Exogenous radar target detection method and system based on sparse model

Abstract

The invention provides an exogenous radar target detection method and system based on a sparse model, and relates to the technical field of radar signal processing, and the method comprises the steps: receiving a time domain monitoring signal of an orthogonal frequency division multiplexing waveform, carrying out the discrete Fourier transform, obtaining a carrier domain monitoring signal, and extracting data at a pilot frequency position; constructing a carrier domain reference signal based on known transmitting end pilot frequency information; carrying out clutter suppression processing on the carrier domain monitoring signal; performing inverse discrete Fourier transform on the carrier domain monitoring signal and the carrier domain reference signal, reconstructing a new time domain reference signal and a new time domain monitoring signal, and performing distance processing to obtain distance domain data; according to the method, only pilot frequency information is utilized to construct reference signals, direct waves do not need to be accurately estimated, and the system complexity is greatly reduced; and clutter suppression is introduced before sparse modeling, so that the problem that strong clutters cover weak targets is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing technology, specifically to a method and system for detecting external radar targets based on a sparse model. Background Technology

[0002] External radar, also known as passive radar, does not emit signals itself. Instead, it uses commercial or broadcast signals (such as digital television signals, FM radio signals, cellular network signals, etc.) transmitted by third parties as an illumination source. It detects, locates, and tracks targets by receiving the signals reflected from them. Compared with traditional active radar, external radar has significant advantages such as no need for spectrum allocation, strong anti-jamming capability, flexible system deployment, low maintenance cost, and no electromagnetic pollution. Therefore, it has received widespread attention in both military and civilian fields.

[0003] Existing methods for detecting external radar targets based on sparse models all revolve around constructing and solving models based on the sparsity of the received signal. However, these methods still have technical shortcomings in practical engineering applications: First, they are highly dependent on the accurate estimation of the direct wave signal: Generally, when constructing the sparse model and performing clutter suppression, it is necessary to use the accurate direct wave signal received by the reference channel as a benchmark. However, the estimation of the direct wave signal is affected by factors such as multipath propagation, channel fading, and noise interference. This not only requires complex estimation algorithms, significantly increasing the computational burden and hardware complexity of the system, but also leads to a decrease in subsequent target detection performance due to estimation errors. Second, they suffer from insufficient performance in detecting weak targets under strong clutter interference: The sparse model solution itself does not consider clutter suppression processing, and existing methods... While some methods add clutter suppression steps before sparse modeling, they only employ simple average subtraction and basic cancellation algorithms. In environments with strong zero-frequency multipath clutter (clutter signal-to-noise ratio CNR can reach 60dB or higher), the clutter suppression effect is poor. Strong clutter can completely mask the echo signal of weak targets (target signal-to-noise ratio SNR as low as -30dB), making the target undetectable. Thirdly, it is difficult to balance accuracy and computational complexity: some methods use unsegmented full-signal range correlation processing to improve target detection accuracy, resulting in an exponential increase in computational load, which cannot meet the real-time requirements of engineering. On the other hand, some methods that simplify computation reduce complexity by sacrificing detection accuracy, making them unsuitable for high-precision target detection scenarios. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide an external radar target detection method and system based on a sparse model, thereby solving at least one of the problems in the background technology.

[0005] A first aspect of the present invention is to provide a method for detecting external radar targets based on a sparse model, the method comprising: The system receives the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, performs a discrete Fourier transform on it to obtain the carrier domain monitoring signal, and extracts the data at its pilot position. Based on the known transmitter pilot information and the data at the pilot location, a carrier domain reference signal containing the pilot information is constructed. In the carrier domain, the carrier domain reference signal is used to perform zero-frequency clutter suppression processing on the carrier domain monitoring signal to obtain a clutter-suppressed carrier domain monitoring signal. The clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal are subjected to inverse discrete Fourier transform to reconstruct a new time domain reference signal and a new time domain monitoring signal. Then, the two are subjected to piecewise distance correlation processing to obtain distance domain data. Based on the range domain data, a sparse model of the range-Doppler spectrum is established and a corresponding optimization problem is formed. The optimization problem is solved to obtain the target detection results.

[0006] According to one aspect of the above technical solution, the steps of receiving the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, performing a discrete Fourier transform on it to obtain the carrier-domain monitoring signal, and extracting the data at its pilot position specifically include: Time-domain monitoring signals include direct waves, zero-frequency multipath clutter, target echoes, and receiver thermal noise, represented as: , in, To monitor the time-domain monitoring signals received by the monitoring channel, For discrete time delay is The subsequent direct wave, It is the total number of zero-frequency multipath clutter. The number of targets to be detected. , They represent the first i The complex envelope amplitude and discrete time delay of a zero-frequency multipath clutter. , and They represent the first q Complex envelope amplitude, discrete time delay, and normalized Doppler frequency shift of the target echo. Indicates the direct wave latency as Doppler frequency shift is A copy, To monitor discrete noise vectors within the channel, n For discrete-time indexing, N The total discrete time of the time-domain monitoring signal; The time-domain monitoring signal is divided into several groups of orthogonal frequency division multiplexing (OFDM) symbols according to the symbol period. Each group of OFDM symbols includes a valid data segment and a cyclic prefix, represented as follows: , , in, Indicates the first l In the group of orthogonal frequency division multiplexing symbols, the first Complex data at each subcarrier position This represents the number of sampling points for valid data in a set of orthogonal frequency division multiplexing symbols. This represents the total number of sampling points in a set of orthogonal frequency division multiplexing symbols. For sampling function, This represents the total number of orthogonal frequency division multiplexing (OFDM) symbol groups. No. Group of orthogonal frequency division multiplexing symbols, These are the basis functions of the inverse discrete Fourier transform. m The index is used for summation, where j is the imaginary unit; For each group of orthogonal frequency division multiplexing symbols, the cyclic prefix is ​​removed, and a discrete Fourier transform is performed to convert the time-domain monitoring signal to the carrier domain, resulting in the carrier-domain monitoring signal, expressed as: , , in, For carrier domain monitoring signal vectors, For the first l Carrier domain monitoring signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols. For the first l In the group of orthogonal frequency division multiplexing symbols, the first Monitoring data of the position of the nth subcarrier, if the nth When the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. If the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. When the subcarrier position is a non-pilot position, the monitoring data is zero.

