Parameter estimation method and system for cross-field ultrahigh frequency wireless channel

Abstract

The invention provides a parameter estimation method and system for a cross-field ultrahigh-frequency wireless channel, and belongs to the field of wireless communication, and the method comprises the steps: constructing an ultrahigh-frequency terahertz channel measurement system, and obtaining a channel transmission function; establishing a cross-field domain channel signal model; executing a cross-field domain parameter estimation and optimization process by adopting a parameter estimation algorithm based on a maximum likelihood criterion; and carrying out channel characterization and analysis based on the estimation parameters. According to the method, near-field and far-field multipath components are intelligently distinguished by introducing the Bayesian information criterion, parameter estimation is performed by adopting the spherical wavefront model and the plane wavefront model respectively, and the problem of estimation deviation caused by neglecting a near-field effect and spatial non-stationarity in a traditional algorithm is effectively solved by combining iterative optimization and visible region modeling. According to the method, the nonlinear phase change, the multipath birth and death phenomenon and the frequency domain characteristic in the ultrahigh-frequency multi-antenna channel can be accurately captured, the parameter estimation precision and the residual performance are remarkably improved, and a key technical support and an experimental basis are provided for channel modeling, large-scale antenna array design and perception communication integration of a 6G ultrahigh-frequency terahertz communication system.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wireless communication, and particularly relates to a parameter estimation method and system for a cross-field ultra-high frequency wireless channel. BACKGROUND

[0002] The sixth generation (6G) wireless system aims to achieve full-spectrum, global coverage and full-scene coverage, has ultra-high transmission capacity, provides strong security protection, and supports the vision of future intelligent communication. The ultra-high frequency terahertz communication system becomes a core candidate technology for realizing this goal due to the huge available bandwidth in the terahertz frequency band (0.1 THz-10 THz). However, the ultra-high frequency terahertz wave has unique propagation characteristics, including high isotropic free space path loss, strong molecular absorption, severe frequency selectivity and non-stationarity, which are significantly different from low-frequency bands such as microwaves and millimeter waves. In order to compensate for high path loss and improve spectral efficiency, large-scale antenna arrays are widely introduced, which makes the channel characteristics more complex. More importantly, in actual ultra-high frequency terahertz communication scenarios, the channel response often contains both near-field and far-field characteristics, forming a typical cross-field propagation environment. Therefore, the accurate analysis and characterization of the channel are the basis for the design and optimization of the communication system. Therefore, the research on the cross-field channel parameter estimation in the ultra-high frequency terahertz band has fundamental significance for breaking through the bottleneck of ultra-high frequency terahertz communication technology and ensuring the efficiency and reliability of the system.

[0003] Recent research has made significant progress in the field of terahertz channel measurement and characterization. Current terahertz channel measurement mainly adopts three methods to cope with different propagation challenges: a frequency-domain vector network analyzer (VNA) based system as a benchmark for high-precision static channel characterization, which measures the path loss and multiple path components (MPCs) through S parameters. For example, N. A. Abbasi quantifies the bidirectional urban path loss within 100 meters by using a 140-300 GHz radio frequency light coupling extended VNA; Y. Chen uses a three-dimensional horn antenna scanning technology to draw the angular distribution map of 140 GHz signals in an indoor environment. The time-domain sliding correlator (TDS) realizes dynamic channel analysis through PN sequence correlation, supporting mobility and real-time applications. For example, Y. Xing and T. S. Rappaport, S. Ju use phased array probes to capture the spatial consistency of 142 GHz urban microcells, and J. M. Eckhardt characterizes the 300 GHz data center shielding dynamics with sub-nanosecond time resolution. Terahertz-TDS systems focus on ultra-wideband material interaction and near-field effect research: S. Priebe extracts the 300 GHz rough surface scattering coefficient through pulse measurement; J. F. Federici quantifies weather-induced pulse distortion for attenuation modeling. These methods collectively reveal the unique nature of terahertz wave propagation.

[0004] Although there are some progresses in the measurement of terahertz channel, it is still a challenge to develop an efficient and accurate algorithm for estimating the parameters of terahertz channel. Traditional algorithms such as Bartlett spectrum, Capon spectrum, signal parameter single estimation based on rotational invariance technique (ESPRIT), multiple signal classification (MUSIC), RiMAX and expectation maximization (EM) are mainly based on far-field assumption and are suitable for low frequency band, and it is difficult to fully capture the unique characteristics of large-scale antenna array terahertz channel. In addition, the spatial alternating generalized expectation maximization (SAGE) algorithm based on EM algorithm effectively reduces the computational complexity by dividing the channel parameters into several subsets. When the number of antennas increases to form a large-scale multiple-input multiple-output (MIMO) system, the spherical wavefront phenomenon caused by near-field effect begins to appear. Under this condition, the plane wavefront (far-field) model deviates significantly from the actual measured wavefront, resulting in serious estimation error or even estimation failure. The research of Ma and Zhang enhances the MUSIC algorithm by introducing near-field scatterers, but this method still cannot deal with the wideband and spatial non-stationary characteristics. At the same time, X. Yin proposes an algorithm for estimating wideband three-dimensional near-field channel parameters based on electromagnetic principles. In the research of Y. Li, the (DSS)-o-SAGE algorithm is developed for direction scanning detection, which effectively solves the phase instability problem in millimeter wave and terahertz band channel parameter estimation and reduces the computational complexity. In addition, Zhou et al. propose a new SAGE algorithm focusing on estimating wideband spatial non-stationary wireless channel parameters containing antenna polarization.

[0005] Therefore, the existing ultra-high frequency terahertz channel parameter estimation methods all focus on the design of single field characteristics, and cannot adapt to the cross-field propagation environment where near-field and far-field coexist, resulting in large parameter estimation deviation or even estimation failure in actual scenarios, which cannot meet the demand of reliable operation of ultra-high frequency terahertz communication system. SUMMARY

[0006] In order to solve the above problems, the present application provides a parameter estimation method for cross-field ultra-high frequency wireless channel, which can distinguish near-field and far-field multipath components (MPCs) in indoor scenarios and perform parameter estimation and channel characterization.

[0007] In order to achieve the above purpose, the present application provides the following technical solutions: A parameter estimation method for cross-field ultra-high frequency wireless channel, comprising the following steps: Channel response data of cross-field channels in the UHF terahertz band within the coverage area of ​​the actual base station is captured. The device's own interference information in the channel response data is removed to obtain the channel transfer function (CTF) containing only the wireless channel characteristics. The CTF is presented as an M×K complex value matrix, where M is the number of antenna elements at the receiving end and K is the number of frequency points. Each element of the CTF matrix is ​​decomposed into the superposition of the near-field multipath component representing spherical waves and the far-field multipath component representing plane waves, resulting in a single-input multiple-output SIMO cross-field model; the CTF is aggregated into a superposition of sub-functions containing corresponding parameter sets. Based on the decomposition structure of the near-field and far-field multipath components in the SIMO cross-field model, a spherical wave range difference term and a plane wave angle term are embedded in the objective function to align the multidimensional parameter sets of the two types of components. The multidimensional parameter sets are decomposed based on the aggregation logic of CTF, and the initial parameters of each multipath component are extracted in descending order of power based on the maximum likelihood criterion and the narrowband far-field assumption. Based on the initial parameters of the multipath components, the multipath delay and angle of arrival are initially estimated. The near-field and far-field categories are distinguished by the BIC criterion. Based on the discrimination results, the final near-field MPC parameter set and far-field MPC parameter set are obtained by using a spherical wave or plane wave model.

