Indoor scene line-of-sight channel identification method and device based on polarization and K factor

Abstract

The invention provides an indoor scene line-of-sight channel identification method and device based on polarization and a K factor. The method provided by the invention comprises the following steps: collecting channel data in different polarization modes in an indoor environment; analyzing the channel data, and calculating the time delay characteristic and K factor of the channel; based on the time delay characteristic and the K factor, analyzing the time similarity and the space similarity between channels, and dividing the whole measurement time period and the transmitting end antenna array into a plurality of areas; based on the time delay characteristics and the K factors of all the channels, respectively calculating a first average value of the overall time delay characteristics and the K factors, and calculating a second average value of the time delay characteristics and the K factors of all the channels in each divided region; comparing the second average value of each region with the first average value, and determining a preliminary region state in combination with the time delay characteristic similarity of various channel data; and determining a dynamic threshold based on the cumulative distribution function of the K factor, and optimizing the initial region state based on the dynamic threshold to obtain a sight distance classification result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of line-of-sight channel identification, and in particular to an indoor scene line-of-sight channel identification method and device based on polarization and K factors. BACKGROUND

[0002] With the rapid development of 6G communication technology, the fine management of signal propagation characteristics in indoor scenes, as an important application environment for wireless communication, has become increasingly prominent. Indoor line-of-sight (LOS) channels, due to their small propagation loss, high signal strength, and stable path, ensure positioning accuracy and communication quality. The accurate identification of LOS and non-line-of-sight (NLOS) channels is a key to improving the performance of indoor wireless communication systems and user experience, and it also provides an important basis for data analysis in indoor near-field environments. Therefore, it is of great significance to carry out research on indoor line-of-sight channel identification.

[0003] Currently, the identification of LOS and NLOS in the academic field mainly focuses on channel characteristic parameters and intelligent algorithms. Traditional methods are mostly based on K factors, time of arrival (TOA), phase difference, etc. For example, a detector is designed by deriving the probability density function of the K factor, or the variance of the phase difference is used for discrimination. With the development of machine learning, algorithms such as convolutional neural networks (CNN) are widely used to improve identification accuracy by extracting channel impulse response (CIR) features and combining time-frequency domain analysis. The identification accuracy of some solutions in specific scenarios can reach more than 98%. However, existing identification methods have obvious limitations. On the one hand, traditional methods based on K factors and other parameters rely on fixed thresholds, but in 6G near-field communication and large-scale MIMO systems, the K factor threshold of indoor dynamic scenes will change due to very short communication distances. It is difficult to reflect the real characteristics of the channel by identifying LOS and NLOS with a fixed threshold specified by a certain standard. On the other hand, existing methods lack systematic and dynamic comprehensive consideration, making it difficult to adapt to complex and changing indoor environments.

[0004] Therefore, there is an urgent need for a method to solve the adaptability problem of existing methods in dynamic scenarios and achieve accurate classification of line-of-sight, non-line-of-sight, and obstructed line-of-sight channels. SUMMARY

[0005] Therefore, the present application provides an indoor scene line-of-sight channel identification method and device based on polarization and K factors to solve the adaptability problem of existing methods in dynamic scenarios and achieve accurate classification of line-of-sight, non-line-of-sight, and obstructed line-of-sight channels.

[0006] Specifically, the present application is implemented through the following technical solutions:

[0007] The first aspect of the present application provides an indoor scene line-of-sight channel identification method based on polarization and K factors, which comprises:

[0008] Channel data under different polarization modes were collected in a laboratory environment.

[0009] The channel data is analyzed to calculate the channel's delay characteristics and K-factor;

[0010] Based on the aforementioned delay characteristics and K-factor, the temporal and spatial similarities between channels are analyzed. Based on the temporal and spatial similarities, the entire measurement period and the transmitting antenna array are divided into multiple regions.

[0011] Based on the calculated delay characteristics and K-factor of all channels, the first average value of the overall delay characteristics and K-factor is calculated respectively. For each divided region, the second average value of the delay characteristics and K-factor of all channels in each region is calculated.

[0012] The second average value of each region is compared with the corresponding first average value, and the preliminary region status is determined by combining the similarity of the time delay characteristics of multiple channel data.

[0013] A dynamic threshold is determined based on the cumulative distribution function of the K factor, and the preliminary region state is optimized based on the dynamic threshold to obtain the line-of-sight classification result.

[0014] The second aspect of this application provides an indoor scene line-of-sight channel recognition device based on polarization and K-factor, the device comprising an acquisition module, a calculation module, a partitioning module, a determination module and an optimization module;

[0015] The acquisition module is used to acquire channel data under different polarization modes in a laboratory indoor environment.

[0016] The calculation module is used to analyze the channel data and calculate the channel's delay characteristics and K-factor.

[0017] The partitioning module is used to analyze the temporal and spatial similarity between channels based on the time delay characteristics and the K factor, and to divide the entire measurement period and the transmitting antenna array into multiple regions based on the temporal and spatial similarity.

[0018] The calculation module is also used to calculate the first average value of the overall delay characteristics and K factor based on the calculated delay characteristics and K factor of all channels, and to calculate the second average value of the delay characteristics and K factor of all channels in each region after division.

[0019] The determining module is used to compare the second average value of each region with the corresponding first average value, and combine the similarity of the delay characteristics of multiple channel data to determine the preliminary region status.

[0020] The optimization module is used to determine a dynamic threshold based on the cumulative distribution function of the K factor, and to optimize the preliminary region state based on the dynamic threshold to obtain the line-of-sight classification result.

[0021] The line-of-sight channel recognition method and apparatus for indoor scenes based on polarization and K-factor provided in this application achieves line-of-sight recognition through an overall process of "multi-dimensional feature extraction → region division → preliminary judgment → dynamic optimization". First, channel data with different polarization modes are collected. The feature dimensions are enriched by utilizing the differences in polarization characteristics (such as the different performance of different polarizations in obstacle reflection). Then, the time delay characteristics and K-factor are calculated. Regions with similar propagation characteristics are divided by time and space similarity. The preliminary state judgment is completed based on the time delay characteristics of the region and the whole, the difference of the average value of K-factor, and the similarity of time delay characteristics. This process integrates multiple features and polarization information to reduce single-dimensional bias. Then, a dynamic threshold is determined based on the cumulative distribution function of K-factor for secondary adjustment. The dynamic threshold can adapt to the actual distribution law of K-factor under different scenarios and avoid the limitations of fixed threshold. The initial judgment reduces basic misjudgments by supplementing with multi-feature cross-validation and polarization characteristics, laying a reliable foundation for secondary optimization. For example, the consistency of multi-polarization data can eliminate random errors under single polarization. Secondary optimization of dynamic thresholds can correct misjudgments in boundary areas in the initial judgment, adapt to changes in the indoor environment (such as changes in the K-factor distribution caused by the movement of obstacles), and ultimately improve the accuracy of line-of-sight recognition and adaptability to complex indoor scenes, forming a closed loop of "comprehensive feature capture → precise correction of deviations", making the classification results more consistent with the actual physical environment. Attached Figure Description

[0022] Figure 1 A flowchart of the indoor scene line-of-sight channel recognition method based on polarization and K-factor provided in Embodiment 1 of this application;

[0023] Figure 2 This is a schematic diagram of the structure of the indoor scene line-of-sight channel recognition device based on polarization and K-factor provided in Embodiment 2 of this application. Detailed Implementation

[0024] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0025] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0026] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0027] The following specific embodiments are given to illustrate the technical solution of this application in detail.

