Harmonic radar multi-target parallel detection method and system

Abstract

The application relates to the technical field of radar detection, and particularly discloses a harmonic radar multi-target parallel detection method and system, which comprises the following steps: through a receiver of a harmonic radar, superimposing echo signals reflected by multiple targets to form aliasing echo data in a preset detection period, and performing time-domain windowing and framing on the aliasing echo data to obtain time-domain echo frames; taking characteristic parameters of the time-domain echo frames as separation basis, performing blind source demixing on the aliasing echo data to obtain single-target echo signals; performing harmonic component analysis on the single-target echo signals to obtain harmonic characteristic vectors; associating and matching the harmonic characteristic vectors with preset detection space domain division information to determine detection wave positions corresponding to the harmonic characteristic vectors; synchronously calculating spatial position parameters of multiple targets relative to the harmonic radar to generate a multi-target parallel detection result; and the application can improve the multi-target parallel detection efficiency of the harmonic radar.

Description

Technical Field

[0001] This invention relates to the field of radar detection technology, and in particular to a method and system for parallel detection of multiple targets using harmonic radar. Background Technology

[0002] In multi-target detection scenarios, traditional harmonic radar suffers from overlapping echo signals due to simultaneous reflections from multiple targets. This makes it difficult for the receiver to directly distinguish the independent echo components of different targets, resulting in low signal separation efficiency and hindering effective decomposition and identification of multi-target signals. Existing detection methods often employ a target-by-target processing approach, making it difficult to simultaneously analyze multiple targets. This significantly reduces the overall detection response speed and fails to meet the real-time requirements of parallel multi-target detection.

[0003] Existing harmonic radar technology suffers from insufficient adaptability in the harmonic feature extraction and spatial wave position matching stages. It cannot accurately correlate and match the target's harmonic features with the corresponding detection space domain and wave position, which easily leads to target positioning errors. At the same time, the calculation process of target spatial position parameters lacks a unified calibration mechanism, resulting in low accuracy in the calculation of parameters such as distance and angle. Ultimately, this leads to insufficient accuracy in multi-target detection results and weak parallel detection capabilities. Summary of the Invention

[0004] In view of the above problems, the purpose of this invention is to provide a method and system for parallel detection of multiple targets by harmonic radar, so as to solve the problems of low detection efficiency, poor real-time performance and unreliable positioning results caused by the inability to effectively separate echo signals during multi-target detection by traditional harmonic radar, the inability to achieve parallel processing of multiple targets, the low accuracy of matching harmonic features with detection wave positions, and the inaccurate calculation of target spatial positions.

[0005] This invention provides a method for parallel detection of multiple targets using harmonic radar, comprising:

[0006] Step 1: Using the receiver of the harmonic radar, within a preset detection period, the echo signals reflected by multiple targets are superimposed to form aliased echo data, and the aliased echo data is windowed and framed in the time domain to obtain the time domain echo frame of the aliased echo data.

[0007] Step 2: Using the characteristic parameters of the time-domain echo frame as the separation basis, perform blind source demixing on the aliased echo data to obtain the single-target echo signal of the multi-target;

[0008] Step 3: Perform harmonic component analysis on the single-target echo signal to obtain the harmonic characteristic vector of the multi-target;

[0009] Step 4: Associate and match the harmonic feature vector with the preset detection space division information to determine the detection wave position corresponding to the harmonic feature vector;

[0010] Step 5: Based on the detected wave position and the harmonic feature vector, simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar, and generate the multi-target parallel detection results of the harmonic radar.

[0011] Preferably, in step 1, the receiver of the harmonic radar superimposes the echo signals reflected from multiple targets within a preset detection period to form aliased echo data, and performs time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data. The process is as follows:

[0012] Within a preset detection period, the receiver of the harmonic radar is controlled to turn on the receiving channel in sequence to receive echo signals reflected from multiple targets in different spatial directions.

[0013] The echo signals are time-domain aligned, and the aligned echo signals are superimposed on the same time base to obtain the aliased echo data of the multi-target;

[0014] The aliased echo data is sampled and quantized to obtain a discrete digital sequence of the aliased echo data;

[0015] Based on the detection period, the frame length and frame shift used for time-domain segmentation are determined, and based on the frame length and frame shift, the discrete digital sequence is segmented one by one to obtain the continuous overlapping signal segments of the discrete digital sequence.

[0016] Time-domain windowing is applied to the continuous overlapping signal segments to obtain the time-domain echo frame of the aliased echo data.

[0017] Preferably, in step 2, the process of using the characteristic parameters of the time-domain echo frame as the separation basis to perform blind source demixing on the aliased echo data to obtain the single-target echo signal of the multi-target system is as follows:

[0018] Based on the characteristic parameters of the time-domain echo frame, blind source separation initialization is performed on the aliased echo data to obtain the initial separation vector set of the aliased echo data;

[0019] The initial separation vector set is iteratively updated to obtain the optimized separation vector set of the aliased echo data;

[0020] Based on the optimized separation vector set, signal separation is performed on the aliased echo data to obtain the independent echo components of the aliased echo data.

[0021] The independent echo components are reconstructed in the time domain to obtain the single-target echo signal of the multi-target system.

[0022] Preferably, the process of performing signal separation on the aliased echo data based on the optimized separation vector set to obtain the independent echo components of the aliased echo data is as follows:

[0023] Based on the optimized separation vector set, the aliased echo data is separated and projected to obtain the projection component set of the aliased echo data.

[0024] The orthogonality of the projected component set is tested to obtain the orthogonality measure of the components in the projected component set.

[0025] Based on the orthogonality metric, the effective projection components of the aliased echo data in the projection component set are selected.

[0026] The effective projection components are subjected to time-domain waveform extraction to obtain the independent echo components of the aliased echo data.

[0027] Preferably, in step 3, the process of performing harmonic component analysis on the single-target echo signal to obtain the harmonic feature vector of the multi-target signal is as follows:

[0028] The single-target echo signal is subjected to spectral decomposition to obtain the spectral distribution of the single-target echo signal;

[0029] The fundamental frequency component of the spectrum distribution is obtained by performing fundamental frequency localization on the spectrum distribution.

[0030] Based on the fundamental frequency component, harmonic sequence extraction is performed on the spectral distribution to obtain the harmonic components of the single-target echo signal;

[0031] The harmonic components are characterized by characteristic parameters to obtain the harmonic characteristic vector of the multi-target.

[0032] Preferably, the process of characterizing the harmonic components using characteristic parameters to obtain the harmonic feature vector of the multi-target is as follows:

[0033] The harmonic components are normalized to obtain the normalized amplitude spectrum of the harmonic components.

[0034] Peak detection is performed on the normalized amplitude spectrum to obtain the main peak amplitude value of the harmonic component;

[0035] Based on the main peak amplitude value, the amplitude distribution relationship between the harmonic components is determined, and the amplitude feature vector of the multi-target is obtained;

[0036] Phase feature extraction is performed on the harmonic components to obtain the phase feature vector of the multi-target;

[0037] The amplitude values ​​of each harmonic in the amplitude feature vector and the phase values ​​of each harmonic in the phase feature vector are vector-fused to obtain the characteristic elements of the harmonic components. The fusion formula for the characteristic elements is as follows:

[0038] ;

[0039] In the formula, Indicates the first Characteristic elements of first harmonic components, Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, Represents the natural constant. Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, This represents the preset coupling coefficient. This indicates the total order of the harmonic components. Represents the cosine function. Represents the imaginary unit;

[0040] The characteristic elements of the harmonic components are arranged according to their order to obtain the harmonic characteristic vector of the multi-target.