[0008] Preferably, the terahertz channel measurement system captures channel response data of the UHF terahertz band cross-field channel within the coverage area of ​​the actual base station. The terahertz channel measurement system includes a computer as a control platform, a displacement platform constituting a virtual antenna array, and a detection platform for acquiring the channel response. The detection platform includes a terahertz transmit (Tx) module, a receive (Rx) module, and a vector network analyzer (VNA). The signal amplitude and phase values ​​of multiple frequency points and multiple antenna elements are collected by VNA frequency domain scanning to obtain the raw channel response data containing wireless channel characteristics and equipment interference. Then, the frequency domain response of the measurement system is extracted by back-to-back calibration, and the equipment interference is eliminated by amplitude and phase normalization. Finally, the channel transfer function (CTF) in M×K matrix form, which only reflects the wireless channel characteristics, is obtained.

[0009] Preferably, the Channel Transfer Function (CTF) is: ; In the formula, These are the measured signal amplitude and phase values. and These represent the frequency domain responses of the receiving and transmitting antennas, respectively. This represents the frequency domain response caused by the VNA and the cable.

[0010] Preferably, the step of decomposing each element of the CTF matrix into a superposition of a near-field multipath component representing a spherical wave and a far-field multipath component representing a plane wave to obtain a single-input multiple-output (SIMO) cross-domain model is as follows: Using a matrix to perform CTF on multipath channels This means that each element in H is written as the superposition of near-field multipath components and far-field multipath components, resulting in a single-input multiple-output (SIMO) cross-domain model: ; In the formula, m is the antenna element index, and k is the frequency index. Antenna spacing, Represents the set composed of near-field multipath components. For MPC serial number, For the first The complex gain of MPC on the first antenna element For the m-th antenna element The birth and death coefficient of each MPC, It is the first The time delay of each MPC It is the first in bandwidth k The frequency of each carrier wave, =c / yes The wavelength; considering the distance difference of the spherical wavefront. It means that, among them From the first The last reflected scatterer of each MPC to the m-th The propagation distance of the antenna element, From the first The propagation distance from the last reflector of an MPC to the first Rx antenna element; The aggregation of CTFs into a superposition of sub-functions containing corresponding parameter sets is specifically as follows: ; in Represents K frequency points, It contains all the parameters for MPC.

[0011] Preferably, based on the decomposition structure of the near-field and far-field multipath components in the SIMO cross-field model, a spherical wave range difference term and a plane wave angle term are embedded in the objective function to align the multidimensional parameter sets of the two types of components; relying on the aggregation logic of CTF to decompose the multidimensional parameter set, and based on the maximum likelihood criterion and the narrowband far-field assumption, the initial parameters of each multipath component are extracted in descending order of power, including: Based on the decomposition structure of the near-field and far-field multipath components in the SIMO cross-field model, a spherical wave range difference term and a plane wave angle term are embedded in the objective function to align the multidimensional parameter sets of the near-field and far-field. Based on the aggregation logic of CTF, the high-dimensional parameter set is decomposed into an independent parameter subset of a single multipath component; Based on the maximum likelihood criterion, a log-likelihood function is constructed, and combined with the narrowband far-field assumption, the initial parameters of each multipath component are extracted in descending order of power. For each multipath component, the expected value of the component is separated from the measured CTF by a successive interference cancellation framework, and its parameters are solved by maximizing the objective function. Define a steering vector that adapts to near-field and far-field characteristics, integrate complex gain and phase factor, and solve for the initial values ​​of the single multipath component parameters.

[0012] Preferably, based on the initial parameters of the multipath components, the multipath delay and angle of arrival are initially estimated; the near-field and far-field categories are determined using the BIC criterion; and the final near-field MPC parameter set and far-field MPC parameter set are obtained using a spherical wave or plane wave model based on the determination results, including: Based on the initial parameters of the multipath components, under the assumptions of narrowband and far-field, the initial time delay and angle of arrival of each multipath component are calculated; The phase vector of each multipath component is extracted and fitted using both linear and quadratic models. Calculate the BIC values ​​corresponding to the two models, and determine whether the multipath component belongs to the near-field or far-field category by comparing the magnitude of the BIC values. If it is identified as a near-field multipath component, the steering vector related to the spherical wave model is used to accurately estimate its angle of arrival, scatterer distance, time delay, and complex gain; If it is determined to be a far-field multipath component, the steering vector related to the plane wave model is used to accurately estimate its angle of arrival, time delay and complex gain. The parameters of all multipath components are iteratively updated in the order of angle of arrival, scatterer distance, time delay, birth and death coefficients, and complex amplitude. After each iteration, the log-likelihood function value is calculated until the increment of the likelihood function before and after the iteration is less than the threshold or the maximum number of iterations is reached, and the final near-field and far-field MPC parameter sets are output.

[0013] Preferably, it further includes: The CTF corresponding to the final near-field and far-field MPC parameters is converted into a time-domain response by Fourier transform, and the time-delay power spectral density is analyzed. Using a Close-in model, the path loss exponent, transmission distance, reference distance, and carrier frequency are substituted to fit the path loss characteristics; The number of multipath components in the near field and far field is statistically identified, and the proportion of near-field multipath components is calculated using a proportional formula. Extract the measured CTF of each Tx-Rx horn pair, substitute it with the spatial difference and frequency difference parameters of the antenna array, calculate the spatial-frequency correlation function, and obtain the correlation quantification results of the channel in the spatial and frequency dimensions.

[0014] The present invention also provides a parameter estimation system for cross-field ultra-high frequency wireless channels, comprising: The channel measurement module is used to capture channel response data of cross-field channels in the UHF terahertz band within the actual base station coverage area, remove device-specific interference information from the channel response data, and obtain the channel transfer function (CTF) containing only wireless channel characteristics; the CTF is presented as an M×K complex value matrix, where M is the number of antenna elements at the receiving end and K is the number of frequency points; The signal modeling module decomposes each element of the CTF matrix into a superposition of near-field multipath components representing spherical waves and far-field multipath components representing plane waves, resulting in a single-input multiple-output (SIMO) cross-field model; and aggregates the CTF into a superposition of sub-functions containing corresponding parameter sets. The first estimation module is used to embed spherical wave range difference terms and plane wave angle terms into the objective function based on the decomposition structure of near-field and far-field multipath components in the SIMO cross-field model, and align the multidimensional parameter sets of the two types of components. It decomposes the multidimensional parameter sets based on the aggregation logic of CTF, and extracts the initial parameters of each multipath component in descending order of power based on the maximum likelihood criterion and the narrowband far-field assumption. The first estimation module is used to initially estimate the multipath delay and angle of arrival based on the initial parameters of the multipath components; it then uses the BIC criterion to distinguish the near-field and far-field categories, and uses a spherical wave or plane wave model to obtain the final near-field MPC parameter set and far-field MPC parameter set based on the discrimination results.

