User time delay estimation method for IRS-assisted multicarrier ISAC system

Abstract

The invention discloses a user time delay estimation method of an IRS-assisted multicarrier ISAC system. In the method, a plurality of users send pilot signals, and each IRS reflects the user signals and transmits the user signals to a remote radio access unit (RRH); the RRH uploads the received signal to a centralized baseband processing unit pool (BBU) through a forward link; as the parameters of the link from the IRS to the RRH can be acquired in advance, the method carries out time delay estimation on the link from the user to the IRS; the BBU constructs an observation matrix of a pilot signal on a joint dimension of a plurality of subcarriers and an IRS unit and utilizes a multi-snapshot effect provided by a plurality of receiving antennas to realize high-precision time delay estimation from a user to an IRS link; according to the method, sparse structures of a frequency domain and a space domain are fully utilized, and the time delay resolution and the estimation precision are effectively improved, so that the user positioning capability in a complex environment is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-carrier ISAC system, and particularly relates to a cloud access network intelligent reflecting surface assisted multi-carrier ISAC system user time delay estimation method. BACKGROUND

[0002] With the increasing demand for high-precision positioning and ranging capability of emerging applications such as intelligent manufacturing, Internet of Vehicles and unmanned systems, integrated sensing and communication (ISAC) technology has gradually become a core function of the next generation of wireless systems. The cloud access network (C-RAN) architecture realizes the unified scheduling of communication and sensing resources through centralized baseband units (BBU) and distributed remote radio units (RRH), and has large-scale deployment capability. However, high-precision time delay estimation relies on large bandwidth signals, which exacerbates the pressure and processing complexity of the front-haul link in the C-RAN system, restricting its scalability in sensing tasks.

[0003] To improve the ranging accuracy and time resolution capability of the system in complex environments, the present application introduces an intelligent reflecting surface (IRS) as a relay sensing node. As a low-cost, low-power and easy-to-deploy new technology, IRS can intelligently reconstruct the signal propagation path by adjusting the phase of its passive reflecting elements. By constructing high-quality reflection links, IRS can significantly improve the quality of echo signals while enhancing signal coverage and alleviating non-line-of-sight (NLOS) obstruction problems, thereby improving the accuracy of target time delay estimation. SUMMARY

[0004] The present application is a multi-carrier ISAC system that can achieve positioning of target users with the assistance of intelligent reflecting surfaces in the context of cloud access networks. By constructing an observation matrix of pilot signals in the joint dimension of multiple subcarriers and IRS elements, and utilizing the snapshots provided by the receiving antennas, high-precision time delay estimation of the user-to-IRS link is achieved.

[0005] The present application targets a typical scenario: the user is in the RRH shadow area and actively sends known multi-carrier OFDM pilot signals; the IRS reflects the signals to multiple sensing RRHs through the LOS link. Since the geometric position and propagation parameters between the IRS and the RRH can be obtained in advance, the main sensing task of the system is transformed into estimating the signal propagation time delay between the user and the IRS. By constructing an observation model in the frequency domain dimension that combines IRS reflecting elements and subcarriers, and utilizing the statistical properties of multi-snapshot received signals, the present application can achieve high-precision estimation of the user-to-IRS link time delay, thereby improving the distance resolution and positioning accuracy in complex environments.

[0006] The technical solutions of the present application are as follows:

[0007] A user delay estimation method for an IRS-assisted multi-carrier ISAC system is disclosed. U users to be estimated transmit known OFDM pilot signals, which are reflected by the IRS to the RRH. The BBU pool processes the received signals from each receiving RRH. The BBU constructs an observation matrix of the pilot signals in the joint dimension of multiple subcarriers and IRS units, and utilizes the multi-snapshot effect and atomic norm minimization methods provided by multiple receiving antennas to achieve high-precision delay estimation of the user-to-IRS link. The specific steps are as follows:

[0008] 1.1) The IRS-assisted user delay estimation system of the multi-carrier communication sensing integrated system in the cloud access network includes a BBU pool, P RRHs, and J IRSs; the BBU pool and each RRH communicate via a wired fronthaul link, and each receiving RRH and passive IRS is equipped with N B and N I The uniform linear antenna array ULA, composed of three antennas, is responsible for sensing the target user's angle information. The BBU pool is aware of the location information of the line-of-sight (LOS) channel from RRH to IRS. The user and RRH are blocked, and there are non-line-of-sight (NLOS) channels between the user and IRS and between IRS and RRH.