[0041] Preferably, in step 4, the process of associating and matching the harmonic feature vector with preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector is as follows:

[0042] Wave position feature encoding is performed on the preset detection airspace division information to obtain the wave position feature template set of the detection wave position;

[0043] The wave position feature template set is subjected to feature space mapping to obtain the spatial mapping features of the wave position feature template set;

[0044] Based on the spatial mapping features, the harmonic feature vectors are correlated with similarity to obtain the correlation distribution between the harmonic feature vectors and the probe wave position.

[0045] Peak location is performed on the correlation distribution to determine the probe wave position corresponding to the harmonic feature vector.

[0046] Preferably, in step 5, the process of simultaneously calculating the spatial position parameters of the multiple targets relative to the harmonic radar based on the probe wave position and the harmonic feature vector, and generating the multi-target parallel detection results of the harmonic radar, is as follows:

[0047] Based on the detected wave position, the harmonic feature vector is subjected to wave position matching calibration to obtain the wave position calibration feature of the harmonic feature vector.

[0048] Spatial pointing analysis is performed on the probe wave position to obtain the spatial pointing parameters of the probe wave position;

[0049] Based on the spatial pointing parameters, the wave position calibration features are mapped by distance and angle to obtain the spatial mapping parameters of the multi-target;

[0050] The spatial mapping parameters are used to calculate the target positions to obtain the spatial position parameters of the multiple targets;

[0051] The detection results of the spatial position parameters are integrated to generate the multi-target parallel detection results of the harmonic radar.

[0052] Preferably, the process of calculating the target position of the spatial mapping parameters to obtain the spatial position parameters of the multiple targets is as follows:

[0053] Extract the distance mapping values ​​and angle mapping values ​​corresponding to the multiple targets from the spatial mapping parameters;

[0054] The distance mapping value is corrected for propagation delay to obtain the radial distance estimate of the multiple targets;

[0055] Beam pointing deviation compensation is performed on the angle mapping value to obtain the spatial azimuth estimate of the multi-target;

[0056] Based on the estimated radial distance and the estimated spatial azimuth, the spatial coordinates of the multiple targets are determined using polar coordinate positioning relationships. The spatial coordinates of the multiple targets are determined using the following formula:

[0057] ;

[0058] In the formula, Indicates the first The spatial coordinates of the target This represents the mapping relationship from polar coordinate space to rectangular coordinate space. Indicates the first Radial distance estimates for each target. Indicates the first Estimated spatial azimuth angles of each target;

[0059] The spatial coordinates of the multiple targets are correlated and integrated to generate the spatial position parameters of the multiple targets.

[0060] The present invention also provides a harmonic radar multi-target parallel detection system, the system comprising:

[0061] The echo acquisition and framing module is used to superimpose the echo signals reflected by multiple targets to form aliased echo data within a preset detection period through the receiver of the harmonic radar, and to perform time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data.

[0062] The blind source signal separation module is used to perform blind source demixing on the aliased echo data based on the characteristic parameters of the time-domain echo frame to obtain the single-target echo signal of the multi-target;

[0063] The harmonic feature extraction module is used to perform harmonic component analysis on the single-target echo signal to obtain the harmonic feature vector of the multi-target.

[0064] The wave position association matching module is used to associate and match the harmonic feature vector with the preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector.

[0065] The target localization and calculation module is used to simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar based on the detection wave position and the harmonic feature vector, and generate the multi-target parallel detection results of the harmonic radar.

[0066] As can be seen from the above technical solution, the harmonic radar multi-target parallel detection method and system provided by the present invention can accurately separate multi-target single echo signals by performing time-domain windowing and framing and blind source demixing on aliased echo data. Combined with harmonic component analysis and wave position correlation matching, it can achieve accurate extraction of multi-target harmonic features and efficient matching of spatial wave positions, and simultaneously complete the calculation of multi-target spatial position parameters, significantly improving the efficiency and real-time performance of harmonic radar multi-target parallel detection, while improving target positioning accuracy and detection result reliability. Attached Figure Description

[0067] Other objects and results of the invention will become more apparent and readily understood by referring to the following description taken in conjunction with the accompanying drawings, and with a more complete understanding of the invention. In the drawings:

[0068] Figure 1 This is a flowchart illustrating a multi-target parallel detection method for harmonic radar according to an embodiment of the present invention.

[0069] Figure 2 This is a functional block diagram of a harmonic radar multi-target parallel detection system according to an embodiment of the present invention; Detailed Implementation

[0070] Traditional harmonic radar suffers from multiple targets, resulting in aliasing of echo signals, difficulty in effective separation, inability to process multiple targets in parallel, low accuracy in matching harmonic features with the detection wave position, and inaccurate calculation of target spatial position. Consequently, it leads to low detection efficiency, poor real-time performance, and unreliable positioning results.

[0071] To address the aforementioned problems, this invention provides a method and system for parallel detection of multiple targets using harmonic radar. Specific embodiments of this invention will be described in detail below with reference to the accompanying drawings.

[0072] To illustrate the harmonic radar multi-target parallel detection method and system provided by this invention Figure 1 An exemplary illustration is provided for a harmonic radar multi-target parallel detection method according to an embodiment of the present invention; Figure 2 An exemplary illustration is provided for a harmonic radar multi-target parallel detection system according to an embodiment of the present invention.

[0073] The following description of exemplary embodiments is merely illustrative and is in no way intended to limit the invention or its application or use. Techniques and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques and equipment should be considered part of the specification.

[0074] Reference Figure 1 The diagram shown is a flowchart illustrating a method for parallel multi-target detection using harmonic radar according to an embodiment of the present invention. In this embodiment, the method includes:

[0075] Step 1: Using the receiver of the harmonic radar, within a preset detection period, the echo signals reflected by multiple targets are superimposed to form aliased echo data, and the aliased echo data is windowed and framed in the time domain to obtain the time domain echo frame of the aliased echo data.

[0076] In this embodiment of the invention, in step 1, the receiver of the harmonic radar superimposes the echo signals reflected from multiple targets within a preset detection period to form aliased echo data, and performs time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data. The process is as follows:

[0077] Within a preset detection period, the receiver of the harmonic radar is controlled to turn on the receiving channel in sequence to receive echo signals reflected from multiple targets in different spatial directions.

[0078] The echo signals are time-domain aligned, and the aligned echo signals are superimposed on the same time base to obtain the aliased echo data of the multi-target;

[0079] The aliased echo data is sampled and quantized to obtain a discrete digital sequence of the aliased echo data;

[0080] Based on the detection period, the frame length and frame shift used for time-domain segmentation are determined, and based on the frame length and frame shift, the discrete digital sequence is segmented one by one to obtain the continuous overlapping signal segments of the discrete digital sequence.

[0081] Time-domain windowing is applied to the continuous overlapping signal segments to obtain the time-domain echo frame of the aliased echo data.

[0082] The receiver of the harmonic radar operates according to a strict time plan throughout the entire preset detection period. This plan specifies the opening time and duration of each receiving channel. The receiver opens each receiving channel sequentially according to the plan. Each opened channel only receives signals from a specific spatial direction pointed to by its antenna, thereby collecting harmonic echo signals reflected back from targets in that direction separately within that time period. Finally, at the end of the detection period, the receiver completes the acquisition of signals from all predetermined spatial directions and obtains a set of original echo signals from multiple targets in different directions.