[0015] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement any of the steps in the parameter estimation method for cross-field ultra-high frequency wireless channels.

[0016] The present invention also provides a computer-readable storage medium storing a computer program that, when loaded by a processor, is capable of executing any of the steps in the parameter estimation method for cross-field ultra-high frequency wireless channels.

[0017] The parameter estimation method for cross-field UHF wireless channels provided by this invention has the following beneficial effects: This invention first eliminates equipment interference to obtain a clean CTF matrix, laying a reliable data foundation for subsequent analysis. Then, it innovatively decomposes the CTF into a superposition of near-field spherical wave and far-field plane wave components, constructing a SIMO cross-field model, breaking the limitations of traditional single-field models. Subsequently, initial parameters are extracted based on the maximum likelihood criterion and the narrowband far-field assumption. The BIC criterion is used to intelligently identify the field category of multipath components, and a suitable adaptation model is employed to complete precise parameter estimation. These operations effectively achieve accurate differentiation and parameter calculation of cross-field multipath components, significantly reducing estimation bias and improving channel model reliability, providing key technical support for the stable operation of ultra-high frequency terahertz communication systems. This scheme overcomes the limitations of traditional single-field adaptation algorithms, achieving accurate differentiation and parameter estimation of cross-field multipath components, effectively solving the problem of excessive estimation bias or failure in practical scenarios, significantly improving parameter estimation accuracy and channel model reliability, and providing key support for the reliable operation of ultra-high frequency terahertz communication systems. Attached Figure Description

[0018] To more clearly illustrate the embodiments and design schemes of the present invention, the accompanying drawings required for this embodiment will be briefly described below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of the parameter estimation method for cross-field ultra-high frequency wireless channels proposed in this invention.

[0020] Figure 2 This is a schematic diagram of the terahertz channel measurement system of the present invention; Figure 3 This is a schematic diagram of the indoor channel measurement scenario layout for this invention; Figure 4 This is a comparison chart of the measurement, estimation, and residual of the time delay power spectral density in this invention; Figure 5 The measured values ​​of the indoor terahertz channel path loss and the fitting results of the CI model are shown in this invention. Figure 6 This is a schematic diagram showing the proportions of near-field multipath components at different frequency bands and distances according to the present invention; Figure 7 This is a schematic diagram of the spatial-frequency correlation when the antenna is aligned with a Tx-Rx distance of 3m according to the present invention; Figure 8 This is a schematic diagram of the spatial-frequency correlation when the antenna is aligned at a Tx-Rx distance of 6m according to the present invention; Figure 9 This is a schematic diagram of the space-frequency correlation when the Tx-Rx distance is 6m and the antenna is not aligned, according to the present invention. Figure 10This is a schematic diagram comparing the frequency correlation functions of different frequency bands in this invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0022] This invention provides a parameter estimation method and system for cross-field ultra-high frequency wireless channels, specifically designed for next-generation terahertz wireless communication systems. The method solves key technical problems such as near-field spherical wave effects, spatial non-stationarity, and traditional algorithm model mismatch introduced by large-scale antenna arrays through an intelligent near-field and far-field multipath component differentiation and joint estimation mechanism, significantly improving the accuracy of channel parameter estimation and the reliability of channel models.

[0023] This invention breaks through the limitations of traditional algorithms that only focus on the near or far field, achieving intelligent differentiation and joint estimation of near-field and far-field multipath components (MPCs) in terahertz channels. Through an improved SAGE algorithm, a Bayesian information criterion (BIC) discriminator, and visible region modeling, it accurately captures the unique characteristics of terahertz channels, such as spherical wavefront effects and spatial non-stationarity. It quantifies the relationship between the near-field multipath ratio and transmission distance and carrier frequency, comprehensively characterizing key statistical properties of the channel, such as delay power spectral density, path loss, and spatial-frequency domain correlation. It provides an engineering-feasible measurement system and implementation scheme, supporting multiple frequency bands such as 260GHz, 330GHz, and 380GHz, adapting to different indoor scenarios, and providing reliable technical support for the research and development of 6G terahertz communication systems.

[0024] First, this invention proposes a terahertz channel measurement system suitable for indoor scenarios, where the communication terminal is equipped with a large-scale antenna array, rendering the traditional far-field assumption invalid and causing the channel to exhibit complex cross-field characteristics. Therefore, a complete terahertz channel parameter extraction solution is designed, based on an improved Spatial Alternating Generalized Expectation-Maximization (SAGE) algorithm, capable of simultaneously and accurately estimating both near-field and far-field multipath components. Assuming the existence of L multipath components in the channel, each component can be intelligently divided into two categories based on its wavefront characteristics: near-field multipath components (whose parameters need to be estimated using a spherical wave model) and far-field multipath components (whose parameters can be estimated using a plane wave model). Near-field multipath components, as complex components in the channel, require estimation of their extended parameter set, including scatterer distance and visible region, i.e., they have a higher parameter dimension. Far-field multipath components, with lower communication requirements, are estimated using a traditional parameter set for unbiased estimation. To ensure the fundamental accuracy of channel modeling, the complex gain, delay, angle of arrival, and other common parameters of all multipath components must be accurately estimated. Due to the sparse nature of the terahertz band, spectrum resources are extremely valuable. To ensure system capacity and communication rates, the accurate characterization of each multipath component has the highest priority.

[0025] Based on this, the core of this invention lies in introducing a near-field / far-field discriminator based on the Bayesian Information Criterion (BIC). This discriminator automatically and accurately classifies each multipath component by analyzing the phase variation characteristics (linear or quadratic) of the multipath components along the antenna array axis, thereby selecting the correct physical model for parameter estimation. Furthermore, addressing the unique spatial non-stationarity (i.e., the "birth and death" phenomenon) of near-field multipath components, this invention proposes a visible region estimation module. This module uses an alternating optimization algorithm to accurately determine the start and end positions of each near-field multipath component on the array, thus fully characterizing its spatial variation characteristics.

[0026] Based on the precisely estimated multipath parameters described above, this invention can also achieve in-depth characterization of key statistical characteristics of terahertz channels, including but not limited to: time delay power spectral density, path loss, the relationship between near-field multipath ratio and frequency and distance, scatterer distribution, and space-frequency correlation. These characterization results provide crucial theoretical basis and data support for the design of future terahertz communication systems, such as antenna array optimization, beam management, and resource allocation.

[0027] The following is an explanation of the proper nouns involved in this invention: CTF (Channel Transfer Function) is a core function that describes the relationship between the input and output signals of a terahertz signal after it propagates through a wireless channel. It directly quantifies the effects of the channel on signal amplitude attenuation, phase shift, multipath superposition, and other factors.

[0028] SIMO (Single-Input Multiple-Output) is an antenna configuration and channel operating mode in wireless communication. Its core feature is a signal transmission architecture of "one transmitter and multiple receivers".

[0029] MPC (Multipath Component) refers to the signal components corresponding to multiple independent propagation paths generated by phenomena such as reflection, scattering, and diffraction during the propagation of a signal from the transmitter (Tx) to the receiver (Rx).

