A target positioning method and system based on a single reference station

Abstract

The present invention discloses a target positioning method and system based on a single reference station, which relates to the field of wireless positioning. The method includes generating a vortex electromagnetic wave carrying orbital angular momentum at the base station transmitter, generating a time-varying phase modulated transmission signal and a metasurface time-varying coding sequence through a programmable metasurface; the time-varying phase modulated transmission signal forms a multipath superposition effect in the propagation environment, and the receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end performs OAM mode separation on the mixed signal of the direct wave and the reflected wave through a spiral phase plate and an annular array to obtain signal components in different modes; and reconstructing an equivalent virtual array response based on the signal components and the metasurface time-varying coding sequence using a compressed sensing algorithm. By innovatively integrating orbital angular momentum mode separation with metasurface dynamic coding technology, the present invention achieves high-precision real-time direction finding in complex multipath environments under a single base station architecture.

Description

Technical Field

[0001] The present invention relates to the field of wireless positioning, and in particular to a target positioning method and system based on a single reference station. Background Art

[0002] In the field of wireless positioning, direction-finding technology based on array signal processing has been widely used. Traditional methods mainly rely on multi-base station collaboration or large-scale antenna arrays to achieve high-precision positioning. MIMO radar systems improve angular resolution through a multi-input multi-output architecture. In recent years, single-base station positioning technology has attracted attention due to its ease of deployment. It achieves target detection through digital beamforming, and its direction-finding accuracy can reach 0.5° (azimuth) at a distance of 100 meters. This type of technology relies on physical aperture expansion, and its performance degrades significantly in complex multipath environments.

[0003] The limitations of existing single-base station technology are mainly reflected in its insufficient dynamic multipath suppression capability. When strong reflectors are present, traditional spatial filtering algorithms have difficulty distinguishing between direct waves and multipath reflection signals, resulting in increased direction-finding errors. In urban canyon environments, multipath delay spread can reach 200ns, causing the TDOA-based positioning method to have an error of more than 5°. Although some studies have attempted to combine machine learning to optimize beam pointing, their high computational complexity and reliance on prior environmental data make it difficult to achieve real-time adaptive optimization. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a target positioning method based on a single reference station to solve the problem of real-time high-precision direction finding of a single base station in a complex multipath environment.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a target positioning method based on a single reference station, which includes generating a vortex electromagnetic wave carrying orbital angular momentum based on a transmitting end of the reference station, generating a time-varying phase modulated transmission signal and a metasurface time-varying coding sequence through a programmable metasurface;

[0008] The time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment. The receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end uses a spiral phase plate and a ring array to perform OAM mode separation on the mixed signal of the direct wave and the reflected wave to obtain signal components under different modes.

[0009] Based on the signal components and the metasurface time-varying coding sequence, the equivalent virtual array response is reconstructed using a compressed sensing algorithm.

[0010] Conduct three-dimensional spatial spectrum estimation on the virtual array response and calculate the target azimuth and elevation angles;

[0011] The target azimuth and elevation angles are fed back to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

[0012] As a preferred solution of the target positioning method based on a single reference station described in the present invention, wherein: based on the base station transmitter, a vortex electromagnetic wave carrying orbital angular momentum is generated, and a time-varying phase modulated transmission signal and a metasurface time-varying coding sequence are generated through a programmable metasurface, including the following steps:

[0013] The base station transmitter loads the center frequency and bandwidth parameters, and configures the geometric structure parameters and feeding phase of the dual circularly polarized helical antenna array;

[0014] The dual circularly polarized helical antenna array generates vortex electromagnetic waves carrying orbital angular momentum according to geometric structure parameters and feeding phase;

[0015] The programmable metasurface controller generates a metasurface time-varying coding sequence based on the Gold sequence, controlling 256 liquid crystal units to adjust the phase offset at a period of 1ms.

[0016] The vortex electromagnetic wave carrying orbital angular momentum is combined with time-varying phase encoding to obtain a time-varying scanning beam through the array factor.

[0017] The time-varying scanning beam is subjected to carrier modulation to generate a time-varying phase modulated transmission signal.

[0018] As a preferred solution of the target positioning method based on a single reference station described in the present invention, wherein: the time-varying phase modulated transmission signal forms a multipath superposition effect in the propagation environment, and the receiving end obtains a mixed signal of the direct wave and the reflected wave, including the following steps:

[0019] Generate channel impulse responses for the direct path and the reflected path based on the reflector position;

[0020] The time-varying phase modulated transmission signal is convolved with the channel impulse response to obtain a time domain signal superimposed with multipath effects;

[0021] The 16-element annular array captures the time domain signal with superimposed multipath effects and generates a time-space sampling matrix.

[0022] The spiral phase plate is used to perform step rotation, project the spatiotemporal sampling matrix, and output the separated modal matrix;

[0023] Matched filtering is performed on the orbital angular momentum mode in the separated modal matrix to obtain a mixed signal of the direct wave and the reflected wave.

[0024] As a preferred solution of the target positioning method based on a single reference station of the present invention, wherein: the receiving end performs OAM mode separation on the mixed signal of the direct wave and the reflected wave through a spiral phase plate and a ring array to obtain signal components in different modes, including the following steps:

[0025] Perform carrier frequency offset compensation and discrete Fourier transform on the spatiotemporal sampling matrix to output a frequency domain aligned signal;

[0026] The spiral phase plate is rotated in steps to perform OAM mode projection operations at 36 azimuth angles on the frequency-domain aligned signal to generate a mode-azimuth matrix.

[0027] Perform singular value decomposition on the modal-orientation matrix and reconstruct the enhanced modal matrix;

[0028] The orthogonal components of the vortex electromagnetic wave of the positive first-order orbital angular momentum and the vortex electromagnetic wave of the negative second-order orbital angular momentum are separated from the enhanced modal matrix to obtain signal components under different modes.

