Orthogonal waveform design method for near-field MIMO radar system

Abstract

The invention provides an orthogonal waveform design method for a near-field MIMO radar system. Relates to the technical field of radar sensing. The method comprises the following steps: acquiring a first signal transmitted by a near-field multiple-input multiple-output (MIMO) radar system, wherein the first signal comprises a second signal transmitted by each array element of the radar system; determining an autocorrelation value of each second signal and a cross-correlation value between any two second signals, and according to the first signal, the plurality of autocorrelation values and the plurality of cross-correlation values, determining an objective function and a constraint condition associated with the orthogonality of the waveform; determining first phase information corresponding to the first signal according to the first signal, the target function and the constraint condition; inputting a target function, a constraint condition and a plurality of pieces of first phase information into the first model to obtain second phase information corresponding to a first signal associated with the spatial-temporal characteristics; and according to the second phase information, the target signal transmitted by each array element is determined, the optimization efficiency is improved, and the system performance of the MIMO radar is improved.

Description

Technical Field

[0001] This application relates to the field of radar sensing technology, and in particular to an orthogonal waveform design method for near-field MIMO radar systems. Background Technology

[0002] With the widespread application of Multiple Input Multiple Output (MIMO) radar in imaging, detection and surveillance, and the increasing demands for radar resolution and anti-jamming capabilities, waveform orthogonality design has become a key aspect of improving MIMO radar performance.

[0003] In related technologies, discrete phase encoding is typically performed on MIMO radar waveforms, and swarm intelligence optimization algorithms are used to optimize the encoded phase matrix in order to improve the performance of MIMO radar.

[0004] However, the above process requires exhaustively searching for the optimal value of the phase coding matrix, which makes the computational complexity high and consumes a lot of computing resources, resulting in low efficiency of orthogonal waveform design for MIMO radar. Summary of the Invention

[0005] This application provides an orthogonal waveform design method for near-field MIMO radar systems to solve the technical problem of low efficiency in orthogonal waveform design for MIMO radar.

[0006] In a first aspect, embodiments of this application provide an orthogonal waveform design method for near-field MIMO radar systems, including:

[0007] Acquire the first signal transmitted by the near-field multiple-input multiple-output (MIMO) radar system, the first signal including the second signal transmitted by each array element of the radar system;

[0008] Determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and based on the first signal, multiple autocorrelation values, and multiple cross-correlation values, determine the objective function and constraints associated with the orthogonality of the waveform;

[0009] Based on the first signal, the objective function, and the constraints, the first phase information corresponding to the first signal is determined. The first phase information includes the time information of the first signal and the position information of each array element.

[0010] By inputting the objective function, constraints, and multiple first phase information into the first model, second phase information corresponding to the first signal associated with spatiotemporal characteristics is obtained;

[0011] Based on the second phase information, the target signal emitted by each array element is determined, and the target function corresponding to the target signal is minimized.

[0012] In one possible implementation, determining the first phase information corresponding to the first signal based on the first signal and the constraint conditions includes:

[0013] Based on the first signal, determine the initial phase information corresponding to the first signal;

[0014] Based on the constraints, the initial phase information is processed to obtain the first phase information corresponding to the first signal.

[0015] In one possible implementation, the initial phase information is processed according to constraints to obtain the first phase information corresponding to the first signal, including:

[0016] Based on the constraints, feature transformation processing is performed on multiple phase values ​​in the initial phase information to obtain the initial phase information after feature transformation processing.

[0017] Determine the spatial position information of each array element and the temporal position information of each second signal;

[0018] The initial phase information after feature transformation, the spatial position information of each array element, and the temporal position information of each second signal are fused to determine the first phase information corresponding to the first signal.

[0019] In one possible implementation, an objective function, constraints, and multiple first phase information are input into a first model to obtain second phase information corresponding to a first signal associated with spatiotemporal characteristics, including:

[0020] Based on the first model, multi-level feature fusion processing is performed on multiple first phase information to obtain the probability score of each element in the first phase information. The probability score is used to indicate the recommendation degree of each phase.

[0021] Based on the probability fractions of each element and the constraints, the second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained.

[0022] In one possible implementation, multi-level feature fusion processing is performed on multiple first phase information to obtain the probability score of each element in the first phase information, including:

[0023] For any given first phase information;

[0024] Based on the first phase information, multiple first features corresponding to the first phase information are determined, and the first features are used to indicate the hierarchical information of the first phase information.

[0025] Multiple first features are aggregated to obtain aggregated second features, and feature extraction is performed on the second features to obtain the probability scores of each element in the first phase information.

[0026] In one possible implementation, based on the probability fractions of each element and the constraints, second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained, including:

[0027] The probability scores of each element are sampled to obtain multiple sampled probability scores, which are used to indicate the recommendation level of each phase after standardization.

[0028] The second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained by multiplying the sampled multiple probability fractions and the phase information in the constraints.

[0029] In one possible implementation, based on a first signal, multiple autocorrelation values, and multiple cross-correlation values, an objective function and constraints associated with the orthogonality of the waveform are determined, including:

[0030] Based on multiple autocorrelation values, first data and second data are determined. The first data is used to indicate the autocorrelation peak sidelobe value in multiple second signals, and the second data is used to indicate the autocorrelation integral sidelobe value of multiple second signals.

[0031] Based on multiple cross-correlation values, a third data and a fourth data are determined. The third data is used to indicate the peak cross-correlation sidelobe value among the multiple second signals, and the fourth data is used to indicate the integral cross-correlation sidelobe value among the multiple second signals.

[0032] Determine the first weight corresponding to the first, second, third, and fourth data points;

[0033] Based on multiple first weights, the first data, second data, third data, and fourth data are weighted to obtain the objective function associated with the orthogonality of the waveform, and the constraint conditions are determined based on the first signal.

[0034] Secondly, embodiments of this application provide an orthogonal waveform design device for near-field MIMO radar systems, comprising: an acquisition module, a first determination module, a second determination module, an input module, and a processing module, wherein...