[0028] Figure 1 This is a flowchart illustrating the indoor scene line-of-sight channel recognition method based on polarization and K-factor provided in Embodiment 1 of this application. Please refer to... Figure 1 The method provided in this embodiment may include:

[0029] S101. Collect channel data under different polarization modes in a laboratory indoor environment.

[0030] Specifically, different polarization modes refer to the vibration direction of the electric field intensity of a single electromagnetic wave. Common polarization modes include vertical polarization and horizontal polarization. Common polarization and cross polarization are polarization matching relationships between the transmitter and receiver. Common polarization usually means that the polarization directions of the transmitter and receiver antennas are the same (such as both being vertically polarized); cross polarization means that the polarization directions of the transmitter and receiver antennas are perpendicular (such as the transmitter being vertically polarized and the receiver being horizontally polarized).

[0031] Furthermore, channel data refers to the signal characteristic parameters recorded when electromagnetic waves pass through a channel (i.e., the propagation path) during propagation. The channel data in this application mainly includes: K-factor, time delay characteristics, etc. Specific channel data will be described in the following embodiments.

[0032] In specific implementation, the acquisition of channel data under different polarization modes in a laboratory indoor environment includes: setting up a fixed antenna array at the transmitting end, deploying array elements in a vertical polarization manner; configuring two antennas with different polarization modes at the receiving end; the antennas include a vertically polarized antenna and a horizontally polarized antenna; connecting a channel measuring instrument to the transmitting end and the receiving end; the transmitting end continuously transmits signals through the antenna array, and the antenna at the receiving end is fixed on a movable device and moves at a constant speed along a preset route; during the movement of the receiving end, the vertically polarized antenna and the horizontally polarized antenna synchronously receive the signals emitted by the transmitting end, and the channel measuring instrument collects and records channel data under the two polarization modes in real time.

[0033] Specifically, a fixed antenna array is set up at the transmitting end, with array elements deployed in a vertical polarization manner. Two antennas with different polarizations, a vertically polarized antenna and a horizontally polarized antenna, are configured at the receiving end. The transmitting and receiving ends are connected to a channel measurement instrument. The transmitting end continuously transmits signals through the antenna array, while the two antennas at the receiving end are fixed to a movable device that moves at a constant speed along a preset route. During the movement of the receiving end, the vertically polarized and horizontally polarized antennas synchronously receive the signals transmitted from the transmitting end, and the channel measurement instrument collects and records channel data in real time under both vertical and horizontal polarization modes.

[0034] For example, in one embodiment, the measurement was conducted in the Marine Communication Laboratory of Zhejiang Ocean University, in an indoor laboratory environment. The carrier frequency band was 3.4 GHz. The transmitter used a fixed, stationary 4×32-element VMI array, with all elements arranged in vertical polarization. The receiver used two antennas arranged with different polarizations, RX1 in vertical polarization and RX2 in horizontal polarization. The receiving antenna was fixed on a trolley and moved at a constant speed in a straight line along a predetermined route. A pillar existed between the transmitter and receiver, creating an obstruction in the propagation path. A high-precision channel measurement instrument was used to collect experimental data in an indoor laboratory setting for the 3.4 GHz VMI array application. Table 1 shows the laboratory indoor environment configuration provided in this application.

[0035] Table 1

[0036] Carrier frequency 3.4 GHz Signal interval 80 μs Signal bandwidth 100 MHz Time delay resolution 10 ns On-top transmit power 14.7 dBm TX antenna Triangular periodic log antenna TX antenna number 128 TX antenna gain 6.5 dBi RX antenna number 2 RX antenna gain 5.6 dBi TX antenna polarization Vertical polarization RX antenna polarization Horizontal and vertical polarization TX antenna height [0.6, 1.35] m RX antenna height 1.25 m Course distance 6 m Average speed 0.6 m / s File size 0.261 GB

[0037] S102. Analyze the channel data and calculate the channel delay characteristics and K-factor.

[0038] The delay characteristics include average delay and root mean square delay.

[0039] Specifically, delay characteristics are parameters describing the differences in propagation time of multipath signals. In this application, delay characteristics include average delay and root mean square (RMS) delay. Average delay refers to the average time it takes for all multipath signals to arrive at the receiver, reflecting the overall propagation delay level of the multipath signals. RMS delay refers to the degree of dispersion of the multipath signal delay relative to the average delay, reflecting the degree of dispersion of the multipath signals (the larger the value, the more dispersed the multipath propagation). The Rician K-factor is a parameter that measures the relative strength of the line-of-sight (LOS) component and the multipath component. It is defined as the ratio of the power of the LOS signal to the total power of all non-line-of-sight (NLOS) multipath signals. The larger the K-factor, the higher the proportion of the LOS component, and the less the channel is affected by multipath interference; conversely, the lower the K-factor, the higher the proportion of the multipath component.

[0040] In specific implementation, the analysis of the channel data and the calculation of the channel delay characteristics include: determining each channel path from the channel data, obtaining the power delay spectrum and path delay corresponding to each channel path; using the power delay spectrum of each channel path as a weight, performing a weighted summation of the path delay of the corresponding channel path, and then dividing by the sum of the power delay spectra of all channel paths to obtain the average delay of the channel; wherein, the power delay spectrum is obtained by averaging the energy of the channel impulse response function over a time interval.

