Hybrid field full-space source perception method and system based on star-ris

Abstract

The application relates to the technical field of wireless sensing, in particular to a mixed-field full-space source sensing method and system based on STAR-RIS. A STAR-RIS-assisted mixed-field full-space source sensing model is constructed. The model comprises STAR-RIS arranged on the surface of a building, a plurality of signal sources distributed on both sides of the STAR-RIS and located in a near-field and far-field mixed area, and a base station for receiving and processing signals. The transmission coefficient and the reflection coefficient of the STAR-RIS are decoupled and normalized to reduce the model dimension, including fixing the phase difference of the transmission and reflection coefficients, making the amplitudes equal, and performing normalization processing on the transmission side channel. Based on the reduced model, the STAR-RIS is controlled to change the unit state according to a preset mode multiple times within the channel coherence time, the received signals are collected, and distance-angle two-dimensional parameter joint estimation is performed to simultaneously obtain the position information of each signal source in the full space. The application realizes high-precision and full-coverage positioning of near-field and far-field signal sources in the full space indoors and outdoors.

Description

Technical Field

[0001] This invention relates to the field of wireless sensing technology, and in particular to a hybrid field full-space source sensing method and system based on STAR-RIS, which is suitable for full-space high-precision position sensing in a hybrid near-field and far-field environment in 6G communication scenarios. Background Technology

[0002] Wireless sensing technology relies on efficient sensing algorithms and communication devices with high degrees of freedom in receiving signals. By analyzing direct, reflected, and scattered wireless signals, it senses target objects or environmental information, completing functions such as positioning, ranging, imaging, detection, and environmental reconstruction, thus realizing the perception and exploration of the physical world. As B5G / 6G systems move towards millimeter-wave / terahertz bands and deploy ultra-large-scale arrays, the number of antennas increases and the array aperture enlarges, extending the Rayleigh distance. The far-field planar wavefront assumption no longer satisfies the non-stationary propagation characteristics of near-field space; therefore, the spherical wavefront assumption should be used in near-field sensing. The range-angle coupling in the near-field spherical wavefront assumption brings a new dimension to wireless sensing. In recent years, with the interdisciplinary application of information science and materials science, reconfigurable intelligent surfaces (RIS) possess the ability to flexibly control electromagnetic wave propagation characteristics, creating a controllable electromagnetic environment. This not only enhances communication signal coverage and alleviates sensing interruption problems but also breaks through the upper limits of the degrees of freedom of existing devices.

[0003] However, the expansion of the near-field region does not mean the disappearance of the far-field region; the far field still dominates, and the near and far fields coexist in the channel. Furthermore, traditional RIS (Rapid Identification and Ranging) algorithms can only modulate electromagnetic waves on one side and cannot effectively handle beams on the opposite side, thus achieving full-space perception of indoor and outdoor environments remains a challenge. In addition, existing research has applied algorithms such as Newtonized Orthogonal Matching Pursuit (NOMP) to perception, but their models have not been specifically optimized for the unique signal modulation mechanism (transmission and reflection coupling) of Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surface (STAR-RIS) and for the estimation of two-dimensional (distance-angle) parameters in hybrid NF / FF scenarios. Summary of the Invention

[0004] To address the limitations of traditional reconfigurable smart surfaces that can only modulate electromagnetic waves on one side, thus failing to simultaneously sense signal sources throughout indoor and outdoor spaces; and the significant expansion of the near-field region in 6G scenarios using large-scale arrays, existing sensing models fail to uniformly and accurately characterize the physical characteristics of mixed near-field and far-field signal propagation, resulting in complex models and high parameter dimensionality; existing parameter estimation algorithms are not optimized for the aforementioned high-dimensional mixed-field models, limiting the accuracy and computational efficiency of full-space source location sensing. This invention proposes a mixed-field full-space source sensing method and system based on STAR-RIS. By constructing a unified STAR-RIS sensing model that integrates the transmission and reflection sides and is compatible with the propagation characteristics of near-field and far-field signals, a specific coefficient decoupling and normalization strategy is adopted to reduce the dimensionality of model parameters. An efficient two-dimensional joint parameter estimation algorithm adapted to this sensing model is designed, achieving high-precision, full-coverage positioning of near-field and far-field signal sources throughout indoor and outdoor spaces, effectively overcoming the grid quantization error of traditional algorithms and improving system processing efficiency.