[0035] The acquisition module is used to acquire the first signal transmitted by the near-field multiple-input multiple-output (MIMO) radar system, the first signal including the second signal transmitted by each array element of the radar system;

[0036] The first determining module is used to determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and to determine the objective function and constraint conditions associated with the orthogonality of the waveform based on the first signal, multiple autocorrelation values ​​and multiple cross-correlation values.

[0037] The second determining module is used to determine the first phase information corresponding to the first signal based on the first signal, the objective function and the constraint conditions. The first phase information includes the time information of the first signal and the position information of each array element.

[0038] The input module is used to input the objective function, constraints and multiple first phase information into the first model to obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics;

[0039] The processing module is used to determine the target signal emitted by each array element based on the second phase information, and the target function corresponding to the target signal is minimized.

[0040] In one possible implementation, the second determining module is specifically used for:

[0041] Based on the first signal, determine the initial phase information corresponding to the first signal;

[0042] Based on the constraints, the initial phase information is processed to obtain the first phase information corresponding to the first signal.

[0043] In one possible implementation, the second determining module is specifically used for:

[0044] Based on the constraints, feature transformation processing is performed on multiple phase values ​​in the initial phase information to obtain the initial phase information after feature transformation processing.

[0045] Determine the spatial position information of each array element and the temporal position information of each second signal;

[0046] The initial phase information after feature transformation, the spatial position information of each array element, and the temporal position information of each second signal are fused to determine the first phase information corresponding to the first signal.

[0047] In one possible implementation, the input module is specifically used for:

[0048] Based on the first model, multi-level feature fusion processing is performed on multiple first phase information to obtain the probability score of each element in the first phase information. The probability score is used to indicate the recommendation degree of each phase.

[0049] Based on the probability fractions of each element and the constraints, the second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained.

[0050] In one possible implementation, the input module is specifically used for:

[0051] For any given first phase information;

[0052] Based on the first phase information, multiple first features corresponding to the first phase information are determined, and the first features are used to indicate the hierarchical information of the first phase information.

[0053] Multiple first features are aggregated to obtain aggregated second features, and feature extraction is performed on the second features to obtain the probability scores of each element in the first phase information.

[0054] In one possible implementation, the input module is also used for:

[0055] The probability scores of each element are sampled to obtain multiple sampled probability scores, which are used to indicate the recommendation level of each phase after standardization.

[0056] The second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained by multiplying the sampled multiple probability fractions and the phase information in the constraints.

[0057] In one possible implementation, the first determining module is further configured to:

[0058] Based on multiple autocorrelation values, first data and second data are determined. The first data is used to indicate the autocorrelation peak sidelobe value in multiple second signals, and the second data is used to indicate the autocorrelation integral sidelobe value of multiple second signals.

[0059] Based on multiple cross-correlation values, a third data and a fourth data are determined. The third data is used to indicate the peak cross-correlation sidelobe value among the multiple second signals, and the fourth data is used to indicate the integral cross-correlation sidelobe value among the multiple second signals.

[0060] Determine the first weight corresponding to the first, second, third, and fourth data points;

[0061] Based on multiple first weights, the first data, second data, third data, and fourth data are weighted to obtain the objective function associated with the orthogonality of the waveform, and the constraint conditions are determined based on the first signal.

[0062] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0063] The memory stores the instructions that the computer executes;

[0064] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0065] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0066] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0067] This application provides an orthogonal waveform design method for near-field MIMO radar systems. The method involves acquiring a first signal transmitted by the near-field MIMO radar system, which includes second signals transmitted by each array element of the radar system; determining the autocorrelation value of each second signal and the cross-correlation value between any two second signals; and determining an objective function and constraints associated with the orthogonality of the waveform based on the first signal, multiple autocorrelation values, and multiple cross-correlation values; determining first phase information corresponding to the first signal based on the first signal, the objective function, and the constraints, whereby the first phase information includes the time information of the first signal and the position information of each array element; and then... The model takes an objective function, constraints, and multiple first phase information as input and obtains second phase information corresponding to the first signal associated with spatiotemporal features. Based on the second phase information, the target signal emitted by each array element is determined, and the objective function corresponding to the target signal is minimized. In the above method, the electronic device can transform the waveform orthogonality optimization problem into an end-to-end mapping task based on spatiotemporal feature learning. By using the first model to model the global dependencies in the phase matrix, it can effectively approximate the global optimal solution while avoiding traditional exhaustive search, significantly reducing the computational complexity of waveform design, improving optimization efficiency, and enhancing the system performance of MIMO radar. Attached Figure Description

[0068] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0069] Figure 1 A schematic diagram illustrating the process of an orthogonal waveform design method for a near-field MIMO radar system provided in this application;

[0070] Figure 2 A schematic diagram illustrating the process of determining second phase information provided in the embodiments of this application;

[0071] Figure 3 A flowchart illustrating another orthogonal waveform design method for MIMO radar systems provided in this application embodiment;

[0072] Figure 4A schematic diagram of an orthogonal waveform design device for a near-field MIMO radar system provided in this application;

[0073] Figure 5 A schematic diagram of the structure of the electronic device provided in this application.

[0074] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0075] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0076] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0077] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0078] In existing technologies, discrete phase encoding is typically performed on MIMO radar waveforms, and swarm intelligence optimization algorithms are used to optimize the encoded phase matrix to improve MIMO radar performance. However, this process requires exhaustively searching for the optimal value of the phase encoding matrix, resulting in high computational complexity and consuming significant computational resources, thus leading to low efficiency in orthogonal waveform design for MIMO radar.

