N p N m × N p N m identity matrix; b represents an adjustable parameter for controlling the convergence rate; represents [λ] i the result of the previous iteration; represents [λ] i the result after the next iteration; Tr represents the trace of the matrix; represents the noise power.

[0047] As an improvement of the above method, the estimation method of the noise power is:

[0048]

[0049] where, denotes the support set of λ, i.e., the indices corresponding to the largest N elements in λ Tar elements; denotes the pseudo-inverse.

[0050] As an improvement to the above method, step 5 includes:

[0051] The joint spectrum Spec of the azimuth angle θ and the radial velocity v is obtained by estimating and rearranging the matrix into That is:

[0052]

[0053] where denotes arranging with |Θ| elements as one column to form a column matrix arranged as .

[0054] This application also provides an angle-velocity joint estimation device for a broadband system, which is implemented based on the above method. The device includes:

[0055] A dechirping pulse compression module for performing dechirping pulse compression processing on the received signal to obtain a pulse-compressed signal;

[0056] A time-domain sampling module for performing time-domain sampling on the pulse-compressed signal to convert the analog signal into a discrete-time signal;

[0057] A module for constructing a multi-dictionary sparse recovery problem, which is used to arrange the discrete-time signal, design a corresponding over-complete dictionary, and construct a corresponding multi-dictionary sparse recovery problem;

[0058] A module for solving the multi-dictionary sparse recovery problem, which is used to solve the multi-dictionary sparse recovery problem using the multi-dictionary sparse Bayesian algorithm; and

[0059] An angle-velocity joint estimation module for rearranging the solution result to obtain an angle-velocity spectrum and realizing angle-velocity joint estimation.

[0060] Compared with the prior art, the advantages of this application are:

[0061] 1. Compared with the traditional subspace method, the present invention can achieve high-resolution and accurate angle-velocity joint estimation under the condition of small samples;

[0062] 2. Compared with the l1-optimized basis pursuit method, the present invention has a higher solution speed and can meet the real-time requirements;

[0063] 3. Compared with existing narrowband methods, the present invention can achieve accurate joint angle-velocity estimation in the presence of fast-slow time coupling and when the spatial steering vector changes with frequency;

[0064] 4. The present invention can also improve the problem of velocity ambiguity and achieve velocity estimation for high-speed targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 The figure shows a schematic diagram of a uniform linear array (ULA) receiving sonar echo signals reflected by a target;

[0066] Figure 2 The figure shows a flow chart of a joint angle-velocity estimation method for a broadband system;

[0067] Figure 3(a) shows the effect of joint angle-velocity estimation in the absence of noise (narrowband MUSIC algorithm);

[0068] Figure 3(b) shows the effect of joint angle-velocity estimation in the absence of noise (ISSM algorithm);

[0069] Figure 3(c) shows the effect of joint angle-velocity estimation in the absence of noise (the algorithm of the present application);

[0070] Figure 4(a) shows a graph of the relationship between the RMSE of angle estimation and SNR;

[0071] Figure 4(b) shows a graph of the relationship between the RMSE of velocity estimation and SNR. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0072] The technical solutions of the present application will be described in detail below with reference to the accompanying drawings.

[0073] The present invention proposes a joint angle-velocity estimation method and device for a broadband system, which is implemented based on a multi-dictionary SBL algorithm and a joint angle-velocity estimation algorithm for a broadband FMCW sonar. By deriving and analyzing the signal model, the present invention converts the joint angle-velocity estimation problem into a sparse recovery problem of multiple dictionaries, and uniformly solves the problems of fast-slow time coupling, velocity ambiguity, and the change of the spatial steering vector with frequency brought by broadband signals. To improve the solution speed, the present invention uses the multi-dictionary SBL algorithm to solve the sparse recovery problem.

[0074] 1. Principle Explanation

[0075] The signal processed by the method of the present application is a sonar echo signal reflected by a target received by a uniform linear array (ULA), where the transmitted signal is a broadband frequency-modulated continuous wave signal, and its function is to obtain the angle and velocity estimation of the target. The schematic diagram is as Figure 1 shown.