[0030] This invention proposes a parameter estimation method for cross-field ultra-high frequency wireless channels, comprising the following steps: Step 1: Capture the channel response data of the UHF terahertz band cross-field channel within the actual base station coverage area, remove the device's own interference information from the channel response data, and obtain the channel transfer function (CTF) that only contains the wireless channel characteristics; the CTF is presented as an M×K complex value matrix, where M is the number of antenna elements at the receiver and K is the number of frequency points.

[0031] Furthermore, the present invention captures channel response data of cross-field channels in the UHF terahertz band within the coverage area of ​​an actual base station based on a terahertz channel measurement system. The terahertz channel measurement system includes a computer as a control platform, a displacement platform constituting a virtual antenna array, and a detection platform for acquiring channel response. The detection platform includes a terahertz transmit (Tx) module, a receive (Rx) module, and a vector network analyzer (VNA).

[0032] The signal amplitude and phase values ​​of multiple frequency points and multiple antenna elements are collected by VNA frequency domain scanning to obtain the raw channel response data containing wireless channel characteristics and equipment interference. Then, the frequency domain response of the measurement system is extracted by back-to-back calibration, and the equipment interference is eliminated by amplitude and phase normalization. Finally, the channel transfer function (CTF) in M×K matrix form, which only reflects the wireless channel characteristics, is obtained.

[0033] The Channel Transfer Function (CTF) is: ; In the formula, These are the measured signal amplitude and phase values. and These represent the frequency domain responses of the receiving and transmitting antennas, respectively. This represents the frequency domain response caused by the VNA and the cable.

[0034] Step 2: Decompose each element of the CTF matrix into the superposition of the near-field multipath component representing spherical waves and the far-field multipath component representing plane waves to obtain a single-input multiple-output SIMO cross-field model; aggregate the CTF into a superposition of sub-functions containing corresponding parameter sets.

[0035] Furthermore, this invention uses a matrix to represent the CTF of multipath channels. This means that each element in H is written as the superposition of near-field multipath components and far-field multipath components, resulting in a single-input multiple-output (SIMO) cross-domain model: ; In the formula, m is the antenna element index, and k is the frequency index. Antenna spacing, Represents the set composed of near-field multipath components. For MPC serial number, For the first The complex gain of MPC on the first antenna element For the m-th antenna element The birth and death coefficient of each MPC, It is the first The time delay of each MPC It is the first in bandwidth k The frequency of each carrier wave, =c / yes The wavelength; considering the distance difference of the spherical wavefront. It means that, among them From the first The last reflected scatterer of each MPC to the m-th The propagation distance of the antenna element, From the first The propagation distance from the last reflector of an MPC to the first Rx antenna element.

[0036] The CTF is aggregated into a superposition of sub-functions containing corresponding parameter sets, specifically as follows: ; in Represents K frequency points, It contains all the parameters for MPC.

[0037] Step 3: Based on the decomposition structure of the near-field and far-field multipath components in the SIMO cross-field model, embed the spherical wave range difference term and the plane wave angle term into the objective function to align the multidimensional parameter sets of the two types of components; decompose the multidimensional parameter sets using the aggregation logic of CTF, and extract the initial parameters of each multipath component in descending order of power based on the maximum likelihood criterion and the narrowband far-field assumption, specifically including: Based on the decomposition structure of the near-field and far-field multipath components in the SIMO cross-field model, a spherical wave range difference term and a plane wave angle term are embedded in the objective function to align the multidimensional parameter sets of the near-field and far-field.

[0038] Based on the aggregation logic of CTF, the high-dimensional parameter set is decomposed into independent parameter subsets of a single multipath component.

[0039] Based on the maximum likelihood criterion, a log-likelihood function is constructed, and combined with the narrowband far-field assumption, the initial parameters of each multipath component are extracted in descending order of power.

[0040] For each multipath component, the expected value of the component is separated from the measured CTF by a successive disturbance cancellation framework, and its parameters are solved by maximizing the objective function.

[0041] Define a steering vector that adapts to near-field and far-field characteristics, integrate complex gain and phase factor, and solve for the initial values ​​of the single multipath component parameters.

[0042] Step 4: Based on the initial parameters of the multipath components, first estimate the multipath delay and angle of arrival; determine the near-field and far-field categories using the BIC criterion, and obtain the final near-field MPC parameter set and far-field MPC parameter set using a spherical wave or plane wave model based on the discrimination results, specifically including: Based on the initial parameters of the multipath components, under the assumptions of narrowband and far-field, the initial time delay and angle of arrival of each multipath component are calculated.

[0043] The phase vector of each multipath component is extracted and fitted using both linear and quadratic models.

[0044] Calculate the BIC values ​​for the two models, and determine whether the multipath component belongs to the near-field or far-field category by comparing the magnitude of the BIC values.

[0045] If it is identified as a near-field multipath component, the steering vector associated with the spherical wave model is used to accurately estimate its angle of arrival, scatterer distance, time delay, and complex gain; if it is identified as a far-field multipath component, the steering vector associated with the plane wave model is used to accurately estimate its angle of arrival, time delay, and complex gain.

[0046] The parameters of all multipath components are iteratively updated in the order of angle of arrival, scatterer distance, time delay, birth and death coefficients, and complex amplitude.

[0047] After each iteration, the log-likelihood function value is calculated until the increment of the likelihood function before and after the iteration is less than the threshold or the maximum number of iterations is reached, and the final near-field and far-field MPC parameter sets are output.

[0048] Step 5: Convert the CTF corresponding to the final near-field and far-field MPC parameters into the time-domain response using Fourier transform and analyze the delay power spectral density; use a Close-in model, substituting the path loss exponent, transmission distance, reference distance, and carrier frequency to fit the path loss characteristics; count the number of multipath components identified as near-field and far-field, and calculate the proportion of near-field multipath components according to the proportional formula; extract the measured CTF of each Tx-Rx horn pair, substitute the antenna array spatial difference and frequency difference parameters, calculate the spatial-frequency correlation function, and obtain the correlation quantification results of the channel in the spatial and frequency dimensions.

[0049] The channel parameter estimation method proposed in this invention will be further illustrated below through specific embodiments.

[0050] Example 1 This invention provides a parameter estimation method for cross-field ultra-high frequency wireless channels, specifically as follows: Figure 1 As shown, it includes the following steps: Step 1: Build a terahertz channel measurement system, perform channel measurements using the system, process the measurement results, and obtain the Channel Transfer Function (CTF) containing only the characteristics of the wireless channel. Specifically, this includes: Channel response data of cross-field channels in the UHF terahertz band within the actual base station coverage area is captured. The device's own interference information in the channel response data is removed to obtain the channel transfer function (CTF) containing only the wireless channel characteristics. The CTF is presented as an M×K complex value matrix, where M is the number of antenna elements at the receiver and K is the number of frequency points.

[0051] like Figure 2 As shown, the terahertz channel measurement system includes a computer (PC) as a control platform, a displacement platform constituting a virtual antenna array, and a detection platform for acquiring channel responses. The detection platform includes a terahertz transmit (Tx) module, a receive (Rx) module, and a Ceyear 3672C VNA (vector network analyzer).