[0079] This application provides an orthogonal waveform design method for near-field MIMO radar systems. By acquiring the second signals emitted by each array element in the near-field MIMO radar system, determining the autocorrelation value of each signal and the cross-correlation value between any two signals, an objective function and constraints related to waveform orthogonality are constructed. Then, based on the objective function and constraints, first phase information containing time and array element position information is extracted and input into a first model for spatiotemporal feature fusion. Finally, the target signal that minimizes the objective function is output. In this method, electronic devices can transform the waveform orthogonality optimization problem into an end-to-end mapping task based on deep learning. The first model is used to model the global spatiotemporal dependencies in the phase matrix, effectively approximating the global optimum while avoiding traditional exhaustive search. This method explicitly captures the spatial relationships between array elements and the temporal characteristics of signals through spatiotemporal position encoding, significantly improving the orthogonality and resolution performance of waveforms in near-field scenarios. Furthermore, the optimization result can be obtained through a single forward computation of the deep learning model, overcoming the limitations of high computational complexity and susceptibility to local optima in traditional methods, greatly reducing computational resource consumption and improving waveform design efficiency. This method can adapt to system configurations with different array element sizes and sequence lengths, enhancing the practicality and applicability of the scheme. While ensuring waveform performance, it significantly improves the efficiency and feasibility of MIMO radar waveform design.

[0080] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0081] Figure 1 A schematic diagram illustrating the process of an orthogonal waveform design method for near-field MIMO radar systems provided in this application is shown below. Figure 1 As shown, the method includes:

[0082] S101. Acquire the first signal transmitted by the near-field multiple-input multiple-output (MIMO) radar system.

[0083] The execution subject of this application embodiment can be an electronic device or an orthogonal waveform design device installed in an electronic device. The orthogonal waveform design device can be implemented through software or a combination of software and hardware. The electronic device can be a terminal device or a server.

[0084] The first signal includes the second signal transmitted by each array element of the radar system.

[0085] The first signal is used to indicate the set of transmitted signals of all array elements at the transmitter of the MIMO radar system. For example, the first signal includes the second signal transmitted individually by each array element, that is, the first signal is the set of the second signals of each array element.

[0086] The second signal is used to indicate the transmitted signal of each element in the MIMO radar system; that is, any second signal includes the phase modulation sequence of that element over multiple pulse periods.

[0087] In some embodiments, if the MIMO radar system has M virtual array elements, and each array element transmits N pulse signals, then the second signal transmitted by the m-th array element can be expressed as:

[0088]

[0089] in, This represents the second signal emitted by the m-th array element. This represents the Nth pulse signal emitted by the m-th array element, where N is the length of the signal sequence.

[0090] It can be obtained that the first signal of the MIMO radar system can be expressed as:

[0091]

[0092] in, This represents the set of signals emitted by each array element. This represents the second signal emitted by the Mth array element.

[0093] S102. Determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and determine the objective function and constraints associated with the orthogonality of the waveform based on the first signal, multiple autocorrelation values ​​and multiple cross-correlation values.

[0094] The autocorrelation value is used to indicate the degree of similarity between the second signals at different time delays.

[0095] Cross-correlation values ​​are used to indicate the degree of similarity between different second signals at different time delays.

[0096] In some embodiments, this application focuses on aperiodic correlation functions. Electronic devices can determine multiple autocorrelation values ​​using autocorrelation functions and multiple cross-correlation values ​​using cross-correlation functions. For example, the second signal transmitted by the m-th array element... and the The second signal emitted by each element The cross-correlation function can be expressed as:

[0097]

[0098] in, Indicates taking the conjugate. This represents the second signal emitted by the m-th array element. Indicates the first The second signal emitted by each array element after an increment of k, where k represents the delay of the cross-correlation function. and , and .

[0099] exist When the autocorrelation function of the second signal emitted by the same array element is expressed as:

[0100]

[0101] The objective function indicates the optimization objective for waveform orthogonality; that is, it is necessary to minimize the objective function through optimization variables to achieve global optimization of waveform performance.

[0102] Constraints are used to indicate physical constraints in the optimization process.

[0103] In some embodiments, the electronic device may determine an objective function and constraints associated with the orthogonality of a waveform based on a first signal, multiple autocorrelation values, and multiple cross-correlation values, as follows: determining first data and second data based on multiple autocorrelation values; determining third data and fourth data based on multiple cross-correlation values; determining first weights corresponding to the first data, second data, third data, and fourth data; weighting the first data, second data, third data, and fourth data according to multiple first weights to obtain the objective function associated with the orthogonality of the waveform, and determining the constraints based on the first signal.

[0104] The first data is used to indicate the autocorrelation peak sidelobe value of the multiple second signals, and the second data is used to indicate the autocorrelation integral sidelobe value of the multiple second signals.

[0105] In some embodiments, the electronic device can, for each second signal, iterate through all autocorrelation values ​​in its autocorrelation function except for zero delay to determine the maximum absolute value of the sidelobe corresponding to that signal; after obtaining the maximum absolute values ​​of the sidelobe of all second signals, the maximum value is selected as the first data, i.e., the autocorrelation peak sidelobe level. The specific process can be expressed as follows:

[0106]

[0107] In some embodiments, the electronic device can, for each second signal, iterate through all autocorrelation values ​​in its autocorrelation function except for zero delay, and calculate the sum of the absolute values ​​of all sidelobes corresponding to that signal; after obtaining the sum of the absolute values ​​of all sidelobes of the second signals, these sums are accumulated as the second data, i.e., the autocorrelation integral sidelobe level. The specific process can be expressed as follows:

[0108]

[0109] The third data is used to indicate the peak sidelobe value of the cross-correlation among the multiple second signals, and the fourth data is used to indicate the integral sidelobe value of the cross-correlation among the multiple second signals.

[0110] In some embodiments, the electronic device can, for any two different second signals, traverse all cross-correlation values ​​in their cross-correlation functions to determine the maximum absolute cross-correlation value corresponding to the signal pair; after obtaining the maximum absolute cross-correlation values ​​of all signal pairs, the maximum value is selected as the third data, i.e., the cross-correlation peak sidelobe level. The specific process can be expressed as follows:

[0111]

[0112] In some embodiments, the electronic device can, for any two different second signals, iterate through all cross-correlation values ​​in their cross-correlation functions and calculate the sum of all absolute cross-correlation values ​​corresponding to the signal pair; after obtaining the sum of the absolute cross-correlation values ​​of all signal pairs, these sums are accumulated as the fourth data, i.e., the cross-correlation integral sidelobe level. The specific process can be expressed as follows:

[0113]

[0114] The first weight is used to indicate the degree of influence of each data point on the objective function; that is, the first weight can represent the different degrees of importance that the objective function attaches to each indicator.