[0041] Specifically, from the collected channel data, all existing channel paths are identified and determined. For each identified channel path, its corresponding power delay spectrum and path delay are extracted. The energy of the channel impulse response function is averaged over a specific time interval to obtain the power delay spectrum of each channel path. Based on the determined power delay spectrum, the corresponding path delay is further matched. Using the power delay spectrum of each channel path as a weight, the path delay of that channel path is weighted and calculated. The weighted results of all channel paths are summed to obtain a weighted sum. The sum of the power delay spectra of all channel paths is calculated. The obtained weighted sum is divided by the sum of the power delay spectra to obtain the average channel delay. The specific implementation process for extracting the power delay spectrum and path delay of channel paths can be found in the descriptions in related technologies, and will not be repeated here.

[0042] For example, in one embodiment, the calculation process for the average latency can be expressed as follows:

[0043] ;

[0044] Among them, the The average delay; The number of channel paths; The power delay spectrum; The sum of the power delay spectra of all channel paths. The This refers to path delay.

[0045] Optionally, the step of analyzing the channel data and calculating the channel delay characteristics includes: calculating the difference between the path delay of each channel path and the average delay, squaring each difference; using the power delay spectrum of each channel path as weights, weighting and summing the squared differences; dividing the summation result by the sum of the power delay spectra of all channel paths, and taking the square root of the result to obtain the root mean square delay of the channel.

[0046] Specifically, based on the determined channel paths, the path delay of each channel path is extracted, and the calculated average delay is retrieved. For each channel path, the difference between its path delay and the average delay is calculated. Each calculated difference is squared to obtain the squared difference result for each channel path. Using the power delay spectrum of each channel path as a weight, this weight is multiplied by the squared difference result of the corresponding path to obtain the weighted squared value of each channel path. The weighted squared values ​​of all channel paths are summed to obtain the weighted sum of squares. The sum of the power delay spectra of all channel paths is calculated. The weighted sum of squares is divided by the sum of the power delay spectra to obtain a quotient. The square root of the quotient is then calculated, and the result is the root mean square delay of the channel.

[0047] For example, in one embodiment, the calculation process of the root mean square delay can be expressed as follows:

[0048] ;

[0049] Among them, the The root mean square delay; The number of channel paths; The power delay spectrum; The sum of the power delay spectra of all channel paths. The For path delay; the This represents the average latency.

[0050] Optionally, the step of analyzing the channel data and calculating the K-factor of the channel includes: performing cluster analysis on the time delay and energy information in the channel data to determine the energy of the earliest arrival path and the energy of the scattering path; the energy of the scattering path is the difference between the sum of the power delay spectra of all channel paths and the energy of the earliest arrival path; taking the logarithm of the ratio of the energy of the earliest arrival path to the energy of the scattering path to obtain the K-factor in decibels.

[0051] Specifically, the time delay and energy information of each channel path in the channel data are extracted. Cluster analysis is performed on the extracted time delay and energy information to distinguish paths with different characteristics. From the cluster analysis results, the earliest arrival path is identified and determined, and its corresponding energy is extracted as the energy of the earliest arrival path. The sum of the power delay spectra of all channel paths is calculated, and the energy of the earliest arrival path is subtracted from the sum of the power delay spectra of all channel paths to obtain the energy of the scattering path. The ratio between the energy of the earliest arrival path and the energy of the scattering path is calculated, and the logarithm of the obtained ratio is performed. The result is the K-factor in decibels. The specific implementation process of cluster analysis can be found in descriptions in related technologies and will not be elaborated here.

[0052] For example, in one embodiment, the calculation process of the K factor can be expressed as follows:

[0053] ;

[0054] Among them, the The K factor; The energy of the earliest arriving path; the This is the sum of the power delay spectra of all channel paths.

[0055] The method provided in this embodiment calculates the channel's delay characteristics (average delay, root mean square delay) and K-factor, providing a reliable basis for line-of-sight identification by quantifying channel propagation characteristics in multiple dimensions. Specifically: the delay characteristics quantify the temporal distribution of multipath propagation by calculating the delay of each path, average delay, and root mean square delay—in line-of-sight scenarios, multipaths are fewer and more concentrated, with shorter average delays and smaller root mean square delays; in non-line-of-sight scenarios, multipaths are more dispersed, with longer average delays and larger root mean square delays, thus distinguishing the channel state from a time dimension; the K-factor determines the energy of the earliest arrival path (usually the line-of-sight path) and the scattering path through cluster analysis, and calculates the logarithm of their ratio, directly reflecting the energy proportion of the line-of-sight component and the multipath component—the K-factor is larger in line-of-sight scenarios and smaller in non-line-of-sight scenarios, providing a basis for judgment from an energy dimension. The combination of the two forms a complementary relationship: the temporal distribution characteristics of the delay feature can compensate for the bias of the K factor due to accidental strong multipath interference, and the energy proportion information of the K factor can correct the anomalies of the delay feature affected by special environments. Through multi-dimensional joint analysis, the misjudgment of a single indicator can be effectively avoided, and the accurate identification of line-of-sight, non-line-of-sight, and occluded line-of-sight states can be achieved, thereby improving the accuracy and reliability of the classification results.

[0056] S103. Based on the aforementioned delay characteristics and K-factor, analyze the temporal and spatial similarities between channels, and based on the aforementioned temporal and spatial similarities, divide the entire measurement period and the transmitting antenna array into multiple regions.

[0057] Specifically, temporal similarity refers to the degree of correlation between the characteristics (such as delay characteristics and K-factor) of the same channel at different points in time. In other words, it refers to the similarity or consistency of channel characteristics as they change over time. If the differences in channel characteristics at different times are small, the temporal similarity is high; otherwise, it is low. Spatial similarity refers to the degree of correlation between the channel characteristics corresponding to different spatial locations (specifically, different antennas in the transmitter's antenna array). In other words, it refers to the similarity in delay characteristics, K-factor, etc., of the channels experienced by different antennas. If the channel characteristics of different antennas are similar, the spatial similarity is high; otherwise, it is low.

[0058] Furthermore, regions are divided based on temporal and spatial similarity. Temporal similarity reflects the stability of the channel over time; channel characteristics within the same temporally related region exhibit consistent patterns of change over time and can be considered to have similar temporal evolution characteristics. Spatial similarity reflects the distribution pattern of the channel in the spatial dimension; antennas within the same spatially related region experience similar channel characteristics and can be considered to be in similar propagation environments. By combining temporal and spatial similarity to divide regions, the entire measurement period and the transmitting antenna array can be divided into multiple sub-regions with relatively consistent channel characteristics. The channel characteristics within each sub-region are more stable and regular, facilitating subsequent precise channel analysis, modeling, or line-of-sight identification for different regions. This reduces analysis errors caused by excessive differences in channel characteristics and improves the overall effectiveness and accuracy of the processing.