[0052] The VNA outputs local oscillator (LO) and radio frequency (RF) reference signals. The RF signal is up-converted by ×27 to reach the terahertz carrier frequency. The LO signal is up-converted by ×24 to generate a hybrid intermediate frequency (IF) signal at a frequency of 7×27 to reach the terahertz carrier frequency. The LO signal is then multiplied by ×24 to generate a hybrid IF signal at a frequency of 7.6MHz. The reference and test IF signals are returned to the vector network analyzer via the transmitter and receiver, and their ratio constitutes the terahertz channel transfer function (CTF). The calculated channel response includes not only the wireless channel characteristics but also the effects of devices, cables, and waveguides. Indoor terahertz measurements were performed at 260, 330, and 380 GHz using a channel detector based on the vector network analyzer. Each band covers a 20 GHz detection bandwidth, with measurements taken at 1001 sampling points. Measurements were performed at 20 MHz intervals. =20MHz, can achieve a time resolution of 50ps =1 / =50ns, and a path resolution of 1.5cm. ∆L=1.5cm. Therefore, the maximum detectable path length is =15m, sufficient to meet the channel measurement needs of the conference room. Both the transmitter and receiver are equipped with horn antennas with a gain of 25dBi and a half-power beamwidth (HPBW) of 15 degrees, enabling a high measurement dynamic range. The receiver is mounted on a stepping rotary table, while the transmitter needs to be deployed in different locations within the conference room. The test signal power is 0.5mW, and the background noise of the measurement platform is -145dBm.

[0053] Channel measurements were conducted in a typical conference room, with the room layout and diagram shown below. Figure 3 As shown, the meeting room is a rectangular area measuring 10m x 5.6m. A small conference table measuring 1.4m x 0.6m and 0.74m in height is placed in the center of the room. In a 10m x 5.5m meeting room, six seats are arranged around the central table.

[0054] Prior to measurement, the present invention calibrates the settings to eliminate system response. All data is then collected via a VNA-based frequency domain scan. The VNA directly outputs the signal amplitude and phase values ​​(i.e., S-parameters) and directly displays the multipath amplitude through its time-domain functionality. A dynamic range of up to 140 dB is achieved within the 260-400 GHz frequency range. To obtain accurate... To eliminate the influence of the frequency domain response caused by antennas, VNAs, and cables on the measurement data, back-to-back calibration of the terahertz detection system is necessary. By directly connecting the T-receive module and the S-transmit module with a 50dB attenuator, this invention can obtain the overall frequency domain response of the measurement system itself. The influence can then be eliminated through amplitude and phase normalization processing in the VNA. The calibration CTF for each antenna pair can be expressed as:

[0055] (1) In the formula, These are the measured S-parameters. and These represent the frequency domain responses of the receiving and transmitting antennas, respectively. This represents the frequency domain response caused by the VNA and the cable. The VNA measures the channel response at each frequency point on the bandwidth sequentially, and the CTF of the entire measurement bandwidth is the set of channel responses at all frequency points.

[0056] Step 2: Establish a single-input multiple-output (SIMO) cross-field model, decomposing the CTF into a superposition model of near-field components representing spherical waves and far-field components representing plane waves.

[0057] Specifically, this invention decomposes each element of the CTF matrix into a superposition of a near-field multipath component representing a spherical wave and a far-field multipath component representing a plane wave, to obtain a single-input multiple-output (SIMO) cross-field model; and aggregates the CTF into a superposition of sub-functions containing corresponding parameter sets.

[0058] First, define a wideband single-input multiple-output (SIMO) channel. The CTF of a multipath channel can be calculated using a matrix. H is a matrix where each element is a complex number with dimensions M×K. Each element can be written as a superposition of near-field MPC and far-field MPC.

[0059] (2) In the formula, m is the antenna element index, and k is the frequency index. Antenna spacing, Represents the set composed of near-field multipath components. For MPC serial number, For the first The complex gain of MPC on the first antenna element For the m-th antenna element The birth and death coefficients of each MPC, where 1 indicates visibility and 0 indicates invisibility. It is the first The time delay of each MPC It is the first in bandwidth k The frequency of each carrier wave, =c / yes The wavelength. Considering the distance difference of the spherical wavefront... It means that, among them From the first The last reflected scatterer of each MPC to the m-th The propagation distance of the antenna element, From the first The propagation distance from the last reflector of each MPC to the first Rx antenna element. This invention sets the normal of the Rx antenna array to an azimuth angle of 0 degrees and the angle of arrival to... Therefore, according to Taylor expansion, the distance difference can be derived as follows: The near-field MPC parameters that need to be extracted from the measurement data can be obtained using a set. It means that, among them and These are the start and end points of the visible region on the Rx antenna array. For far-field MPC, the distance parameter is no longer needed. and birth and death parameters and The extracted parameter set can be simplified to: Therefore, CTF H can be rewritten as:

[0060] (3) in express K Each frequency point, It contains all the parameters for MPC.

[0061] Secondly, considering the unavoidable measurement noise, the measured CTF can be expressed as: (4) in Let represent the noise matrix, assuming that each element follows an independent and identically distributed standard complex Gaussian distribution with zero mean and unit variance. Furthermore, This represents the variance of the noise matrix.

[0062] Step 3: Using the improved SAGE algorithm, with the maximum likelihood criterion as the core and combined with successive interference cancellation technology, the initial parameters of each multipath component are extracted in descending order of power.

[0063] Specifically, this invention is based on the decomposition structure of the near-field multipath components and the far-field multipath components in the SIMO cross-field model. It embeds the spherical wave range difference term and the plane wave angle term into the objective function to align the multidimensional parameter sets of the two types of components. It decomposes the multidimensional parameter sets based on the aggregation logic of CTF and extracts the initial parameters of each multipath component in descending order of power based on the maximum likelihood criterion and the narrowband far-field assumption.

[0064] Here, this invention utilizes the maximum likelihood criterion to extract the parameters of each MPC. The expected value of the log-likelihood function can be obtained by the following formula:

[0065] (5) in [·] represents the expectation operation, and vec{·} represents the vectorization operation. It refers to the number of receiving antenna elements. The number of frequency sampling points This represents the actual measured Channel Transfer Function (CTF) data matrix. The parameter set is estimated through maximum likelihood maximization in (5), and in this invention... The parameters are estimated values ​​and can be obtained using the following formula:

[0066] (6) Due to the high dimension of Γ, the direct maximization of Equation (6) is computationally expensive. Instead, this invention employs a successive interference cancellation (SIC) framework, which operates under the assumption of orthogonality between the various MPCs. Within the SIC framework, components are extracted in descending power order, thereby mitigating leakage from the dominant MPC to the weaker MPCs. Due to the inherently high delay and angular resolution of wideband massive MIMO channels, any pair of MPCs whose delay or angular separation exceeds the system resolution limit can be considered mutually orthogonal. Therefore, this work employs SIC technology to retrieve the parameters of each MPC sequentially. Specifically, the first... Each component is obtained by maximizing it.