[0115] In some embodiments, a second weight corresponding to each data point can be pre-set according to the actual application process of the MIMO radar system. Different weight combinations reflect the degree of emphasis on different performance indicators. For example, the second weight corresponding to the first data point can be... The second weight corresponding to the second data can be The second weight corresponding to the third data can be The second weight corresponding to the fourth data can be .

[0116] In some embodiments, the electronic device can perform weighted summation on the first data, second data, third data, and fourth data according to a preset second weight to construct an objective function associated with waveform orthogonality, while simultaneously determining corresponding constraint conditions based on the physical characteristics of the first signal. The specific process can be represented as follows:

[0117]

[0118]

[0119]

[0120] Among them, constraint C1 means that the transmitted signal must meet the constant modulus constraint to ensure power efficiency; constraint C2 means that the phase value of each transmitted signal can be selected from the preset discrete code symbol set to meet the implementation requirements of the system.

[0121] It can be seen that the optimization problem constructed in this application aims to minimize the weighted sum (objective function) of the four autocorrelation peak sidelobe levels (APSL), autocorrelation integral sidelobe levels (AISL), cross-correlation peak sidelobe levels (CPSL), and cross-correlation integral sidelobe levels (CISL). This optimization problem is both non-convex and non-smooth, and traditional optimization methods are difficult to directly obtain the global optimal solution.

[0122] Minimizing APSL can effectively improve the radar system's ability to detect weak targets and reduce inter-symbol interference of the signal itself; minimizing CPSL helps improve the radar system's resolution performance in multi-target scenarios; minimizing AISL can enhance the anti-interference capability of a single signal, making it better adaptable to clutter suppression scenarios; minimizing CISL can significantly reduce inter-channel interference and improve the coexistence capability of multiple signals, thus making the designed waveform more in line with the actual needs of the MIMO radar system.

[0123] S103. Determine the first phase information corresponding to the first signal based on the first signal, the objective function, and the constraints.

[0124] The first phase information includes the time information of the first signal and the position information of each array element.

[0125] The first phase information is used to indicate the phase matrix representation after feature embedding and position encoding, providing a structured input containing spatiotemporal context for the subsequent waveform optimization process.

[0126] In some embodiments, the electronic device may determine the first phase information corresponding to the first signal based on the first signal, the objective function, and the constraints in the following manner: determining the initial phase information corresponding to the first signal based on the first signal; and processing the initial phase information according to the constraints to obtain the first phase information corresponding to the first signal.

[0127] The initial phase information is used to indicate the original discrete phase matrix corresponding to the first signal.

[0128] In some embodiments, the electronic device can extract the discrete phase coding sequence of each second signal based on the second signals emitted by each array element in the first signal, and construct an initial phase matrix. Specifically, as shown in S301, the electronic device obtains the phase modulation information of each array element from the first signal, and constructs a phase matrix Y of dimension N×M based on this information, where the first signal is:

[0129]

[0130] The initial phase matrix of the first signal can be expressed as:

[0131]

[0132] Wherein, the phase matrix corresponds to the m-th element. The pulse signals emitted by each array element All satisfy the constant modulus constraint, that is , This represents the discrete phase value corresponding to the nth pulse signal of the mth array element. Its value comes from a predefined set of discrete code symbols and can be represented as:

[0133]

[0134] Where B represents the preset number of discrete phases.

[0135] In some embodiments, the electronic device may process the initial phase information according to constraints to obtain the first phase information corresponding to the first signal based on the following implementation: performing feature transformation processing on multiple phase values ​​in the initial phase information according to constraints to obtain the initial phase information after feature transformation processing; determining the spatial position information of each array element and the temporal position information of each second signal; and performing fusion processing on the feature-transformed initial phase information, the spatial position information of each array element, and the temporal position information of each second signal to determine the first phase information corresponding to the first signal.

[0136] In some embodiments, the electronic device employs a Transformer-based deep learning model to process the initial phase information according to constraints to obtain optimized first phase information. This processing mainly includes feature embedding and position encoding, configuring spatiotemporal structure information for the initial phase matrix to construct the input to the Transformer encoder. The specific process is as follows:

[0137] Electronic devices perform feature transformation (i.e., feature embedding) on ​​the initial phase matrix, mapping it from a discrete numerical space to a high-dimensional feature space. Specifically, this can be achieved through trainable network layers that replace multiple phase values ​​in the initial phase matrix p with d-dimensional feature vectors, thereby elevating the N×M phase matrix into a feature tensor of N×M×d dimensions, which can be represented as:

[0138]

[0139] in, This is a network layer with updatable parameters and a dimension of B×d, meaning that each row is a d-dimensional feature vector of an element in the code symbol set, and the row index is consistent with the code symbol set index. The operation performed by this network layer is to use each phase value in the input phase matrix p as the row index of this network layer, and the indexed row is used as the feature vector to replace each phase value, thus completing the feature embedding and dimension expansion of the initial phase matrix.

[0140] To extract the positional relationships of each element in the initial phase matrix after feature transformation, the electronic device can perform position encoding for different array elements in the spatial dimension. By distinguishing the physical positions of different array elements, the spatial positional information of each array element is obtained, enabling the model to distinguish signals from array elements at different positions. The encoded value can be calculated from the array element index m, specifically as follows:

[0141]

[0142] in, Let d represent the spatial coding component of the m-th element at feature position (i, k), where k represents the k-th component of the coding vector, and d represents the feature dimension. As a frequency factor, it can increase rapidly with increasing k, causing the wavelength of the trigonometric function to grow geometrically, thus forming a change from high frequency to low frequency in different dimensions.

[0143] Electronic devices can perform position encoding on different pulse signals in the time dimension. By identifying the transmission order of the pulses, the temporal position information of each second signal can be obtained, enabling the model to understand the temporal sequence and dependency of the pulse signals. The encoded value can be calculated from the pulse index n, specifically as follows:

[0144]

[0145] in, Let d represent the time-coded component of the nth pulse signal at feature position (i, k), where k represents the kth component of the coding vector, and d represents the feature dimension. As a frequency factor, it can increase rapidly with increasing k, causing the wavelength of the trigonometric function to grow geometrically, thus forming a change from high frequency to low frequency in different dimensions.