[0007] According to one aspect of the above technical solution, the step of constructing a carrier domain reference signal containing pilot information based on known transmitter pilot information and data at the pilot location specifically includes: Based on the known transmitter pilot information, the pilot positions are assigned the known pilot values ​​from the transmitter pilot information, and the non-pilot positions are set to zero. A carrier domain reference signal containing pilot information is constructed, represented as: , , in, For the carrier domain reference signal vector, No. l Carrier domain reference signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols For the first l In the group of orthogonal frequency division multiplexing symbols, the first Reference data for the position of the nth subcarrier, if the nth When the position of the first subcarrier is the pilot position, the reference data is the pilot value; if the first... When the subcarrier position is a non-pilot position, the reference data is zero.

[0008] According to one aspect of the above technical solution, in the carrier domain, the step of performing zero-frequency clutter suppression processing on the carrier domain monitoring signal using the carrier domain reference signal to obtain a clutter-suppressed carrier domain monitoring signal specifically includes: Extracting the first group of orthogonal frequency division multiplexing symbols from several groups The monitoring data of each subcarrier constitutes a monitoring data vector; Extract the first orthogonal frequency division multiplexing symbol from the corresponding group. The reference data of each subcarrier constitutes a reference data vector; Based on the monitoring data vector and the reference data vector, with the goal of minimizing the energy of the signal after clutter suppression, the adaptive cancellation coefficient is calculated using the following formula: , in, For adaptive offset coefficient, , These are the monitoring data vector and the reference data vector, respectively. The extended cancellation algorithm is used to monitor the carrier domain signal. Each subcarrier is independently processed for clutter suppression to obtain the clutter-suppressed carrier domain monitoring signal. The calculation formula is as follows: , in, For the first The carrier domain monitoring signal vector after subcarrier clutter suppression.

[0009] According to one aspect of the above technical solution, the steps of performing inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively to reconstruct a new time-domain reference signal and a new time-domain monitoring signal, and then performing piecewise distance correlation processing on the two to obtain distance-domain data, specifically include: The clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal are respectively subjected to inverse discrete Fourier transform according to their signs, and a cyclic prefix is ​​added to reconstruct a new time-domain reference signal and a new time-domain monitoring signal. The new time-domain reference signal and the new time-domain monitoring signal are each divided into several segments with a preset segment length, as follows: , , in, t For continuous time variables, For continuous time variables t The time-domain reference signal below, For continuous time variables t The time-domain monitoring signal below, For the first b Each segment corresponds to a reference signal segment. For the first b Each segment corresponds to a monitoring signal segment. The total number of segments. It is a preset segment length. For a rectangular pulse function, when At that time, the rectangular pulse function is 1, and the rest are 0; Range compression is performed on the time-domain reference signal and time-domain monitoring signal corresponding to each segment to obtain range-domain data, represented as follows: , in, For the first b Each segment experiences a time delay. The distance domain data afterwards This is the preset maximum detectable delay. For the first b The time-domain monitoring signal corresponding to each segment For the first b The time-domain reference signal corresponding to each segment is delayed. And take the signal component after complex conjugation, * is the complex conjugation operation.

[0010] According to one aspect of the above technical solution, the step of establishing a sparse model of the range-Doppler spectrum based on the range domain data specifically includes: Based on the velocity range and operating parameters of the target to be detected, the Doppler frequency search interval is determined. For each segment of range domain data, the Doppler frequency is used as the search variable to construct a segmented mutual ambiguity function, expressed as: , in, For the first b A segmented mutual ambiguity function, The frequency is the Doppler frequency. The global mutual ambiguity function is obtained by weighted summation of all segmented mutual ambiguity functions according to their slow-time indices, and is expressed as: , in, , For the first b Slow-time indexes for each segment; When the product of the segment length and the maximum Doppler frequency in the Doppler frequency search interval satisfies a preset small condition, the phase of the sampling point within the segment is approximated as the phase at the midpoint of the segment. The approximate piecewise mutual ambiguity function is obtained, expressed as: , Introducing the approximate piecewise mutual ambiguity function into the global mutual ambiguity function yields the global approximate mutual ambiguity function, expressed as: ; Based on the global approximate mutual ambiguity function, the time delay and Doppler frequency are discretized to construct the Doppler shift Fourier transform matrix and the range domain echo matrix; Based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a sparse model of the range-Doppler spectrum is established.

[0011] According to one aspect of the above technical solution, the steps of discretizing the time delay and Doppler frequency based on the global approximate mutual ambiguity function to construct the Doppler shift Fourier transform matrix and the range domain echo matrix specifically include: The global approximate mutual ambiguity function is split into a time delay discrete part and a Doppler frequency discrete part; Based on the aforementioned discrete delay portion, the delay interval [0, ... Discretize the data into a first preset number of grids, construct a range-domain echo matrix, and represent it as follows: , in, For the first preset quantity, The range domain signal echo matrix, The matrix is ​​represented as OK, A complex matrix of columns; The Doppler frequency search interval is discretized into a second preset number of grids, and based on the slow-time index and the discrete portion of the Doppler frequency, a Doppler shift Fourier transform matrix is ​​constructed, expressed as: , Among them, the Doppler frequency range is , This is the Doppler shift Fourier transform matrix. For the second preset quantity, The matrix is ​​represented as OK, A complex matrix of columns, For the first Slow-time index for each segment.