[0067] (7) E represents the mathematical expectation. Represents the log-likelihood function, where The CTF of the measurement is represented by the first... The expected value of each MPC can be calculated using the following formula: (8) These are estimated values ​​for the MPC parameters. Take from 1 -1, the strongest MPC starts from Extract from. Then from Subtract the reconstructed channel response given in (8) from the middle, and use the residual to estimate the subsequent MPC. This process continues until all L significant MPCs are identified. Furthermore, it can be shown that maximizing the log-likelihood function in (7) is equivalent to maximizing the objective function. :

[0068] (9) (10) in Indicates no The A set of parameters for each MPC, where and as well as It can be calculated using the following formula: (11) in This indicates the conjugate transpose. This represents the steering vector of the Rx antenna array.

[0069] (12) in It is the mask matrix that defines the visible region of MPC, that is: (13) in and Let D represent a vector with all zeros and a vector with all ones, each containing n entries. Furthermore, D( The delay matrix is ​​defined by the following formula.

[0070] (14) Step 4: Based on the initial parameters of the multipath components, first estimate the multipath delay and angle of arrival (AoA); use the BIC criterion to distinguish between near and far fields, then estimate the parameters accordingly, and iteratively optimize the parameters in the order of AoA, distance, and delay until convergence.

[0071] Specifically, this invention uses the initial parameters of the multipath components as a basis to first estimate the multipath delay and angle of arrival; then, it uses the BIC criterion to distinguish the near-field and far-field categories, and uses a spherical wave or plane wave model to obtain the final near-field MPC parameter set and far-field MPC parameter set based on the discrimination results.

[0072] Step 4 will be explained in further detail below: To reduce computational load, a coarse-to-fine search strategy was adopted during initialization. The coarse stage first calculated delay and angle under narrowband and far-field models. Therefore, the parameter spaces of delay and angle satisfy the Orthogonal Random Measurement (OSM) condition, thus achieving the original SAGE processing. The coarse-to-fine search process ignores the birth and death of MPCs in the spatial domain, indicating... It is set as a full matrix. In the next step, the delay and angle parameters are estimated according to the principles of machine learning.

[0073] 1) Coarse parameter estimation for narrowband and far-field assumptions: In this step, the first... The initial delay and angle of arrival (AoA) of each MPC. The initial estimated time delay of each MPC, i.e. It can be calculated using the following formula:

[0074] (15) Refers to the first The path in the first The frequency response observed on each antenna element. Then, the initial estimate of AoA. Calculated by the following formula:

[0075] (16) in, This represents the array steering vector of Rx used in the coarse estimation step, calculated based on the far-field assumption.

[0076] 2) Identification of near-field and far-field MPC: Use the Bayesian Information Criterion (BIC) to determine whether MPC belongs to the near field or the far field. Extract the first... Phase vector of MPC The BIC values ​​were calculated by fitting the data using both linear and quadratic models.

[0077] (17) (18) in and This represents the estimated error variance after fitting a linear phase model and a quadratic (spherical wave) phase model, respectively, when using the Bayesian Information Criterion (BIC) for model evaluation. If... < If the result is positive, it is classified as far-field MPC; otherwise, it is classified as near-field MPC.

[0078] 3) Parameter precise estimation: In this step, the search will be refined by considering both spherical and planar wavefront assumptions to improve estimation accuracy.

[0079] When the When each MPC is a near-field MPC, that is The near-field parameters can be estimated by solving the following problem. and as follows: (19) in It is the steering vector under the near-field assumption, that is: (20) Then, proceed by solving the following problems. Detailed explanation: (twenty one) It is the first The time delay of the MPC, finally, The estimate is obtained by the following formula: (twenty two) When the Each MPC is a far-field MPC, i.e. At that time, the near-field steering vector in (19)-(22) can be used to guide the near-field. Replace with far-field steering vector To estimate far-field parameters and .

[0080] 4) Iterative optimization So far, the estimated parameter set has been obtained. Then to Perform iterative refinement until convergence. In each iteration, the parameter set... The elements in the equation are updated in the following order: AoA, distance between the scatterer and Rx, delay, birth-death coefficient, and complex amplitude. The process is as follows:

[0081] (twenty three) (twenty four) (25) (26) in Represents the first of the given parameters i The estimation results of the second iteration It is and The first estimate generated by substituting into (13) Represented as At the end of each iteration cycle, the value of the likelihood function in (5) is calculated and compared with the likelihood value in the previous iteration. If the increase in the likelihood function compared with the previous iteration is less than a certain threshold, the iteration is considered to have converged.

[0082] (27) here, This represents the relative likelihood improvement threshold. The iteration will stop once the number of iterations exceeds a certain threshold.

[0083] Step 5: Based on the optimized parameters, calculate the time delay power spectral density through Fourier transform, fit the path loss using the CI model, calculate the near-field ratio according to the proportion of near-field multipaths, and output the complete parameter set.

[0084] Step 5 specifically includes: Time-delay power spectral density analysis: The CTF is converted to a time-domain response using Fourier transform, revealing the power distribution in the time-delay domain. This method can accurately extract multipath components, such as... Figure 4As shown, the residual power is significantly reduced.

[0085] Path loss analysis: The signal attenuation characteristics were analyzed using a Close-in (CI) model for fitting. (28) in This represents the path loss index. This represents the distance between Tx and Rx. This represents the reference distance between Tx and Rx, set to 1m. Indicates the carrier frequency. The shadow fading (in dB) follows a zero-mean Gaussian distribution. The analysis results are as follows: Figure 5 As shown, the path loss exponent increases with increasing frequency.

[0086] Near-field multipath ratio calculation: Quantifying the proportion of near-field MPC and analyzing its relationship with transmission distance and carrier frequency: The percentage of near-field MPC quantifies the proportion of spatially resolvable near-field MPCs among all MPCs. It is defined as a ratio:

[0087] (29) in This indicates the number of MPCs identified as near-field paths. This indicates the far-field MPC count.

[0088] Analysis results as follows Figure 6 As shown, the near-field proportion increases with increasing frequency and decreasing distance.

[0089] Spatial-frequency correlation analysis: Calculate the spatial-frequency correlation function to obtain the quantitative results of the channel's correlation in the spatial and frequency dimensions, and evaluate the channel's variation characteristics in the spatial and frequency domains. (30) in, It is the measured CTF for each Tx-Rx speaker pair. Indicates the spatial difference on the antenna array. This indicates the frequency difference spanning a 20GHz scan bandwidth.

[0090] The correlation quantification results of the channel in the spatial and frequency dimensions include: The correlation coefficient under different combinations of antenna spacing (Δd) and frequency difference (Δf) intuitively reflects the variation law of the channel in the spatial frequency domain.

[0091] Differences in space-frequency correlation under antenna alignment / misalignment, different transmission distances, and different terahertz frequency bands (260GHz / 330GHz / 380GHz).

[0092] By quantifying the correlation results of channels in the spatial and frequency dimensions, the spatial consistency and frequency selectivity of channels are quantified, providing data support for system optimization such as beam management and subchannel allocation.

[0093] Analysis results as follows Figures 7-10 As shown, the correlation is affected by antenna alignment and frequency.