[0076] The method first performs de-chirping and pulse compression processing on the received signal, and then samples the pulse-compressed signal in the time domain to convert the analog signal into a discrete-time signal for processing. Next, the sampled discrete-time signal is arranged in a certain way and an over-complete dictionary is designed to construct a corresponding multi-dictionary sparse recovery problem. Then, the multi-dictionary sparse Bayesian algorithm is used to quickly solve the multi-dictionary sparse recovery problem. Finally, the solution result is rearranged to obtain the angle-velocity spectrum, realizing joint angle-velocity estimation. The method flow is as Figure 2 shown.

[0077] 2. Method description

[0078] Assume there are a total of N Tar targets to be measured. The radial distance, radial velocity, and azimuth angle of the i-th target from the sonar are R i , v i , and θ i , respectively, where i ∈ {1, 2, 3,..., N Tar}}. The transmitting array element emits a string of Chirp signals in a coherent processing interval. Its pulse width is B, the starting frequency is f c , the pulse duration is T c , the pulse repetition interval is T p , the number of pulses is N p , and the CPI is T CPI = T p N p . The transmitted signal can be expressed as:

[0079]

[0080] where t m = mT p , m ∈ {1, 2,..., N p}}, and μ = B / T c .

[0081] The transmitted signal is reflected by the target and received by the receiving array. The receiving array is a uniform linear array consisting of N m array elements with an element spacing of d. Among them

[0082] To describe the received signal, first, a brief analysis of the time delay of the received signal is made. Take the speed of sound in water c = 1500 m / s. Let the time delay of the signal transmitted at time t after being reflected by the i-th target and reaching the receiving array be τ i (t), which satisfies:

[0083] R i - v i t = (v i + c)τi / 2, (2)

[0084] Thus, it can be obtained that:

[0085]

[0086] The corresponding reflected signal received by the receiving array can be expressed as:

[0087] r i (t + τ i (t)) = α i s(t), (4)

[0088] where is the loss coefficient after the transmitted signal propagates and is reflected. There exists such that:

[0089]

[0090] It is easy to prove Therefore, there is:

[0091] r i (t) = α i s(t - τ i (t)), (6)

[0092] That is:

[0093]

[0094] Substituting (1) into (7), it can be obtained that:

[0095]

[0096] where:

[0097]

[0098] Thus, the instantaneous frequency f i of r i is:

[0099]

[0100] Let the reflected signal received by the l-th array element from the i-th target be Since the time delay of the signal reflected by the i-th target reaching the l-th array element relative to the reference array element is Therefore, there is:

[0101]

[0102] Combining (10), it can be obtained that:

[0103]

[0104] To achieve dechirp pulse compression, r i,l is correlated with the reference signal s to obtain the difference frequency output as:

[0105]

[0106] When μT c dN m / c << 1, that is at this time, the in (13) can be ignored, and the signal can be approximated by the narrowband model. When that is at this time, the distance that the target moves within one CPI does not exceed one range cell, and the in (13) can also be ignored. However, in order to obtain higher range resolution, the bandwidth B of the transmitted signal also needs to be made wider, making the above two terms unable to be ignored, resulting in the inability to directly apply the classical narrowband processing method.

[0107] For the convenience of subsequent processing, (13) is rearranged into the following form:

[0108]

[0109] where

[0110] Let the output obtained after dechirp pulse compression of the signal received by the l-th array element be There is:

[0111]

[0112] where n l is noise. The above signal is sampled in the time domain at a sampling rate of f s , and substituting t m = mT p , the sampled signal is:

[0113]

[0114] Subsequently, will be directly processed to estimate the parameter to be estimated.

[0115] It is worth mentioning that processing based on the signal model such as (16) can also improve the situation of velocity ambiguity. The traditional processing method ignores the coupling between fast time and slow time, and the obtained velocity Doppler term is: At this time, in order to avoid velocity ambiguity, the velocity to be measured should not exceed T in the broadband signalp It is usually relatively large, resulting in a very narrow measurable speed range and making it difficult to meet the requirements for speed measurement of high-speed targets in practice. In the present invention, the coupling of fast time and slow time is considered, and the resulting velocity Doppler term is To avoid velocity ambiguity, it is required that the maximum velocity satisfies:

[0116]

[0117] where LCM represents the least common multiple, and N t is the number of fast-time sampling points. To simplify the discussion, here it is considered that f c is an integer multiple of. It is easy to prove that the maximum measurable velocity at this time greatly improves the measurable velocity range and alleviates velocity ambiguity.