[0029] As a preferred solution of the target positioning method based on a single reference station of the present invention, wherein: based on the signal component and the metasurface time-varying coding sequence, the equivalent virtual array response is reconstructed using a compressed sensing algorithm, including the following steps:

[0030] The separated modal matrix and the hypersurface time-varying coding sequence are combined to construct a joint observation matrix through Kronecker product;

[0031] The K-SVD algorithm is used to train the overcomplete dictionary to optimize the sparse representation performance;

[0032] Use the compressed sensing algorithm to process the constrained sparse reconstruction problem and obtain the sparse coefficient vector;

[0033] The sparse coefficient vector and the complete dictionary are Toeplitz-corrected to obtain the equivalent virtual array response.

[0034] As a preferred solution of the target positioning method based on a single reference station of the present invention, wherein: performing three-dimensional spatial spectrum estimation on the virtual array response and calculating the target azimuth and elevation angle include the following steps:

[0035] The optimized virtual array response matrix is ​​divided into five overlapping sub-arrays, and the covariance matrix of each sub-array is processed to generate a full-rank covariance matrix;

[0036] Perform eigenvalue decomposition on the full-rank covariance matrix and estimate the number of sources using the MDL criterion;

[0037] Target azimuth and elevation are calculated in parallel on the GPU.

[0038] As a preferred solution of the target positioning method based on a single reference station of the present invention, wherein: the target azimuth and elevation angles are fed back to the reference station control unit, and the metasurface coding strategy and OAM mode power allocation are dynamically optimized, comprising the following steps:

[0039] Calculate the variance, deviation and virtual array SNR in the sliding window based on the target azimuth and elevation angles, and output the error feature vector;

[0040] The error characteristics are input into the DDPG network, which outputs the metasurface phase update step size and OAM power ratio;

[0041] The drive voltages of 256 liquid crystal cells are adjusted according to the metasurface phase update step size to generate a new coding sequence and dynamically optimize the metasurface coding strategy.

[0042] Power is allocated according to the OAM power ratio to calibrate the feed phase and dynamically optimize the OAM mode power distribution.

[0043] In a second aspect, the present invention provides a direction-finding and positioning system based on a single reference station, comprising a metasurface control module, which generates a vortex electromagnetic wave carrying orbital angular momentum based on the transmitting end of the reference station, and generates a time-varying phase modulated transmission signal and a metasurface time-varying coding sequence through a programmable metasurface;

[0044] In the modal separation module, the time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment. The receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end performs OAM mode separation on the mixed signal of the direct wave and the reflected wave through a spiral phase plate and a ring array to obtain signal components under different modes;

[0045] The virtual array reconstruction module reconstructs the equivalent virtual array response using a compressed sensing algorithm based on the signal components and the metasurface time-varying coding sequence;

[0046] The three-dimensional spatial spectrum estimation module performs three-dimensional spatial spectrum estimation on the virtual array response and calculates the target azimuth and elevation angle;

[0047] The optimization module feeds back the target azimuth and elevation angles to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

[0048] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the target positioning method based on a single reference station as described in the first aspect of the present invention is implemented.

[0049] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the target positioning method based on a single reference station as described in the first aspect of the present invention.

[0050] The beneficial effects of the present invention are as follows: by innovatively integrating orbital angular momentum modal separation and metasurface dynamic coding technology, high-precision real-time direction finding in complex multipath environments is achieved under a single base station architecture, vortex waves carrying different modes are generated through a dual circularly polarized antenna array, and time-varying coding is combined to form a transmission signal with three-dimensional space-time-modal characteristics, significantly improving the equivalent array aperture and angular resolution, and using a compressed sensing algorithm to reconstruct the virtual array. Through an over-complete dictionary and matrix correction, the computing efficiency and noise resistance performance are greatly improved, and an intelligent-driven closed-loop optimization system is constructed to achieve dynamic adjustment of metasurface coding and power allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 Flowchart of the target positioning method based on a single reference station.

[0053] Figure 2 Schematic diagram of the target positioning method system based on a single reference station.

[0054] Figure 3 Flowchart of time-varying phase modulation transmission signal.

[0055] Figure 4 Flowchart for separating orthogonal components. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0059] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a target positioning method based on a single reference station, comprising the following steps:

[0060] S1. Generate vortex electromagnetic waves carrying orbital angular momentum based on the base station transmitter, and generate time-varying phase modulation transmission signals and metasurface time-varying coding sequences through the programmable metasurface.

[0061] S1.1. The base station transmitter loads the center frequency and bandwidth parameters, and configures the geometric structure parameters and feeding phase of the dual circularly polarized helical antenna array.

[0062] Furthermore, the base station transmitter configures the geometric parameters and feeding phase of the dual circularly polarized helical antenna array based on the preset center frequency and bandwidth parameters. The dual circularly polarized helical antenna array generates vortex electromagnetic waves carrying orbital angular momentum based on the geometric parameters and feeding phase. The geometric parameters include the array element spacing, spiral radius, and axial length, and the feeding phase includes the initial excitation phase difference of each array element. The programmable metasurface controller generates a metasurface time-varying coding sequence based on the Gold sequence, controlling 256 liquid crystal units to adjust the phase offset at a fixed period, with the phase offset adjustment range of 0 to 2π. The vortex electromagnetic wave carrying orbital angular momentum and the metasurface time-varying coding sequence are synthesized through array factor calculation, and the synthesized time-varying scanning beam is carrier modulated using orthogonal frequency division multiplexing to ultimately generate a time-varying phase modulated transmission signal.

[0063] S1.2. The dual circularly polarized helical antenna array generates vortex electromagnetic waves carrying orbital angular momentum according to geometric structure parameters and feeding phase.

[0064] Furthermore, the dual circularly polarized helical antenna array generates vortex electromagnetic waves carrying orbital angular momentum based on geometric structure parameters and feeding phase, where the geometric structure parameters include array element spacing, spiral radius and axial length, and the feeding phase includes the initial excitation phase difference of each array element. The dual circularly polarized helical antenna array generates orbital angular momentum modes by adjusting the feeding phase difference of adjacent array elements. For example, the positive first-order orbital angular momentum mode requires the phase difference between adjacent array elements to be 2π / N, and the negative second-order orbital angular momentum mode requires the phase difference to be -4π / N, where N represents the total number of array elements. The radiation field distribution of the dual circularly polarized helical antenna array is determined by the product of the array factor and the unit radiation pattern. The spatial position and feeding phase of each array element are taken into account in the calculation of the array factor. The unit radiation pattern reflects the circular polarization characteristics of the helical antenna. The generated vortex electromagnetic wave has a spiral phase wavefront, the rotation direction of the phase wavefront is positively or negatively correlated with the order of the orbital angular momentum mode, and the amplitude distribution presents a ring feature.