[0059] In specific implementation, based on the aforementioned delay characteristics and K-factor, the temporal and spatial similarities between channels are analyzed. Based on the temporal and spatial similarities, the entire measurement period and the transmitting antenna array are divided into multiple regions, including: extracting delay characteristics and K-factor data at different times, calculating the correlation coefficient of data at adjacent times, and analyzing the temporal similarity of the channels; extracting delay characteristics and K-factor data corresponding to different antenna elements in the transmitting antenna array, calculating the correlation coefficient of data between array elements, and analyzing the spatial similarity of the channels; dividing the entire measurement period into several time segments when the temporal similarity is lower than a preset time threshold, based on the strength of the temporal similarity; dividing the transmitting antenna array into several spatial subarrays when the spatial similarity is lower than a preset spatial threshold, based on the distribution characteristics of the spatial similarity; and combining the time segments and spatial subarrays to form multiple spatiotemporal regions.

[0060] Specifically, from the calculated channel data, the time delay characteristics (average time delay, root mean square time delay) and K-factor data corresponding to different times are extracted. For the extracted data at different times, the correlation coefficient between the time delay characteristics and K-factor data of two adjacent times is calculated, and the time similarity of the channel is analyzed using this coefficient. From the transmitting antenna array, the time delay characteristics and K-factor data corresponding to each different antenna element are extracted. The correlation coefficient between the time delay characteristics and K-factor data of different antenna elements in the transmitting antenna array is calculated, and the spatial similarity of the channel is analyzed using this coefficient. Based on the strength variation of the analyzed time similarity, a preset time threshold is set. When the time similarity between adjacent times is lower than the preset time threshold, the entire measurement period is segmented into several time sub-segments. Based on the distribution characteristics of the analyzed spatial similarity, a preset spatial threshold is set. When the spatial similarity between different antenna elements in the transmitting antenna array is lower than the preset spatial threshold, the antenna array is divided into blocks, and the array is divided into several spatial subarrays. The various time segments and spatial subarrays obtained from the division are combined one by one to form multiple spatiotemporal regions containing time and spatial information.

[0061] S104. Based on the calculated delay characteristics and K-factor of all channels, calculate the first average value of the overall delay characteristics and K-factor respectively. For each divided region, calculate the second average value of the delay characteristics and K-factor of all channels in each region.

[0062] Specifically, the latency characteristics (including average latency and root mean square latency) and K-factor data of all channels are collected. The average latency of all channels is summed and divided by the total number of channels to obtain the first average of the overall average latency. The root mean square latency of all channels is summed and divided by the total number of channels to obtain the first average of the overall root mean square latency. The K-factor of all channels is summed and divided by the total number of channels to obtain the first average of the overall K-factor. For each divided region, the latency characteristics (average latency and root mean square latency) and K-factor data of all channels within that region are extracted. The average latency of all channels within that region is summed and divided by the number of channels within that region to obtain the second average of the average latency of that region. The root mean square latency of all channels within that region is summed and divided by the number of channels within that region to obtain the second average of the root mean square latency of that region. The K-factor of all channels within that region is summed and divided by the number of channels within that region to obtain the second average of the K-factor of that region.

[0063] For example, this application uses a common-polarized antenna and a cross-polarized antenna as examples. The acquired data includes: the second average value of the common-polarization K-factor and the second average value of the cross-polarization K-factor for each region, as well as the first average value of the overall common-polarization K-factor and the first average value of the overall cross-polarization K-factor; the second average value of the common-polarization delay characteristics and the second average value of the cross-polarization delay characteristics for each region, as well as the first average value of the overall common-polarization delay characteristics and the first average value of the overall cross-polarization delay characteristics.

[0064] S105. Compare the second average value of each region with the corresponding first average value, and determine the preliminary region status by combining the similarity of the time delay characteristics of multiple channel data.

[0065] Specifically, the preliminary region status refers to the preliminary attribute judgment made for each divided region based on the comparison between the second average value of each region (the average value of the delay characteristics and K-factor within the region) and the first average value of the whole (the average value of the delay characteristics and K-factor of all channels), combined with the similarity between the delay characteristics within the region and the delay characteristics of the whole. This status reflects the overall characteristics of the channel in terms of multipath propagation in the region, such as whether it belongs to the line-of-sight region, the non-line-of-sight region, or the obstructed line-of-sight region.

[0066] In specific implementation, the second average value of the co-polarization K-factor in each region is compared with the first average value of the overall co-polarization K-factor to obtain the first branch and the second branch; based on the second average value of the cross-polarization K-factor in each region of the first and second branches, it is compared with the overall cross-polarization K-factor to obtain the third and fourth branches corresponding to the first branch, and the fifth and sixth branches corresponding to the second branch; wherein, the second average value of the cross-polarization K-factor in each region of the third and sixth branches is not greater than the first average value of the overall cross-polarization K-factor, and the third branch is the line-of-sight region. The domain is divided into six branches, which are non-line-of-sight regions. For the fourth and fifth branches, the comparison signs of the regional co-polarization delay characteristics and the global co-polarization delay characteristics, as well as the comparison signs of the regional cross-polarization delay characteristics and the global cross-polarization delay characteristics, are calculated respectively. If the comparison signs of the delay characteristics under the two polarizations are consistent, the delay characteristics are determined to be similar. For regions with similar delay characteristics under the fourth and fifth branches, they are determined to be occluded line-of-sight regions. For regions with dissimilar delay characteristics under the fourth branch, they are determined to be line-of-sight regions. For regions with dissimilar delay characteristics under the fifth branch, they are determined to be non-line-of-sight regions.