[0146] The initial phase matrix after feature embedding is obtained. The spatial position information of each array element and the temporal position information of each second signal The electronic device can then add the three data points together and fuse them into the input X of the Transformer encoder, which is the first phase information, and can be represented as:

[0147]

[0148] In this way, by introducing spatial and temporal location coding, the input dimension of the Transformer can be expanded to model the spatiotemporal dependencies between array elements and pulses. This effectively solves the problem that traditional methods do not fully consider the spatial distribution of array elements and the temporal sequence relationship of pulses in waveform design, resulting in a lack of robustness in the optimization results. That is, by using spatial-temporal dual-dimensional coding, the model can capture global spatiotemporal features, thereby improving radar resolution and anti-interference capabilities.

[0149] S104. Input the objective function, constraints and multiple first phase information into the first model to obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics.

[0150] The second phase information is used to indicate the optimized discrete phase information. Specifically, the second phase information is the globally optimal solution for the objective function (e.g., peak sidelobe level, integral sidelobe level, etc.) achieved through deep feature learning, based on the model satisfying constant modulus and discrete phase constraints. This phase information is the optimal baseband representation of the array's transmitted waveform under specific spatiotemporal characteristics (e.g., beam pointing, interference suppression direction, etc.), and can be used to generate signals transmitted by the radar system to achieve high-efficiency radar performance.

[0151] S105. Determine the target signal transmitted by each array element based on the second phase information.

[0152] Among them, the objective function corresponding to the target signal is the smallest.

[0153] The target signal is used to indicate the physical waveform emitted by each array element, generated by the optimized second phase information. In other words, the target signal is the final transmitted waveform that satisfies all constraints and optimizes the system performance indicators.

[0154] In some embodiments, after the electronic device obtains the second phase information, the electronic device outputs the second phase information based on the model's processing of the input parameters. And construct the transmitted signal based on the second phase information. The target signal, specifically, is represented as:

[0155]

[0156] In this way, the electronic device processes the initialized waveform parameters and other data based on the trained model, and outputs the optimal phase matrix. and constant mode transmission signal It can be seen that the signal achieves low autocorrelation sidelobe values ​​and low cross-correlation values, and has good orthogonality.

[0157] This application provides an orthogonal waveform design method for near-field MIMO radar systems. By acquiring the first signal containing signals from each array element transmitted by the near-field MIMO radar system, the objective function and constraints associated with waveform orthogonality are determined. Combined with first phase information containing spatiotemporal position features, a first model is used for optimization to generate second phase information that minimizes the objective function, thus determining the final waveform transmitted by each array element. In this method, the electronic device constructs the waveform design problem as a constrained optimization model and uses deep learning technology to deeply fuse the spatiotemporal structure of the waveform for end-to-end solution. This enables the efficient design of high-performance orthogonal waveforms. After optimization, the designed waveforms possess both sufficiently low autocorrelation sidelobe peaks to effectively reduce channel self-interference and sufficiently low cross-correlation values ​​to significantly suppress inter-channel interference. This significantly improves the MIMO radar's ability to distinguish and detect dense multi-target targets. Furthermore, the model can fully preserve the spatiotemporal structure information of the waveform through feature embedding and position encoding, without relying on the dependence on the initial solution or complex iterative update rules of traditional optimization methods, significantly reducing computational resource consumption.

[0158] Based on any of the above embodiments, the following, in conjunction with Figure 2 The method for determining the second phase information is described in detail.

[0159] Figure 2 This is a schematic diagram illustrating the process of determining second phase information according to an embodiment of this application. Please refer to... Figure 2 The method may include:

[0160] S201. Based on the first model, perform multi-level feature fusion processing on multiple first phase information to obtain the probability score of each element in the first phase information.

[0161] The probability score is used to indicate the degree of recommendation for each phase.

[0162] In some embodiments, after receiving an input X that integrates phase features and spatiotemporal location information, the electronic device inputs it into a pre-trained Transformer encoder. The encoder performs deep global context modeling on the input first phase information through its unique multi-head self-attention mechanism and feedforward neural network, and calculates a probability score for each element in the first phase information through output layer mapping to evaluate the quality of the phase value at that location.

[0163] In some embodiments, the electronic device may perform multi-level feature fusion processing on multiple first phase information to obtain the probability score of each element in the first phase information based on the following implementation: for any first phase information; determine multiple first features corresponding to the first phase information based on the first phase information; perform feature aggregation processing on the multiple first features to obtain the aggregated second features, and perform feature extraction on the second features to obtain the probability score of each element in the first phase information.

[0164] The first feature is used to indicate the hierarchical information of the first phase information.

[0165] In some embodiments, the electronic device can extract features at different levels from the first phase information. Specifically, it establishes a structured association model for the global optimization of the phase matrix by constructing query, key, and value vector groups, and obtains three sets of feature vectors for each position in the sequence: query... ,key Sum , can be represented as:

[0166]

[0167] in, , and All of these are the results of applying different linear transformations to the first phase information X, in order to extract spatiotemporal dependent features in different directions. Let represent the query matrix of the h-th attention head, serving as a reference feature to characterize the query needs of each position to pay attention to other positions. Let represent the key matrix of the h-th attention head, which serves as a matching feature, characterizing the key information at each position that can be retrieved by other positions. Let represent the value matrix of the h-th attention head, which serves as the weighted available feature to be aggregated with K as the key. It contains the original feature information to be aggregated at each position, and h represents the attention head index. That is, by using multiple attention heads, the model can capture diverse spatiotemporal dependencies in different feature subspaces in parallel. , and All of these are network layers with updatable parameters, used to extract high-dimensional features from the input X, in order to extract deep discriminative features for calculating attention weights, thereby providing an information basis for subsequent feature aggregation.