[0083] The acquired raw echo signals are aligned in the time domain. This alignment is achieved by identifying and compensating for the relative time delay between the signals. Specifically, a reference channel signal is selected as the baseline, and the time offset of each other channel signal relative to this baseline signal is calculated. Then, the entire signal waveform of each non-reference channel is shifted along the time axis by an amount equal to the calculated time offset, so that the points representing the reflection event at the same moment in all channels are aligned on the time axis. After that, the aligned signal waveforms of all channels are added one by one according to the time points, that is, the amplitude values ​​of all channels at the same moment are added together to generate a single composite signal containing target information from all directions. This composite signal is the multi-target aliased echo data.

[0084] The obtained analog aliased echo data is sampled and quantized to convert it into digital form. This process is completed by the analog-to-digital converter (ADC) hardware circuit. The ADC measures the input echo voltage signal at a fixed and extremely high clock frequency, and records the instantaneous voltage value of the signal at the moment each clock pulse arrives. This measurement is called sampling. Then, the ADC maps this continuous voltage value to the nearest discrete digital level. This mapping is called quantization. After such continuous sampling and quantization operations, the original continuous waveform is converted into a long string of numbers arranged in chronological order. This string of numbers is the discrete digital sequence of the aliased echo data.

[0085] The frame length and frame shift used for time-domain segmentation are determined based on the total duration of the entire detection cycle. The frame length is directly set to be equal to the duration of the detection cycle to ensure that each frame of data contains a complete detection cycle. The frame shift is set to a fixed time value that is slightly smaller than the frame length to ensure that there is some overlap between frames. These two parameters are used to segment the discrete digital sequence. When segmenting, starting from the first number of the sequence, a subsequence is taken out with a number equal to the frame length divided by the total number of numbers corresponding to the sampling time interval. Then the starting point is moved backward by the number of numbers corresponding to the frame shift time length, and another subsequence of the same length is taken out. This process is repeated until the end of the sequence. The series of short digital subsequences obtained in this way are the continuous overlapping signal segments of the discrete digital sequence.

[0086] For each consecutive overlapping signal segment, time-domain windowing is performed. Time-domain windowing is accomplished by multiplying a pre-designed window function coefficient sequence with the signal segment point by point. The length of the window function coefficient sequence is exactly the same as the number of numbers contained in the signal segment. Each coefficient represents a weight value between zero and one. Multiplying each number in the signal segment with the coefficient at the corresponding position in the window function sequence produces a new number. This operation effectively adjusts the amplitude of the data at both ends of the signal segment while keeping the middle part unchanged. After weighted calculation of all points, the original signal segment is modified into a new signal segment with both ends smoothly transitioning to zero. This new signal segment is the time-domain echo frame of the final processed aliased echo data.

[0087] Step 2: Using the characteristic parameters of the time-domain echo frame as the separation basis, perform blind source demixing on the aliased echo data to obtain the single-target echo signal of the multi-target;

[0088] In this embodiment of the invention, in step 2, the aliased echo data is blind-source demixed using the characteristic parameters of the time-domain echo frame as the separation basis to obtain the single-target echo signal of the multi-target system. The process is as follows:

[0089] Based on the characteristic parameters of the time-domain echo frame, blind source separation initialization is performed on the aliased echo data to obtain the initial separation vector set of the aliased echo data;

[0090] The initial separation vector set is iteratively updated to obtain the optimized separation vector set of the aliased echo data;

[0091] Based on the optimized separation vector set, signal separation is performed on the aliased echo data to obtain the independent echo components of the aliased echo data.

[0092] The independent echo components are reconstructed in the time domain to obtain the single-target echo signal of the multi-target system.

[0093] Based on the optimized separation vector set, signal separation is performed on the aliased echo data to obtain the independent echo components of the aliased echo data. The process is as follows:

[0094] Based on the optimized separation vector set, the aliased echo data is separated and projected to obtain the projection component set of the aliased echo data.

[0095] The orthogonality of the projected component set is tested to obtain the orthogonality measure of the components in the projected component set.

[0096] Based on the orthogonality metric, the effective projection components of the aliased echo data in the projection component set are selected.

[0097] The effective projection components are subjected to time-domain waveform extraction to obtain the independent echo components of the aliased echo data.

[0098] Blind source separation initialization of aliased echo data is performed based on the statistical characteristic parameters of time-domain echo frames. Principal component analysis is used to obtain the initial separation vector set. The specific implementation process is as follows: first, the sampling point data of all time-domain echo frames are collected, and the covariance of these data points in different time dimensions is calculated to form a covariance matrix. Then, eigenvalue decomposition is performed on this matrix. Eigenvalue decomposition is to obtain its eigenvalues ​​and corresponding eigenvectors by solving the characteristic equation of the matrix. Finally, these calculated eigenvectors are arranged in order of their corresponding eigenvalues. This set of arranged eigenvectors is the initial separation vector set of the required aliased echo data.

[0099] The initial separation vector set is iteratively updated to obtain the optimized separation vector set. This process uses a fixed-point iterative algorithm. Starting from the initial separation vector set, in each iteration, the algorithm uses the current separation vector to perform a linear projection on the aliased echo data, and then calculates the higher-order statistics of the projected signal, usually a fourth-order cumulant. Then, the direction of the current vector is adjusted according to the magnitude and sign of this higher-order statistic, thereby generating a new, better-estimated separation vector, which replaces the current vector. This adjustment and replacement step is repeated until the angle between the newly generated vector and the previous vector changes so little that it can be ignored. At this point, the iteration terminates, and the final stable and convergent vector set is the optimized separation vector set of the aliased echo data.

[0100] Based on the optimized separation vector set, the aliased echo data is separated and projected. This step is achieved through vector space projection calculation. Specifically, the complete aliased echo data sequence is treated as a multidimensional time series. Then, each vector in the optimized separation vector set is extracted sequentially, and the entire aliased echo data sequence is multiplied by the vector. The dot product operation is to multiply the multidimensional data point at each time point in the data sequence with the corresponding vector and then sum them to obtain a scalar time series describing the intensity change of the data in the direction of the vector. This dot product operation is repeated for all vectors in the optimized separation vector set to obtain a set of scalar time series. This set of sequences together constitutes the projection component set of the aliased echo data.

[0101] Orthogonality tests are performed on the projected component sets to obtain orthogonality measures. This is accomplished by calculating a cross-correlation matrix. Specifically, for any two different scalar time series in the projected component sets, one series is fixed, and the other series slides point by point on the time axis. At each sliding position, the sum of the corresponding product of the overlapping parts of the two series is calculated. After traversing all possible sliding positions, the maximum value of the product sum is found. This maximum value is defined as the cross-correlation coefficient between the two component series. The above calculation is performed on all possible combinations of series pairs to obtain a matrix that records the cross-correlation coefficients between all component pairs. All coefficient values ​​of the upper or lower triangular parts extracted from this matrix constitute the orthogonality measure set of the components in the projected component set.

[0102] The effective projection components are filtered based on the obtained set of orthogonality metrics. The filtering process sets a specific orthogonality threshold that is close to zero. Then, the cross-correlation coefficient of each projection component with all other components is checked, i.e., the orthogonality metric value. If the absolute value of the cross-correlation coefficient of a projection component with every other component is lower than the set orthogonality threshold, it is determined that the component is highly uncorrelated with other components, i.e., approximately orthogonal, and is retained. Conversely, if the cross-correlation coefficient of a projection component with any other component is higher than the threshold, it is removed. Through this method of comparison and judgment, the set of all retained projection components is the effective projection component of the aliased echo data in the projection component set.