[0009] 1.2) Each frame is divided into two parts: the first T... P The remaining time slots constitute the uplink pilot subframe, and the remaining TT... p Time slots constitute downlink subframes for data transmission; a comb pilot pattern is applied for uplink angle estimation. Specifically, the total bandwidth is divided into K subcarriers with an index set. K P Subcarriers are selected for pilot transmission, where Δ = K / K p This indicates the pilot subcarrier spacing; on each pilot subcarrier, the user transmits different pilot symbols to achieve frequency domain identification. Meanwhile, the reflection coefficient of the intelligent reflective surface is time-domain tuned in each pilot time slot, and the phase of the reflection matrix changes independently in each time slot.

[0010] 1.3) In the system, the user ID is u, and there is an uplink L between the user and the intelligent reflective surface IRS. UI Multipath components; UE-IRS channel frequency response of the k-th subcarrier of the u-th user at the J-th IRS. The p-th received RHH comes from the IRS-BS channel frequency response at the j-th IRS. As shown below:

[0011]

[0012] Accordingly, the downlink between the IRS and the receiving remote radio unit RRH has LIB A multipath; of which, the lth UI The path from the user to the IRS is determined by the path fading coefficient. Path delay and the angle of arrival The characterization, with a sampling frequency of f s (l UI =1,2,...,L UI (j = 1, 2, ..., J); In the link from IRS to RRH, the line-of-sight (LOS) path parameters are as follows: and These represent the fading coefficient, delay, angle of arrival (AOA), and angle of departure (AOD) of the line-of-sight (LOS) channel from the IRS to the receiving RRH.

[0013] In addition, the l IB The non-line-of-sight NLOS path parameters from IRS to RRH are as follows: and These are the lth channels from the IRS to the received RRH channel. IB Fading coefficient, delay, AOA and AOD of each path IB =2,…,L IB ;j=1,2,...,J); where and The number of antennas representing the response vector of the array are N and N respectively. B and N I ;

[0014] The array response vector (j = 1, 2, ..., J) from the u-th user to the j-th IRS angle of arrival is:

[0015]

[0016] in d represents the angle of arrival from user u to the IRS j, and d represents the distance between two adjacent reflective elements of the IRS.

[0017] 1.4) In the IRS-BS channel, where It is the lth IB The fading coefficient from the j-th IRS to the p-th RRH under a non-line-of-sight path, (l IB =2,...,L IB );

[0018] The array response vectors for the p-th RRH arrival angle and the j-th IRS departure angle are:

[0019]

[0020] Where, d RRHd represents the distance between two adjacent antennas of RRH. IRS λ represents the distance between two adjacent reflective elements of the IRS; λ represents the carrier wavelength. and Let these represent the departure angle and arrival angle from the j-th IRS to the p-th RRH, respectively; where Let represent the reflection coefficient matrix of the IRS in time slot t, where and The phase shift of the i-th reflection unit in the t-th time slot is represented; the pilot signals transmitted by u users in the t-th time slot are defined as... at last It is additive white Gaussian noise, σ 2 Indicates IRS noise power. It is the identity matrix;

[0021] 1.5) The signal of the k-th subcarrier received in the p-th RRH of the t-th time slot:

[0022]

[0023] in This represents the channel length s of the k-th subcarrier along the path from the u-th user to the j-th IRS. u,t This represents the pilot signal transmitted by the u-th user in the t-th time slot, v k,t H represents the noise generated by the signal in the k-th subcarrier of the t-th time slot. IB,j,p,k It includes the line-of-sight and non-line-of-sight paths of the signal reaching the RRH after reflection from the IRS, and is expressed in the following form:

[0024]

[0025] The entire system is modeled as follows:

[0026]

[0027] In the system and Let represent the NLOS and LOS channel matrices of the j-th IRS to p-th RRH on the k-th subcarrier, respectively; all receiving remote radio units (RRHs) synchronously transmit the received signals to the baseband processing unit (BBU) pool via a wired fronthaul link in each time slot to achieve centralized signal processing; specifically, for the t-th time slot, the k-th subcarrier, and the u-th user, the global stacked signal received by the BBU pool is denoted as . Represented as:

[0028]