[0012] According to one aspect of the above scheme, the step of establishing a sparse model of the range-Doppler spectrum based on the Doppler shift Fourier transform matrix and the range domain echo matrix specifically includes: Based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a discretized mutually fuzzy function matrix is ​​established, expressed as: , in, Represents the discretized mutual fuzzy function matrix. The matrix is ​​represented as OK, A complex matrix of columns; By transposing the mutually fuzzy function matrix and combining it with the unitary property of the Doppler shift Fourier transform matrix, a sparse model is obtained, expressed as: , in, This is the conjugate transpose of the Doppler shift Fourier transform matrix.

[0013] According to one aspect of the above technical solution, from the sparse model Due to the sparsity of the sparse model, the sparse model is transformed into an optimization problem, expressed as: , in, Denotes the F-norm of a matrix. Denotes the L1 norm of a matrix. This is the preset error tolerance threshold; The optimization problem is solved by the orthogonal matching pursuit algorithm to obtain the target detection results.

[0014] Another aspect of the present invention provides a method and system for detecting external radar targets based on a sparse model. The system is used to implement the aforementioned method for detecting external radar targets based on a sparse model, and the system includes: The monitoring signal acquisition module is used to receive the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, and to perform a discrete Fourier transform on it to obtain the carrier domain monitoring signal. The reference signal acquisition module is used to extract data at the pilot position from the carrier domain monitoring signal and construct a carrier domain reference signal containing pilot information based on the known transmitter pilot information. The clutter suppression module is used to perform zero-frequency clutter suppression processing on the carrier domain monitoring signal in the carrier domain using the carrier domain reference signal, so as to obtain the clutter-suppressed carrier domain monitoring signal. The distance correlation processing module is used to perform inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively, reconstruct a new time domain reference signal and a new time domain monitoring signal, and perform segmented distance correlation processing on the two to obtain distance domain data; The sparse model solving module is used to establish a sparse model of the range-Doppler spectrum based on the range domain data and form a corresponding optimization problem, solve the optimization problem, and obtain the target detection results.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Existing sparse model-based methods generally require high-precision direct wave signal estimation as a prerequisite for modeling and clutter suppression, which increases the complexity and computational burden of processing. This scheme abandons the reliance on estimating the complete direct wave signal and instead directly utilizes the inherent and known pilot information in the transmitted signal to construct the reference signal and perform clutter suppression. The entire processing flow does not require complex channel estimation and signal reconstruction steps, which reduces the stringent requirements of the algorithm on the quality of the front-end signal and improves the engineering practicality and robustness of the method.

[0016] 2. Traditional sparse model methods are prone to masking weak target signals in the context of strong direct waves and multipath clutter, resulting in a decrease in detection performance. Before constructing the sparse model, this scheme innovatively introduces a pilot-based extended phase cancellation zero-frequency clutter suppression step in the carrier domain. This step can effectively remove strong clutter components related to the direct wave in the monitoring signal, thereby highlighting the weak target echo that was originally masked.

[0017] 3. This scheme transforms the calculation of continuous mutually fuzzy functions into piecewise approximations through piecewise distance correlation processing, and establishes a time-invariant sparse model based on this. The structure of this model is independent of the data accumulation time and only depends on the preset distance-Doppler grid number, making the computational complexity controllable. At the same time, it combines efficient sparse recovery algorithms such as orthogonal matching pursuit for solving, which can significantly reduce the amount of computation and memory consumption while ensuring high-resolution imaging (detection accuracy). Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 The RD (range-Doppler) spectra of targets with different signal-to-noise ratios were detected in a simulation experiment according to Embodiment 1 of the present invention. Figure 2This is a PSLR (peak-to-sidelobe ratio) curve from a simulation experiment of Embodiment 1 of the present invention; Figure 3 This is the ISLR (integral sidelobe ratio) curve of Embodiment 1 of the present invention in a simulation experiment; Figure 4 The RD (range-Doppler) spectra of targets with different signal-to-noise ratios were detected in a field experiment according to Embodiment 1 of the present invention. Figure 5 The image shows the RD (range-Doppler) spectrum after two-dimensional matched filtering following the use of the extended phase cancellation zero-frequency clutter suppression method in the experimental test of Embodiment 1 of the present invention. Detailed Implementation

[0019] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present invention will be more thorough and complete.

[0020] Example 1 Embodiment 1 of the present invention provides an external radar target detection method based on a sparse model, the method comprising steps S1-S7: Step S1: Receive the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, perform a discrete Fourier transform on it to obtain the carrier domain monitoring signal, and extract the data at its pilot position. Time-domain monitoring signals include direct waves, zero-frequency multipath clutter, target echoes, and receiver thermal noise, represented as: , in, To monitor the time-domain monitoring signals received by the monitoring channel, For discrete time delay is The subsequent direct wave, It is the total number of zero-frequency multipath clutter. The number of targets to be detected. , They represent the first i The complex envelope amplitude and discrete time delay of a zero-frequency multipath clutter. , and They represent the first q Complex envelope amplitude, discrete time delay, and normalized Doppler frequency shift of the target echo. Indicates the direct wave latency as Doppler frequency shift is A copy, To monitor discrete noise vectors within the channel, i , q All are summation indices. n For discrete-time indexing, N The total discrete time of the time-domain monitoring signal; The time-domain monitoring signal is divided into several groups of orthogonal frequency division multiplexing (OFDM) symbols according to the symbol period. Each group of OFDM symbols includes a valid data segment and a cyclic prefix, represented as follows: , , in, Indicates the first l In the group of orthogonal frequency division multiplexing symbols, the first Complex data at each subcarrier position This represents the number of sampling points for valid data in a set of orthogonal frequency division multiplexing symbols. This represents the total number of sampling points in a set of orthogonal frequency division multiplexing symbols. For sampling function, This represents the total number of orthogonal frequency division multiplexing (OFDM) symbol groups. No. Group of orthogonal frequency division multiplexing symbols, These are the basis functions of the inverse discrete Fourier transform. m The index is used for summation, where j is the imaginary unit; For each group of orthogonal frequency division multiplexing symbols, the cyclic prefix is ​​removed, and a discrete Fourier transform is performed to convert the time-domain monitoring signal to the carrier domain, resulting in the carrier-domain monitoring signal, expressed as: , , in, For carrier domain monitoring signal vectors, For the first l Carrier domain monitoring signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols. For the first l In the group of orthogonal frequency division multiplexing symbols, the first Monitoring data of the position of the nth subcarrier, if the nth When the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. If the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. When the subcarrier position is a non-pilot position, the monitoring data is zero.