[0103] Independent echo components are obtained by extracting time-domain waveforms from the effective projection components. The extraction process is divided into two stages: peak detection and waveform integration. In the peak detection stage, the amplitude time series of each effective projection component is analyzed, an amplitude threshold based on the global noise level is set, and all local maxima points in the sequence whose amplitudes exceed the threshold are identified and marked as effective peak points. In the waveform integration stage, a waveform segment of a fixed time length is extracted from the original amplitude sequence with each detected peak point as the center. Then, those waveform segments that appear continuously in time and whose intervals are less than a specific value are determined to belong to the same radiation source, and these segments are smoothly connected end to end in chronological order to form a complete and continuous waveform. The series of waveforms obtained after processing each effective projection component in this way are the independent echo components of the aliased echo data.

[0104] Time-domain waveform reconstruction is performed on the independent echo components to obtain the single-target echo signal. Reconstruction is the process of restoring each independent echo component to the original signal space. Specifically, each independent echo component is treated as an amplitude modulation sequence and multiplied by the optimized separation vector that generated the component. This multiplication multiplies the scalar value of each moment in the amplitude sequence by the value of the entire separation vector, thereby generating a multidimensional time series segment. This sequence segment represents the contribution of the target to each original receiving channel. Then, these multidimensional segments at all time points are arranged in chronological order and summed. If there are multiple independent components corresponding to the same physical target, they are added first. The final summation results in a complete multidimensional time series, which is the single-target echo signal recovered after separation, corresponding to the single physical target before aliasing.

[0105] Step 3: Perform harmonic component analysis on the single-target echo signal to obtain the harmonic characteristic vector of the multi-target;

[0106] In this embodiment of the invention, step 3, which involves performing harmonic component analysis on the single-target echo signal to obtain the harmonic feature vector of the multi-target signal, is as follows:

[0107] The single-target echo signal is subjected to spectral decomposition to obtain the spectral distribution of the single-target echo signal;

[0108] The fundamental frequency component of the spectrum distribution is obtained by performing fundamental frequency localization on the spectrum distribution.

[0109] Based on the fundamental frequency component, harmonic sequence extraction is performed on the spectral distribution to obtain the harmonic components of the single-target echo signal;

[0110] The harmonic components are characterized by characteristic parameters to obtain the harmonic characteristic vector of the multi-target.

[0111] The process of characterizing the harmonic components using characteristic parameters to obtain the harmonic feature vector of the multi-target is as follows:

[0112] The harmonic components are normalized to obtain the normalized amplitude spectrum of the harmonic components.

[0113] Peak detection is performed on the normalized amplitude spectrum to obtain the main peak amplitude value of the harmonic component;

[0114] Based on the main peak amplitude value, the amplitude distribution relationship between the harmonic components is determined, and the amplitude feature vector of the multi-target is obtained;

[0115] Phase feature extraction is performed on the harmonic components to obtain the phase feature vector of the multi-target;

[0116] The amplitude values ​​of each harmonic in the amplitude feature vector and the phase values ​​of each harmonic in the phase feature vector are vector-fused to obtain the characteristic elements of the harmonic components. The fusion formula for the characteristic elements is as follows:

[0117] ;

[0118] In the formula, Indicates the first Characteristic elements of first harmonic components, Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, Represents the natural constant. Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, This represents the preset coupling coefficient. This indicates the total order of the harmonic components. Represents the cosine function. Represents the imaginary unit;

[0119] The characteristic elements of the harmonic components are arranged according to their order to obtain the harmonic characteristic vector of the multi-target.

[0120] To obtain the spectral distribution of a single-target echo signal, spectral decomposition is performed. This process is achieved through Discrete Fourier Transform (DFT). Specifically, each sampling point of the discrete-time sequence of the single-target echo signal is sequentially extracted. Simultaneously, a set of complex exponential basis functions, equal in number to the total number of sampling points, is prepared. Each basis function corresponds to a specific discrete frequency, increasing at uniform frequency intervals from zero until approaching half the sampling frequency. For each specific frequency point to be calculated, all discrete values ​​of the corresponding complex exponential basis function on the time axis are generated. These values ​​are complex cosine and sine function values. Then, the real value of each sampling point in the signal sequence is multiplied by the complex number corresponding to the basis function at that point. This multiplication operation calculates the product of the real part with the real part and the imaginary part with the real part. Then, the product results of all sampling points are accumulated in the complex field, that is, the real parts of all results are added to obtain the sum of the real parts, and the imaginary parts of all results are added to obtain the sum of the imaginary parts. Finally, a complex number corresponding to that frequency point is obtained. The above point-by-point multiplication and complex number accumulation operation is performed completely for each preset discrete frequency point. The final sequence composed of the complex numbers of all frequency points is the spectral distribution of the single target echo signal.

[0121] To obtain the fundamental frequency component, the fundamental frequency component is located in the amplitude spectrum of the spectrum distribution. This method involves finding the frequency corresponding to the maximum peak in the amplitude spectrum of the spectrum distribution. Specifically, the amplitude of the complex number corresponding to each discrete frequency point in the spectrum distribution is calculated by taking the square of the real part of the complex number and the square of the imaginary part, and then taking the square root of the sum. This yields a sequence of amplitude values ​​that correspond one-to-one with the frequency points. Then, the DC component, i.e., the amplitude corresponding to zero frequency, is excluded from this sequence. Starting from the first positive frequency point, the amplitude value of each frequency point is compared with the previously recorded maximum amplitude value. If the amplitude value of the current point is greater than the recorded maximum value, the recorded maximum value is updated with the current amplitude value, and the frequency value corresponding to the current point is recorded. This process continues until the preset highest analysis frequency point is reached. The frequency value that is finally recorded is the fundamental frequency component of the spectrum distribution.

[0122] Harmonic sequence extraction is performed on the spectral distribution based on the obtained fundamental frequency components. Harmonic sequence extraction is accomplished by calculating integer multiples of the fundamental frequency and extracting the spectral values ​​at the corresponding positions. Specifically, the value of the fundamental frequency component is used as the reference interval, and its value is calculated sequentially as multiples of two, three, and up to the pre-set highest harmonic order multiples to obtain a series of theoretical values ​​of harmonic frequencies. Then, in the discrete frequency point sequence of the spectral distribution, the nearest discrete frequency point index is found for each calculated theoretical harmonic frequency value. The search method is to compare the difference between the theoretical frequency value and each discrete frequency point, and select the index with the smallest absolute value of the difference. The complex number value at the corresponding position is extracted from the spectral distribution using this index. This complex number contains two parts: a real part and an imaginary part. This search and extraction operation is performed on all orders of harmonic frequencies. The extracted complex numbers of each order are arranged into a list in order of harmonic order from low to high. This list is the harmonic components of the single-target echo signal.

[0123] Amplitude normalization of harmonic components yields a normalized amplitude spectrum. Amplitude normalization involves dividing the amplitude value of each harmonic by its maximum value. Specifically, the amplitude value of the complex number corresponding to each harmonic in the harmonic component list is first calculated by summing the squares of the real and imaginary parts of the complex number, and then taking the square root of the sum to obtain the original amplitude value of each harmonic. Next, all these original amplitude values ​​are iterated through, and the maximum value is found by comparing them one by one. Then, the original amplitude value of each harmonic is divided by this maximum value. This division operation uses the original amplitude value as the dividend and the maximum value as the divisor to calculate the quotient. All the calculated quotients form a new sequence in the range of zero to one. This new sequence is the normalized amplitude spectrum of the harmonic components.