[0067] Specifically, the second average value of the co-polarization K-factor in each region is compared with the first average value of the overall co-polarization K-factor, resulting in two branches: if the second average value of the regional co-polarization K-factor is greater than the first average value of the overall co-polarization K-factor, it belongs to the first branch; if the second average value of the regional co-polarization K-factor is less than or equal to the first average value of the overall co-polarization K-factor, it belongs to the second branch. For each region in the first branch, its second average value of the cross-polarization K-factor is compared with the first average value of the overall cross-polarization K-factor: if the second average value of the regional cross-polarization K-factor is less than or equal to the first average value of the overall cross-polarization K-factor, it belongs to the third branch; if the second average value of the regional cross-polarization K-factor is greater than the first average value of the overall cross-polarization K-factor, it belongs to the fourth branch. For each region in the second branch, its second average value of the cross-polarization K-factor is compared with the first average value of the overall cross-polarization K-factor: if the second average value of the regional cross-polarization K-factor is greater than the first average value of the overall cross-polarization K-factor, it belongs to the fifth branch; if the second average value of the regional cross-polarization K-factor is less than or equal to the first average value of the overall cross-polarization K-factor, it belongs to the sixth branch. All regions corresponding to the third branch are directly classified as line-of-sight regions, and all regions corresponding to the sixth branch are classified as non-line-of-sight regions. For each region in the fourth and fifth branches, the comparison sign between the second average value of the region's co-polarization delay characteristic and the first average value of the overall co-polarization delay characteristic is calculated (if the region value > the overall value, the sign is "+"; if the region value ≤ the overall value, the sign is "-"), as well as the comparison sign between the second average value of the region's cross-polarization delay characteristic and the first average value of the overall cross-polarization delay characteristic (signing rules are the same as above). If the comparison sign of the co-polarization delay characteristic of a region is consistent with the comparison sign of the cross-polarization delay characteristic, the region is determined to have similar delay characteristics: regions with similar delay characteristics in the fourth branch are classified as occluded line-of-sight regions, and regions with similar delay characteristics in the fifth branch are classified as occluded line-of-sight regions. If the comparison sign of the co-polarization delay characteristic of a certain region is inconsistent with the comparison sign of the cross-polarization delay characteristic, the delay characteristics of that region are determined to be dissimilar: the regions with dissimilar delay characteristics in the fourth branch are determined to be line-of-sight regions, and the regions with dissimilar delay characteristics in the fifth branch are determined to be non-line-of-sight regions.

[0068] The method provided in this embodiment first divides the first and second branches by comparing the co-polarization K-factor with the global co-polarization K-factor. Then, based on the two branches, the regional cross-polarization K-factor is compared with the global cross-polarization K-factor to obtain the third to sixth branches. The third branch is directly determined to be the line-of-sight region and the sixth branch is the non-line-of-sight region. The signs of the regional co-polarization and global co-polarization time delay characteristics, and the regional cross-polarization and global cross-polarization time delay characteristics are further compared for the fourth and fifth branches. The similarity of the time delay characteristics is judged based on whether the signs are consistent, and then the occlusion line-of-sight, line-of-sight, or non-line-of-sight region is determined. Using a hierarchical comparison of dual-polarization K-factors as the core approach, this method first uses the co-polarization K-factor (a key parameter reflecting the strength of the direct path) to initially screen the approximate range of channel states. Then, the cross-polarization K-factor (reflecting polarization deflection and scattering) further narrows the decision boundary, reducing the one-sidedness of judging by a single polarization parameter. Simultaneously, for branches that cannot be directly determined after comparing dual-polarization K-factors, a sign consistency verification based on time delay characteristics is introduced. This combines the K-factor's ability to characterize the strength of direct / scattered paths with the time delay characteristic's reflection of multipath propagation consistency, achieving complementary verification of multi-dimensional parameters and effectively avoiding... Misjudgments can be caused by the non-stationarity of a single parameter (such as a traditional fixed K-factor threshold) in the channel characteristic space. Furthermore, the hierarchical comparison logic makes the judgment process progressive and the boundaries clear. First, it directly locks the line-of-sight (third branch) and non-line-of-sight (sixth branch) regions with clear features. Then, it focuses on the fourth and fifth branches with ambiguous features for fine-grained verification. This not only improves the judgment efficiency, but also ensures that the judgment results are highly matched with the actual physical occlusions in the laboratory indoor scene (such as LOS / OLOS / NLOS switching caused by obstacles such as pillars and conference tables) through multi-parameter collaboration. Finally, it ensures the accuracy of the regional state judgment.

[0069] S106. Determine a dynamic threshold based on the cumulative distribution function of the K factor, and optimize the preliminary region state based on the dynamic threshold to obtain the line-of-sight classification result.

[0070] Specifically, the cumulative distribution function (CDF) describes the probability that a random variable (K-factor) takes a value less than or equal to a certain specific value. For the K-factor, its CDF reflects the probability distribution of K-factor values ​​less than or equal to a certain threshold across the whole or a region, and can intuitively present the statistical distribution characteristics of the K-factor (such as the range of numerical concentration, distribution pattern, etc.). The dynamic threshold is a judgment criterion calculated based on the cumulative distribution function of the K-factor and adjusted according to changes in data distribution characteristics. It is not a fixed value, but is dynamically determined according to the statistical laws of the actual K-factor (such as distribution interval, probability critical point, etc.), and can adapt to the distribution differences of the K-factor in different scenarios. The line-of-sight classification result is the final state judgment of each region after optimization by the dynamic threshold. It is a correction and confirmation of the preliminary region state (line-of-sight region, non-line-of-sight region, occluded line-of-sight region), and finally clarifies the specific category of each region as line-of-sight, non-line-of-sight, or occluded line-of-sight, providing accurate classification conclusions for channel line-of-sight identification.

[0071] Furthermore, this application considers that the preliminary regional state is determined by comparing average values ​​and the similarity of time delay characteristics, which may lead to misjudgments due to statistical biases (such as local data fluctuations) or fixed judgment logic (such as atypical edge region characteristics). The dynamic threshold is generated based on the global cumulative distribution function of the K-factor, which can reflect the global statistical regularity of the data. Optimizing the preliminary results using the dynamic threshold can correct errors caused by local feature biases in the preliminary judgment, making the regional state division more consistent with the statistical characteristics of the overall channel environment, and improving the robustness and accuracy of the classification.

[0072] In specific implementation, the dynamic threshold is determined based on the cumulative distribution function of the K-factor, and the preliminary region state is optimized based on the dynamic threshold to obtain the sight distance classification result. This includes: calculating the K-factor distribution of the sight distance region, the occluded sight distance region, and the non-sight distance region in the preliminary region state, and constructing the cumulative distribution function of the overall K-factor; determining a first dynamic threshold to distinguish between the sight distance region and the occluded sight distance region, and a second dynamic threshold to distinguish between the occluded sight distance region and the non-sight distance region, based on the K-factor probability density characteristics of the three regions in the cumulative distribution function; and classifying the region as a sight distance region as a non-sight distance region. The second average value of the K-factor is compared with the first dynamic threshold. If the threshold condition is met, the viewing distance state is retained; otherwise, it is corrected to the occluded viewing distance state. The second average value of the K-factor initially determined to be a non-viewing distance region is compared with the second dynamic threshold. If the threshold condition is met, the non-viewing distance state is retained; otherwise, it is corrected to the occluded viewing distance state. The second average value of the K-factor initially determined to be an occluded viewing distance region is compared with both the first and second dynamic thresholds. If it is between the two thresholds, the occluded viewing distance state is maintained; otherwise, it is corrected to the corresponding state according to the threshold boundary. The verification results are integrated to obtain the final viewing distance classification result.