[0094] The parameter estimation method for cross-field UHF wireless channels provided by this invention has the following advantages: 1. Cross-field adaptive estimation capability: By intelligently identifying near-field and far-field multipath components through the Bayesian information criterion, and using spherical wave model and plane wave model for parameter estimation respectively, it effectively solves the estimation bias problem caused by model mismatch in cross-field environment of traditional algorithm and significantly improves the accuracy of parameter estimation.

[0095] 2. Precise characterization of spatial non-stationarity: The system innovatively introduces visible region parameters to model near-field multipath components. By estimating the start and end antenna indices, it accurately describes the spatial birth and death phenomena of multipath components, thereby achieving a precise characterization of the spatial non-stationarity of terahertz channels.

[0096] 3. Quantitative characterization of near-field effect: For the first time, the intrinsic relationship between near-field multipath ratio and transmission distance and carrier frequency is accurately quantified, revealing the core principle that "the shorter the transmission distance and the higher the carrier frequency, the more significant the near-field effect", providing key theoretical basis for the design of 6G ultra-high frequency terahertz system.

[0097] 4. Scatterer localization and sensing capability: Based on the estimated multipath component parameters (distance, angle), the two-dimensional localization of the main scatterers in the physical propagation environment is realized, which not only deepens the understanding of the channel propagation mechanism, but also provides underlying support for the integrated sensing and communication technology.

[0098] 5. Comprehensive spatial-frequency domain characterization: Based on accurately estimated channel parameters, the spatial-frequency correlation function is calculated to accurately reflect the channel's variation characteristics in the spatial-frequency domain, providing reliable data support for the optimization of key technologies such as beam management and sub-channel allocation, and significantly improving the system's spectral efficiency.

[0099] 6. Iterative optimization of convergence performance: An improved spatial alternation generalized expectation maximization algorithm framework is adopted. By continuously eliminating disturbances and optimizing parameters alternately through the maximum likelihood criterion, the algorithm can still converge quickly in complex environments across fields, and the residual power is significantly reduced.

[0100] 7. Multi-scenario adaptability and practicality: Supports channel measurement and parameter estimation in multiple terahertz frequency bands such as 260GHz, 330GHz and 380GHz. It eliminates equipment response through system calibration, ensuring that clean channel response data can be obtained in different indoor scenarios, and has wide engineering applicability.

[0101] The combined effect of these superior characteristics makes the terahertz cross-field channel parameter estimation method provided by this invention exhibit significant technical advantages and application value in 6G communication system design, providing reliable technical support for terahertz channel modeling, system design and performance optimization.

[0102] Based on the same inventive concept, the present invention also provides a parameter estimation system for cross-field ultra-high frequency wireless channels, comprising: The channel measurement module is used to capture channel response data of cross-field channels in the UHF terahertz band within the actual base station coverage area, remove the device's own interference information from the channel response data, and obtain the channel transfer function (CTF) containing only the wireless channel characteristics; the CTF is presented as an M×K complex value matrix, where M is the number of antenna elements at the receiving end and K is the number of frequency points.

[0103] The signal modeling module decomposes each element of the CTF matrix into a superposition of near-field multipath components representing spherical waves and far-field multipath components representing plane waves, resulting in a single-input multiple-output (SIMO) cross-field model; and aggregates the CTF into a superposition of sub-functions containing corresponding parameter sets.

[0104] The first estimation module is used to embed spherical wave range difference terms and plane wave angle terms into the objective function based on the decomposition structure of near-field and far-field multipath components in the SIMO cross-field model, and align the multidimensional parameter sets of the two types of components. It decomposes the multidimensional parameter sets based on the aggregation logic of CTF, and extracts the initial parameters of each multipath component in descending order of power based on the maximum likelihood criterion and the narrowband far-field assumption.

[0105] The first estimation module is used to initially estimate the multipath delay and angle of arrival based on the initial parameters of the multipath components; it then uses the BIC criterion to distinguish the near-field and far-field categories, and uses a spherical wave or plane wave model to obtain the final near-field MPC parameter set and far-field MPC parameter set based on the discrimination results.

[0106] The various modules in the aforementioned parameter estimation system for cross-field UHF wireless channels can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware within or independently of the processor in a computer device, or stored in software within the memory of the computer device, so that the processor can call and execute the corresponding operations of each module.

[0107] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps in an embodiment of a parameter estimation method for cross-field ultra-high frequency wireless channels. Specific implementation methods can be found in the method embodiments, and will not be repeated here.

[0108] Furthermore, the present invention also provides a non-transitory computer-readable storage medium containing instructions on which a computer program is stored. For example, a memory containing instructions that can be executed by a processor of a computer device to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. When the computer program is executed by the processor, it can implement the steps in an embodiment of a parameter estimation method for cross-field ultra-high frequency wireless channels. Specific implementation methods can be found in the method embodiments, which will not be repeated here.

[0109] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0110] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0111] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0112] These computer program instructions can also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0113] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail in this specification and embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention; and all technical solutions and improvements that do not depart from the spirit and scope of the present invention are covered within the protection scope of the present invention patent. No reference numerals in the claims should be construed as limiting the scope of the claims. Any simple variations or equivalent substitutions of technical solutions that can be readily obtained by those skilled in the art within the scope of the technology disclosed in the present invention are within the protection scope of the present invention.

Claims

1. A method for parameter estimation across a field region ultra-high frequency wireless channel, characterized in that, The method comprises the following steps: capturing channel response data of a hyper-frequency terahertz band cross-field channel in an actual base station coverage area, removing device self-interference information in the channel response data, and obtaining a channel transfer function (CTF) containing only wireless channel characteristics; the CTF is in the form of an M×K complex value matrix, M is the number of receiving end antenna elements, and K is the number of frequency points; decomposing each element of the CTF matrix into a superposition of a near-field multipath component representing a spherical wave and a far-field multipath component representing a plane wave, and obtaining a single-input multiple-output (SIMO) cross-field model; and aggregating the CTF into a sub-function superposition form containing a corresponding parameter set; based on the decomposition structure of the near-field multipath component and the far-field multipath component in the SIMO cross-field model, embedding a spherical wave distance difference term and a plane wave angle term in an objective function, and aligning the multi-dimensional parameter sets of the two types of components; relying on the aggregation logic of the CTF to disassemble the multi-dimensional parameter sets, extracting initial parameters of each multipath component in power descending order based on a maximum likelihood criterion and a narrowband far-field assumption; based on the initial parameters of the multipath components, preliminarily estimating multipath time delays and angles of arrival; discriminating near-field and far-field categories through a BIC criterion, and obtaining final near-field MPC parameter sets and far-field MPC parameter sets according to the discrimination results by using a spherical wave model or a plane wave model.

2. The method for parameter estimation over cross-field domain UHF wireless channel according to claim 1, characterized in that, A terahertz channel measurement system is used to capture channel response data of a hyper-frequency terahertz band cross-field channel in an actual base station coverage area, the terahertz channel measurement system comprising a computer as a control platform, a displacement platform constituting a virtual antenna array, and a detection platform for acquiring channel responses; the detection platform comprises a terahertz transmitting (Tx) module, a receiving (Rx) module, and a vector network analyzer (VNA); The VNA is used to scan and collect signal amplitude and phase values of multiple frequency points and multiple antenna elements, and to obtain original channel response data containing wireless channel characteristics and device self-interference; then, the self-frequency domain response of the measurement system is extracted through back-to-back calibration, and the device interference is removed through amplitude and phase normalization processing, and finally, a channel transfer function (CTF) in the form of an M×K matrix is obtained, which only reflects the wireless channel characteristics.