[0029] 1.6) After receiving the signal from the target, each RRH quantizes it and then sends the quantized signal to the BBU pool. First, the Nth signal of the p-th RRH is quantized. B The signal received by the antenna is represented as follows:

[0030]

[0031] The quantized signal is calculated from the normalized signal as follows:

[0032]

[0033] in It is a scaling factor. Represents the integer closest to x. This indicates that the quantizer is applied to the Nth RRH of the p-th RRH. B The resolution of the received signal at the root antenna during quantization is then used to send the quantized signal to the BBU pool via an error-free fronthaul link; the final signal received by the BBU pool is:

[0034]

[0035] Furthermore, in the IRS-assisted user-to-delay estimation system of the multi-carrier communication sensing integrated system, the BBU knows the location information of the IRS and the receiving RRH. Therefore, the angle and delay information are known in the LOS channel from the IRS to the RRH, i.e., the RRH angle of arrival. Angle of departure from IRS and the delay from IRS to RRH It is necessary to estimate the latency between all users and all IRSs.

[0036] The specific steps are as follows:

[0037] 2.1) For the non-line-of-sight NLOS cascaded path from the user via the IRS to the receiving RRH, a two-dimensional atomic norm model is used for characterization, where the atom is composed of two dimensions: angle and pilot subcarrier. The two-dimensional modeling method allows each atom to carry both angle information and frequency domain features. In the angle estimation stage, the two-dimensional sparse structure information related to the angle can be obtained simultaneously by solving the two-dimensional atomic norm minimization problem.

[0038] 2.2) The LOS path portion of the cascaded channel from the reflector to the receiving RRH, where... This represents the concatenated channel of the LOS path from the reflector to the receiving RRH, which is represented as follows:

[0039]

[0040] The BBU has the location information of the IRS and the base station. The channel representation is as follows:

[0041] Define composition Z 1D,j The atomic norm is represented by a one-dimensional atom set A. 1D,j (j = 1, 2, ..., J) are the required components to form the 1D atomic norm. In the following formula, column index n corresponds to the nth reflection unit of the IRS, and the corresponding formula is: The values ​​are 0, ..., N I -1,

[0042]

[0043] K p Indicates pilot subcarrier, Indicates K p Γ j,p,k,t Delay information in Combined vectors:

[0044]

[0045] Γ j,p,t The known part in the representation Its form is as follows:

[0046]

[0047] Define composition Z 1D,j Atom set A 1D,j (j = 1, 2, ..., J) can be represented in the following form:

[0048]

[0049] The user sends a signal to the IRS, which reflects it to the cascaded channel Γ on the RRH line-of-sight path. j,p,t It can be represented in the following form:

[0050]

[0051] 2.3) In order to estimate the received signal The delay in receiving the signal Rewritten as:

[0052]

[0053] In the formula In the middle, Γ j,p,t middle It is known. These are obtained by minimizing the two-dimensional atomic norm in the angle estimation stage, and they are only related to the IRS reflection matrix and the information of the path from the IRS to the sensing RRH; where s u,t This represents the pilot signal sent by the u-th user in the t-th time slot.

[0054] This leads to the following optimization problem:

[0055]

[0056] Where x 1,j Variables to be optimized The matrix to be optimized, where λ1 represents the regularization term and Tr represents the trace of the matrix. To represent the 1D atomic norm using the semidefinite programming problem SDP, a Hermitian Tolliterz matrix, T, is used. 1D,j It is a dimension K p ×K p Topulitz matrix, T 1D Written as:

[0057]

[0058] in Where T 1D,0 Denotes the Hermitian matrix for i = 1, 2, ..., K p -1,T 1D,i Let represent the Toplitz matrix; the above time delay estimation problem is modeled as an SDP, whose objective function and constraints are convex, and therefore can be solved efficiently using standard convex optimization tools;

[0059] 2.4) After obtaining the optimal solution, the user's time delay information is extracted by performing eigenvalue decomposition or spectral estimation on the optimal Toeplitz matrix.

[0060] The design concept of this invention is as follows:

[0061] This invention involves multiple users transmitting pilot signals, with each IRS reflecting user signals to a remote radio access unit (RRH). The RRH then transmits the received signals to a centralized baseband processing unit (BBU) via a fronthaul link. The BBU performs joint processing on the received signals from multiple RRHs. By constructing a joint observation model of the IRS reflection unit and subcarriers in the frequency domain and utilizing the statistical characteristics of multi-shot received signals, this invention can achieve high-precision estimation of the user-to-IRS link delay, thereby improving distance resolution and positioning accuracy in complex environments. Using this invention, high-precision user positioning can be achieved in complex environments.