[0005] To achieve the above objectives, the technical solution adopted is:

[0006] This invention provides a hybrid field full-space source sensing method based on STAR-RIS, comprising the following steps:

[0007] S1: Construct a STAR-RIS-assisted hybrid field full-space source sensing model. This model includes STAR-RIS deployed on the surface of a building, multiple signal sources distributed on both sides of the STAR-RIS and located in a mixed near-field and far-field region, and a base station for receiving and processing signals. The signals emitted by the signal sources reach the base station after being reflected or transmitted through the STAR-RIS. The STAR-RIS is controlled by the base station.

[0008] S2: Decouple and normalize the transmission and reflection coefficients of STAR-RIS to reduce the model dimension, including fixing the phase difference between the transmission and reflection coefficients, making their amplitudes equal, and normalizing the transmission side channel.

[0009] S3: Based on the dimensionality-reduced model, STAR-RIS is controlled to change its unit state multiple times in a preset mode during the channel coherence time, collect received signals, and perform joint estimation of range-angle two-dimensional parameters to simultaneously obtain the location information of each source in the entire space.

[0010] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S1, the STAR-RIS operates in energy segmentation mode, and the transmission coefficient of its m-th unit is... and reflection coefficient They are represented as follows: and ,in , These represent the amplitude modulation of the transmission coefficient and the amplitude modulation of the reflection coefficient of the m-th STAR-RIS array element, respectively. , These are the phase modulations of the transmission coefficient and reflection coefficient of the m-th STAR-RIS array element, respectively, and they satisfy the energy conservation constraint: and phase constraints: .

[0011] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S2, decoupling the transmission coefficient and reflection coefficient of STAR-RIS includes: fixing the phase difference between the transmission coefficient and the reflection coefficient to... and order Thus , It is the imaginary unit.

[0012] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S2, the normalization processing of the transmission coefficient and reflection coefficient of STAR-RIS includes: mirroring the arrival angle of the incident signal from the transmission side, i.e. ,in, This indicates the angle of arrival on the transmission side of the signal source to STAR-RIS. This indicates the angle of arrival on the reflection side from the source to STAR-RIS; and the channel is normalized. , This represents the transmission-side channel gain from the source to STAR-RIS. This represents the channel gain on the reflection side from the source to STAR-RIS. It is the imaginary unit.

[0013] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S3, STAR-RIS changes its transmission and reflection coefficient states according to a preset state matrix that satisfies the incoherence characteristic within the coherence time. Next, among them It is greater than or equal to half of the number of STAR-RIS array elements M.

[0014] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, the joint parameter estimation in step S3 further includes the following sub-steps:

[0015] S31: Greedy search: Based on the residual signal, a greedy search is performed on a preset two-dimensional parameter discrete grid to obtain the initial estimates of the distance and angle of arrival of each source.

[0016] S32: Single refinement: Starting from the initial estimate, Newton's iteration method is used to perform local optimization in the continuous parameter space to eliminate grid quantization error;

[0017] S33: Iterative refinement: Interference cancellation and Newton iteration re-optimization are performed sequentially on the estimated parameters of multiple sources to improve the accuracy of multi-source estimation;

[0018] S34: Gain Update: Based on Optimized Distance and Angle Parameters The joint channel gain estimate of all sources is updated using the least squares method. .

[0019] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S32, the execution condition of the Newton iteration method is that the second-order reciprocal matrix of the optimization objective function is a negative definite matrix, and the optimization objective function is to minimize ,in This represents the total received signal after vectorization. For channel gain, Distance and angle The corresponding hybrid steering vector.