[0065] S1.3. The programmable metasurface controller generates a metasurface time-varying coding sequence based on the Gold sequence, and controls 256 liquid crystal units to adjust the phase offset in a period of 1ms.

[0066] Furthermore, the programmable metasurface controller generates a metasurface time-varying coding sequence based on a 31-bit Gold sequence, with each symbol corresponding to a phase control instruction within a 1ms time period. The programmable metasurface controller maps the Gold sequence symbols to phase offsets for 256 liquid crystal cells. The phase offset adjustment range is 0 to 2π, with a phase resolution of 10 bits. The 256 liquid crystal cells synchronously update their phase states in a 1ms cycle, and the phase update instructions are transmitted to each liquid crystal cell driver circuit via a serial bus. The liquid crystal cell driver circuit adjusts the applied voltage based on the received phase instruction. The voltage is linearly related to the phase offset. For example, a phase offset of π corresponds to a 3.5V drive voltage. After phase adjustment, the liquid crystal cells form a specific wavefront modulation, enabling dynamic control of beam pointing.

[0067] S1.4. Combine the vortex electromagnetic wave carrying orbital angular momentum with the time-varying phase encoding and obtain the time-varying scanning beam through the array factor.

[0068] Furthermore, the vortex electromagnetic waves carrying orbital angular momentum are combined with time-varying phase encoding through array factor calculation. The array factor calculation takes into account the position coordinates of each element in the dual circularly polarized helical antenna array and the instantaneous phase offset generated by the programmable metasurface controller. The array factor is a discrete summation, consisting of the amplitude weighting and phase superposition of the radiation field of each element. The phase term is composed of the path difference phase determined by the spatial position of the array element and the time-varying phase offset applied by the programmable metasurface controller. The generation of the time-varying scanning beam is achieved by real-time updating the phase offset in the array factor calculation. The phase offset is updated from the programmable metasurface controller at a 1ms cycle. The beam scanning range is determined by the geometric dimensions and operating wavelength of the dual circularly polarized helical antenna array, and the scanning accuracy depends on the phase resolution of the programmable metasurface controller. The generated time-varying scanning beam maintains the orbital angular momentum characteristics of the original vortex electromagnetic wave while also having time-varying beam pointing capability.

[0069] S1.5. Carrier modulate the time-varying scanning beam to generate a time-varying phase modulated transmission signal.

[0070] Furthermore, the time-varying scanning beam is carrier-modulated using orthogonal frequency division multiplexing (OFDM). Orthogonal frequency division multiplexing (OFDM) modulation maps the baseband signal of the time-varying scanning beam to multiple orthogonal subcarriers. The orbital angular momentum characteristics and beam pointing information of the time-varying scanning beam are retained during the carrier modulation process. The center frequency of the modulated RF signal is determined by the center frequency parameter loaded by the base station transmitter. The subcarrier spacing of the OFDM modulation is based on the bandwidth parameter loaded by the base station transmitter, ensuring that each subcarrier maintains orthogonality. The time-varying phase modulated transmission signal after carrier modulation is amplified by a power amplifier and radiated by a dual circularly polarized spiral antenna array. The phase modulation information of the time-varying phase modulated transmission signal comes from the time-varying phase coding sequence generated by the programmable metasurface controller. The time-varying phase modulated transmission signal appears as an electromagnetic wave with a continuously changing phase in the time domain and presents a time-varying scanning beam pattern in the spatial domain.

[0071] S2. The time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment, and the receiving end obtains a mixed signal of the direct wave and the reflected wave.

[0072] S2.1. Generate channel impulse responses of the direct path and the reflected path based on the reflector position.

[0073] Furthermore, channel impulse responses for both the direct and reflected paths are generated based on the reflector's location. This information includes distance, azimuth, and altitude coordinates. The direct path's channel impulse response is determined by the straight-line distance between the transmitter and receiver, while the reflection path's channel impulse response calculation considers the geometric relationship between the reflector and the transmitter and receiver. The channel impulse response expressions for each path include the amplitude attenuation coefficient, propagation delay, and Doppler shift. The amplitude attenuation coefficient is based on the free-space propagation model and the reflection coefficient. The propagation delay is calculated by dividing the path length by the speed of light, and the Doppler shift is determined by the relative velocity of the transmitter and receiver and the carrier wavelength. The reflection path's channel impulse response also considers the effect of the reflector's material properties on signal amplitude and phase. The reflection coefficient is calculated using the Fresnel equation. The generated channel impulse responses for both the direct and reflected paths are used in subsequent multipath signal synthesis, with the duration of the channel impulse response covering the maximum multipath delay spread.

[0074] S2.2. Convolve the time-varying phase modulated transmit signal with the channel impulse response to obtain a time domain signal with multipath effects superimposed on it.

[0075] Furthermore, the time-varying phase modulated transmission signal and the channel impulse response are superimposed with multipath effects through discrete convolution operations. Discrete convolution operations use the overlap-preserving method to ensure operational efficiency. The sampling rate of the time-varying phase modulated transmission signal is consistent with the resolution of the channel impulse response, ensuring that the convolution operation results accurately reflect the actual propagation characteristics. During the convolution operation, each sampling point of the time-varying phase modulated transmission signal is multiplied and accumulated with each path component of the channel impulse response. The complex multiplication includes amplitude products and phase addition. The time domain signal superimposed with multipath effects retains the modulation format and bandwidth characteristics of the original time-varying phase modulated transmission signal, while also reflecting the delay spread and frequency selective fading introduced by multipath propagation. The time domain signal output by the convolution operation has a time length equal to the sum of the time-varying phase modulated transmission signal duration and the channel impulse response duration minus one. The signal amplitude fluctuates due to multipath interference.