[0118] Fix the fast-time sampling index k of and let:

[0119]

[0120] where Let [Y k l,m = Y k (l,m), then there is:

[0121]

[0122] where

[0123]

[0124] Vectorize Y k and we can get:

[0125]

[0126] where represents the Kronecker product,

[0127] It can be seen from the last equation of (20) that by expanding A k into an overcomplete dictionary, the parameters to be estimated can be obtained by the compressive sensing algorithm. Specifically, first divide the value ranges of v and θ into fine grids to form sets and set Θ respectively. Let where |·| represents the cardinality of the set. The overcomplete dictionary obtained by expanding A k is denoted by​ It is shown that there is:

[0128]

[0129] When the grid is fine enough, such that:

[0130]

[0131] At this time is a sparse vector, and its support set can be obtained by the compressive sensing algorithm, and the estimates of the parameters v and θ to be estimated are obtained therefrom, where Let:

[0132]

[0133] where N t is the total number of fast-time samplings. It can be seen that has group sparsity, so the following optimization problem can be constructed to solve:

[0134]

[0135] where, ∈ k represents the noise threshold; ||·||2 represents the 2-norm of the matrix; ||·|| 2,1 represents the 2,1-norm of the matrix.

[0136] Although the above optimization problem is solvable, it is complex to solve due to the large problem scale. To meet the need for real-time performance, the multi-dictionary SBL method is used below to implement the estimation of the parameters.

[0137] Assume k = 1, 2,..., N t are independent of each other and all follow a zero-mean circularly symmetric Gaussian distribution. Since has group sparsity, it can be assumed that k = 1, 2,..., N t satisfies the prior of the covariance matrix, that is, the covariance matrices of all are the same, all being Λ = diag(λ), where Therefore, there is:

[0138]

[0139] Assume The posterior probability can be obtained as:

[0140]

[0141] Combining (25) and (26) gives:

[0142]

[0143] Taking the logarithm of (27) gives:

[0144]

[0145] An estimate of λ can be obtained by maximizing the above logarithmic probability density That is:

[0146]

[0147] (29) can be solved by an iterative algorithm. Specifically,

[0148]

[0149] where represents N p N m × N p N m identity matrix, b is an adjustable parameter that controls the convergence rate, is [λ] i the result of the previous iteration, is [λ] i the result after the next iteration. The noise power is required in (30). When it is unknown, it can be estimated as follows:

[0150]

[0151] where is the support set of λ, that is, the indices corresponding to the largest N Tar elements in λ, represents the pseudo-inverse.

[0152] Finally, the joint spectrum Spec of the azimuth angle θ and the velocity v can be obtained from the estimated rearranged into matrix, that is:

[0153]

[0154] where means arranging with |Θ| elements as a column, arranged into columns, arranged as matrix.

[0155] The algorithm flow table is shown in Table 1.

[0156] Table 1 Wideband System Angle-Velocity Joint Estimation Method

[0157]

[0158] 3. Simulation Experiments

[0159] Let the number of targets N Tar = 3, the distances of the targets from the sonar are 4m, 19m, and 10m respectively, the azimuth angles are 11°, 12°, and 13° respectively, and the radial velocities are 10m / s, -10m / s, and 9m / s respectively. Set the starting frequency f of the transmitted signal c to be 1500Hz, the bandwidth is 1500Hz, the pulse duration is 0.1s, the pulse repetition interval is 0.5s, the number of pulses is 5, the receiving array is a uniform linear array, the number of array elements is 5, and the element spacing is d = c / (2f c ). It can be seen that if the coupling between fast time and slow time is ignored, the maximum velocity that can be estimated at this time is v max = 0.5m / s, and velocity ambiguity will occur. After considering the coupling term, according to the description of the method of the present application, the maximum velocity that can be estimated is v max = 20m / s, and there is no velocity ambiguity at this time. When no noise is considered, the results of using the narrowband MUSIC algorithm, the ISSM algorithm, and the proposed wideband system angle-velocity joint estimation method are shown in Figures 3(a), 3(b), and 3(c) respectively.

[0160] As can be seen from Figure 3(a), the narrowband MUSIC algorithm cannot resolve the signals at all, indicating that the coupling term in (13) cannot be ignored. In Figure 3(b), due to insufficient number of snapshots in the wideband ISSM algorithm, there are a large number of false peaks in the estimation results, and it cannot resolve the two targets when their parameters are close. In contrast, the algorithm proposed in the present invention in Figure 3(c) can achieve accurate estimation of the target azimuth angle and velocity, and the peak is the sharpest, the sidelobe is the lowest, and the resolution is the highest.