[0021] It should be noted that symbol segmentation and cyclic prefix removal are performed on the time-domain monitoring signal to eliminate inter-symbol interference in the orthogonal frequency division multiplexing waveform, obtaining independent and clean single-symbol valid data, laying the foundation for subsequent domain conversion and pilot extraction. Furthermore, the time-domain monitoring signal is converted to the carrier domain using Discrete Fourier Transform, utilizing the inherent distribution characteristics of data at pilot positions in the carrier domain to provide a suitable signal form for subsequent targeted clutter suppression in the carrier domain. In addition, data at pilot positions is extracted, and data at non-pilot positions is set to zero, eliminating invalid components in the carrier-domain monitoring signal, reducing the computational load of subsequent signal processing, while retaining core pilot information, providing data support for constructing a reference signal without direct wave dependence.

[0022] Step S2: Based on the known transmitter pilot information and the data at the pilot location, construct a carrier domain reference signal containing the pilot information; Specifically, based on the known transmitter pilot information, the pilot positions are assigned the known pilot values ​​from the transmitter pilot information, and the non-pilot positions are set to zero, thus constructing a carrier domain reference signal containing pilot information, represented as: , , in, For the carrier domain reference signal vector, No. l Carrier domain reference signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols For the first l In the group of orthogonal frequency division multiplexing symbols, the first Reference data for the position of the nth subcarrier, if the nth When the position of the first subcarrier is the pilot position, the reference data is the pilot value; if the first... When the subcarrier position is a non-pilot position, the reference data is zero.

[0023] In other words, based on the known pilot information at the transmitting end, a time-domain reference signal with the same structure as the carrier domain monitoring signal is constructed to ensure the matching between the time-domain reference signal and the time-domain monitoring signal in subsequent clutter suppression processing. Furthermore, the construction of the time-domain reference signal only utilizes the known pilot information at the transmitting end, without relying on the estimation of the direct wave signal of the reference channel, completely eliminating the dependence on the accurate estimation of the direct wave in traditional methods and greatly simplifying the construction process of the time-domain reference signal. In addition, the method of zeroing non-pilot positions is consistent with that of the carrier domain monitoring signal, ensuring that only the effective signal at the pilot position is processed during clutter suppression, avoiding interference from invalid data on the clutter suppression effect.

[0024] Step S3: In the carrier domain, the carrier domain monitoring signal is subjected to zero-frequency clutter suppression processing using the carrier domain reference signal to obtain the clutter-suppressed carrier domain monitoring signal. Specifically, this involves extracting the first element from several groups of orthogonal frequency division multiplexing symbols. The monitoring data of each subcarrier constitutes a monitoring data vector; Extract the first orthogonal frequency division multiplexing symbol from the corresponding group. The reference data of each subcarrier constitutes a reference data vector; Based on the monitoring data vector and the reference data vector, with the goal of minimizing the energy of the signal after clutter suppression, the adaptive cancellation coefficient is calculated using the following formula: , in, For adaptive offset coefficient, , These are the monitoring data vector and the reference data vector, respectively. The extended cancellation algorithm is used to monitor the carrier domain signal. Each subcarrier is independently processed for clutter suppression to obtain the clutter-suppressed carrier domain monitoring signal. The calculation formula is as follows: , in, For the first The carrier domain monitoring signal vector after subcarrier clutter suppression.

[0025] In other words, by processing each subcarrier independently, and considering the multi-carrier structural characteristics of orthogonal frequency division multiplexing (OFDM) signals, precise suppression of zero-frequency clutter on each subcarrier is achieved, avoiding mutual interference between subcarriers. Furthermore, with the goal of minimizing the energy of the signal after clutter suppression, an adaptive cancellation coefficient is calculated, making the clutter suppression process adaptive. This allows for dynamic adjustment of the cancellation degree based on the clutter intensity of different subcarriers, improving the targeting and effectiveness of clutter suppression. Finally, extended destructive clutter suppression is performed in the carrier domain, directly canceling the carrier domain components of zero-frequency clutter. Compared to traditional time-domain clutter suppression methods, this is more suitable for OFDM signals and offers superior clutter suppression performance. In addition, clutter suppression is completed before sparse modeling, eliminating strong zero-frequency clutter components in advance and preventing clutter from masking weak target signals in subsequent signal processing, thus ensuring the detection of weak targets.

[0026] Step S4: Perform inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively to reconstruct a new time domain reference signal and a new time domain monitoring signal, and perform segmented distance correlation processing on the two to obtain distance domain data; Specifically, the carrier domain monitoring signal and the carrier domain reference signal after clutter suppression are respectively subjected to inverse discrete Fourier transform according to their signs, and a new time-domain reference signal and a new time-domain monitoring signal are reconstructed by adding a cyclic prefix respectively. The new time-domain reference signal and the new time-domain monitoring signal are each divided into several segments with a preset segment length, as follows: , , in, t As a continuous-time variable, it represents the instantaneous value of the signal in the time dimension. For continuous time variables t The time-domain reference signal below, For continuous time variables t The time-domain monitoring signal below, For the first b Each segment corresponds to a reference signal segment. For the first b Each segment corresponds to a monitoring signal segment. The total number of segments. It is a preset segment length. For a rectangular pulse function, when At that time, the rectangular pulse function is 1, and the rest are 0; Furthermore, The selection of the target is related to the approximate accuracy of the target and the computational complexity of the method. The shorter the segment (the more segments), the higher the approximate accuracy, but the computational complexity also increases accordingly. The selection of a target needs to take into account both the accuracy of target detection and the computational complexity.