3. The method for parameter estimation over cross-field domain UHF wireless channel according to claim 2, characterized in that, The channel transfer function (CTF) is: ； wherein are measured signal amplitude and phase values, and denote the frequency domain responses of the receive and transmit side antennas, respectively, denote the frequency domain responses of the VNA and cable.

4. The method for parameter estimation over a cross-field domain UHF wireless channel according to claim 3, wherein, each element of the CTF matrix is decomposed into a superposition of a near-field multipath component representing a spherical wave and a far-field multipath component representing a plane wave, and a single-input multiple-output (SIMO) cross-field model is obtained; CTF of a multipath channel is represented by a matrix where each element in H is written as a superposition of near-field and far-field multipath components, resulting in a single-input multiple-output (SIMO) cross-field domain model: ； In the formula, m is the antenna element index, and k is the frequency index. Antenna spacing, Represents the set composed of near-field multipath components. For MPC serial number, For the first The complex gain of MPC on the first antenna element For the m-th antenna element The birth and death coefficient of each MPC, It is the first The time delay of each MPC It is the first in bandwidth k The frequency of each carrier wave, =c / yes The wavelength; considering the distance difference of the spherical wavefront. It means that, among them From the first The last reflected scatterer of each MPC to the m-th The propagation distance of the antenna element, From the first The propagation distance from the last reflector of an MPC to the first Rx antenna element; the CTF is aggregated into a sub-function superposition form containing a corresponding parameter set; ； wherein represents K frequency points, , containing parameters of all MPCs.

5. The method for parameter estimation over a cross-field domain UHF wireless channel according to claim 4, wherein, based on the decomposition structure of the near-field multipath component and the far-field multipath component in the SIMO cross-field model, a spherical wave distance difference term and a plane wave angle term are embedded in an objective function, and the multi-dimensional parameter sets of the two types of components are aligned; relying on the aggregation logic of the CTF to disassemble the multi-dimensional parameter sets, extracting initial parameters of each multipath component in power descending order based on a maximum likelihood criterion and a narrowband far-field assumption; based on the decomposition structure of the near-field multipath component and the far-field multipath component in the SIMO cross-field model, a spherical wave distance difference term and a plane wave angle term are embedded in an objective function, and the multi-dimensional parameter sets of the two types of components are aligned; relying on the aggregation logic of the CTF to disassemble the multi-dimensional parameter sets, extracting initial parameters of each multipath component in power descending order based on a maximum likelihood criterion and a narrowband far-field assumption; Based on the CTF aggregation logic, the high-dimensional parameter set is decomposed into independent parameter subsets of single MPC; Based on the maximum likelihood criterion, the log-likelihood function is constructed, combined with the narrowband far-field assumption, and the initial parameters of each MPC are extracted in descending order of power; For each MPC, the expected value of the component is separated from the measured CTF through the successive interference cancellation framework, and its parameters are solved by maximizing the objective function; Define the steering vector that adapts to the near-field and far-field characteristics, integrate the complex gain and phase factor, and solve the initial value of the single MPC parameter.

6. The method for parameter estimation across inter-field region UHF wireless channels according to claim 4, wherein, Based on the initial parameters of the MPC, the MPC time delay and angle of arrival are preliminarily estimated; the BIC criterion is used to distinguish between near-field and far-field categories, and the spherical wave or plane wave model is used to obtain the final near-field MPC parameter set and far-field MPC parameter set according to the results, including: Based on the initial parameters of the MPC, the initial time delay and angle of arrival of each MPC are calculated under the narrowband and far-field assumption; Extract the phase vector of each MPC and fit it with a linear model and a quadratic model, respectively; Calculate the BIC values corresponding to the two models, and compare the BIC values to determine whether the MPC belongs to the near-field or far-field category; If it is determined to be a near-field MPC, use the steering vector related to the spherical wave model to accurately estimate its angle of arrival, scatterer distance, time delay, and complex gain; If it is determined to be a far-field MPC, use the steering vector related to the plane wave model to accurately estimate its angle of arrival, time delay, and complex gain; Iteratively update the parameters of all MPCs in the order of angle of arrival, scatterer distance, time delay, life and death coefficient, and complex amplitude; Calculate the log-likelihood function value after each iteration until the likelihood function increment before and after iteration is less than a threshold or the maximum number of iterations is reached, and output the final near-field and far-field MPC parameter set.

7. The method for parameter estimation over a cross-field domain UHF wireless channel according to claim 6, wherein, Also includes: For the final near-field and far-field MPC parameters, the CTF is converted to the time domain response by Fourier transform, and the time delay power spectral density is analyzed; Use the Close-in model to fit the path loss characteristics by substituting the path loss exponent, transmission distance, reference distance, and carrier frequency; Statistically count the number of near-field and far-field MPCs, and calculate the proportion of near-field MPCs according to the proportion formula. Extract the measured CTF of each Tx-Rx horn pair, substitute the antenna array spatial difference and frequency difference parameters, calculate the spatial-frequency correlation function, and obtain the correlation quantization results of the channel in the spatial and frequency dimensions.

8. A system for parameter estimation across a field region ultra-high frequency wireless channel, characterized by, It includes: A channel measurement module is used to capture the channel response data of the ultra-high frequency terahertz band cross-field domain channel in the actual base station coverage area, remove the device self-interference information in the channel response data, and obtain a channel transfer function CTF containing only wireless channel characteristics; the CTF is presented in the form of an M×K complex value matrix, where M is the number of receiving antenna elements and K is the number of frequency points; A signal modeling module is used to decompose each element of the CTF matrix into the superposition of near-field MPCs representing spherical waves and far-field MPCs representing plane waves, obtaining a single-input multiple-output (SIMO) cross-field domain model; and the CTF is aggregated into a sub-function superposition form containing corresponding parameter sets. The first estimation module is configured to embed a spherical wave distance difference term and a plane wave angle term in a target function based on a decomposition structure of near-field multipath components and far-field multipath components in a SIMO cross-field model, and align multi-dimensional parameter sets of the two types of components; rely on an aggregation logic of CTF to disassemble the multi-dimensional parameter sets, extract initial parameters of each multipath component in power descending order based on a maximum likelihood criterion and a narrowband far-field assumption; and preliminarily estimate multipath time delays and angles of arrival based on the initial parameters of the multipath components. The first estimation module is configured to preliminarily estimate multipath time delays and angles of arrival based on the initial parameters of the multipath components; determine near-field and far-field categories by using a BIC criterion, and obtain final near-field MPC parameter sets and far-field MPC parameter sets by using a spherical wave model or a plane wave model according to a determination result.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program comprises instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-8. The processor executes the computer program to implement steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is loaded by the processor to implement steps of the method in any one of claims 1 to 7.

Images