[0161] To verify the noise robustness of the proposed algorithm, Figures 4(a) and 4(b) show the estimation accuracies of the angle and distance at different signal-to-noise ratios (SNRs). Add Gaussian white noise, and when the SNR is 0 - 50dB, 100 Monte Carlo simulations are performed, and the root mean square error (RMSE) of the angle and velocity estimations is shown in Figure 4.

[0162] As can be seen from Figures 4(a) and 4(b), when the SNR is higher than 20dB, the proposed method can achieve high accuracy in estimating the angle and velocity, demonstrating the robustness of the proposed algorithm.

[0163] The present application also provides an angle-velocity joint estimation device for a broadband system, which is implemented based on the above method. The device includes:

[0164] A dechirp pulse compression module, configured to perform dechirp pulse compression processing on a received signal to obtain a pulse-compressed signal;

[0165] A time-domain sampling module, configured to perform time-domain sampling on the pulse-compressed signal to convert an analog signal into a discrete-time signal;

[0166] A multi-dictionary sparse recovery problem construction module, configured to sample and arrange the discrete-time signal, design a corresponding over-complete dictionary, and construct a corresponding multi-dictionary sparse recovery problem;

[0167] A multi-dictionary sparse recovery problem solving module, configured to solve the multi-dictionary sparse recovery problem using a multi-dictionary sparse Bayesian algorithm; and

[0168] An angle-velocity joint estimation module, configured to rearrange the solution result to obtain an angle-velocity spectrum, and implement angle-velocity joint estimation.

[0169] The present application may also provide a computer device, including: at least one processor, a memory, at least one network interface, and a user interface. Each component in the device is coupled together through a bus system. It can be understood that the bus system is used to realize the connection and communication between these components. In addition to a data bus, the bus system further includes a power bus, a control bus, and a status signal bus.

[0170] Among them, the user interface may include a display, a keyboard, or a pointing device. For example, a mouse, a trackball, a touchpad, or a touch screen, etc.

[0171] It can be understood that the memory in the disclosed embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), or a flash memory. The volatile memory can be a Random Access Memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memories described herein are intended to include but not be limited to these and any other suitable types of memories.

[0172] In some embodiments, the memory stores the following elements, executable modules, or data structures, or subsets or supersets thereof: an operating system and application programs.

[0173] Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., and is used to implement various basic services and handle hardware-based tasks. The application programs include various application programs, such as a Media Player, a Browser, etc., and are used to implement various application services. The program for implementing the method of the disclosed embodiments of the present application can be included in the application programs.

[0174] In the above-mentioned embodiments, by calling the programs or instructions stored in the memory, specifically, the programs or instructions stored in the application programs, the processor is configured to:

[0175] Execute the steps of the above method.

[0176] The above method can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with the ability to process signals. During implementation, the steps of the above method can be completed by the integrated logic circuit of the hardware in the processor or instructions in the form of software. The above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed above. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. Combining the steps of the above-disclosed method can be directly embodied as being executed and completed by a hardware decoding processor, or by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0177] It can be understood that these embodiments described in the present application can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application, or a combination thereof.

[0178] For software implementation, the technology of the present application can be implemented by executing the functional modules of the present application (such as procedures, functions, etc.). The software code can be stored in the memory and executed by the processor. The memory can be implemented inside or outside the processor.

[0179] The present application may also provide a non-volatile storage medium for storing a computer program. When the computer program is executed by a processor, each step in the above method embodiments can be implemented.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit them. Although the present application has been described in detail with reference to the embodiments, those of ordinary skill in the art should understand that any modification or equivalent replacement of the technical solutions of the present application does not depart from the spirit and scope of the technical solutions of the present application, and they should all be covered within the scope of the claims of the present application.

Claims

1. A broadband system angle-velocity joint estimation method for estimating the target angle and velocity according to the received signal by a uniform linear array, including: Step 1: Perform dechirp pulse compression processing on the received signal to obtain a pulse-compressed signal; Step 2: Perform time-domain sampling on the pulse-compressed signal to convert the analog signal into a discrete-time signal; Step 3: Arrange the discrete-time signal, design a corresponding overcomplete dictionary, and construct a corresponding multi-dictionary sparse recovery problem; Step 4: Solve the multi-dictionary sparse recovery problem using the multi-dictionary sparse Bayesian algorithm; Step 5: Rearrange the solution result to obtain an angle-velocity spectrum, and realize angle-velocity joint estimation.