[0124] Peak detection is performed on the normalized amplitude spectrum to obtain the main peak amplitude value. Peak detection identifies local maxima in the normalized amplitude spectrum sequence and finds the maximum value. Specifically, it sequentially traverses each amplitude value from the first element to the last element in the normalized amplitude spectrum. For an amplitude value in the middle of the sequence, it reads the values ​​of the preceding and following elements simultaneously. It compares the current value with both the preceding and following values. If the current value is strictly greater than both the preceding and following values, it is marked as a peak point, and its amplitude value is recorded. After completely traversing the entire sequence and identifying all peak points, all recorded peak point amplitude values ​​are compared together. By comparing them one by one, the largest amplitude value is found. This largest amplitude value is the main peak amplitude value of the harmonic component.

[0125] The amplitude characteristic vector is obtained by determining the amplitude distribution relationship between harmonic components based on the amplitude value of the main peak. The relationship is determined by calculating the ratio of the normalized amplitude of each harmonic to the amplitude of the main peak. Specifically, the normalized amplitude value corresponding to each harmonic in the normalized amplitude spectrum is read sequentially, and this value is divided by the amplitude value of the main peak. That is, the normalized amplitude value of the harmonic is used as the dividend, and the amplitude value of the main peak is used as the divisor to calculate a specific ratio. This ratio reflects the proportional relationship between the relative intensity of the harmonic and the strongest peak value. Then, the calculated ratios of each harmonic are arranged in order from low to high harmonic order to form a new ordered sequence. This ordered sequence is the amplitude characteristic vector of the multi-objective harmonic.

[0126] Phase feature extraction is performed on harmonic components to obtain phase feature vectors. Phase feature extraction involves calculating the phase angles corresponding to each order of harmonic complex numbers. Specifically, each complex number in the harmonic component list is processed sequentially. For each complex number, the values ​​of its imaginary and real parts are extracted, and the quotient of the imaginary part value divided by the real part value is calculated. Then, the four-quadrant arctangent function is calculated on this quotient value. This calculation is achieved by looking up a pre-stored function table that maps tangent values ​​to angle values. The input is the ratio of the imaginary part to the real part, and the output is a phase angle value within a specific angle interval. The above division and table lookup operations are performed on each order of harmonic to obtain its phase value. Then, the phase values ​​of all orders of harmonics are arranged in order from the first fundamental wave to the highest order harmonic into a list. This list is the multi-objective phase feature vector.

[0127] The characteristic elements are obtained by vector fusion of the amplitude values ​​of each harmonic in the amplitude feature vector and the phase values ​​of each harmonic in the phase feature vector. Vector fusion combines the amplitude ratio and phase value of the same order harmonic into an ordered pair. Specifically, starting from the lowest order harmonic, the amplitude ratio value corresponding to that order harmonic in the amplitude feature vector is read, and the phase angle value corresponding to that order harmonic in the phase feature vector is read. These two values ​​are combined as a data pair, usually with the amplitude ratio as the first element and the phase angle as the second element, forming the characteristic element of that order harmonic. Then, the same reading and combination operation is performed on the next order harmonic until all orders of harmonics have been processed. The final set of a series of data pairs ordered by order is the characteristic element of the harmonic component.

[0128] In the characteristic element fusion formula of harmonic components, the parameters Sourced from the first Measurement or calculation of the amplitude of first harmonic components; parameters Sourced from the first Measurement or calculation of the phase value of the first harmonic component; parameters These are natural constants defined in mathematics, and are fixed constants; parameters Sourced from the first Measurement or calculation of the amplitude of first harmonic components; parameters Sourced from the first Measurement or calculation of the phase value of the first harmonic component; parameters Parameters are either preset externally or determined according to system requirements. The parameters are obtained by statistically analyzing the total number of harmonic components. It is the imaginary unit defined in mathematics, and is a fixed constant.

[0129] The formula is used to calculate the first The characteristic elements of the first harmonic component are calculated by using the logic of the first harmonic component. The amplitude and phase reproduction term of the first harmonic component itself Multiply by a term that includes the effects of coupling with other harmonics, where "1" represents the fundamental characteristic of the property when there is no coupling. Used to control the degree of coupling effect, the summation operation is applied to the case of the first digit, excluding the second digit. All other harmonics outside the first order By calculating the amplitudes of other harmonics and the first harmonic... The ratio of the amplitudes of the first harmonics and the cosine value of the phase difference And sum them up to obtain the other harmonic pairs. The sum of the effects of first-order harmonic coupling is ultimately multiplied by its own recurrence term as the first-order harmonic coupling effect. Characteristic elements of first harmonic components.

[0130] When the preset coupling coefficient At that time, the coupling effect term disappears, and the first... The characteristic elements of a first harmonic component are determined solely by its amplitude value. and phase value Decision, manifested as At this time, there is no coupling effect from other harmonics; when As the value increases, the degree of control over the coupling effect strengthens, and the overall value including the coupling term increases. The characteristic elements of a first harmonic component are amplified by the influence of other harmonics; if a certain first harmonic component... Amplitude value of the first harmonic component Increase, all other things being equal. As the value increases, if the phase difference Corresponding cosine value If the value is positive, the contribution of the coupling term to the characteristic element increases; conversely, if the value is negative, the contribution decreases. When the total order of the harmonic components... As the value increases, the range of harmonic orders involved in the coupling calculation expands. If there is a newly added [number of harmonic orders]... For the first harmonic, the summation term of the coupling terms may increase, thereby increasing the first harmonic. The characteristic elements of the first harmonic component are affected by the coupling of more harmonics; when the phase difference When the cosine value changes The sign and magnitude of change accordingly, directly affecting the contribution of the coupling term to the feature element, and thus causing the th The characteristic elements of the first harmonic component change.

[0131] The harmonic eigenvector is obtained by arranging the characteristic elements of the harmonic components according to their order. This process involves sequentially integrating all characteristic element data pairs into a unified vector structure. Specifically, it starts with the characteristic element data pair of the first harmonic, adding it as the first element to an initially empty list structure. Then, in ascending order, the characteristic elements of the second harmonic, the third harmonic, and so on, are appended to the end of the list, ultimately forming a complete list. Each element in this list is itself a data pair containing two values, representing the amplitude ratio and phase angle of the corresponding harmonic order, respectively. This final, ordered list structure is the multi-objective harmonic eigenvector.

[0132] Step 4: Associate and match the harmonic feature vector with the preset detection space division information to determine the detection wave position corresponding to the harmonic feature vector;

[0133] In this embodiment of the invention, in step 4, the process of associating and matching the harmonic feature vector with preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector is as follows:

[0134] Wave position feature encoding is performed on the preset detection airspace division information to obtain the wave position feature template set of the detection wave position;

[0135] The wave position feature template set is subjected to feature space mapping to obtain the spatial mapping features of the wave position feature template set;

[0136] Based on the spatial mapping features, the harmonic feature vectors are correlated with similarity to obtain the correlation distribution between the harmonic feature vectors and the probe wave position.

[0137] Peak location is performed on the correlation distribution to determine the probe wave position corresponding to the harmonic feature vector.