[0076] Specifically, the expression is,

[0077] ;

[0078] in, For time The received signal, For time and The channel coefficient of delay, For events, is the transmission delay.

[0079] S2.3. Capture the time domain signal with superimposed multipath effects based on a 16-element ring array and generate a space-time sampling matrix.

[0080] Furthermore, a 16-element circular array uses synchronous sampling to capture time-domain signals with superimposed multipath effects. The element spacing of the 16-element circular array is set to half a wavelength to avoid grating lobe effects. Each element is connected to a receiving channel consisting of a low-noise amplifier, a bandpass filter, and an analog-to-digital converter (ADC), whose sampling rate meets the Nyquist criterion. The time-domain signals acquired simultaneously by the 16 elements are arranged in chronological order to form a space-time sampling matrix. The rows of the space-time sampling matrix correspond to the element numbers, and the columns correspond to the sampling time points. The space-time sampling matrix is ​​stored in a complex matrix format, with the real and imaginary parts representing the in-phase and quadrature components of the signal, respectively. The length of the time dimension of the space-time sampling matrix is ​​determined by the signal duration and sampling rate, while the spatial dimension is fixed at 16, corresponding to the number of elements. The resulting space-time sampling matrix preserves the time-domain waveform characteristics and spatial phase relationships of the original signal, allowing for subsequent modal separation.

[0081] S2.4. Use the spiral phase plate to perform step rotation, project the spatiotemporal sampling matrix, and output the separated modal matrix.

[0082] Furthermore, the spiral phase plate is mechanically rotated in 10° step angles, and a modal projection operation is performed on the spatiotemporal sampling matrix at each rotation position. The modal projection operation performs complex multiplication and accumulation on the spatiotemporal sampling matrix and the phase distribution corresponding to the current angle of the spiral phase plate. The phase distribution contains the topological charge characteristics and rotation angle information of the spiral phase plate. The projection operation output contains the complex amplitudes of the positive first-order orbital angular momentum mode and the negative second-order orbital angular momentum mode components, which are arranged in order of rotation angle to form a modal matrix. The rows of the modal matrix correspond to different rotation angles, the columns correspond to the separated orbital angular momentum mode types, and the matrix elements represent the projection intensity of each mode at the corresponding rotation angle. The generated modal matrix retains the spatial distribution characteristics of different orbital angular momentum modes in the original signal and is used for subsequent signal reconstruction and parameter estimation.

[0083] S2.5. Perform matched filtering on the orbital angular momentum mode in the separated modal matrix to obtain a mixed signal of the direct wave and the reflected wave.

[0084] Furthermore, the OAM modes in the separated modal matrix are subjected to matched filtering. The reference signal for the matched filtering uses the mathematical model of an ideal positive first-order OAM mode and a negative second-order OAM mode. The matched filtering operation is performed by calculating the cross-correlation function between each element of the modal matrix and the reference signal. The cross-correlation function calculation involves amplitude multiplication and phase alignment. The matched filtering of the positive first-order OAM mode outputs the direct wave component, while the matched filtering of the negative second-order OAM mode outputs the reflected wave component. After the matched filtering, the direct and reflected wave components are time-aligned and superimposed to form a mixed signal containing the direct and reflected waves. The mixed signal retains the time-domain waveform characteristics and multipath delay information of the original signal, and the signal amplitude reflects the attenuation characteristics of each path. During the matched filtering process, the orthogonality of the OAM modes ensures the effective separation of the direct and reflected wave components.

[0085] S3. At the receiving end, the OAM mode separation of the mixed signal of the direct wave and the reflected wave is performed through a spiral phase plate and a ring array to obtain signal components in different modes.

[0086] S3.1. Perform carrier frequency offset compensation and discrete Fourier transform on the space-time sampling matrix to output a frequency domain aligned signal.

[0087] Furthermore, the space-time sampling matrix undergoes digital down-conversion to compensate for carrier frequency offset. This compensation is achieved through complex mixing and low-pass filtering. The local oscillator frequency of the complex mixer aligns with the center frequency parameter loaded by the base station transmitter. The compensated signal is converted to a frequency domain representation via a fast Fourier transform (FFT). The number of FFT points is determined by the time length and sampling rate of the space-time sampling matrix. The frequency resolution of the frequency-domain aligned signal is determined by the number of FFT points. The spectral amplitude reflects the distribution of signal energy in the frequency domain. The frequency-domain aligned signal retains the frequency domain characteristics of the original signal, eliminating the spectral shift caused by carrier frequency offset, facilitating subsequent modal separation. The spectral range of the frequency-domain aligned signal is limited by the bandwidth parameter loaded by the base station transmitter. Frequency components outside the bandwidth range are suppressed by the digital filter.

[0088] S3.2. Perform step rotation on the spiral phase plate and perform OAM mode projection operation at 36 azimuth angles on the frequency domain aligned signal to generate a mode-azimuth matrix.

[0089] Furthermore, the spiral phase plate is mechanically rotated in 10-degree increments, and an orbital angular momentum modal projection operation is performed on the frequency-domain aligned signal at 36 azimuth positions. The projection operation at each azimuth position performs a complex inner product of the frequency-domain aligned signal with the phase distribution of the spiral phase plate at the current angle. The projection results are arranged in azimuth order to form a modal-azimuth matrix. The rows of the matrix correspond to the 36 azimuth angles from 0° to 350°, and the columns correspond to the separated orbital angular momentum mode types. The complex elements of the modal-azimuth matrix represent the energy distribution of each orbital angular momentum mode at different azimuth angles. The amplitude information of the matrix reflects the spatial characteristics of the signal, while the phase information preserves the propagation delay. The resulting modal-azimuth matrix is ​​used in subsequent singular value decomposition processing to ensure the effective separation of the orthogonal components of the different orbital angular momentum modes.

[0090] S3.3. Perform singular value decomposition on the modal-orientation matrix and reconstruct the enhanced modal matrix.