2. The angle-velocity joint estimation method for a broadband system according to claim 1, characterized in that The said Step 1 includes: The pulse compression signal obtained by performing dechirp pulse compression on the signal received by the \(l\)th array element is Among them, N Tar represents the number of targets to be measured; represents noise; The transmitting array element emits a string of Chirp signals in a coherent processing interval, T c represents the pulse duration of the signal, T p represents the pulse repetition interval of the signal, N p represents the number of pulses of the signal, B represents the pulse width of the signal, f c represents the starting frequency of the signal; t m = mT p , m ∈ {1, 2,..., N p}; The reflected signal received by the l-th array element from the i-th target The difference-frequency output obtained by correlating with the reference signal s: Among them, represents the loss coefficient after the transmitted signal propagates and reflects; R i represents the radial distance of the i-th target from the sonar; θ i represents the azimuth angle of the i-th target from the sonar; v i represents the radial velocity of the i-th target from the sonar; c represents the speed of sound in water; μ = B / T c ; 3. The angle-velocity joint estimation method for a broadband system according to claim 2, wherein The said Step 2 includes: Time-domain sampling of the pulse pressure signal is performed at a sampling rate of f s , and the discrete-time signal after sampling is as follows: Wherein, k represents the fast-time sampling index.

4. The angle-velocity joint estimation method for a broadband system according to claim 3, characterized in that The said Step 3 includes: Let There is / are: Among them, N m represents the number of array elements; T represents transpose; For Y k Vectorize to obtain: Among them, represents the Kronecker product, Expand A k into an overcomplete dictionary, and the parameter to be estimated is obtained by the compressive sensing algorithm. Specifically, first divide the value ranges of the radial velocity v and the azimuth angle θ into fine grids to form sets and set Θ respectively; let |·| denote the cardinality of a set; for A k the overcomplete dictionary obtained by expansion is denoted by and there is: Construct a corresponding multi-dictionary sparse recovery problem: Among them, is a sparse vector, and its support set is obtained by the compressive sensing algorithm, and the estimates of the parameters v and θ to be estimated are obtained therefrom; N t represents the total number of fast-time samplings; ∈ k represents the noise threshold; ||·||2 represents the 2-norm of the matrix; ||·|| 2,1 represents the 2,1-norm of the matrix.

5. The broadband system angle-velocity joint estimation method according to claim 4, characterized in that The said Step 4 includes: All have the same covariance matrix, which is Λ = diag(λ), where Estimation of λ is as follows: Wherein, H represents the conjugate matrix; Solve through an iterative algorithm to obtain: Among them, represents N p N m ×N p N m identity matrix; b represents an adjustable parameter for controlling the convergence rate; represents [λ] i result of the previous iteration; represents [λ] i result after the next iteration; Tr represents the trace of a matrix; represents the noise power.

6. The broadband system angle-velocity joint estimation method according to claim 5, wherein The estimated method for the noise power is as follows: Among them, represents the support set of λ, that is, the indices corresponding to the largest N Tar elements in λ; represents the pseudo-inverse.

7. The angle-velocity joint estimation method for a broadband system according to claim 5, characterized in that The said Step 5 includes: The joint spectrum Spec of the azimuth angle θ and the radial velocity v is obtained by estimating rearranged as the matrix of, that is: Among them, means that takes |Θ| elements as a column and arranges them into columns, and the arrangement is matrix.

8. A broadband system angle-velocity joint estimation device, implemented based on the method according to any one of claims 1-7, characterized in that, The said device includes: A dechirp pulse compression module, used to perform dechirp pulse compression processing on the received signal to obtain a pulse-compressed signal; A time-domain sampling module, used to perform time-domain sampling on the pulse-compressed signal to convert the analog signal into a discrete-time signal; A module for constructing a multi-dictionary sparse recovery problem, used to arrange the discrete-time signal, design a corresponding overcomplete dictionary, and construct a corresponding multi-dictionary sparse recovery problem; A module for solving the multi-dictionary sparse recovery problem, used to solve the multi-dictionary sparse recovery problem using the multi-dictionary sparse Bayesian algorithm; and An angle-velocity joint estimation module, used to rearrange the solution result to obtain an angle-velocity spectrum and realize angle-velocity joint estimation.

Images