[0138] The pre-defined detection airspace division information is encoded using wave position features to obtain a wave position feature template set. The detection airspace division information includes the number of each detection wave position, the beam center pointing angle, and the beamwidth. The encoding process generates a feature vector for each wave position independently. This vector consists of three parts of numerical values ​​connected in sequence. The first part is the real value converted from the wave position number itself. The second part is the four values ​​obtained by transforming the azimuth and elevation angles of the beam center pointing angle using sine and cosine functions. The third part is the single value converted from the beamwidth value. The above encoding operation is performed on all wave positions one by one. The feature vectors generated for each wave position are arranged into a set according to the wave position number. This set is the wave position feature template set of the detection wave positions.

[0139] To obtain spatially mapped features, a principal component analysis (PCA) method is used to perform feature space mapping on the wave position feature template set. First, the average vector of all eigenvectors in the wave position feature template set is calculated. Then, this average vector is subtracted from each eigenvector to obtain a centered vector. Next, the covariance matrix of all centered vectors is calculated, and eigenvalues ​​are obtained by eigenvalue decomposition. The eigenvectors are arranged in descending order of their corresponding eigenvalues, and the eigenvectors corresponding to the first few largest eigenvalues ​​are selected as the projection principal axes. Finally, each centered eigenvector in the wave position feature template set is subjected to a dot product operation with these selected projection principal axes. Each dot product result yields a scalar. All dot product results are arranged in order to form a new low-dimensional vector. This new vector is the spatially mapped feature of the wave position template. The set of spatially mapped features of all wave position templates constitutes the spatially mapped feature of the wave position feature template set.

[0140] The similarity association of harmonic feature vectors based on spatial mapping features yields the correlation distribution. This similarity association is achieved by calculating the cosine similarity between the harmonic feature vector and the spatial mapping feature of each wave position. Specifically, the harmonic feature vector is first mapped to a low-dimensional space identical to the spatial mapping feature by subtracting the same average vector from the wave position feature template set and performing the same projection principal axis dot product operation. Then, for each wave position's spatial mapping feature, its dot product with the mapped harmonic feature vector is calculated. Simultaneously, the magnitude of the spatial mapping feature and the magnitude of the mapped harmonic feature vector are calculated. The dot product is the sum of the product of corresponding values ​​of the two vectors, and the magnitude is the square root of the sum of the squares of the vector values. Finally, the dot product is divided by the product of the two magnitudes, and the quotient is the correlation degree between the wave position and the harmonic feature vector. This calculation is repeated for all wave positions, and the resulting sequence of correlation values ​​arranged in wave position number order is the correlation distribution between the harmonic feature vector and the probe wave position.

[0141] Peak location of the correlation distribution is used to determine the probe position corresponding to the harmonic feature vector. This process involves finding the maximum value and its position in the correlation distribution. Specifically, it involves sequentially traversing each correlation value in the correlation distribution sequence, initializing a maximum value to a very small number and recording its index. Starting from the correlation value corresponding to the first probe position, the current correlation value is compared with the recorded maximum value. If the current value is greater than the recorded maximum value, the recorded maximum value is updated with the current value, and the recorded index is updated with the current probe position number. The comparison continues with the correlation value corresponding to the next probe position until all values ​​in the sequence have been traversed. Finally, the probe position number pointed to by the index is recorded. This probe position is the probe position where the peak value in the correlation distribution is located, and this probe position is determined as the probe position corresponding to the harmonic feature vector.

[0142] Step 5: Based on the detected wave position and the harmonic feature vector, simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar, and generate the multi-target parallel detection results of the harmonic radar.

[0143] In this embodiment of the invention, in step 5, the spatial position parameters of the multiple targets relative to the harmonic radar are calculated simultaneously based on the detection wave position and the harmonic feature vector to generate the multi-target parallel detection results of the harmonic radar. The process is as follows:

[0144] Based on the detected wave position, the harmonic feature vector is subjected to wave position matching calibration to obtain the wave position calibration feature of the harmonic feature vector.

[0145] Spatial pointing analysis is performed on the probe wave position to obtain the spatial pointing parameters of the probe wave position;

[0146] Based on the spatial pointing parameters, the wave position calibration features are mapped by distance and angle to obtain the spatial mapping parameters of the multi-target;

[0147] The spatial mapping parameters are used to calculate the target positions to obtain the spatial position parameters of the multiple targets;

[0148] The detection results of the spatial position parameters are integrated to generate the multi-target parallel detection results of the harmonic radar.

[0149] The process of calculating the target position of the spatial mapping parameters to obtain the spatial position parameters of the multiple targets is as follows:

[0150] Extract the distance mapping values ​​and angle mapping values ​​corresponding to the multiple targets from the spatial mapping parameters;

[0151] The distance mapping value is corrected for propagation delay to obtain the radial distance estimate of the multiple targets;

[0152] Beam pointing deviation compensation is performed on the angle mapping value to obtain the spatial azimuth estimate of the multi-target;

[0153] Based on the estimated radial distance and the estimated spatial azimuth, the spatial coordinates of the multiple targets are determined using polar coordinate positioning relationships. The spatial coordinates of the multiple targets are determined using the following formula:

[0154] ;

[0155] In the formula, Indicates the first The spatial coordinates of the target This represents the mapping relationship from polar coordinate space to rectangular coordinate space. Indicates the first Radial distance estimates for each target. Indicates the first Estimated spatial azimuth angles of each target;

[0156] The spatial coordinates of the multiple targets are correlated and integrated to generate the spatial position parameters of the multiple targets.

[0157] Wave position matching calibration is performed on the harmonic feature vector based on the probe wave position to obtain the wave position calibration feature. This calibration is achieved by retrieving the feature template of the corresponding wave position from the wave position feature template set and performing a difference operation with the harmonic feature vector. Specifically, the determined probe wave position number is used as an index to find the feature template vector with the same number in the wave position feature template set. Then, the harmonic feature vector and the found wave position feature template vector are subtracted at corresponding positions. That is, the value of each dimension in the harmonic feature vector is subtracted from the value of the same dimension in the wave position feature template vector. The differences of all dimensions are arranged in the original order to form a new vector. This new vector is the wave position calibration feature of the harmonic feature vector.

[0158] Spatial pointing parameters are obtained by analyzing the spatial pointing of the probe wave position. This analysis is performed by querying a preset probe airspace division information table. This information table records the spatial pointing angle and beamwidth of the beam center corresponding to each wave position number. Specifically, the probe wave position number is used as the query keyword to perform an exact match search in this information table. The data row corresponding to the number is found, and the azimuth angle and elevation angle values ​​of the beam center are extracted from the row. At the same time, the beamwidth value is extracted. These two angle values ​​and the beamwidth value are combined in a fixed order, such as azimuth angle, elevation angle, and beamwidth, to form a three-dimensional parameter group. This parameter group is the spatial pointing parameter of the probe wave position.

[0159] Spatial mapping parameters are obtained by mapping the wave position calibration features to range and angle based on the spatial pointing parameters. This mapping is accomplished through a pre-trained linear mapping matrix, which defines the transformation relationship from calibration features to range and angle parameters. Specifically, the wave position calibration feature vector is multiplied by this linear mapping matrix. The multiplication is the dot product operation between each row of the matrix and the vector. Each row of the matrix corresponds to an output parameter. By calculating the weighted sum of all elements of the calibration feature vector, scalar values ​​related to range and angle are obtained respectively. Typically, a new vector containing elements such as range, azimuth offset, and pitch offset is output. This new vector is the spatial mapping parameter for the multi-target system.