[0027] Range compression is performed on the time-domain reference signal and time-domain monitoring signal corresponding to each segment to obtain range-domain data, represented as follows: , in, For the first b Each segment experiences a time delay. The distance domain data afterwards This is the preset maximum detectable delay. For the first b The time-domain monitoring signal corresponding to each segment For the first b The time-domain reference signal corresponding to each segment is delayed. And take the signal component after complex conjugation, * is the complex conjugation operation.

[0028] It should be noted that the carrier domain signal is converted back to the time domain by inverse discrete Fourier transform, realizing the reverse conversion of the signal domain. At the same time, the processing effect of the previous carrier domain clutter suppression is preserved, providing a suitable time domain signal form for subsequent distance correlation processing. Furthermore, the addition of a cyclic prefix restores the complete structure, eliminates inter-symbol interference in the time domain transmission and processing of the signal, and ensures the integrity and effectiveness of the time domain signal.

[0029] Furthermore, the segmented processing decomposes the overall range correlation into multiple segmented range correlations, which greatly simplifies the computational complexity of the subsequent mutual ambiguity function. At the same time, the approximate accuracy and computational load of target detection can be flexibly balanced by adjusting the segment length. In addition, range compression processing is performed on each segment to convert the time domain signal into range domain data, realizing the preliminary extraction of target range information and laying the foundation for the subsequent construction of a sparse model by combining Doppler information.

[0030] Step S5: Based on the range domain data, establish a sparse model of the range-Doppler spectrum and form a corresponding optimization problem. Solve the optimization problem to obtain the target detection result.

[0031] Specifically, based on the velocity range and operating parameters of the target to be detected, the Doppler frequency search interval is determined. For each segment of range domain data, the Doppler frequency is used as the search variable to construct a segmented mutual ambiguity function, expressed as: , in, For the first b A piecewise mutual ambiguity function that simultaneously characterizes time delay. (Distance) and Doppler frequency (Velocity) dimension of signal energy, For Doppler frequency, Doppler frequency shift in continuous time variable t The corresponding phase term; The global mutual ambiguity function is obtained by weighted summation of all segmented mutual ambiguity functions according to their slow-time indices, and is expressed as: , in, , For the first b The slow-time index of the segment represents the first segment. b The position of each segment on the global timeline Indexing the Doppler frequency shift in slow time The corresponding phase term; Wherein, slow time is defined as .

[0032] When the product of the segment length and the maximum Doppler frequency in the Doppler frequency search interval satisfies a preset small condition, the phase of the sampling point within the segment is approximated as the phase at the midpoint of the segment. The approximate piecewise mutual ambiguity function is obtained, expressed as: , Among them, satisfying the preset small quantity condition means that the product is small, for example, much less than 1.

[0033] Introducing the approximate piecewise mutual ambiguity function into the global mutual ambiguity function yields the global approximate mutual ambiguity function, expressed as: ; Among them, a piecewise mutual ambiguity function is constructed and phase approximation is performed, which greatly simplifies the computational complexity of the mutual ambiguity function while ensuring detection accuracy. At the same time, a global approximate mutual ambiguity function is obtained by combining slow time index, so as to achieve joint representation of target distance and Doppler information.

[0034] Based on the global approximate mutual ambiguity function, the time delay and Doppler frequency are discretized to construct the Doppler shift Fourier transform matrix and the range domain echo matrix; Specifically, the global approximate mutual ambiguity function is split into a time delay discrete part and a Doppler frequency discrete part; Based on the aforementioned discrete delay portion, the delay interval [0, ... Discretize the data into a first preset number of grids, construct a range-domain echo matrix, and represent it as follows: , in, For the first preset quantity, The range domain signal echo matrix, The matrix is ​​represented as OK, A complex matrix of columns; The Doppler frequency search interval is discretized into a second preset number of grids, and based on the slow-time index and the discrete portion of the Doppler frequency, a Doppler shift Fourier transform matrix is ​​constructed, expressed as: , Among them, the Doppler frequency range is , This is the Doppler shift Fourier transform matrix. For the second preset quantity, The matrix is ​​represented as OK, A complex matrix of columns, For the first Slow-time index for each segment.

[0035] In other words, by discretizing the time delay and Doppler frequency, the continuous Doppler spectral domain is transformed into a discrete grid space, providing a discretized mathematical basis for the construction of sparse models. At the same time, the number of discrete grids can be flexibly adjusted according to the detection accuracy requirements.

[0036] Next, based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a sparse model of the range-Doppler spectrum is established.

[0037] Specifically, based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a discretized mutually fuzzy function matrix is ​​established, expressed as: , in, Represents the discretized mutual fuzzy function matrix. The matrix is ​​represented as OK, A complex matrix of columns; By transposing the mutually fuzzy function matrix and combining it with the unitary property of the Doppler shift Fourier transform matrix, a sparse model is obtained, expressed as: , in, This is the conjugate transpose of the Doppler shift Fourier transform matrix.

[0038] Finally, from the sparse model Due to the sparsity of the sparse model, the sparse model is transformed into an optimization problem, expressed as: , in, Denotes the F-norm of a matrix. Denotes the L1 norm of a matrix. This is the preset error tolerance threshold; The optimization problem is solved by the orthogonal matching pursuit algorithm to obtain the target detection results.