[0168] Querying via electronic device ( ),key( ),value( After the vector is generated, the electronic device performs global feature aggregation through a scaling dot product attention mechanism, i.e. and The two are matched using a scaled dot product method. For The (to be weighted) available (actual) features for the key. By computing a weighted context vector, where the weights are determined by the query vector and the key vector, the output at each position can be focused on the information most relevant to it in the sequence.

[0169]

[0170] in, Indicates the correlation strength between any two phase elements. Indicates the scaling factor. This represents the normalization exponential function, which normalizes the correlation strength score between any two elements into a probability distribution, i.e., attention weights. This represents the weighted sum of all value vectors, i.e., the first feature.

[0171] In some embodiments, in order to capture the inter-phase dependencies from different perspectives and improve the model's expressive power and the robustness of the optimization process, the features extracted from multiple heads are concatenated along the feature dimension, and then further feature aggregation and extraction are performed to obtain the output Z of the Transformer encoder, i.e., the second feature, which can be represented as:

[0172]

[0173] in, This represents the first feature output by the Hth attention head. This indicates a concatenation operation, which involves concatenating the output matrices of the H attention heads along the feature dimensions. Represents the linear projection weight matrix, which is a network layer with updatable parameters used to reduce dimensionality, fuse, and integrate information from the spliced ​​multidimensional composite features.

[0174] In some embodiments, to improve the model's feature extraction capability and enable the model to more accurately fit the complex nonlinear relationship between input and output, the Transformer encoder output Z is input into a feedforward neural network (FFN) consisting of two linear layers and a nonlinear activation function to further extract spatial and temporal features, as well as dimensional changes. The specific process can be represented as follows:

[0175]

[0176] in, , , and All are network layers with updatable parameters; their dimensions should be chosen to ensure that the final output... That is, the dimension of the last dimension is equal to the number of elements in the code symbol set and corresponds one-to-one with each element, representing the probability score of selecting the corresponding element in the code symbol set.

[0177] S202. Based on the probability fractions of each element and the constraints, obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics.

[0178] In some embodiments, the electronic device may obtain the second phase information corresponding to the first signal associated with the spatiotemporal features based on the probability scores of each element and the constraints, by sampling the probability scores of each element to obtain multiple sampled probability scores; and multiplying the multiple sampled probability scores with the phase information in the constraints to obtain the second phase information corresponding to the first signal associated with the spatiotemporal features.

[0179] The sampled probability score is used to indicate the degree of recommendation for each phase after standardization.

[0180] In some embodiments, to preserve the gradient, the electronic device may employ a phase soft quantization method based on Gumbel-Softmax, by... The last dimension, the probability score, is obtained by sampling elements from the code symbol set, i.e., for each... :

[0181]

[0182] in, This is represented by the i-th component in the probability vector of the approximate one-hot encoding, i.e., the standardized probability distribution. Gumbel noise, which does not participate in backpropagation, is used to introduce randomness into the sampling process, encouraging the model to explore different phase choices and preventing training from getting trapped in local optima. Follows a uniform distribution. The temperature parameter should be chosen to be small enough that the soft sampling vector... It is close to the one-hot encoding form.

[0183] In some embodiments, based on the phase information in the constraints, it can be obtained that, assuming Each element corresponds one-to-one with the element in the code symbol set based on its index. Therefore, each phase value in the output phase matrix (second phase information) can be calculated using the following formula, which can be expressed as:

[0184]

[0185] After obtaining the second phase information Subsequently, an optimal phase matrix may exist. The values ​​of each element in the model are not necessarily strictly equal to the values ​​of the elements in the code symbol set. To strictly adhere to the constraints, the element with the closest value in the code symbol set can be selected to replace the model output based on the nearest principle. The corresponding elements in the [database]. This tiny difference has almost no effect on the orthogonality of the output waveform.

[0186] This application provides an orthogonal waveform design method for near-field MIMO radar systems. By performing multi-level feature fusion processing on multiple first phase information elements according to a first model, a probability score for each phase element is obtained. Then, sampling and fusion are performed in conjunction with constraints to obtain second phase information associated with spatiotemporal characteristics. In this method, the electronic device, through multi-level feature extraction and aggregation, can deeply analyze the hierarchical feature information in the phase matrix and aggregate features at different levels. This allows for precise quantification of the merits of each discrete phase value, generating probability scores reflecting the recommendation level of each phase value. Sampling processing then combines the constraints with the recommendation scores, achieving intelligent selection of the optimal phase combination under strict waveform design constraints. This method effectively solves the problem of traditional waveform optimization methods struggling to efficiently search for the global optimum in discrete phase space. Through guided search based on probability scores, it significantly improves the efficiency and quality of waveform optimization, enabling the design of excellent waveforms with low autocorrelation sidelobe peaks and low cross-correlation values. This effectively reduces channel self-interference and inter-channel interference, improving the radar system's detection performance and multi-target resolution. Furthermore, this method avoids excessive consumption of computing resources and is applicable to MIMO radar systems of different scales. While ensuring waveform orthogonality, it significantly improves the efficiency and practicality of waveform design.

[0187] The following example illustrates an orthogonal waveform design method for MIMO radar systems.

[0188] Figure 3 This is a flowchart illustrating another orthogonal waveform design method for MIMO radar systems provided in this application embodiment. Please refer to... Figure 3 Specifically, it includes the following steps:

[0189] S301. Establish a MIMO radar phase-coded transmission signal model.

[0190] Electronic devices construct a discrete phase-coded signal model for a near-field MIMO radar system. Specifically, this includes: defining a radar system architecture consisting of M transmitting elements (first signal), with each element transmitting a phase-coded sequence containing N symbols (second signal); establishing a mathematical model with the phase matrix as the optimization variable (initial phase matrix), where the transmitted signal of each element satisfies constant modulus constraints and the phase value is taken from a predefined set of discrete code symbols; and determining the mathematical relationship between the signal matrix and the phase matrix to provide a rigorous mathematical foundation for subsequent optimization processes. This model fully considers near-field propagation characteristics and implementation constraints to ensure the feasibility of the designed waveform.