[0073] Specifically, extract K-factor data from three types of regions in the preliminary regional state: extract the second average K-factor of all regions from the determined line-of-sight region to form the line-of-sight region K-factor dataset; extract the second average K-factor of all regions from the occluded line-of-sight region to form the occluded line-of-sight region K-factor dataset; and extract the second average K-factor of all regions from the non-line-of-sight region to form the non-line-of-sight region K-factor dataset. Plot the probability density curve (or frequency distribution histogram) of the overall K-factor cumulative distribution function, and observe the distribution range of the line-of-sight region K-factor data (usually concentrated in the higher value range), the distribution range of the occluded line-of-sight region K-factor data (usually between line-of-sight and non-line-of-sight), and the distribution range of the non-line-of-sight region K-factor data (usually concentrated in the lower value range). Identify the intersection points or probability density abrupt change points of the three types of data distribution: determine the lower limit position of the line-of-sight region data distribution (i.e., the area near the lowest value in the line-of-sight region K-factor data), and the upper limit position of the non-line-of-sight region data distribution (i.e., the area near the highest value in the non-line-of-sight region K-factor data), while observing the overlap boundary between the occluded line-of-sight region data and the first two types of data. Based on the boundary between the K-factor distributions of the viewing distance region and the occluded viewing distance region, a critical value of the overlapping interval of the two data distributions is selected as the first dynamic threshold. For example, if the K-factor data in the viewing distance region is mainly distributed in [K1_min, K1_max], and the data in the occluded viewing distance region is mainly distributed in [K2_min, K2_max], and the overlap interval between the two is [K_overlap_low, K_overlap_high], then the upper limit of the overlap interval (K_overlap_high) is determined as the first dynamic threshold, ensuring that the vast majority of data in the viewing distance region is above this threshold, and the vast majority of data in the occluded viewing distance region is below this threshold. Similarly, based on the boundary between the K-factor distributions of the occluded viewing distance region and the non-viewing distance region, a critical value of the overlapping interval of the two data distributions is determined as the second dynamic threshold. For example, if the distribution range of K-factor data in the occluded viewing distance region is [K2_min, K2_max], and the data in the non-viewing distance region is mainly distributed in [K3_min, K3_max], and the overlap range between the two is [K_overlap_low', K_overlap_high'], then the lower limit of the overlap range (K_overlap_low') is determined as the second dynamic threshold to ensure that the vast majority of data in the non-viewing distance region is below this threshold, and the vast majority of data in the occluded viewing distance region is above this threshold.

[0074] For each region initially identified as a viewing distance region, its second average K-factor is extracted and compared with a first dynamic threshold. If the average is higher than the first dynamic threshold, meeting the threshold condition for a viewing distance region, its viewing distance state is retained; if it is lower than or equal to the first dynamic threshold, its state is corrected to an occluded viewing distance state. For each region initially identified as a non-viewing distance region, its second average K-factor is extracted and compared with a second dynamic threshold. If the average is lower than the second dynamic threshold, meeting the threshold condition for a non-viewing distance region, its non-viewing distance state is retained; if it is higher than or equal to the second dynamic threshold, its state is corrected to an occluded viewing distance state. For each region initially identified as an occluded viewing distance region, its second average K-factor is extracted and compared with both the first and second dynamic thresholds. If the average is between the first and second dynamic thresholds (i.e., lower than the first dynamic threshold and higher than the second dynamic threshold), its occluded viewing distance state is maintained; if the average is higher than or equal to the first dynamic threshold, it is corrected to a viewing distance state; if the average is lower than or equal to the second dynamic threshold, it is corrected to a non-viewing distance state. The final line-of-sight classification result is formed by summarizing the verified and corrected states of all regions.

[0075] The method provided in this embodiment determines the first and second dynamic thresholds based on the K-factor distribution characteristics (probability density curves, distribution intervals, overlapping boundaries, etc.) of three types of preliminary regions, rather than using fixed values. This allows it to adapt to the actual statistical laws of the K-factor in different scenarios. For example, the boundary between the line of sight and the occlusion line of sight will change with the scattering intensity in the environment, and the dynamic threshold can capture this change. At the same time, the correction process verifies the preliminary state by comparing the second average value of the K-factor in each region with the dynamic threshold (the line of sight region needs to be higher than the first threshold, the non-line of sight region needs to be lower than the second threshold, and the occlusion line of sight region needs to be between the two), forming a closed-loop logic of "data distribution → threshold adaptation → boundary verification". This approach first avoids the potential for insufficient adaptability of fixed thresholds in complex environments (e.g., a single threshold cannot cover the differences in K-factors across different propagation scenarios), making the thresholds more closely match the actual data characteristics. Second, by correcting the initial state through dynamic thresholds, it can correct misjudgments caused by local statistical biases (e.g., some areas are initially judged as line-of-sight but the actual K-factor is at the boundary between line-of-sight and occlusion line-of-sight), ensuring that the division of the three categories of line-of-sight, non-line-of-sight, and occlusion line-of-sight regions is more consistent with their true K-factor distribution patterns. Finally, this dynamic adjustment mechanism, which combines data statistical characteristics, significantly improves the accuracy and robustness of line-of-sight classification results, providing a more reliable classification basis for channel line-of-sight identification and enabling it to better adapt to changes in different propagation environments.

[0076] Optionally, after obtaining the sight distance classification result, the method further includes: comparing the sight distance classification result with the position and occlusion range of obstacles in the laboratory indoor physical environment, and calculating the error between the actual proportion of each state and the sight distance classification result; when the error exceeds a preset threshold, readjusting the correlation judgment standard of the region division or the K-factor dynamic threshold, and recalculating the sight distance classification result until the adaptation error between the sight distance classification result and the physical environment is within a preset range.