[0091] Furthermore, the modal-orientation matrix is ​​decomposed into the product of a left singular vector matrix, a singular value matrix, and a right singular vector matrix through singular value decomposition. After singular value decomposition, the components corresponding to the dominant singular values ​​are retained. The dominant singular value selection criterion is greater than 10% of the maximum singular value. The reconstructed enhanced modal matrix is ​​recalculated from the retained singular value components. The row dimension of the enhanced modal matrix is ​​consistent with the original modal-orientation matrix, and the column dimension is reduced to the number of effective modes. The amplitude information of the enhanced modal matrix reflects the concentrated distribution of signal energy in the spatial and modal dimensions, while the phase information retains the original propagation characteristics. During the reconstruction process, small singular values ​​corresponding to noise are suppressed, improving the signal-to-noise ratio and modal separation purity of the signal. The enhanced modal matrix is ​​used for subsequent orthogonal component extraction to ensure the effective separation of different orbital angular momentum modes.

[0092] S3.4. Separate the orthogonal components of the vortex electromagnetic wave of positive first-order orbital angular momentum and the vortex electromagnetic wave of negative second-order orbital angular momentum from the enhanced modal matrix to obtain signal components under different modes.

[0093] Furthermore, the enhanced modal matrix separates the vortex electromagnetic wave components of the positive first-order orbital angular momentum and the vortex electromagnetic wave components of the negative second-order orbital angular momentum through an orthogonal projection operation. This orthogonal projection operation is implemented based on the orthogonal properties of the orbital angular momentum modes. The positive first-order orbital angular momentum modal components are extracted by performing a dot product of the enhanced modal matrix with the reference mode of the positive first-order orbital angular momentum mode. The negative second-order orbital angular momentum modal components are extracted using the same method. The separated positive first-order orbital angular momentum modal components appear as independent complex signals in the time domain, preserving the amplitude and phase information of the original signals. The signal components under different modes have clear physical meanings. The positive first-order orbital angular momentum modal components mainly contain direct wave energy, while the negative second-order orbital angular momentum modal components concentrate reflected wave energy. The orthogonal separation process ensures that the crosstalk between the different orbital angular momentum modal components is less than -30dB, meeting the requirements of subsequent signal processing.

[0094] S4. Based on the signal components and the metasurface time-varying coding sequence, the compressed sensing algorithm is used to reconstruct the equivalent virtual array response.

[0095] S4.1. The separated modal matrix and the hypersurface time-varying coding sequence are combined through Kronecker product to construct a joint observation matrix.

[0096] Furthermore, the separated modal matrix and the metasurface time-varying coding sequence are combined to construct a joint observation matrix via a Kronecker product operation. This Kronecker product operation expands each element of the modal matrix by the outer product of the complete period of the metasurface time-varying coding sequence. The row dimension of the joint observation matrix is ​​equal to the number of rows of the separated modal matrix multiplied by the length of the metasurface time-varying coding sequence, and the column dimension remains the same as the column dimension of the separated modal matrix. The construction of the joint observation matrix preserves the orbital angular momentum modal information of the separated modal matrix and the spatiotemporal modulation characteristics of the metasurface time-varying coding sequence. The generated joint observation matrix reflects the joint sparsity of the signal in the modal and spatiotemporal domains, providing a complete observational basis for subsequent compressed sensing reconstruction. The joint observation matrix is ​​stored in the form of a complex matrix, with the real and imaginary parts corresponding to the in-phase and orthogonal components of the signal, respectively.

[0097] S4.2. Train the overcomplete dictionary through the K-SVD algorithm to optimize the sparse representation performance.

[0098] Furthermore, the K-SVD algorithm optimizes the sparse representation performance of the overcomplete dictionary by iteratively updating dictionary atoms and sparse coefficients. The K-SVD algorithm takes as input the set of column vectors of the joint observation matrix. Each iteration consists of a sparse coding phase and a dictionary update phase. The sparse coding phase uses the orthogonal matching pursuit algorithm to calculate the sparse coefficients for the current dictionary. The dictionary update phase optimizes each atom by performing singular value decomposition on each dictionary atom and its corresponding sparse coefficient, retaining the principal singular vectors as the updated dictionary atoms. The overcomplete dictionary has 1024 columns and 1024 rows, matching the dimensions of the joint observation matrix. Dictionary atoms represent typical angle-delay-Doppler joint features. The trained overcomplete dictionary reduces the sparse representation error of the joint observation matrix to below a preset threshold, improving the accuracy and stability of subsequent compressed sensing reconstruction. The K-SVD algorithm converges when the sparse representation error between consecutive iterations changes by less than 1% or the maximum number of iterations is reached.

[0099] S4.3. Use the compressed sensing algorithm to process the constrained sparse reconstruction problem and obtain a sparse coefficient vector.

[0100] Furthermore, the compressed sensing algorithm solves the constrained sparse reconstruction problem using an improved orthogonal matching pursuit process. At each iteration, the improved orthogonal matching pursuit process calculates the correlation coefficient between the observed signal residual and each atom in the overcomplete dictionary. This correlation coefficient is calculated using a complex inner product operation, taking into account both the amplitude and phase information of the signal. The dictionary atom with the largest correlation coefficient is selected for inclusion in the support set. After the support set is updated, the optimal sparse coefficient vector for the current support set is found using the least squares method. The least squares solution includes regularization to ensure numerical stability. The sparse coefficient vector outputted at each iteration is designed to monotonically decrease the reconstruction error, which is calculated as the L2 norm of the product of the observed signal, the dictionary, and the sparse coefficient vector. The improved orthogonal matching pursuit process terminates when the reconstruction error falls below 5% of the observed signal energy or when the number of iterations reaches 20% of the number of atoms in the overcomplete dictionary. The non-zero elements in the final output sparse coefficient vector correspond to the most relevant eigenatoms in the overcomplete dictionary, and the non-zero values ​​represent the weighted contribution of each eigenatom to the signal reconstruction. The sparsity of the sparse coefficient vector is controlled by the L1 norm constraint, ensuring the sparse representation characteristics of the signal in the overcomplete dictionary.