[0062] The beneficial effects of this invention are as follows:

[0063] This invention addresses the time delay estimation of the user-to-IRS link by utilizing the multi-carrier characteristics of OFDM systems and the spatial aperture provided by the IRS to construct an observation model in the joint frequency-space domain. Specifically, this invention proposes a high-resolution time delay estimation algorithm based on Atom Norm Minimization (ANM). This method models the channel response of the user-to-IRS link as a sparse linear combination on a one-dimensional set of atoms in a continuous time delay domain, where each "atom" represents the joint steering vector corresponding to a specific time delay, spanning all subcarriers and IRS units.

[0064] Ultimately, the time delay estimation problem is formalized as an ANM optimization problem, which can be efficiently solved using semidefinite programming (SDP). This method fully utilizes the dual apertures provided by the frequency domain (multiple subcarriers) and the spatial domain (multiple IRS units), effectively overcoming the resolution limitations of traditional time delay estimation algorithms. Since this method only requires pre-acquiring the link parameters from the IRS to the RRH, without needing complete real-time channel information, the algorithm is robust and particularly suitable for high-precision user positioning tasks in complex obstruction and multipath environments. Compared with other algorithms, this invention achieves higher estimation accuracy and time delay resolution, and has promising engineering application prospects. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the system model;

[0066] Figure 2 These are mean square error plots under different algorithms. Detailed Implementation

[0067] The present invention will be further described below with reference to the accompanying drawings.

[0068] In a user delay estimation method for a cloud access network intelligent reflector-assisted multi-carrier (ISAC) system, multiple users transmit pilot signals, and each IRS reflects the user signals and transmits them to the remote radio access unit (RRH). The RRH then transmits the received signals to the centralized baseband processing unit (BBU) via a fronthaul link. Since the IRS-RRH link parameters can be obtained in advance, this method estimates the delay of the user-to-IRS link. The BBU constructs an observation matrix of the pilot signals across multiple subcarriers and IRS units, and utilizes the multi-shot effect provided by multiple receiving antennas to achieve high-precision delay estimation of the user-to-IRS link. This method fully leverages the sparse structure of the frequency and spatial domains, effectively improving delay resolution and estimation accuracy, thereby enhancing user positioning capabilities in complex environments.

[0069] like Figure 1As shown, a user delay estimation method for an IRS-assisted multi-carrier ISAC system is characterized by: U users to be estimated transmitting known OFDM pilot signals that are reflected by the IRS to the RRH; the BBU pool processing the received signals from each receiving RRH; the BBU constructing an observation matrix of the pilot signals in the joint dimension of multiple subcarriers and IRS units; and utilizing the multi-snapshot effect and atomic norm minimization method provided by multiple receiving antennas to achieve high-precision delay estimation of the user-to-IRS link. The specific steps are as follows:

[0070] 1.1) The IRS-assisted user delay estimation system of the multi-carrier communication sensing integrated system in the cloud access network includes a BBU pool, P RRHs, and J IRSs; the BBU pool and each RRH communicate via a wired fronthaul link, and each receiving RRH and passive IRS is equipped with N B and N I The uniform linear antenna array ULA, composed of three antennas, is responsible for sensing the target user's angle information. The BBU pool is aware of the location information of the line-of-sight (LOS) channel from RRH to IRS. The user and RRH are blocked, and there are non-line-of-sight (NLOS) channels between the user and IRS and between IRS and RRH.

[0071] 1.2) Each frame is divided into two parts: the first T... P The remaining time slots constitute the uplink pilot subframe, and the remaining TT... p Time slots constitute downlink subframes for data transmission; a comb pilot pattern is applied for uplink angle estimation. Specifically, the total bandwidth is divided into K subcarriers with an index set. K P Subcarriers are selected for pilot transmission, where Δ = K / K p This indicates the pilot subcarrier spacing; on each pilot subcarrier, the user transmits different pilot symbols to achieve frequency domain identification. Meanwhile, the reflection coefficient of the intelligent reflective surface is time-domain tuned in each pilot time slot, and the phase of the reflection matrix changes independently in each time slot.