[0077] Specifically, basic data on the laboratory's indoor physical environment is collected, including the specific location coordinates and dimensions of obstacles (such as walls, furniture, and equipment), as well as the occlusion range formed by these obstacles (identifying which areas are completely occluded, partially occluded, or unoccluded). Based on the physical environment data, the actual viewing distance status of each area in the laboratory is determined: unoccluded areas are marked as actual viewing distance areas, completely occluded areas as actual non-viewing distance areas, and partially occluded areas as actual occluded viewing distance areas. The proportions of these three actual states in the overall environment are then calculated (actual viewing distance proportion, actual non-viewing distance proportion, and actual occluded viewing distance proportion). The obtained viewing distance classification results are extracted, and the proportions of viewing distance areas, non-viewing distance areas, and occluded viewing distance areas are calculated (classified viewing distance proportion, classified non-viewing distance proportion, and classified occluded viewing distance proportion). The error between the actual proportion and the classified proportion is calculated: the error for viewing distance status, the error for non-viewing distance status, and the error for occluded viewing distance status are calculated separately, and a comprehensive error (the average error or weighted error of the three) is calculated according to requirements. The calculated error is compared with a preset error threshold to determine if it exceeds the threshold range. If the error exceeds the preset threshold, the adjustment process begins: The correlation criteria for region division are adjusted: the preset threshold for temporal similarity is increased or decreased (changing the granularity of temporal sub-segments), and the preset threshold for spatial similarity is increased or decreased (changing the granularity of spatial sub-arrays). Alternatively, the dynamic threshold of the K-factor is adjusted: the cumulative distribution function of the K-factor is re-analyzed, and the positions of the first dynamic threshold (the boundary between viewing distance and occlusion viewing distance) and the second dynamic threshold (the boundary between occlusion viewing distance and non-viewing distance) are moved. Based on the adjusted correlation criteria or dynamic threshold, the steps of region division, preliminary region state determination, and dynamic threshold optimization are re-executed to generate new viewing distance classification results. The above steps are repeated, and the new classification results are compared with the actual state of the physical environment. The error is statistically analyzed and determined to be within the preset range until the adaptation error between the viewing distance classification results and the laboratory indoor physical environment meets the preset requirements.

[0078] The method provided in this embodiment achieves accurate identification through a closed loop of "multi-dimensional feature extraction → region division → preliminary judgment → dynamic optimization". First, channel data with different polarization modes are collected. The differences in the characteristics of vertical and horizontal polarization in obstacle reflection and scattering enrich the feature dimensions. Then, the time delay characteristics and K-factor are calculated. Regions with similar propagation characteristics are divided based on temporal and spatial similarity. A preliminary judgment is made based on the difference between the average value of the region and the whole, and the similarity of the time delay characteristics. Finally, a second adjustment is made using a dynamic threshold determined by the K-factor cumulative distribution function. The addition of polarization characteristics reduces the bias of a single polarization, the preliminary judgment enables rapid screening of regions with clear features, and the dynamic threshold adapts to environmental changes. The combination of these three aspects improves the accuracy of identification and its adaptability to complex indoor scenes, making the results more consistent with the actual physical environment.

[0079] Secondly, a cumulative distribution function is constructed based on the initial K-factor distribution of the region. Dynamic thresholds (first for distinguishing between line-of-sight distance and occlusion distance, and second for distinguishing between occlusion distance and non-line-of-sight distance) are determined based on the probability density characteristics of the three types of regions. The initial state is then corrected through threshold comparison. Dynamic thresholds avoid the insufficient adaptability of fixed thresholds when the environment changes, and can accurately reflect the true distribution of the K-factor in real time, correcting misjudgments caused by local statistical biases. This makes the division of the three types of regions more consistent with the actual energy distribution characteristics, significantly improving the robustness and accuracy of the classification results.

[0080] Corresponding to the aforementioned embodiment of an indoor scene line-of-sight channel recognition method based on polarization and K-factor, this application also provides an embodiment of an indoor scene line-of-sight channel recognition device based on polarization and K-factor.

[0081] Figure 2 This is a schematic diagram of the indoor scene line-of-sight channel recognition device based on polarization and K-factor provided in Embodiment 2 of this application. Please refer to... Figure 2 The device provided in this embodiment includes a data acquisition module 210, a calculation module 220, a division module 230, a determination module 240, and an optimization module 250.

[0082] The acquisition module 210 is used to acquire channel data under different polarization modes in a laboratory indoor environment.

[0083] The calculation module 220 is used to analyze the channel data and calculate the channel's delay characteristics and K-factor.

[0084] The partitioning module 230 is used to analyze the temporal and spatial similarities between channels based on the time delay characteristics and the K factor, and to divide the entire measurement period and the transmitting antenna array into multiple regions based on the temporal and spatial similarities.

[0085] The calculation module 220 is also used to calculate the first average value of the overall delay characteristics and K factor based on the calculated delay characteristics and K factor of all channels, and to calculate the second average value of the delay characteristics and K factor of all channels in each region after division.

[0086] The determining module 240 is used to compare the second average value of each region with the corresponding first average value, and determine the preliminary region status by combining the time delay characteristics similarity of multiple channel data.

[0087] The optimization module 250 is used to determine a dynamic threshold based on the cumulative distribution function of the K factor, and optimize the preliminary region state based on the dynamic threshold to obtain the line-of-sight classification result.

[0088] The apparatus of this embodiment can be used to perform... Figure 1 The steps of the method embodiment shown are similar in principle and process, and will not be repeated here.

[0089] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0090] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0091] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for line-of-sight channel recognition in indoor scenes based on polarization and K-factor, characterized in that, The method includes: Channel data under different polarization modes were collected in a laboratory environment. The channel data is analyzed to calculate the channel's delay characteristics and K-factor; Based on the aforementioned delay characteristics and K-factor, the temporal and spatial similarities between channels are analyzed. Based on the temporal and spatial similarities, the entire measurement period and the transmitting antenna array are divided into multiple regions. Based on the calculated delay characteristics and K-factor of all channels, the first average value of the overall delay characteristics and K-factor is calculated respectively. For each divided region, the second average value of the delay characteristics and K-factor of all channels in each region is calculated. The second average value of each region is compared with the corresponding first average value, and the preliminary region status is determined by combining the similarity of the time delay characteristics of multiple channel data. A dynamic threshold is determined based on the cumulative distribution function of the K factor, and the preliminary region state is optimized based on the dynamic threshold to obtain the line-of-sight classification result.

2. The method according to claim 1, characterized in that, The dynamic threshold is determined based on the cumulative distribution function of the K factor, and the preliminary region state is optimized based on the dynamic threshold to obtain the line-of-sight classification result, including: Calculate the K-factor distribution of the line-of-sight region, the occluded line-of-sight region, and the non-line-of-sight region in the preliminary regional state, and construct the cumulative distribution function of the overall K-factor; Based on the K-factor probability density characteristics of the three regions in the cumulative distribution function, a first dynamic threshold for distinguishing between the viewing distance region and the occluded viewing distance region, and a second dynamic threshold for distinguishing between the occluded viewing distance region and the non-viewing distance region are determined. The second average value of the K factor initially determined to be the viewing distance region is compared with the first dynamic threshold. If the threshold condition is met, the viewing distance state is retained; otherwise, it is corrected to the occluded viewing distance state. The second average value of the K factor initially determined to be a non-line-of-sight region is compared with the second dynamic threshold. If the threshold condition is met, the non-line-of-sight state is retained; otherwise, it is corrected to an obstructed line-of-sight state. The second average value of the K factor, which is initially determined to be an area with obstructed viewing distance, is compared with the first and second dynamic thresholds. If it is between the two thresholds, the obstructed viewing distance state is maintained; otherwise, it is corrected to the corresponding state according to the threshold boundary. The verification results are integrated to obtain the final line-of-sight classification result.