[0020] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, further, in step S33, the execution is refined cyclically. Each time, the residual signal is updated as follows: ,in, Let p be the updated residual signal vector corresponding to the p-th source. This represents the total received signal after vectorization. The estimated channel gain for the l-th source is... This is the hybrid steering vector of the l-th source calculated based on the current estimated parameters.

[0021] According to the STAR-RIS-based hybrid field full-space source sensing method of the present invention, in step S3, the location information of the information source is calculated by the following formula: ,in, Let L be the two-dimensional physical location coordinates of the l-th information source. Let l be the estimated time delay index for the l-th source. Let the estimated angle of arrival be for the l-th source. For time-delay resolution, It is the speed of light.

[0022] Furthermore, the present invention also provides a hybrid field full-space source sensing system based on STAR-RIS, comprising:

[0023] STAR-RIS, deployed on the surface of buildings, is used to simultaneously modulate transmitted and reflected signals;

[0024] Multiple signal sources are distributed on both sides of STAR-RIS and are located in a mixed region of near and far fields;

[0025] The base station is used to receive signals reflected or transmitted by STAR-RIS and perform the sensing methods described above to achieve high-precision positioning and parameter estimation of information sources throughout the entire space.

[0026] The beneficial effects achieved by adopting the above technical solution are:

[0027] 1. Achieved full-space perception: By utilizing the physical characteristics of STAR-RIS to simultaneously transmit and reflect electromagnetic waves, this invention breaks through the limitation of traditional RIS that can only serve one side of the space, and enables synchronous perception of sources to be sensed on both sides of the deployment surface (such as indoor and outdoor).

[0028] 2. Improved sensing accuracy and robustness in mixed-field environments: The unified mixed-field full-space source sensing model proposed in this invention is more closely aligned with the actual scenario of near-field region expansion under 6G massive MIMO antenna array deployment, fundamentally improving the model's accuracy. Combined with an iterative optimization process specifically designed for two-dimensional parameter estimation, it effectively overcomes the grid quantization error problem of traditional algorithms, enabling high-precision positioning in complex near-field and far-field mixed environments.

[0029] 3. Optimized system processing efficiency: This invention effectively reduces the dimensionality of the signal model by decoupling the inherent coupling characteristics of STAR-RIS. The proposed estimation process combines low complexity with high accuracy, effectively overcoming the discretization error of traditional grid-based algorithms, and achieving higher computational efficiency while ensuring sensing performance. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. The drawings are merely illustrative of some embodiments of the present invention and are not intended to limit the scope of the present invention to all embodiments.

[0031] Figure 1 This is a flowchart illustrating the hybrid field full-space source sensing method based on STAR-RIS according to an embodiment of the present invention.

[0032] Figure 2 This is a schematic diagram of the structure of the STAR-RIS-based hybrid field full-space source sensing system according to an embodiment of the present invention. Detailed Implementation

[0033] The exemplary solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art.

[0034] To address the limitations of traditional smart surfaces in achieving full-space sensing and the inaccuracy of full-space hybrid field models, this invention discloses a hybrid field full-space source sensing method based on STAR-RIS, such as... Figure 1 As shown, it includes the following steps:

[0035] Step S1: Construct a STAR-RIS-assisted hybrid field full-space source sensing model. This model includes STAR-RIS deployed on the surface of a building, multiple signal sources (users) distributed on both sides of the STAR-RIS (such as indoors and outdoors) and located in a mixed area of ​​near field and far field, and a base station for receiving and processing signals; the signals emitted by the signal sources reach the base station after being reflected or transmitted through the STAR-RIS; the STAR-RIS is controlled by the base station.

[0036] Consider a STAR-RIS-assisted communication scenario where there is no line-of-sight (LoS) link between the signal source and the base station. The signal from the source reaches the base station equipped with B antennas via reflection or transmission through the STAR-RIS, and the system operates in uplink transmission mode. The STAR-RIS is equipped with M elements, forming a uniform linear array (ULA). The transmission coefficient (TC) and reflection coefficient (RC) are defined as follows: and This involves how the TC control signal is transmitted through STAR-RIS, and how the RC control signal is reflected.