[0039] It should be noted that a sparse model is constructed based on the range domain echo matrix and the Doppler shift Fourier transform matrix. The characteristics of the sparse model are used to overcome the resolution limitations of traditional matched filtering methods, achieving high-resolution extraction of target range and Doppler information. The sparse model is transformed into an optimization problem, and the target detection problem is transformed into a sparse matrix reconstruction problem by utilizing the sparsity of the target, which greatly improves the anti-interference capability and accuracy of target detection. The orthogonal matching pursuit algorithm is used to solve the optimization problem. This algorithm has the characteristics of fast solution speed, low computational cost, and high reconstruction accuracy, and can quickly realize the reconstruction of sparse matrix to meet the real-time requirements of radar target detection.

[0040] Simulation experiment: To verify the effectiveness of the proposed method, the following simulation parameters were set: the simulation environment consisted of an Intel Core i5-12400F CPU @4.4GHz, 16GB RAM, and MATLAB 2019b. The illumination source signal was an OFDM-based shortwave digital broadcast signal with a quincunx pilot distribution, 256 subcarriers, and a sampling frequency of 12kHz. Complex Gaussian white noise was added to the signal transmission channel to simulate the actual transmission environment.

[0041] In the method proposed in this invention, for detection scenarios of high-speed targets such as aircraft, the Doppler frequency shift range is set to -10Hz to 10Hz, and the number of Doppler grids is set to 128. Considering that the discrete time delay of the target generally does not exceed the cyclic prefix length of the OFDM symbol, the number of time delay grids is set to 25, and the total number of grids in the range-Doppler spectrum is 3200. The method proposed in this invention adopts a time-invariant sparse model, and the size of the dictionary matrix is ​​only related to the scale of the RD spectrum and is independent of the accumulation time. Therefore, 50 OFDM symbols are selected for experiments. The segment length is taken as the length of 1 OFDM symbol, i.e., 21.33ms, and the time delay search range is taken as the length of 25 sampling points, i.e., 2.08ms, according to the grid size. The orthogonal matching pursuit algorithm is used as the sparse reconstruction algorithm, and the sparsity is set to 10. The reconstructed RD spectrum is processed by ordered thresholding to remove small error peaks. The threshold is set to the average value of all values ​​in the original RD spectrum.

[0042] Experiment 1: Multi-target detection performance test This experiment is set up as a multi-target scenario with three targets. The simulation parameters of the targets and multipath clutter are shown in Table 1 below.

[0043] Table 1:

[0044] like Figure 1 As can be seen, the method proposed in this invention successfully detected all three targets, including weak target 3 with a signal-to-noise ratio of -32dB, and the target parameters were consistent with those set in Table 1. The results show that the method proposed in this invention has good detection performance under strong clutter interference. The performance improvement is attributed to the introduction of a zero-frequency clutter suppression step before sparse modeling, which effectively solves the problem of strong clutter masking weak targets.

[0045] Experiment 2: Quantitative Performance Evaluation This experiment uses two metrics, Peak Sidelobe Ratio (PSLR) and Integrated Sidelobe Ratio (ISLR), to quantitatively evaluate different methods. Clutter parameters are the same as in Experiment 1, using 30 Orthogonal Frequency Division Multiplexing (OFDM) symbols. Each experiment inserts only a single fixed-position target, with a target signal-to-noise ratio ranging from -30 dB to -18 dB and a step size of 3 dB. Fifty Monte Carlo simulations are performed at each signal-to-noise ratio, and the mean PSLR and ISLR values ​​are calculated.

[0046] Depend on Figures 2-3 As can be seen, the PSLR and ISLR of the method proposed in this invention are both at low levels, meeting the target detection requirements. In the target detection performance evaluation, the lower the PSLR and ISLR, the better the detection performance, which is attributed to the effective zero-frequency clutter suppression performed before sparse modeling in this invention. In addition, PSLR and ISLR are inversely proportional to the target signal-to-noise ratio; for every 3dB increase in the signal-to-noise ratio, PSLR and ISLR decrease by approximately 3dB, consistent with the parameter definition, further verifying the correctness of the results.

[0047] Experiment 3: Actual Measurement Experiment To further verify the effectiveness of the proposed method, experiments were conducted using measured data from a passive radar system based on shortwave digital broadcast signals. The radar system comprises a non-cooperative transmitter and a receiver, located on the coasts of Qingdao and Haiyang, respectively. The system operates at a frequency of 10.05 MHz in ground wave mode, and the receiving antenna is a 16-element linear array. The proposed method only requires monitoring signals and pilot information for target detection. To improve the target signal-to-noise ratio, 120 OFDM symbols are used for processing.

[0048] Depend on Figure 4 It can be seen that although residual sea clutter exists at frequencies of approximately 1 Hz and 0.2 Hz, the target still appears clearly at a position with a Doppler frequency of approximately 5 Hz and a range element of approximately 4.5. Since the contrast method cannot effectively detect the target, this experiment uses the classical extended phase-clutter cancellation method combined with the processing results of two-dimensional matched filtering. Figure 5 (This is for reference only.) Figure 4 and Figure 5 The comparative results show that the method proposed in this invention can effectively detect targets in real-world scenarios, verifying the correctness and practicality of the method.

[0049] Example 2 Embodiment 2 of the present invention provides a method and system for detecting external radar targets based on a sparse model, the system comprising: The monitoring signal acquisition module is used to receive the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, and to perform a discrete Fourier transform on it to obtain the carrier domain monitoring signal. The reference signal acquisition module is used to extract data at the pilot position from the carrier domain monitoring signal and construct a carrier domain reference signal containing pilot information based on the known transmitter pilot information. The clutter suppression module is used to perform zero-frequency clutter suppression processing on the carrier domain monitoring signal in the carrier domain using the carrier domain reference signal, so as to obtain the clutter-suppressed carrier domain monitoring signal. The distance correlation processing module is used to perform inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively, reconstruct a new time domain reference signal and a new time domain monitoring signal, and perform segmented distance correlation processing on the two to obtain distance domain data; The sparse model solving module is used to establish a sparse model of the range-Doppler spectrum based on the range domain data and form a corresponding optimization problem, solve the optimization problem, and obtain the target detection results.