[0160] The extraction process involves directly reading elements from a specific position in the spatial mapping parameter vector. Specifically, the spatial mapping parameter vector is an ordered sequence where the first element is defined as the distance mapping value, the second element as the azimuth angle mapping value, and the third element as the pitch angle mapping value. During extraction, the first value is read from the beginning of the vector, which is the distance mapping value. Then, the second and third values ​​are read, and these two values ​​together constitute the angle mapping value.

[0161] The radial range estimate is obtained by correcting the propagation delay of the range mapping value. This correction is achieved by adding a fixed delay deviation caused by the propagation speed of the medium. Specifically, the distance value corresponding to the fixed delay deviation is calculated in advance by calibration measurement of the radar system at a known distance. This distance value is a constant. Then, the obtained range mapping value is directly added to this pre-calibrated constant distance value. The result of the addition operation is the more accurate straight-line distance from the target to the radar after the delay correction. This result is the radial range estimate of multiple targets.

[0162] Beam pointing deviation compensation is performed on the angle mapping value to obtain the spatial azimuth estimate. This compensation is achieved through a deviation lookup table, which stores the system measurement deviation values ​​corresponding to different angle mapping values. Specifically, the angle mapping value is used as input, and an interpolation lookup is performed in the deviation lookup table to find the two entries that are closest to the mapping value. Then, based on the stored deviation values ​​corresponding to these two entries, an accurate deviation compensation amount is calculated through linear interpolation. Finally, the original angle mapping value is added to this calculated deviation compensation amount to obtain the compensated angle value, which is the spatial azimuth estimate of the multi-target target.

[0163] Based on the estimated radial distance and spatial azimuth, the spatial coordinates of multiple targets are determined using polar coordinate positioning relationships. This determination process involves converting polar coordinates to rectangular coordinates. The source is the first The estimated radial distance obtained after measuring or analyzing a target; parameters The source is the first The estimated values ​​of spatial azimuth obtained after measuring or analyzing a target; Corresponding to the The spatial coordinates of a target are supported by the spatial attributes of the target's actual location; The source is a pre-defined system of rules used to convert polar coordinate space to rectangular coordinate space.

[0164] This formula utilizes a pre-defined mapping relationship between polar coordinate space and rectangular coordinate space. , put the first Radial distance estimates for each target Spatial azimuth estimate These two parameters in polar coordinate form are transformed into forms that can represent the first... Rectangular coordinates of the target spatial location This completes the conversion process from polar coordinate parameter system to rectangular coordinate representation.

[0165] when When the value of changes, in With the values ​​remaining unchanged, through the mapping relationship Get the first Spatial coordinates of the target It will cause a positional change in space along the direction away from or towards the origin; when When the value of changes, in With the values ​​remaining unchanged, through the mapping relationship Get the first Spatial coordinates of the target It will rotate in space along a circular direction centered on the origin; and The changes in their respective values ​​cooperate with each other, making the economic... After mapping Spatial coordinates of the target It can present the positional distribution and variation patterns in space, which are determined by both radial distance and azimuth angle.

[0166] The spatial coordinates of multiple targets are correlated and integrated to generate spatial position parameters. This integration process involves organizing the spatial coordinates of all detected targets into a structured list according to the order in which they were discovered. Specifically, each detected target is assigned a unique sequential number. Then, the target number, its corresponding radial distance estimate, the azimuth and elevation angles included in the spatial azimuth estimate, and the calculated spatial coordinates are packaged together into a data unit. Next, all the data units of the targets are arranged in ascending order of their target numbers to form a general data list. This list is the spatial position parameters of the multiple targets.

[0167] The detection results of spatial position parameters are integrated to generate multi-target parallel detection results for harmonic radar. This integration involves encapsulating the spatial position parameters into a final standard output format. Specifically, a result data structure is created, which contains a header recording the total number of targets and the detection timestamp. Then, all target data units from the spatial position parameter list are sequentially filled into the detailed target list of the result data structure. Each target data unit occupies one row in the detailed list, and each row contains fields such as target number, radial distance, azimuth angle, elevation angle, and spatial coordinates. After filling, the entire data structure containing the result header and the detailed target list is the final multi-target parallel detection result of the harmonic radar.

[0168] As can be seen from the above embodiments, the multi-target parallel detection method of harmonic radar provided by the present invention can accurately separate single echo signals of multiple targets by performing time-domain windowing and framing and blind source demixing on aliased echo data. Combined with harmonic component analysis and wave position correlation matching, it can achieve accurate extraction of harmonic features of multiple targets and efficient matching of spatial wave positions, and simultaneously complete the calculation of spatial position parameters of multiple targets. This significantly improves the efficiency and real-time performance of multi-target parallel detection of harmonic radar, while also improving the target positioning accuracy and the reliability of detection results.

[0169] like Figure 2 The diagram shown is a functional block diagram of a harmonic radar multi-target parallel detection system 100 provided in an embodiment of the present invention, including an echo acquisition and framing module 101, a blind source signal separation module 102, a harmonic feature extraction module 103, a wave position correlation matching module 104, and a target positioning calculation module 105.

[0170] In this embodiment, the functions of each module are as follows:

[0171] The echo acquisition and framing module 101 is used to superimpose the echo signals reflected by multiple targets to form aliased echo data within a preset detection period through the receiver of the harmonic radar, and to perform time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data.

[0172] The blind source signal separation module 102 is used to perform blind source demixing on the aliased echo data based on the characteristic parameters of the time-domain echo frame to obtain the single-target echo signal of the multi-target.

[0173] The harmonic feature extraction module 103 is used to perform harmonic component analysis on the single-target echo signal to obtain the harmonic feature vector of the multi-target.

[0174] The wave position association matching module 104 is used to associate and match the harmonic feature vector with the preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector.

[0175] The target positioning and calculation module 105 is used to simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar based on the detection wave position and the harmonic feature vector, and generate the multi-target parallel detection results of the harmonic radar.

[0176] As can be seen from the above embodiments, the harmonic radar multi-target parallel detection system provided by the present invention can accurately separate single echo signals of multiple targets by performing time-domain windowing and framing and blind source demixing on aliased echo data. Combined with harmonic component analysis and wave position correlation matching, it can achieve accurate extraction of harmonic features of multiple targets and efficient matching of spatial wave positions, and simultaneously complete the calculation of spatial position parameters of multiple targets, significantly improving the efficiency and real-time performance of harmonic radar multi-target parallel detection, while improving target positioning accuracy and detection result reliability.