[0037] STAR-RIS operates under an energy splitting (ES) model, controlling bidirectional incoming waves by adjusting the transmission and reflection coefficients. The transmission and reflection coefficients are defined as follows:

[0038]

[0039] in, and These represent the amplitude and phase shift modulation of the m-th STAR-RIS element, respectively, controlled by the base station. The amplitude components of RC and TC satisfy the law of conservation of energy, i.e. The phase part satisfies STAR-RIS can rapidly change TC and RC within the coherence time, enabling the receiver to sense signals arriving at STAR-RIS. The proposed model has multi-user sensing capability, with user separation depending on their unique spatiotemporal characteristics. The baseband signal transmitted by the source can be expressed as:

[0040]

[0041] Where N is the number of subcarriers, n represents the position of the time-domain sampling point, and Ts = T / N is the sampling rate. It is the symbol on the k-th subcarrier. When the baseband signal is modulated onto the carrier frequency and shaped by filtering, it can be represented as:

[0042]

[0043] in, For the effective duration, The duration of the cyclic prefix, where t represents the continuous time variable. It is the carrier frequency. It is the subcarrier spacing. It is a rectangle function.

[0044] To achieve hybrid field full-space sensing, this scheme constructs a STAR-RIS-assisted hybrid field full-space source sensing model, in which STAR-RIS is equipped with a large number of elements, thereby extending the near-field region. In this model, the channels from the source to STAR-RIS are:

[0045]

[0046] in, This represents the channel between the l-th user and STAR-RIS. This represents the time delay from the signal sent by the l-th user to the STAR-RIS reference element. The time delay of the signal arriving at different array elements can be specifically expressed as:

[0047]

[0048] Where m represents the index number of the STAR-RIS array element, and c represents the speed of light. This represents the physical distance of the m-th array element relative to the reference array element. and Let represent the direction of arrival and distance between STAR-RIS and the l-th user, respectively. This time delay variable, which considers both far-field and near-field factors, is crucial for far- and near-field hybrid modeling. Therefore, the signal arriving at the m-th element of STAR-RIS can be expressed as:

[0049]

[0050] At the same time, a similar time delay exists from the m-th element of STAR-RIS to the b-th antenna of the base station, which can be expressed as:

[0051]

[0052] in, and These are the departure angle of STAR-RIS and the arrival angle of the base station, respectively, and the reference time delay of the signal from STAR-RIS to the base station, denoted as . , is a function of the distance between the two nodes. Since the positions of the base station and STAR-RIS cannot be modified in real time, we assume the angle... and ,as well as This is known. Therefore, the signal reaching the b-th antenna of the base station is:

[0053]

[0054] in, It is the phase caused by time delay, that is, the phase from the m-th element of STAR-RIS to the b-th antenna of the base station. This indicates the channel from STAR-RIS to the base station. Represents the nth element in matrix A Line number Column elements, It is additive white Gaussian noise.

[0055] To achieve real-time sensing during communication, it is necessary to rely on received signals. Estimate the DOA at STAR-RIS and simultaneously sense the distance to the corresponding user. However, this model includes not only the distance dimension and the DOA dimension, but also the STAR-RIS state change dimension ( In this model, the transmission coefficient and reflection coefficient are coupled, making it difficult to directly use traditional methods for estimating high-dimensional models. Therefore, it is necessary to reduce the dimensionality of the model and simultaneously estimate the two-dimensional distance-direction variables.

[0056] Step S2: Decouple and normalize the transmission and reflection coefficients of STAR-RIS to reduce the model dimension. This includes fixing the phase difference between the transmission and reflection coefficients, making their amplitudes equal, and normalizing the transmission side channel to simplify the received signal expression.

[0057] Due to the constraint of the law of conservation of energy, the phase difference between TC and RC must be π / 2 or 3π / 2. To eliminate the difference between RC and TC, this phase difference is fixed at... and order To ensure that the observation accuracy on both sides is equal, thereby... , The imaginary unit, , These represent the amplitude modulation of the transmission coefficient and the amplitude modulation of the reflection coefficient of the m-th STAR-RIS array element, respectively.