[0191] S302. Based on the transmitted signal model, calculate the autocorrelation function and cross-correlation function of the waveform, and construct an optimization problem.

[0192] Based on the established signal model, the electronic device calculates the aperiodic autocorrelation function and cross-correlation function of each second signal to quantify the orthogonality performance of the waveform. Specifically, this includes: defining four core performance indicators: autocorrelation peak sidelobe level, autocorrelation integral sidelobe level, cross-correlation peak sidelobe level, and cross-correlation integral sidelobe level; constructing an optimization function (minimum value of the objective function) with the weighted sum of these four indicators as the objective function, and setting the weight coefficients (first weight) according to the specific application requirements of the radar system; and determining the constraints including constant mode constraints and discrete phase constraints to form a complete mathematical optimization problem framework.

[0193] S303. Train an unsupervised deep learning model based on a Transformer encoder to minimize the objective function.

[0194] Electronic devices solve waveform optimization problems by training a deep learning model based on a Transformer encoder. Specifically, this involves: constructing a model architecture that includes a multi-head self-attention mechanism and a feedforward neural network; extracting the global spatiotemporal dependency of the phase matrix through the self-attention mechanism, performing nonlinear feature transformation using the feedforward network, and finally outputting the probability score distribution of each phase; using the Gumbel-Softmax technique to achieve differentiable optimization under discrete phase constraints; and using the optimization objective (objective function) constructed in step two as the loss function for unsupervised learning during training, employing the Adam optimizer for gradient backpropagation and neural network parameter optimization to minimize the objective function, enabling the model to learn to generate high-quality waveforms from any initial phase.

[0195] Before training the model, the training samples need to be initialized, specifically including: initializing Ns training samples, each sample being an N×M matrix, i.e. Each element is a code symbol set. The index value of the random sample, i.e. This indicates that the phase of the nth coded pulse signal emitted by the mth array element is the nth coded symbol set. The system uses N elements and extracts Nb samples in batches from Ns training samples during training, without repetition, to perform gradient calculations and update neural network parameters sequentially, ensuring training stability and efficiency.

[0196] S304. Based on the trained model, carry out orthogonal waveform design.

[0197] The electronic device performs orthogonal waveform design based on the trained model. Specifically, this includes: inputting a randomly initialized phase matrix (first phase information) into the trained model, obtaining the optimized phase probability distribution (probability score) through forward propagation; obtaining soft phase values ​​through differentiable sampling based on the probability distribution, and then obtaining the final phase matrix (second phase information) that meets discrete constraints through hard quantization; generating the target signal emitted by each array element according to the optimized phase matrix, and verifying that it minimizes the objective function; and outputting a high-performance orthogonal waveform with low autocorrelation sidelobes and low cross-correlation values ​​to meet the practical application requirements of near-field MIMO radar systems.

[0198] Figure 4 A schematic diagram of an orthogonal waveform design device for a near-field MIMO radar system provided in this application is shown below. Figure 4 As shown, the orthogonal waveform design device 40 for near-field MIMO radar systems provided in this embodiment includes: an acquisition module 41, a first determination module 42, a second determination module 43, an input module 44, and a processing module 45, wherein...

[0199] Acquisition module 41 is used to acquire the first signal transmitted by the near-field multiple input multiple output MIMO radar system, the first signal including the second signal transmitted by each array element of the radar system;

[0200] The first determining module 42 is used to determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and to determine the objective function and constraint conditions associated with the orthogonality of the waveform based on the first signal, multiple autocorrelation values ​​and multiple cross-correlation values.

[0201] The second determining module 43 is used to determine the first phase information corresponding to the first signal based on the first signal, the objective function and the constraint conditions. The first phase information includes the time information of the first signal and the position information of each array element.

[0202] Input module 44 is used to input the objective function, constraints and multiple first phase information into the first model to obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics;

[0203] The processing module 45 is used to determine the target signal emitted by each array element based on the second phase information, and the target function corresponding to the target signal is minimized.

[0204] The orthogonal waveform design device for near-field MIMO radar systems provided in this application embodiment can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar, and will not be repeated here.

[0205] In one possible implementation, the second determining module 43 is specifically used for:

[0206] Based on the first signal, determine the initial phase information corresponding to the first signal;

[0207] Based on the constraints, the initial phase information is processed to obtain the first phase information corresponding to the first signal.

[0208] In one possible implementation, the second determining module 43 is specifically used for:

[0209] Based on the constraints, feature transformation processing is performed on multiple phase values ​​in the initial phase information to obtain the initial phase information after feature transformation processing.

[0210] Determine the spatial position information of each array element and the temporal position information of each second signal;

[0211] The initial phase information after feature transformation, the spatial position information of each array element, and the temporal position information of each second signal are fused to determine the first phase information corresponding to the first signal.

[0212] In one possible implementation, the input module 44 is specifically used for:

[0213] Based on the first model, multi-level feature fusion processing is performed on multiple first phase information to obtain the probability score of each element in the first phase information. The probability score is used to indicate the recommendation degree of each phase.

[0214] Based on the probability fractions of each element and the constraints, the second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained.

[0215] In one possible implementation, the input module 44 is specifically used for:

[0216] For any given first phase information;

[0217] Based on the first phase information, multiple first features corresponding to the first phase information are determined, and the first features are used to indicate the hierarchical information of the first phase information.

[0218] Multiple first features are aggregated to obtain aggregated second features, and feature extraction is performed on the second features to obtain the probability scores of each element in the first phase information.

[0219] In one possible implementation, the input module 44 is further configured to:

[0220] The probability scores of each element are sampled to obtain multiple sampled probability scores, which are used to indicate the recommendation level of each phase after standardization.

[0221] The second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained by multiplying the sampled multiple probability fractions and the phase information in the constraints.

[0222] In one possible implementation, the first determining module 42 is further configured to:

[0223] Based on multiple autocorrelation values, first data and second data are determined. The first data is used to indicate the autocorrelation peak sidelobe value in multiple second signals, and the second data is used to indicate the autocorrelation integral sidelobe value of multiple second signals.