[0101] S4.4. Perform Toeplitz correction on the sparse coefficient vector and the complete dictionary to obtain the equivalent virtual array response.

[0102] Furthermore, the product of the sparse coefficient vector and the overcomplete dictionary undergoes a Toeplitz correction process to generate an equivalent virtual array response. This process rearranges the vector form of the product result into a symmetric Toeplitz matrix structure. The matrix dimensions of the equivalent virtual array response are determined by the preset virtual array scale, with the number of rows and columns corresponding to the number of equivalent array elements in the virtual array. During the Toeplitz correction process, the elements of the product result vector are filled into the matrix diagonal and parallel diagonal positions according to the recursive relationship of the Toeplitz matrix. The corrected equivalent virtual array response matrix maintains the direction of arrival information of the original signal while satisfying the Hermitian symmetry of the array covariance matrix. The amplitude of the equivalent virtual array response matrix reflects the spatial signal energy distribution, and the phase information preserves the propagation delay characteristics. The Toeplitz correction process improves the spatial smoothing performance of the virtual array, enabling subsequent spatial spectrum estimation to effectively resolve coherent signal sources.

[0103] S5. Perform three-dimensional spatial spectrum estimation on the virtual array response and calculate the target azimuth and elevation.

[0104] S5.1. Divide the optimized virtual array response matrix into five overlapping subarrays, process the covariance matrix of each subarray, and generate a full-rank covariance matrix.

[0105] Furthermore, the optimized virtual array response matrix is ​​divided into five overlapping subarrays using a sliding window approach. Each subarray contains 32 consecutive virtual array elements, and adjacent subarrays overlap by 28 elements. The covariance matrix of each subarray is calculated using complex conjugate transpose multiplication. Before the complex conjugate transpose multiplication, the subarray data is de-averaged to eliminate DC offset. The five subarray covariance matrices are arithmetic averaged to generate a full-rank covariance matrix. This arithmetic averaging process preserves the phase consistency of each subarray. The full-rank covariance matrix has a dimension of 32×32, and the matrix elements reflect the spatial correlation between the elements of the virtual array. The generated full-rank covariance matrix satisfies the positive Hermitian matrix property, providing input data that meets the mathematical requirements of the subsequent eigenvalue decomposition. The overlapping subarray division design ensures the decorrelation capability of the full-rank covariance matrix in coherent signal scenarios, improving the resolution performance of spatial spectrum estimation.

[0106] S5.2. Perform eigenvalue decomposition on the full-rank covariance matrix and estimate the number of information sources using the MDL criterion.

[0107] Furthermore, the full-rank covariance matrix is ​​decomposed into the product of an eigenvector matrix and a diagonal eigenvalue matrix through eigenvalue decomposition of the complex matrix. The Jacobi iteration algorithm is used during the eigenvalue decomposition process to ensure numerical accuracy. The MDL criterion estimates the true number of sources by calculating the information criterion under different hypothetical numbers of sources. The information criterion calculation includes a log-likelihood function term and a model complexity penalty term. The log-likelihood function term is calculated based on the ordered eigenvalues ​​obtained from the eigenvalue decomposition. The model complexity penalty term is proportional to the number of hypothetical sources and the array degrees of freedom. The MDL criterion iterates over all possible hypotheses about the number of sources and selects the hypothesis that minimizes the information criterion value as the final estimate of the number of sources. The eigenvector matrix output by the eigenvalue decomposition is sorted in descending order by eigenvalue. The first K eigenvectors constitute the signal subspace, and the remaining eigenvectors constitute the noise subspace. The estimated number of sources is used in subsequent three-dimensional spatial spectrum calculations to ensure that the dimensionality of the signal subspace matches the actual number of sources. The estimation performance of the MDL criterion is affected by the number of snapshots. When the number of snapshots is insufficient, regularization is used to improve the estimation stability.

[0108] S5.3. Calculate the target azimuth and elevation angles in parallel on the GPU.

[0109] Furthermore, a parallel computing architecture on the GPU enables a joint search for target azimuth and elevation. This architecture divides the 3D spatial spectrum calculation task into multiple thread blocks, each processing different angle intervals. Each thread block computes the spatial spectrum value within a specified azimuth-elevation interval. This spatial spectrum calculation is performed using a projection operation between the noise subspace and the steering vector. The steering vector calculation takes into account the geometry of the virtual array and the current search angle. The noise subspace data is read from global memory. During the parallel computing process, shared memory is used to cache frequently reused data, reducing global memory access latency. The 3D spatial spectrum matrix is ​​written to global memory, with row and column indices corresponding to the discretized azimuth and elevation search grids. The GPU's parallel computing architecture significantly improves the efficiency of 3D spatial spectrum calculations, achieving a speedup of two orders of magnitude compared to serial CPU computation. The peak position of the spatial spectrum matrix is ​​quickly located through subsequent parallel reduction operations, determining the target azimuth and elevation estimates.

[0110] Specifically, the target azimuth angle expression is:

[0111] ;

[0112] in, is the target azimuth, is the simplified steering vector in the azimuth dimension, is the noise subspace matrix after full-rank covariance matrix decomposition, is the conjugate transpose of the noise subspace matrix.

[0113] Specifically, the target elevation angle expression is,

[0114] ;

[0115] in, is the target elevation angle, Simplified steering vector for the elevation dimension.

[0116] S6. Feedback the target azimuth and elevation angles to the base station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

[0117] S6.1. Calculate the variance, bias, and virtual array SNR within the sliding window based on the target azimuth and elevation angles, and output the error feature vector.

[0118] Furthermore, the angle variance and angle deviation are calculated based on historical estimates of the target azimuth and elevation within a sliding window. The angle variance is obtained by statistically analyzing the dispersion of the angle estimates within the window, while the angle deviation is obtained by comparing the current estimate with the reference trajectory. The virtual array SNR is calculated as the ratio of the signal subspace energy to the noise subspace energy of the virtual array response matrix. The signal subspace energy is the sum of the first K largest eigenvalues, while the noise subspace energy is the average of the remaining eigenvalues. The error eigenvector combines the three eigenvalues—angle variance, angle deviation, and virtual array SNR—in a fixed order. The dimension of the error eigenvector is 3, and each element is normalized. The angle variance is calculated using an unbiased estimation form, with the denominator being the sliding window length minus one, which is set to 10 consecutive estimation moments. The angle deviation calculation requires pre-stored or received reference trajectory data, which can be provided via external input or historical smoothing results. The error eigenvector serves as the input state variable for subsequent reinforcement learning decisions, reflecting the real-time performance of the current direction-finding system.