[0058] Assuming the angle of arrival of the signal reaching STAR-RIS from the transmission side is mirror-image, this means ,in, This indicates the angle of arrival on the transmission side of the signal source to STAR-RIS. This indicates the angle of arrival on the reflection side of the signal source to STAR-RIS. Furthermore, to reconcile the differences in reflection and transmission coefficients on opposite sides, [the following is used]: To normalize this difference, This represents the transmission-side channel gain from the source to STAR-RIS. This represents the channel gain on the reflection side from the source to STAR-RIS. The imaginary unit. At this sampling point, the received signal can be expressed as:

[0059]

[0060] in, This represents the vector of signals received by all antennas of the base station at the i-th sampling time. This represents the array response matrix from STAR-RIS to the base station. This represents the state matrix of STAR-RIS. Represents the cascaded channel matrix; This represents the transmitted signal vector of all users at the i-th sampling time. Let represent the additive noise vector at the i-th sampling time. Let represent the manifold vector from the user to STAR-RIS; for the l-th source, its manifold vector is:

[0061]

[0062] Next, we make the following approximation:

[0063]

[0064] in, The propagation delay from the STAR-RIS reference element to the base station reference antenna. The time delay from the signal sent by the l-th user to the STAR-RIS reference element. For the discrete delay index of the l-th user, Let be the time-delay resolution. Under this approximation, the distance can be re-expressed as... ,in , This represents the reference delay index from STAR-RIS to the base station, while It can be represented as:

[0065]

[0066] Furthermore, STAR-RIS switches states according to a given period, assuming the channel remains constant within that period. To simplify the analysis and demonstrate practical applicability, a widely used, pre-stored Hadamard matrix satisfying the mutual incoherence property (MIP) is used as the state matrix. The STAR-RIS switches a total of... Second-rate, If the number of elements M in the STAR-RIS array is greater than or equal to half, the corresponding state for each switch can be represented as follows: The received signal can then be represented as:

[0067]

[0068] in, It is the concatenated channel matrix of the l-th source. This is the STAR-RIS state matrix at the nth state transition. It is an additive white Gaussian noise matrix. It is the signal vector transmitted by the l-th source. It can be seen that, for the entire space, only L users are sparse. The received signal is converted into:

[0069]

[0070] in, .

[0071] Step S3: Based on the dimensionality-reduced model, STAR-RIS is controlled to change its cell state multiple times according to a preset mode within the channel coherence time, collecting received signals and performing joint estimation of range-angle two-dimensional parameters to simultaneously obtain the location information of each source in the entire space. The joint parameter estimation specifically includes four sub-steps: greedy search, single refinement, cyclic refinement, and gain update.

[0072] Step S31: Greedy search: Based on the residual signal, perform a greedy search on a preset two-dimensional parameter discrete grid to obtain the initial estimates of the distance and angle of arrival of each source.

[0073] First, set the number of iterations to L, and the loop variable to... First, remove the values ​​obtained from the previous loop to obtain the residual:

[0074]

[0075] in This represents the residual signal vector at the k-th iteration. Represents the received signal vector. This represents the estimated channel gain for the l-th user. This is the hybrid steering vector of the l-th source calculated based on the current estimated parameters.

[0076] Then, the maximum likelihood method is used to identify the closest result in a predefined grid:

[0077]

[0078] This process is similar to traditional OMP, but a finer search is needed after obtaining the estimate. A smaller grid is set around the coarse result to improve accuracy. The size of the fine grid is 1 / q of the size of the coarse grid, and there are a total of A grid like this. In obtaining and After obtaining the estimated value, the channel can be estimated accordingly.

[0079]

[0080] Step S32: Single refinement: Starting from the initial estimate, Newton's iteration method is used to perform local optimization in the continuous parameter space to eliminate grid quantization error.