[0224] Based on multiple cross-correlation values, a third data and a fourth data are determined. The third data is used to indicate the peak cross-correlation sidelobe value among the multiple second signals, and the fourth data is used to indicate the integral cross-correlation sidelobe value among the multiple second signals.

[0225] Determine the first weight corresponding to the first, second, third, and fourth data points;

[0226] Based on multiple first weights, the first data, second data, third data, and fourth data are weighted to obtain the objective function associated with the orthogonality of the waveform, and the constraint conditions are determined based on the first signal.

[0227] This embodiment provides an orthogonal waveform design device for near-field MIMO radar systems, which can execute the method provided in the above-described method embodiments. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0228] Figure 5 A schematic diagram of the structure of the electronic device provided in this application. Figure 5 As shown, the electronic device 50 provided in this embodiment includes at least one processor 51 and a memory 52. ​​Optionally, the electronic device 50 further includes a communication component 53. The processor 51, memory 52, and communication component 53 are connected via a bus.

[0229] In a specific implementation, at least one processor 51 executes computer execution instructions stored in memory 52, causing at least one processor 51 to perform the above-described method.

[0230] The specific implementation process of processor 51 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0231] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0232] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0233] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0234] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0235] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0236] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0237] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0238] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0239] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0240] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0241] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0242] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0243] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for designing orthogonal waveforms for near-field MIMO radar systems, characterized in that, include: Acquire a first signal transmitted by a near-field multiple-input multiple-output (MIMO) radar system, wherein the first signal includes second signals transmitted by each element of the radar system; Determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and based on the first signal, multiple autocorrelation values, and multiple cross-correlation values, determine the objective function and constraints associated with the orthogonality of the waveform; Based on the first signal, the objective function, and the constraint conditions, the first phase information corresponding to the first signal is determined. The first phase information includes the time information of the first signal and the position information of each array element. The objective function, the constraints, and multiple first phase information are input into the first model to obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics; Based on the second phase information, the target signal emitted by each array element is determined, and the target function corresponding to the target signal is minimized.

2. The method according to claim 1, characterized in that, Based on the first signal and the constraint conditions, the first phase information corresponding to the first signal is determined, including: Based on the first signal, determine the initial phase information corresponding to the first signal; Based on the constraints, the initial phase information is processed to obtain the first phase information corresponding to the first signal.

3. The method according to claim 2, characterized in that, Based on the constraints, the initial phase information is processed to obtain the first phase information corresponding to the first signal, including: According to the constraints, feature transformation processing is performed on multiple phase values ​​in the initial phase information to obtain the feature-transformed initial phase information; Determine the spatial position information of each array element and the temporal position information of each second signal; The initial phase information after feature transformation, the spatial position information of each array element, and the temporal position information of each second signal are fused to determine the first phase information corresponding to the first signal.

4. The method according to claim 1, characterized in that, By inputting the objective function, the constraints, and multiple first phase information into the first model, second phase information corresponding to the first signal associated with spatiotemporal features is obtained, including: Based on the first model, multi-level feature fusion processing is performed on the multiple first phase information to obtain the probability score of each element in the first phase information. The probability score is used to indicate the recommendation degree of each phase. Based on the probability scores of each element and the constraints, the second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained.

5. The method according to claim 4, characterized in that, Multi-level feature fusion processing is performed on the plurality of first phase information to obtain the probability score of each element in the first phase information, including: For any given first phase information; Based on the first phase information, a plurality of first features corresponding to the first phase information are determined, and the first features are used to indicate the hierarchical information of the first phase information. The multiple first features are subjected to feature aggregation processing to obtain aggregated second features, and feature extraction is performed on the second features to obtain the probability score of each element in the first phase information.

6. The method according to claim 4, characterized in that, Based on the probability scores of each element and the constraints, the second phase information corresponding to the first signal associated with the spatiotemporal characteristics is obtained, including: The probability scores of each element are sampled to obtain multiple sampled probability scores, which are used to indicate the recommendation level of each phase after standardization. The sampled probability scores are multiplied by the phase information in the constraints to obtain the second phase information corresponding to the first signal associated with the spatiotemporal characteristics.

7. The method according to claim 1, characterized in that, Based on the first signal, multiple autocorrelation values, and multiple cross-correlation values, determine the objective function and constraints associated with the orthogonality of the waveform, including: Based on the plurality of autocorrelation values, first data and second data are determined, wherein the first data is used to indicate the autocorrelation peak sidelobe value in the plurality of second signals, and the second data is used to indicate the autocorrelation integral sidelobe value of the plurality of second signals; Based on the plurality of cross-correlation values, a third data and a fourth data are determined, wherein the third data is used to indicate the peak cross-correlation sidelobe value of the plurality of second signals, and the fourth data is used to indicate the integral cross-correlation sidelobe value of the plurality of second signals; Determine the first weights corresponding to the first data, the second data, the third data, and the fourth data; Based on the plurality of first weights, the first data, the second data, the third data, and the fourth data are weighted to obtain an objective function associated with the orthogonality of the waveform, and the constraint conditions are determined based on the first signal.

8. An orthogonal waveform design device for near-field MIMO radar systems, characterized in that, include: The module comprises an acquisition module, a first determination module, a second determination module, an input module, and a processing module, wherein... The acquisition module is used to acquire a first signal transmitted by a near-field multiple-input multiple-output (MIMO) radar system, wherein the first signal includes a second signal transmitted by each array element of the radar system. The first determining module is used to determine the autocorrelation value of each second signal and the cross-correlation value between any two second signals, and to determine the objective function and constraint conditions associated with the orthogonality of the waveform based on the first signal, multiple autocorrelation values ​​and multiple cross-correlation values; The second determining module is used to determine the first phase information corresponding to the first signal based on the first signal, the objective function, and the constraint conditions. The first phase information includes the time information of the first signal and the position information of each array element. The input module is used to input the objective function, the constraint conditions, and multiple first phase information into the first model to obtain the second phase information corresponding to the first signal associated with the spatiotemporal features; The processing module is used to determine the target signal emitted by each array element based on the second phase information, wherein the target function corresponding to the target signal is minimized.

9. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

Images