[0177] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for parallel detection of multiple targets using harmonic radar, characterized in that, The method includes: Step 1: Using the receiver of the harmonic radar, within a preset detection period, the echo signals reflected by multiple targets are superimposed to form aliased echo data, and the aliased echo data is windowed and framed in the time domain to obtain the time domain echo frame of the aliased echo data. Step 2: Using the characteristic parameters of the time-domain echo frame as the separation basis, perform blind source demixing on the aliased echo data to obtain the single-target echo signal of the multi-target; Step 3: Perform harmonic component analysis on the single-target echo signal to obtain the harmonic characteristic vector of the multi-target; Step 4: Associate and match the harmonic feature vector with the preset detection space division information to determine the detection wave position corresponding to the harmonic feature vector; Step 5: Based on the detected wave position and the harmonic feature vector, simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar, and generate the multi-target parallel detection results of the harmonic radar. The process is as follows: Based on the detected wave position, the harmonic feature vector is subjected to wave position matching calibration to obtain the wave position calibration feature of the harmonic feature vector. Spatial pointing analysis is performed on the probe wave position to obtain the spatial pointing parameters of the probe wave position; Based on the spatial pointing parameters, the wave position calibration features are mapped by distance and angle to obtain the spatial mapping parameters of the multi-target; The spatial mapping parameters are used to calculate the target positions to obtain the spatial position parameters of the multiple targets. The process is as follows: Extract the distance mapping values ​​and angle mapping values ​​corresponding to the multiple targets from the spatial mapping parameters; The distance mapping value is corrected for propagation delay to obtain the radial distance estimate of the multiple targets; Beam pointing deviation compensation is performed on the angle mapping value to obtain the spatial azimuth estimate of the multi-target; Based on the estimated radial distance and the estimated spatial azimuth, the spatial coordinates of the multiple targets are determined using polar coordinate positioning relationships. The spatial coordinates of the multiple targets are determined using the following formula: ； In the formula, Indicates the first The spatial coordinates of the target This represents the mapping relationship from polar coordinate space to rectangular coordinate space. Indicates the first Radial distance estimates for each target. Indicates the first Estimated spatial azimuth angles of each target; The spatial coordinates of the multiple targets are correlated and integrated to generate the spatial position parameters of the multiple targets; The detection results of the spatial position parameters are integrated to generate the multi-target parallel detection results of the harmonic radar.

2. The method for parallel detection of multiple targets using harmonic radar as described in claim 1, characterized in that, In step 1, the receiver of the harmonic radar superimposes the echo signals reflected from multiple targets within a preset detection period to form aliased echo data, and performs time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data. The process is as follows: Within a preset detection period, the receiver of the harmonic radar is controlled to turn on the receiving channel in sequence to receive echo signals reflected from multiple targets in different spatial directions. The echo signals are time-domain aligned, and the aligned echo signals are superimposed on the same time base to obtain the aliased echo data of the multi-target; The aliased echo data is sampled and quantized to obtain a discrete digital sequence of the aliased echo data; Based on the detection period, the frame length and frame shift used for time-domain segmentation are determined, and based on the frame length and frame shift, the discrete digital sequence is segmented one by one to obtain the continuous overlapping signal segments of the discrete digital sequence. Time-domain windowing is applied to the continuous overlapping signal segments to obtain the time-domain echo frame of the aliased echo data.

3. The method for parallel detection of multiple targets using harmonic radar as described in claim 1, characterized in that, In step 2, the characteristic parameters of the time-domain echo frame are used as the separation basis to perform blind source demixing on the aliased echo data to obtain the single-target echo signal of the multi-target system. The process is as follows: Based on the characteristic parameters of the time-domain echo frame, blind source separation initialization is performed on the aliased echo data to obtain the initial separation vector set of the aliased echo data; The initial separation vector set is iteratively updated to obtain the optimized separation vector set of the aliased echo data; Based on the optimized separation vector set, signal separation is performed on the aliased echo data to obtain the independent echo components of the aliased echo data. The independent echo components are reconstructed in the time domain to obtain the single-target echo signal of the multi-target system.

4. The method for parallel detection of multiple targets using harmonic radar as described in claim 3, characterized in that, Based on the optimized separation vector set, signal separation is performed on the aliased echo data to obtain the independent echo components of the aliased echo data. The process is as follows: Based on the optimized separation vector set, the aliased echo data is separated and projected to obtain the projection component set of the aliased echo data. The orthogonality of the projected component set is tested to obtain the orthogonality measure of the components in the projected component set. Based on the orthogonality metric, the effective projection components of the aliased echo data in the projection component set are selected. The effective projection components are subjected to time-domain waveform extraction to obtain the independent echo components of the aliased echo data.

5. The method for parallel detection of multiple targets using harmonic radar as described in claim 1, characterized in that, In step 3, harmonic component analysis is performed on the single-target echo signal to obtain the harmonic characteristic vector of the multi-target signal. The process is as follows: The single-target echo signal is subjected to spectral decomposition to obtain the spectral distribution of the single-target echo signal; The fundamental frequency component of the spectrum distribution is obtained by performing fundamental frequency localization on the spectrum distribution. Based on the fundamental frequency component, harmonic sequence extraction is performed on the spectral distribution to obtain the harmonic components of the single-target echo signal; The harmonic components are characterized by characteristic parameters to obtain the harmonic characteristic vector of the multi-target.

6. The method for parallel detection of multiple targets using harmonic radar as described in claim 5, characterized in that, The process of characterizing the harmonic components using characteristic parameters to obtain the harmonic feature vector of the multi-target is as follows: The harmonic components are normalized to obtain the normalized amplitude spectrum of the harmonic components. Peak detection is performed on the normalized amplitude spectrum to obtain the main peak amplitude value of the harmonic component; Based on the main peak amplitude value, the amplitude distribution relationship between the harmonic components is determined, and the amplitude feature vector of the multi-target is obtained; Phase feature extraction is performed on the harmonic components to obtain the phase feature vector of the multi-target; The amplitude values ​​of each harmonic in the amplitude feature vector and the phase values ​​of each harmonic in the phase feature vector are vector-fused to obtain the characteristic elements of the harmonic components. The fusion formula for the characteristic elements is as follows: ； In the formula, Indicates the first Characteristic elements of first harmonic components, Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, Represents the natural constant. Indicates the first The amplitude value of the first harmonic component. Indicates the first Phase value of the first harmonic component, This represents the preset coupling coefficient. This indicates the total order of the harmonic components. Represents the cosine function. Represents the imaginary unit; The characteristic elements of the harmonic components are arranged according to their order to obtain the harmonic characteristic vector of the multi-target.

7. The method for parallel detection of multiple targets using harmonic radar as described in claim 1, characterized in that, In step 4, the process of associating and matching the harmonic feature vector with the preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector is as follows: Wave position feature encoding is performed on the preset detection airspace division information to obtain the wave position feature template set of the detection wave position; The wave position feature template set is subjected to feature space mapping to obtain the spatial mapping features of the wave position feature template set; Based on the spatial mapping features, the harmonic feature vectors are correlated with similarity to obtain the correlation distribution between the harmonic feature vectors and the probe wave position. Peak location is performed on the correlation distribution to determine the probe wave position corresponding to the harmonic feature vector.

8. A harmonic radar multi-target parallel detection system, characterized in that, For implementing the harmonic radar multi-target parallel detection method according to any one of claims 1-7, the system comprises: The echo acquisition and framing module is used to superimpose the echo signals reflected by multiple targets to form aliased echo data within a preset detection period through the receiver of the harmonic radar, and to perform time-domain windowing and framing on the aliased echo data to obtain the time-domain echo frame of the aliased echo data. The blind source signal separation module is used to perform blind source demixing on the aliased echo data based on the characteristic parameters of the time-domain echo frame to obtain the single-target echo signal of the multi-target; The harmonic feature extraction module is used to perform harmonic component analysis on the single-target echo signal to obtain the harmonic feature vector of the multi-target. The wave position association matching module is used to associate and match the harmonic feature vector with the preset detection spatial domain division information to determine the detection wave position corresponding to the harmonic feature vector. The target localization and calculation module is used to simultaneously calculate the spatial position parameters of the multiple targets relative to the harmonic radar based on the detection wave position and the harmonic feature vector, and generate the multi-target parallel detection results of the harmonic radar.

Images