[0081] To eliminate estimation errors introduced by the grid, refinement based on Newton's method is used to optimize the estimation results. The optimization objective is to minimize:

[0082]

[0083] in, This represents the total received signal after vectorization. For channel gain, Distance and angle The corresponding hybrid steering vector. Expanding the above equation and eliminating the constant term yields:

[0084]

[0085] in, For the simplified objective function, To obtain the real part of the complex number. Subsequently, the estimated value obtained through greedy search can be updated in the following way:

[0086]

[0087] in and Let each represent the objective function of formula (19). The corresponding first-order partial derivative and Hessian matrix are in the following form:

[0088]

[0089] The first and second partial derivatives can be obtained from the following equation:

[0090]

[0091] The above process is executed Second-rate.

[0092] Step S33: Iterative refinement: For the estimated parameters of multiple sources, interference cancellation and Newton iteration re-optimization are performed sequentially to improve the accuracy of multi-source estimation.

[0093] Based on the results obtained, The residual corresponding to each result is updated using the following formula:

[0094]

[0095] in, Let p be the updated residual signal vector corresponding to the p-th source. This represents the total received signal after vectorization. The estimated channel gain for the l-th source is... This is the guidance vector for the l-th source. Subsequently, it is refined using the same method as in the single-step refinement to improve estimation accuracy. Iterative refinement requires execution of... The loop continues.

[0096] Step S34: Gain Update: Based on Optimized Distance and Angle Parameters The joint channel gain estimate of all sources is updated using the least squares method. .

[0097] The angle and time delay estimates obtained from the preceding steps can be used to further update the channel coefficients. By multiplying the received signal by the subspace spanned by the estimation results using the least squares method, we obtain:

[0098]

[0099] It should be noted that, in order to ensure the convergence of the algorithm based on Newton's method, a validity check must be performed before executing formula (20), which is premised on satisfying the following conditions: ( (where is a negative definite matrix); therefore, regardless of whether it is a single refinement or a cyclic refinement scenario, the update rule corresponding to formula (20) can only be executed when this condition is met.

[0100] After obtaining all the estimation results, the location information of the corresponding information source can be obtained.

[0101]

[0102] in, Let L be the two-dimensional physical location coordinates of the l-th information source. Let l be the estimated time delay index for the l-th source. Let the estimated angle of arrival be for the l-th source. For time-delay resolution, It is the speed of light.

[0103] The method of this invention can avoid the grid error of traditional OMP and its improved algorithms, and can achieve high-precision, full-coverage sensing effect.

[0104] This invention also discloses a hybrid field full-space source sensing system based on STAR-RIS, such as... Figure 2 As shown, it includes:

[0105] STAR-RIS, deployed on the surface of buildings, is used to simultaneously modulate transmitted and reflected signals;

[0106] Multiple signal sources are distributed on both sides of STAR-RIS and are located in a mixed region of near and far fields;

[0107] The base station is used to receive signals reflected or transmitted by STAR-RIS and perform the sensing methods described above to achieve high-precision positioning and parameter estimation of information sources throughout the entire space.

[0108] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A hybrid field full-space source sensing method based on STAR-RIS, characterized in that, Includes the following steps: S1: Construct a STAR-RIS-assisted hybrid field full-space source sensing model. This model includes STAR-RIS deployed on the surface of a building, multiple signal sources distributed on both sides of the STAR-RIS and located in a mixed near-field and far-field region, and a base station for receiving and processing signals. The signals emitted by the signal sources reach the base station after being reflected or transmitted through the STAR-RIS. The STAR-RIS is controlled by the base station. S2: Decouple and normalize the transmission and reflection coefficients of STAR-RIS to reduce the model dimension, including fixing the phase difference between the transmission and reflection coefficients, making their amplitudes equal, and normalizing the transmission side channel. S3: Based on the dimensionality-reduced model, STAR-RIS is controlled to change its unit state multiple times in a preset mode during the channel coherence time, collect received signals, and perform joint estimation of range-angle two-dimensional parameters to simultaneously obtain the location information of each source in the entire space.