[0072] 1.3) In the system, the user ID is u, and there is an uplink L between the user and the intelligent reflective surface IRS. UI Multipath components; UE-IRS channel frequency response of the k-th subcarrier of the u-th user at the J-th IRS. The p-th received RHH comes from the IRS-BS channel frequency response at the j-th IRS. As shown below:

[0073]

[0074] Accordingly, the downlink between the IRS and the receiving remote radio unit RRH has L IB A multipath; of which, the lth UI The path from the user to the IRS is determined by the path fading coefficient. Path delay and the angle of arrival The characterization, with a sampling frequency of f s (l UI =1,2,...,L UI (j = 1, 2, ..., J); In the link from IRS to RRH, the line-of-sight (LOS) path parameters are as follows: and These represent the fading coefficient, delay, angle of arrival (AOA), and angle of departure (AOD) of the line-of-sight (LOS) channel from the IRS to the receiving RRH.

[0075] In addition, the l IB The non-line-of-sight NLOS path parameters from IRS to RRH are as follows: and These are the lth channels from the IRS to the received RRH channel. IB Fading coefficient, delay, AOA and AOD of each path IB =2,...,L IB ;j=1,2,...,J); where and The number of antennas representing the response vector of the array are N and N respectively. B and N I ;

[0076] The array response vector (j = 1, 2, ..., J) from the u-th user to the j-th IRS angle of arrival is:

[0077]

[0078] in d represents the angle of arrival from user u to the IRS j, and d represents the distance between two adjacent reflective elements of the IRS.

[0079] 1.4) In the IRS-BS channel, where It is the lth IB The fading coefficient from the j-th IRS to the p-th RRH under a non-line-of-sight path, (l IB =2,...,L IB );

[0080] The array response vectors for the p-th RRH arrival angle and the j-th IRS departure angle are:

[0081]

[0082] Where, d RRH d represents the distance between two adjacent antennas of RRH. IRS λ represents the distance between two adjacent reflective elements of the IRS; λ represents the carrier wavelength. and Let these represent the departure angle and arrival angle from the j-th IRS to the p-th RRH, respectively; where Let represent the reflection coefficient matrix of the IRS in time slot t, where and The phase shift of the i-th reflection unit in the t-th time slot is represented; the pilot signals transmitted by u users in the t-th time slot are defined as... at last It is additive white Gaussian noise, σ 2 Indicates IRS noise power. It is the identity matrix;

[0083] 1.5) The signal of the k-th subcarrier received in the p-th RRH of the t-th time slot:

[0084]

[0085] in This represents the channel length s of the k-th subcarrier along the path from the u-th user to the j-th IRS. u,t This represents the pilot signal transmitted by the u-th user in the t-th time slot, v k,t H represents the noise generated by the signal in the k-th subcarrier of the t-th time slot. IB,j,p,k It includes the line-of-sight and non-line-of-sight paths of the signal reaching the RRH after reflection from the IRS, and is expressed in the following form:

[0086]

[0087] The entire system is modeled as follows:

[0088]

[0089] In the system and Let represent the NLOS and LOS channel matrices of the j-th IRS to p-th RRH on the k-th subcarrier, respectively; all receiving remote radio units (RRHs) synchronously transmit the received signals to the baseband processing unit (BBU) pool via a wired fronthaul link in each time slot to achieve centralized signal processing; specifically, for the t-th time slot, the k-th subcarrier, and the u-th user, the global stacked signal received by the BBU pool is denoted as . Represented as:

[0090]

[0091] 1.6) After receiving the signal from the target, each RRH quantizes it and then sends the quantized signal to the BBU pool. First, the Nth signal of the p-th RRH is quantized. B The signal received by the antenna is represented as follows:

[0092]

[0093] The quantized signal is calculated from the normalized signal as follows:

[0094]

[0095] in It is a scaling factor. Represents the integer closest to x. This indicates that the quantizer is applied to the Nth RRH of the p-th RRH. B The resolution of the received signal at the root antenna during quantization is then used to send the quantized signal to the BBU pool via an error-free fronthaul link; the final signal received by the BBU pool is:

[0096]

[0097] In the IRS-assisted user-to-delay estimation system of the multi-carrier communication sensing integrated system, the BBU knows the location information of the IRS and the receiving RRH. Therefore, the angle and delay information are known in the LOS channel from the IRS to the RRH, i.e., the angle of arrival of the RRH. Angle of departure from IRS and the delay from IRS to RRH It is necessary to estimate the latency between all users and all IRSs.