[0050] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0052] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this patent should be determined by the appended claims.

Claims

1. A method for detecting external radar targets based on a sparse model, characterized in that, The method includes: The system receives the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, performs a discrete Fourier transform on it to obtain the carrier domain monitoring signal, and extracts the data at its pilot position. Based on the known transmitter pilot information and the data at the pilot location, a carrier domain reference signal containing the pilot information is constructed. In the carrier domain, the carrier domain reference signal is used to perform zero-frequency clutter suppression processing on the carrier domain monitoring signal to obtain a clutter-suppressed carrier domain monitoring signal. The clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal are subjected to inverse discrete Fourier transform to reconstruct a new time domain reference signal and a new time domain monitoring signal. Then, the two are subjected to piecewise distance correlation processing to obtain distance domain data. Based on the range domain data, a sparse model of the range-Doppler spectrum is established and a corresponding optimization problem is formed. The optimization problem is solved to obtain the target detection results.

2. The external radar target detection method based on a sparse model according to claim 1, characterized in that, The steps of receiving the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, performing a discrete Fourier transform on it to obtain the carrier-domain monitoring signal, and extracting the data at its pilot position specifically include: Time-domain monitoring signals include direct waves, zero-frequency multipath clutter, target echoes, and receiver thermal noise, represented as: ， in, To monitor the time-domain monitoring signals received by the monitoring channel, For discrete time delay is The subsequent direct wave, It is the total number of zero-frequency multipath clutter. The number of targets to be detected. , They represent the first i The complex envelope amplitude and discrete time delay of a zero-frequency multipath clutter. , and They represent the first q Complex envelope amplitude, discrete time delay, and normalized Doppler frequency shift of the target echo. Indicates the direct wave latency as Doppler frequency shift is A copy, To monitor discrete noise vectors within the channel, n For discrete-time indexing, N The total discrete time of the time-domain monitoring signal; The time-domain monitoring signal is divided into several groups of orthogonal frequency division multiplexing (OFDM) symbols according to the symbol period. Each group of OFDM symbols includes a valid data segment and a cyclic prefix, represented as follows: ， ， in, Indicates the first l In the group of orthogonal frequency division multiplexing symbols, the first Complex data at each subcarrier position This represents the number of sampling points for valid data in a set of orthogonal frequency division multiplexing symbols. This represents the total number of sampling points in a set of orthogonal frequency division multiplexing symbols. For sampling function, This represents the total number of orthogonal frequency division multiplexing (OFDM) symbol groups. No. Group of orthogonal frequency division multiplexing symbols, These are the basis functions of the inverse discrete Fourier transform. m The index is used for summation, where j is the imaginary unit; For each group of orthogonal frequency division multiplexing symbols, the cyclic prefix is ​​removed, and a discrete Fourier transform is performed to convert the time-domain monitoring signal to the carrier domain, resulting in the carrier-domain monitoring signal, expressed as: ， ， in, For carrier domain monitoring signal vectors, For the first l Carrier domain monitoring signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols. For the first l In the group of orthogonal frequency division multiplexing symbols, the first Monitoring data of the position of the nth subcarrier, if the nth When the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. If the first subcarrier position is a pilot position, the monitoring data is the value of the extracted pilot position after passing through the channel response. When the subcarrier position is a non-pilot position, the monitoring data is zero.

3. The external radar target detection method based on a sparse model according to claim 2, characterized in that, The step of constructing a carrier domain reference signal containing pilot information based on known transmitter pilot information and data at the pilot locations specifically includes: Based on the known transmitter pilot information, the pilot positions are assigned the known pilot values ​​from the transmitter pilot information, and the non-pilot positions are set to zero. A carrier domain reference signal containing pilot information is constructed, represented as: ， ， in, For the carrier domain reference signal vector, No. l Carrier domain reference signal subvectors corresponding to a group of orthogonal frequency division multiplexing symbols For the first l In the group of orthogonal frequency division multiplexing symbols, the first Reference data for the position of the nth subcarrier, if the nth When the position of the first subcarrier is the pilot position, the reference data is the pilot value; if the first... When the subcarrier position is a non-pilot position, the reference data is zero.

4. The external radar target detection method based on a sparse model according to claim 3, characterized in that, In the carrier domain, the step of performing zero-frequency clutter suppression processing on the carrier domain monitoring signal using the carrier domain reference signal to obtain the clutter-suppressed carrier domain monitoring signal specifically includes: Extracting the first group of orthogonal frequency division multiplexing symbols from several groups The monitoring data of each subcarrier constitutes a monitoring data vector; Extract the first orthogonal frequency division multiplexing symbol from the corresponding group. The reference data of each subcarrier constitutes a reference data vector; Based on the monitoring data vector and the reference data vector, with the goal of minimizing the energy of the signal after clutter suppression, the adaptive cancellation coefficient is calculated using the following formula: ， in, For adaptive offset coefficient, , These are the monitoring data vector and the reference data vector, respectively. The extended cancellation algorithm is used to monitor the carrier domain signal. Each subcarrier is independently processed for clutter suppression to obtain the clutter-suppressed carrier domain monitoring signal. The calculation formula is as follows: ， in, For the first The carrier domain monitoring signal vector after subcarrier clutter suppression.