2. The STAR-RIS-based hybrid field full-space source sensing method according to claim 1, characterized in that, In step S1, the STAR-RIS operates in energy-segmented mode, and the transmission coefficient of its m-th unit is... and reflection coefficient They are represented as follows: and ,in , These represent the amplitude modulation of the transmission coefficient and the amplitude modulation of the reflection coefficient of the m-th STAR-RIS array element, respectively. , These are the phase modulations of the transmission coefficient and reflection coefficient of the m-th STAR-RIS array element, respectively, and they satisfy the energy conservation constraint: and phase constraints: .

3. The STAR-RIS-based hybrid field full-space source sensing method according to claim 2, characterized in that, In step S2, decoupling the transmission coefficient and reflection coefficient of STAR-RIS includes: fixing the phase difference between the transmission coefficient and the reflection coefficient to a certain value. and order Thus , It is the imaginary unit.

4. The STAR-RIS-based hybrid field full-space source sensing method according to claim 1, characterized in that, In step S2, normalizing the transmission and reflection coefficients of STAR-RIS includes mirroring the angle of arrival of the incident signal from the transmission side. ,in, This indicates the angle of arrival on the transmission side of the signal source to STAR-RIS. This indicates the angle of arrival on the reflection side from the source to STAR-RIS; and the channel is normalized. , This represents the transmission-side channel gain from the source to STAR-RIS. This represents the channel gain on the reflection side from the source to STAR-RIS. It is the imaginary unit.

5. The STAR-RIS-based hybrid field full-space source sensing method according to claim 1, characterized in that, In step S3, STAR-RIS changes its transmission and reflection coefficient states according to a preset state matrix that satisfies the incoherence characteristic within the coherence time. Next, among them It is greater than or equal to half of the number of STAR-RIS array elements M.

6. The STAR-RIS-based hybrid field full-space source sensing method according to claim 1, characterized in that, The joint parameter estimation in step S3 includes the following sub-steps: S31: Greedy search: Based on the residual signal, a greedy search is performed on a preset two-dimensional parameter discrete grid to obtain the initial estimates of the distance and angle of arrival of each source. S32: Single refinement: Starting from the initial estimate, Newton's iteration method is used to perform local optimization in the continuous parameter space to eliminate grid quantization error; S33: Iterative refinement: Interference cancellation and Newton iteration re-optimization are performed sequentially on the estimated parameters of multiple sources to improve the accuracy of multi-source estimation; S34: Gain Update: Based on Optimized Distance and Angle Parameters The joint channel gain estimate of all sources is updated using the least squares method. .

7. The STAR-RIS-based hybrid field full-space source sensing method according to claim 6, characterized in that, In step S32, the execution condition of the Newton iteration method is that the second-order reciprocal matrix of the optimization objective function is a negative definite matrix, and the optimization objective function is to minimize... ,in This represents the total received signal after vectorization. For channel gain, Distance and angle The corresponding hybrid steering vector.

8. The STAR-RIS-based hybrid field full-space source sensing method according to claim 6, characterized in that, In step S33, the execution is refined cyclically. Each time, the residual signal is updated as follows: ,in, Let p be the updated residual signal vector corresponding to the p-th source. This represents the total received signal after vectorization. The estimated channel gain for the l-th source is... This is the hybrid steering vector of the l-th source calculated based on the current estimated parameters.

9. The STAR-RIS-based hybrid field full-space source sensing method according to claim 1, characterized in that, In step S3, the location information of the information source is calculated using the following formula: ,in, Let L be the two-dimensional physical location coordinates of the l-th information source. Let l be the estimated time delay index for the l-th source. Let the estimated angle of arrival be for the l-th source. For time-delay resolution, It is the speed of light.

10. A hybrid field full-space source sensing system based on STAR-RIS, characterized in that, include: STAR-RIS, deployed on the surface of buildings, is used to simultaneously modulate transmitted and reflected signals; Multiple signal sources are distributed on both sides of STAR-RIS and are located in a mixed region of near and far fields; A base station is used to receive signals reflected or transmitted by STAR-RIS and to execute the sensing method as described in any one of claims 1-9 to achieve high-precision positioning and parameter estimation of information sources throughout the entire space.

Images