[0119] S6.2. Input the error characteristics into the DDPG network and output the metasurface phase update step and OAM power ratio.

[0120] Furthermore, the error feature vector is input into the state input layer of the DDPG network, which contains a fully connected neural network structure to extract features from the error feature vector. The actor network of the DDPG network outputs two continuous action quantities: the metasurface phase update step and the OAM power ratio. The metasurface phase update step is limited to the range of 0.1π to 0.5π, and the OAM power ratio is limited to the range of 0.5 to 2.0. The critic network evaluates the Q value of the current state and action. The Q value calculation considers a weighted combination of angular variance, angular deviation, and virtual array SNR. The DDPG network training process uses an experience replay mechanism to store transition samples. A transition sample consists of a four-tuple consisting of state, action, reward, and next state. The target network stabilizes the training process through soft updates. The target network's update coefficient is set to 0.01. The metasurface phase update step and OAM power ratio output by the DDPG network directly control the parameter adjustment of the programmable metasurface and power divider, forming a closed-loop optimization. The reward function of the DDPG network is designed as a weighted sum of the inverse of the angle variance and the exponential decay term of the angle deviation. The reward function value increases with the improvement of direction finding accuracy.

[0121] S6.3. Adjust the driving voltage of 256 liquid crystal units according to the metasurface phase update step, generate a new coding sequence, and dynamically optimize the metasurface coding strategy.

[0122] Furthermore, the metasurface phase update step size is mapped to the drive voltage adjustment for each of the 256 liquid crystal cells using a voltage-phase conversion relationship. The slope of this voltage-phase conversion relationship is determined by the electro-optical properties of the liquid crystal material. The new drive voltage value for each liquid crystal cell is superimposed on the existing voltage, with the voltage adjustment range limited to the liquid crystal operating voltage range. The updated 256 drive voltage values ​​are synchronously written to the liquid crystal cell driver circuit via a serial bus, which converts the digital voltage instructions into analog drive signals. The updated drive voltages cause the liquid crystal cells to produce corresponding refractive index changes, with the refractive index change exhibiting a linear relationship with the drive voltage. The refractive index distribution of the 256 liquid crystal cells forms a new phase modulation pattern, which is arranged in a Gold sequence to generate a new coding sequence. The period of the new coding sequence remains constant at 1ms, and the sequence update process is strictly synchronized with the transmit signal time slot. The metasurface's coding strategy can be adjusted in real time during dynamic optimization, optimizing beam characteristics by periodically reconfiguring the coding sequence. The response speed of the liquid crystal cells ensures that phase state switching is completed within a 1ms period, meeting real-time control requirements.

[0123] S6.4. Allocate power according to the OAM power ratio to calibrate the feed phase and dynamically optimize the OAM mode power allocation.

[0124] Furthermore, the OAM power ratio is converted into specific power values ​​for the positive first-order orbital angular momentum mode and the negative second-order orbital angular momentum mode using a power allocation formula. The power allocation formula calculates the allocation ratio for each mode based on the total transmit power and the OAM power ratio. The feed network adjusts the excitation amplitude of each element in the dual circularly polarized helical antenna array, with the excitation amplitude proportional to the square root of the allocated power, enabling precise control of the power ratio. Feed phase calibration measures the phase difference between the positive first-order orbital angular momentum mode and the negative second-order orbital angular momentum mode. This phase difference measurement uses coherent detection to extract the pilot signal phase information. The calibrated feed phase compensation value is written to the beam control memory, which distributes the phase compensation data to each element feed channel via a serial bus. The closed-loop adjustment capability of the OAM mode power allocation during dynamic optimization adapts to channel variations by periodically updating the power ratio parameter. The power divider's response time is synchronized with the metasurface coding sequence update period, ensuring coordinated power adjustment and beamforming. The accuracy of feed phase calibration affects modal orthogonality, and the phase compensation residual error is controlled within π / 20 radians.

[0125] This embodiment further provides a direction finding and positioning system based on a single reference station, comprising:

[0126] The metasurface control module generates vortex electromagnetic waves carrying orbital angular momentum based on the base station transmitter, and generates time-varying phase modulation transmission signals and metasurface time-varying coding sequences through the programmable metasurface;

[0127] In the modal separation module, the time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment. The receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end performs OAM mode separation on the mixed signal of the direct wave and the reflected wave through a spiral phase plate and a ring array to obtain signal components under different modes;

[0128] The virtual array reconstruction module reconstructs the equivalent virtual array response using a compressed sensing algorithm based on the signal components and the metasurface time-varying coding sequence;

[0129] The three-dimensional spatial spectrum estimation module performs three-dimensional spatial spectrum estimation on the virtual array response and calculates the target azimuth and elevation angle;

[0130] The optimization module feeds back the target azimuth and elevation angles to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

[0131] This embodiment also provides a computer device, which is applicable to the case of a target positioning method based on a single reference station, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the target positioning method based on a single reference station proposed in the above embodiment.

[0132] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0133] This embodiment also provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the program implements the target positioning method based on a single reference station as proposed in the above embodiment. The 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.

[0134] In summary, the present invention innovatively integrates orbital angular momentum modal separation and metasurface dynamic coding technology to achieve high-precision real-time direction finding in complex multipath environments under a single base station architecture. Vortex waves carrying different modes are generated through a dual circularly polarized antenna array, and time-varying coding is combined to form a transmission signal with three-dimensional space-time-modal characteristics, which significantly improves the equivalent array aperture and angular resolution. The compressed sensing algorithm is used to reconstruct the virtual array, and the computational efficiency and noise resistance are greatly improved through over-complete dictionaries and matrix corrections. An intelligent-driven closed-loop optimization system is constructed to achieve dynamic adjustment of metasurface coding and power allocation.