5. The external radar target detection method based on a sparse model according to claim 4, characterized in that, The steps of performing inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively to reconstruct new time-domain reference signals and new time-domain monitoring signals, and then performing piecewise distance correlation processing on the two to obtain range-domain data, specifically include: The clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal are respectively subjected to inverse discrete Fourier transform according to their signs, and a cyclic prefix is ​​added to reconstruct a new time-domain reference signal and a new time-domain monitoring signal. The new time-domain reference signal and the new time-domain monitoring signal are each divided into several segments with a preset segment length, as follows: ， ， in, t For continuous time variables, For continuous time variables t The time-domain reference signal below, For continuous time variables t The time-domain monitoring signal below, For the first b Each segment corresponds to a reference signal segment. For the first b Each segment corresponds to a monitoring signal segment. The total number of segments. It is a preset segment length. For a rectangular pulse function, when At that time, the rectangular pulse function is 1, and the rest are 0; Range compression is performed on the time-domain reference signal and time-domain monitoring signal corresponding to each segment to obtain range-domain data, represented as follows: ， in, For the first b Each segment experiences a time delay. The distance domain data afterwards This is the preset maximum detectable delay. For the first b The time-domain monitoring signal corresponding to each segment For the first b The time-domain reference signal corresponding to each segment is delayed. And take the signal component after complex conjugation, * is the complex conjugation operation.

6. The external radar target detection method based on a sparse model according to claim 5, characterized in that, The steps for establishing a sparse model of the range-Doppler spectrum based on the range domain data specifically include: Based on the velocity range and operating parameters of the target to be detected, the Doppler frequency search interval is determined. For each segment of range domain data, the Doppler frequency is used as the search variable to construct a segmented mutual ambiguity function, expressed as: ， in, For the first b A segmented mutual ambiguity function, The frequency is the Doppler frequency. The global mutual ambiguity function is obtained by weighted summation of all segmented mutual ambiguity functions according to their slow-time indices, and is expressed as: ， in, , For the first b Slow-time indexes for each segment; When the product of the segment length and the maximum Doppler frequency in the Doppler frequency search interval satisfies a preset small condition, the phase of the sampling point within the segment is approximated as the phase at the midpoint of the segment. The approximate piecewise mutual ambiguity function is obtained, expressed as: ， Introducing the approximate piecewise mutual ambiguity function into the global mutual ambiguity function yields the global approximate mutual ambiguity function, expressed as: ； Based on the global approximate mutual ambiguity function, the time delay and Doppler frequency are discretized to construct the Doppler shift Fourier transform matrix and the range domain echo matrix; Based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a sparse model of the range-Doppler spectrum is established.

7. The external radar target detection method based on a sparse model according to claim 6, characterized in that, Based on the aforementioned global approximate mutual ambiguity function, the steps of discretizing the time delay and Doppler frequency to construct the Doppler shift Fourier transform matrix and the range domain echo matrix specifically include: The global approximate mutual ambiguity function is split into a time delay discrete part and a Doppler frequency discrete part; Based on the aforementioned discrete delay portion, the delay interval [0, ... Discretize the data into a first preset number of grids, construct a range-domain echo matrix, and represent it as follows: ， in, For the first preset quantity, The range domain echo matrix, The matrix is ​​represented as OK A complex matrix of columns; The Doppler frequency search interval is discretized into a second preset number of grids, and based on the slow-time index and the discrete portion of the Doppler frequency, a Doppler shift Fourier transform matrix is ​​constructed, expressed as: ， Among them, the Doppler frequency range is , This is the Doppler shift Fourier transform matrix. For the second preset quantity, The matrix is ​​represented as OK, A complex matrix of columns, For the first Slow-time index for each segment.

8. The external radar target detection method based on a sparse model according to claim 7, characterized in that, The steps for establishing a sparse model of the range-Doppler spectrum based on the Doppler shift Fourier transform matrix and the range domain echo matrix specifically include: Based on the Doppler shift Fourier transform matrix and the range domain echo matrix, a discretized mutually fuzzy function matrix is ​​established, expressed as: ， in, Represents the discretized mutual fuzzy function matrix. The matrix is ​​represented as OK, A complex matrix of columns; By transposing the mutually fuzzy function matrix and combining it with the unitary property of the Doppler shift Fourier transform matrix, a sparse model is obtained, expressed as: ， in, This is the conjugate transpose of the Doppler shift Fourier transform matrix.

9. The external radar target detection method based on a sparse model according to claim 8, characterized in that, From the sparse model Due to the sparsity of the sparse model, the sparse model is transformed into an optimization problem, expressed as: ， in, Denotes the F-norm of a matrix. Denotes the L1 norm of a matrix. This is the preset error tolerance threshold; The optimization problem is solved by the orthogonal matching pursuit algorithm to obtain the target detection results.

10. A method system for detecting external radar targets based on a sparse model, characterized in that, The system is used to implement the external radar target detection method based on a sparse model as described in any one of claims 1 to 9, and the system comprises: The monitoring signal acquisition module is used to receive the time-domain monitoring signal from the external radar monitoring channel based on the orthogonal frequency division multiplexing waveform, and to perform a discrete Fourier transform on it to obtain the carrier domain monitoring signal. The reference signal acquisition module is used to extract data at the pilot position from the carrier domain monitoring signal and construct a carrier domain reference signal containing pilot information based on the known transmitter pilot information. The clutter suppression module is used to perform zero-frequency clutter suppression processing on the carrier domain monitoring signal in the carrier domain using the carrier domain reference signal, so as to obtain the clutter-suppressed carrier domain monitoring signal. The distance correlation processing module is used to perform inverse discrete Fourier transform on the clutter-suppressed carrier domain monitoring signal and the carrier domain reference signal respectively, reconstruct a new time domain reference signal and a new time domain monitoring signal, and perform segmented distance correlation processing on the two to obtain distance domain data; The sparse model solving module is used to establish a sparse model of the range-Doppler spectrum based on the range domain data and form a corresponding optimization problem, solve the optimization problem, and obtain the target detection results.

Images