3. The method according to claim 1, characterized in that, The step of comparing the second average value of each region with the corresponding first average value, and determining the preliminary region state by combining the similarity of the time delay characteristics of multiple channel data, includes: The second average value of the copolarization K-factor in each region is compared with the first average value of the overall copolarization K-factor to obtain the first branch and the second branch; Based on the second average value of the cross-polarization K-factor of each region in the first branch and the second branch, it is compared with the overall cross-polarization K-factor to obtain the third branch and the fourth branch under the first branch, and the fifth branch and the sixth branch under the second branch; wherein, the second average value of the cross-polarization K-factor of each region in the third branch and the sixth branch is not greater than the first average value of the overall cross-polarization K-factor, the third branch is the line-of-sight region, and the sixth branch is the non-line-of-sight region. For the fourth and fifth branches, the comparison signs of the regional co-polarization delay characteristics and the global co-polarization delay characteristics are calculated respectively, as well as the comparison signs of the regional cross-polarization delay characteristics and the global cross-polarization delay characteristics. If the signs of the time delay characteristics under the two polarizations are consistent, the time delay characteristics are determined to be similar. For regions with similar time delay characteristics under the fourth and fifth branches, they are determined to be occlusion range regions; for regions with dissimilar time delay characteristics under the fourth branch, they are determined to be range regions; and for regions with dissimilar time delay characteristics under the fifth branch, they are determined to be non-range regions.

4. The method according to claim 1, characterized in that, Based on the aforementioned delay characteristics and K-factor, the temporal and spatial similarities between channels are analyzed. Based on these temporal and spatial similarities, the entire measurement period and the transmitting antenna array are divided into multiple regions, including: Extract the overall time delay characteristics and K-factor data of the transmitting antenna array at different times, calculate the correlation coefficient of data at adjacent times, and analyze the time similarity of the channel; Extract the time delay characteristics and K-factor data of different antenna elements in the transmitting antenna array at the same time, calculate the correlation coefficient between the data of array elements, and analyze the spatial similarity of the channel; Based on the changes in the strength of the time similarity, when the time similarity is lower than a preset time threshold, the entire measurement period is divided into several time segments. Based on the distribution characteristics of the spatial similarity, when the spatial similarity is lower than a preset spatial threshold, the antenna array is divided into several spatial subarrays. By combining time segments with spatial subarrays, multiple spatiotemporal regions are formed.

5. The method according to claim 1, characterized in that, The time delay characteristics include average time delay and root mean square time delay. The analysis of the channel data and the calculation of the channel's time delay characteristics include: Each channel path is determined from the channel data, and the power delay spectrum and path delay corresponding to each channel path are obtained; The power delay spectrum of each channel path is used as a weight to sum the path delay of the corresponding channel path, and then divided by the sum of the power delay spectra of all channel paths to obtain the average delay of the channel; wherein, the power delay spectrum is obtained by averaging the energy of the channel impulse response function over the time interval.

6. The method according to claim 5, characterized in that, The analysis of the channel data and the calculation of the channel delay characteristics include: Calculate the difference between the path delay of each channel path and the average delay, and square each difference. The power delay spectrum of each channel path is used as the weight, and the squared differences are summed in a weighted manner. Divide the summation result by the sum of the power delay spectra of all channel paths, and take the square root of the result to obtain the root mean square delay of the channel.

7. The method according to claim 1, characterized in that, The analysis of the channel data and the calculation of the channel's K-factor include: Cluster analysis is performed on the time delay and energy information in the channel data to determine the energy of the earliest arrival path and the energy of the scattering path; the energy of the scattering path is the difference between the sum of the power delay spectra of all channel paths and the energy of the earliest arrival path. The K-factor, expressed in decibels, is obtained by taking the logarithm of the ratio of the energy of the earliest arriving path to the energy of the scattering path.

8. The method according to claim 1, characterized in that, The acquisition of channel data under different polarization modes in a laboratory indoor environment includes: A fixed antenna array is set up at the transmitting end, and the array elements are deployed in a vertical polarization manner; Two antennas with different polarizations are configured at the receiving end; the antennas include a vertically polarized antenna and a horizontally polarized antenna; A channel measuring instrument is connected to the transmitting end and the receiving end. The transmitting end continuously transmits signals through an antenna array, and the antenna of the receiving end is fixed on a movable device and moves at a constant speed along a preset route. During the movement of the receiving end, the vertically polarized antenna and the horizontally polarized antenna synchronously receive the signal emitted by the transmitting end, and the channel measurement instrument collects and records the channel data under the two polarization modes in real time.

9. The method according to claim 1, characterized in that, After obtaining the distance classification result, the method further includes: By comparing the line-of-sight classification results with the location and occlusion range of obstacles in the laboratory's indoor physical environment, the actual proportion of each state and the error between the line-of-sight classification results are statistically analyzed. When the error exceeds a preset threshold, the correlation judgment criteria for region division or the dynamic threshold of the K factor is readjusted, and the line-of-sight classification result is recalculated until the line-of-sight classification result and the physical environment adaptation error are within the preset range.

10. An indoor scene line-of-sight channel recognition device based on polarization and K-factor, characterized in that, The device includes an acquisition module, a calculation module, a partitioning module, a determination module, and an optimization module; The acquisition module is used to acquire channel data under different polarization modes in a laboratory indoor environment. The calculation module is used to analyze the channel data and calculate the channel's delay characteristics and K-factor. The partitioning module is used to analyze the temporal and spatial similarity between channels based on the time delay characteristics and the K factor, and to divide the entire measurement period and the transmitting antenna array into multiple regions based on the temporal and spatial similarity. The calculation module is also used to calculate the first average value of the overall delay characteristics and K factor based on the calculated delay characteristics and K factor of all channels, and to calculate the second average value of the delay characteristics and K factor of all channels in each region after division. The determining module is used to compare the second average value of each region with the corresponding first average value, and combine the similarity of the delay characteristics of multiple channel data to determine the preliminary region status. The optimization module is used to determine a dynamic threshold based on the cumulative distribution function of the K factor, and to optimize the preliminary region state based on the dynamic threshold to obtain the line-of-sight classification result.

Images