[0098] The specific steps are as follows:

[0099] 2.1) For the non-line-of-sight NLOS cascaded path from the user via the IRS to the receiving RRH, a two-dimensional atomic norm model is used for characterization, where the atom is composed of two dimensions: angle and pilot subcarrier. The two-dimensional modeling method allows each atom to carry both angle information and frequency domain features. In the angle estimation stage, the two-dimensional sparse structure information related to the angle can be obtained simultaneously by solving the two-dimensional atomic norm minimization problem.

[0100] 2.2) The LOS path portion of the cascaded channel from the reflector to the receiving RRH, where... This represents the concatenated channel of the LOS path from the reflector to the receiving RRH, which is represented as follows:

[0101]

[0102] The BBU has the location information of the IRS and the base station. The channel representation is as follows:

[0103] Define composition Z 1D,j The atomic norm is represented by a one-dimensional atom set A. 1D,j (j = 1, 2, ..., J) are the required components to form the 1D atomic norm. In the following formula, column index n corresponds to the nth reflection unit of the IRS, and the corresponding formula is: The values ​​are 0, ..., N I -1,

[0104]

[0105] K p Indicates pilot subcarrier, Indicates K p Γ j,p,k,t Delay information in Combined vectors:

[0106]

[0107] Γ j,p,t The known part in the representation Its form is as follows:

[0108]

[0109] Define composition Z 1D,j Atom set A 1D,j (j = 1, 2, ..., J) can be represented in the following form:

[0110]

[0111] The user sends a signal to the IRS, which reflects it to the cascaded channel Γ on the RRH line-of-sight path. j,p,t It can be represented in the following form:

[0112]

[0113] 2.3) In order to estimate the received signal The delay in receiving the signal Rewritten as:

[0114]

[0115] In the formula In the middle, Γ j,p,t middle It is known. These are obtained by minimizing the two-dimensional atomic norm in the angle estimation stage, and they are only related to the IRS reflection matrix and the information of the path from the IRS to the sensing RRH; where s u,t This represents the pilot signal sent by the u-th user in the t-th time slot.

[0116] This leads to the following optimization problem:

[0117]

[0118] Where x 1,j Variables to be optimized The matrix to be optimized, λ1 represents the regularization term, and Tr represents the trace of the matrix; to represent the 1D atomic norm using the semidefinite programming problem SDP, a Hermitian Tolliterz matrix, T, is used. 1D,j It is a dimension K p ×K p Topulitz matrix, T 1D Written as:

[0119]

[0120] in Where T 1D,0 Denotes the Hermitian matrix for i = 1, 2, ..., K p -1,T 1D,i Let represent the Toplitz matrix; the above time delay estimation problem is modeled as an SDP, whose objective function and constraints are convex, and therefore can be solved efficiently using standard convex optimization tools;

[0121] 2.4) After obtaining the optimal solution, the user's time delay information is extracted by performing eigenvalue decomposition or spectral estimation on the optimal Toeplitz matrix.

[0122] To verify the performance of the algorithm of this invention, it was compared with the classical sparse recovery method based on the L1 norm. Figure 2 The results show a comparison of the root mean square error (RMSE) of time delay estimation between the two algorithms under different transmit power and signal snapshot number (T).

[0123]

[0124] Simulation results clearly demonstrate that the algorithm of this invention consistently and significantly outperforms the L1 norm method under all test conditions. Its core advantage lies in its ability to maintain high-accuracy estimation even under low transmit power or low snapshot (T) conditions, exhibiting a smooth performance curve and strong robustness. This is primarily due to the algorithm's effective utilization of the joint processing gain of subcarriers and IRS units, thereby achieving higher accuracy in sparse recovery.

[0125] In summary, this invention represents a significant technological advancement in delay estimation. It not only offers higher accuracy but also reduces system overhead, providing a reliable technical solution for achieving high-precision, low-resource-consumption user positioning in complex NLOS environments.