[0135] 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 the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A target positioning method based on a single reference station, characterized by: include, Based on the base station transmitter, a vortex electromagnetic wave carrying orbital angular momentum is generated, and a time-varying phase modulated transmission signal and a metasurface time-varying coding sequence are generated through a programmable metasurface. The time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment. The receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end uses a spiral phase plate and a ring array to perform OAM mode separation on the mixed signal of the direct wave and the reflected wave to obtain signal components under different modes. Perform carrier frequency offset compensation and discrete Fourier transform on the spatiotemporal sampling matrix to output a frequency domain aligned signal; The spiral phase plate is rotated in steps to perform OAM mode projection operations at 36 azimuth angles on the frequency-domain aligned signal to generate a mode-azimuth matrix. Perform singular value decomposition on the modal-orientation matrix and reconstruct the enhanced modal matrix; The orthogonal components of the vortex electromagnetic wave of the positive first-order orbital angular momentum and the vortex electromagnetic wave of the negative second-order orbital angular momentum are separated from the enhanced modal matrix to obtain signal components under different modes; Based on the signal components and the metasurface time-varying coding sequence, the equivalent virtual array response is reconstructed using a compressed sensing algorithm. The separated modal matrix and the hypersurface time-varying coding sequence are combined to construct a joint observation matrix through Kronecker product; The K-SVD algorithm is used to train the overcomplete dictionary to optimize the sparse representation performance; Use the compressed sensing algorithm to process the constrained sparse reconstruction problem and obtain the sparse coefficient vector; The sparse coefficient vector and the complete dictionary are Toeplitz-corrected to obtain the equivalent virtual array response; Conduct three-dimensional spatial spectrum estimation on the virtual array response and calculate the target azimuth and elevation angles; The target azimuth and elevation angles are fed back to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

2. The target positioning method based on a single reference station according to claim 1, wherein: Based on the base station transmitter, a vortex electromagnetic wave carrying orbital angular momentum is generated, and a time-varying phase modulation transmission signal and a metasurface time-varying coding sequence are generated through a programmable metasurface, including the following steps: The base station transmitter loads the center frequency and bandwidth parameters, and configures the geometric structure parameters and feeding phase of the dual circularly polarized helical antenna array; The dual circularly polarized helical antenna array generates vortex electromagnetic waves carrying orbital angular momentum according to geometric structure parameters and feeding phase; The programmable metasurface controller generates a metasurface time-varying coding sequence based on the Gold sequence, controlling 256 liquid crystal units to adjust the phase offset at a period of 1ms. The vortex electromagnetic wave carrying orbital angular momentum is combined with time-varying phase encoding to obtain a time-varying scanning beam through the array factor. The time-varying scanning beam is subjected to carrier modulation to generate a time-varying phase modulated transmission signal.

3. The target positioning method based on a single reference station according to claim 2, wherein: The time-varying phase modulated transmission signal forms a multipath superposition effect in the propagation environment, and the receiving end obtains a mixed signal of the direct wave and the reflected wave, including the following steps: Generate channel impulse responses for the direct path and the reflected path based on the reflector position; The time-varying phase modulated transmission signal is convolved with the channel impulse response to obtain a time domain signal superimposed with multipath effects; The 16-element annular array captures the time domain signal with superimposed multipath effects and generates a time-space sampling matrix. The spiral phase plate is used to perform step rotation, project the spatiotemporal sampling matrix, and output the separated modal matrix; Matched filtering is performed on the orbital angular momentum mode in the separated modal matrix to obtain a mixed signal of the direct wave and the reflected wave.

4. The target positioning method based on a single reference station according to claim 3, wherein: The three-dimensional spatial spectrum of the virtual array response is estimated to calculate the target azimuth and elevation, including the following steps: The optimized virtual array response matrix is ​​divided into five overlapping sub-arrays, and the covariance matrix of each sub-array is processed to generate a full-rank covariance matrix; Perform eigenvalue decomposition on the full-rank covariance matrix and estimate the number of sources using the MDL criterion; Target azimuth and elevation are calculated in parallel on the GPU.

5. The target positioning method based on a single reference station according to claim 4, wherein: Feedback the target azimuth and elevation angles to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation, including the following steps: Calculate the variance, deviation and virtual array SNR in the sliding window based on the target azimuth and elevation angles, and output the error feature vector; The error characteristics are input into the DDPG network, which outputs the metasurface phase update step size and OAM power ratio; The drive voltages of 256 liquid crystal cells are adjusted according to the metasurface phase update step size to generate a new coding sequence and dynamically optimize the metasurface coding strategy. Power is allocated according to the OAM power ratio to calibrate the feed phase and dynamically optimize the OAM mode power distribution.

6. A direction-finding and positioning system based on a single reference station, based on the target positioning method based on a single reference station according to any one of claims 1 to 5, characterized in that: include, The metasurface control module generates vortex electromagnetic waves carrying orbital angular momentum based on the base station transmitter, and generates time-varying phase modulation transmission signals and metasurface time-varying coding sequences through the programmable metasurface; In the modal separation module, the time-varying phase modulated transmitted signal forms a multipath superposition effect in the propagation environment. The receiving end obtains a mixed signal of the direct wave and the reflected wave. The receiving end performs OAM mode separation on the mixed signal of the direct wave and the reflected wave through a spiral phase plate and a ring array to obtain signal components under different modes; The virtual array reconstruction module reconstructs the equivalent virtual array response using a compressed sensing algorithm based on the signal components and the metasurface time-varying coding sequence; The three-dimensional spatial spectrum estimation module performs three-dimensional spatial spectrum estimation on the virtual array response and calculates the target azimuth and elevation angle; The optimization module feeds back the target azimuth and elevation angles to the reference station control unit to dynamically optimize the metasurface coding strategy and OAM mode power allocation.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the target positioning method based on a single reference station according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the target positioning method based on a single reference station according to any one of claims 1 to 5 are implemented.

Images