Claims

1. A user delay estimation method for an IRS-assisted multi-carrier ISAC system, characterized in that: U users to be estimated transmit known OFDM pilot signals, which are reflected by the IRS to the RRH; the BBU pool processes the received signals from each receiving RRH; the BBU constructs an observation matrix of the pilot signals in the joint dimension of multiple subcarriers and IRS units, and utilizes the multi-snapshot effect and atomic norm minimization method provided by multiple receiving antennas to achieve high-precision delay estimation of the user-to-IRS link; the specific steps are as follows: 1.1) The IRS-assisted user delay estimation system of the multi-carrier communication sensing integrated system in the cloud access network includes a BBU pool, P RRHs, and J IRSs; the BBU pool and each RRH communicate via a wired fronthaul link, and each receiving RRH and passive IRS is equipped with N B and N I The uniform linear antenna array ULA, composed of three antennas, is responsible for sensing the target user's angle information. The BBU pool is aware of the location information of the line-of-sight (LOS) channel from RRH to IRS. The user and RRH are blocked, and there are non-line-of-sight (NLOS) channels between the user and IRS and between IRS and RRH. 1.2) Each frame is divided into two parts: the first T... P The remaining time slots constitute the uplink pilot subframe, and the remaining TT... p Time slots constitute downlink subframes for data transmission; a comb pilot pattern is applied for uplink angle estimation. Specifically, the total bandwidth is divided into K subcarriers with an index set. K P Subcarriers are selected for pilot transmission, where Δ = K / K p This indicates the pilot subcarrier spacing; on each pilot subcarrier, the user transmits different pilot symbols to achieve frequency domain identification. Meanwhile, the reflection coefficient of the intelligent reflective surface is time-domain tuned in each pilot time slot, and the phase of the reflection matrix changes independently in each time slot. 1.3) In the system, the user ID is u, and there is an uplink L between the user and the intelligent reflective surface IRS. UI Multipath components; UE-IRS channel frequency response of the k-th subcarrier of the u-th user at the J-th IRS. The p-th received RHH comes from the IRS-BS channel frequency response at the j-th IRS. As shown below: Accordingly, the downlink between the IRS and the receiving remote radio unit RRH has L IB A multipath; of which, the lth UI The path from the user to the IRS is determined by the path fading coefficient. Path delay and the angle of arrival The characterization, with a sampling frequency of f s (l UI =1,2,…,L UI (j = 1, 2, ..., J); In the link from IRS to RRH, the line-of-sight (LOS) path parameters are as follows: and These represent the fading coefficient, delay, angle of arrival (AOA), and angle of departure (AOD) of the line-of-sight (LOS) channel from the IRS to the receiving RRH. In addition, the l IB The non-line-of-sight NLOS path parameters from IRS to RRH are as follows: and These are the lth channels from the IRS to the received RRH channel. IB Fading coefficient, delay, AOA and AOD of each path IB =2,...,L IB ;j=1,2,…,J); where and The number of antennas representing the response vector of the array are N and N respectively. B and N I ; The array response vector (j = 1, 2, ..., J) from the u-th user to the j-th IRS angle of arrival is: in d represents the angle of arrival from user u to the IRS j, and d represents the distance between two adjacent reflective elements of the IRS. 1.4) In the IRS-BS channel, where It is the lth IB The fading coefficient from the j-th IRS to the p-th RRH under a non-line-of-sight path, (l IB =2,...,L IB ); The array response vectors for the p-th RRH arrival angle and the j-th IRS departure angle are: Where, d RRH d represents the distance between two adjacent antennas of RRH. IRS λ represents the distance between two adjacent reflective elements of the IRS; λ represents the carrier wavelength. and Let these represent the departure angle and arrival angle from the j-th IRS to the p-th RRH, respectively; where Let represent the reflection coefficient matrix of the IRS in time slot t, where and The phase shift of the i-th reflection unit in the t-th time slot is represented; the pilot signals transmitted by u users in the t-th time slot are defined as... at last It is additive white Gaussian noise, σ 2 Indicates IRS noise power. It is the identity matrix; 1.5) The signal of the k-th subcarrier received in the p-th RRH of the t-th time slot: in This represents the channel length s of the k-th subcarrier along the path from the u-th user to the j-th IRS. u,t This represents the pilot signal transmitted by the u-th user in the t-th time slot, v k,t H represents the noise generated by the signal in the k-th subcarrier of the t-th time slot. IB,j,p,k It includes the line-of-sight and non-line-of-sight paths of the signal reaching the RRH after reflection from the IRS, and is expressed in the following form: The entire system is modeled as follows: In the system and Let represent the NLOS and LOS channel matrices of the j-th IRS to p-th RRH on the k-th subcarrier, respectively; all receiving remote radio units (RRHs) synchronously transmit the received signals to the baseband processing unit (BBU) pool via a wired fronthaul link in each time slot to achieve centralized signal processing; specifically, for the t-th time slot, the k-th subcarrier, and the u-th user, the global stacked signal received by the BBU pool is denoted as . Represented as: 1.6) After receiving the signal from the target, each RRH quantizes it and then sends the quantized signal to the BBU pool. First, the Nth signal of the p-th RRH is quantized. B The signal received by the antenna is represented as follows: The quantized signal is calculated from the normalized signal as follows: in It is a scaling factor. Represents the integer closest to x. This indicates that the quantizer is applied to the Nth RRH of the p-th RRH. B The resolution of the received signal at the root antenna during quantization is then used to send the quantized signal to the BBU pool via an error-free fronthaul link; the final signal received by the BBU pool is:

2. The intelligent reflector-assisted user delay estimation method in a cloud access network multi-carrier ISAC system according to claim 1, characterized in that, In the IRS-assisted user-to-delay estimation system of the multi-carrier communication sensing integrated system, the BBU knows the location information of the IRS and the receiving RRH. Therefore, the angle and delay information are known in the LOS channel from the IRS to the RRH, i.e., the angle of arrival of the RRH. Angle of departure from IRS and the delay from IRS to RRH It is necessary to estimate the latency between all users and all IRSs. The specific steps are as follows: 2.1) For the non-line-of-sight NLOS cascaded path from the user via the IRS to the receiving RRH, a two-dimensional atomic norm model is used for characterization, where the atom is composed of two dimensions: angle and pilot subcarrier. The two-dimensional modeling method allows each atom to carry both angle information and frequency domain features simultaneously. In the angle estimation stage, the angle-related two-dimensional sparse structure information can be obtained simultaneously by solving the two-dimensional atomic norm minimization problem. 2.2) The LOS path portion of the cascaded channel from the reflector to the receiving RRH, where... This represents the concatenated channel of the LOS path from the reflector to the receiving RRH, which is represented as follows: The BBU has the location information of the IRS and the base station. The channel representation is as follows: Define composition Z 1D,j The atomic norm is represented by the one-dimensional atomic set A. 1D,j (j = 1, 2, ..., J) are the required components to form the 1D atomic norm. In the following formula, column index n corresponds to the nth reflection unit of the IRS, and the corresponding formula is: The values ​​are 0, ..., N I -1, K p Indicates pilot subcarrier, Indicates K p Γ j,p,k,t Delay information in Combined vectors: Γ j,p,t The known part in the representation Its form is as follows: Define composition Z 1D,j Atom set A 1D,j (j = 1, 2, ..., J) can be represented in the following form: The user sends a signal to the IRS, which reflects it to the cascaded channel Γ on the RRH line-of-sight path. j,p,t It can be represented in the following form: 2.3) In order to estimate the received signal The delay in receiving the signal Rewritten as: In the formula In the middle, Γ j,p,t middle It is known. These are obtained by minimizing the two-dimensional atomic norm in the angle estimation stage, and they are only related to the IRS reflection matrix and the information of the path from the IRS to the sensing RRH; where s u,t This represents the pilot signal sent by the u-th user in the t-th time slot. This leads to the following optimization problem: Where x 1,j Variables to be optimized The matrix to be optimized, λ1 represents the regularization term, and Tr represents the trace of the matrix; to represent the 1D atomic norm using the semidefinite programming problem SDP, a Hermitian Tolliterz matrix, T, is used. 1D,j It is a dimension K p ×K p Toplitz matrix, T 1D Written as: in Where T 1D,0 Denotes the Hermitian matrix for i = 1, 2, ..., K p -1,T 1D,i Represents the Toplitz matrix; The aforementioned time delay estimation problem is modeled as an SDP, whose objective function and constraints are convex, and therefore can be solved efficiently using standard convex optimization tools. 2.4) After obtaining the optimal solution, the user's time delay information is extracted by performing eigenvalue decomposition or spectral estimation on the optimal Toeplitz matrix.

Images