WO2023005746A1 - 一种信道建模方法、装置、存储介质及电子装置 - Google Patents

一种信道建模方法、装置、存储介质及电子装置 Download PDF

Info

Publication number
WO2023005746A1
WO2023005746A1 PCT/CN2022/106543 CN2022106543W WO2023005746A1 WO 2023005746 A1 WO2023005746 A1 WO 2023005746A1 CN 2022106543 W CN2022106543 W CN 2022106543W WO 2023005746 A1 WO2023005746 A1 WO 2023005746A1
Authority
WO
WIPO (PCT)
Prior art keywords
channel
measurement data
external field
sampling points
multipath
Prior art date
Application number
PCT/CN2022/106543
Other languages
English (en)
French (fr)
Inventor
吕星哉
余泽浩
许靖
芮华
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2023005746A1 publication Critical patent/WO2023005746A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/309—Measuring or estimating channel quality parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/309—Measuring or estimating channel quality parameters
    • H04B17/336—Signal-to-interference ratio [SIR] or carrier-to-interference ratio [CIR]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel

Definitions

  • Embodiments of the present disclosure relate to the communication field, and in particular, to a channel modeling method, device, storage medium, and electronic device.
  • the commonly used modeling method for wireless channel modeling is the method of deriving a multipath scattering model from typical scene measurement data.
  • the channel model TR38.901 commonly used in 5G systems.
  • special wireless signals are used to measure and collect wireless channel characteristics. After data analysis, it is summarized into different multipath time domain and spatial domain distribution characteristics, and the corresponding parameters are calculated.
  • the channel model first select the required scene, and then use the previously calculated parameters to generate the model in the corresponding scene. Its disadvantage is that it can only be used for generalized research on certain types of classic scenes. Once the channel in a specific specific scene is simulated, it will be due to the large difference between the specific channel and the classic scene, or the variance inside the classic scene. deviation.
  • Embodiments of the present disclosure provide a channel modeling method, device, storage medium, and electronic device, to at least solve the problem of simulating a channel in a specific specific scenario in the related art because the specific channel is greatly different from the classic scenario, or The variance inside the classic scene produces a problem of large deviation.
  • a channel modeling method including:
  • a channel modeling device including:
  • the acquisition module is configured to obtain the external field measurement data when the external field communication equipment of multiple sampling points receives signals, and obtain multiple external field measurement data;
  • the feature extraction module is configured to perform feature extraction on the plurality of external field measurement data respectively to obtain channel characteristics of the plurality of sampling points;
  • the processing module is configured to perform model regression processing according to the channel characteristics of the plurality of sampling points.
  • a computer-readable storage medium where a computer program is stored in the storage medium, wherein the computer program is set to execute any one of the above method embodiments when running in the steps.
  • an electronic device including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to perform any of the above Steps in the method examples.
  • FIG. 1 is a block diagram of a hardware structure of a mobile terminal of a channel modeling method according to an embodiment of the present disclosure
  • FIG. 2 is a flow chart of a channel modeling method according to an embodiment of the present disclosure
  • Fig. 3 is a block diagram of the channel construction system according to the present embodiment.
  • FIG. 4 is a block diagram of a channel modeling device according to the present embodiment.
  • FIG. 5 is a first block diagram of a channel modeling device according to this preferred embodiment
  • Fig. 6 is a second block diagram of a channel modeling device according to this preferred embodiment.
  • Fig. 7 is a third block diagram of a channel modeling device according to this preferred embodiment.
  • FIG. 1 is a block diagram of the hardware structure of the mobile terminal of the channel modeling method of the embodiment of the present disclosure.
  • the mobile terminal may include one or more (only shown in FIG. 1 1)
  • Processor 102 may include but not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the above-mentioned mobile terminal may also include a communication function
  • the transmission device 106 and the input and output device 108 may be understood that the structure shown in FIG. 1 is only for illustration, and it does not limit the structure of the above mobile terminal.
  • the mobile terminal may also include more or fewer components than those shown in FIG. 1 , or have a different configuration from that shown in FIG. 1 .
  • the memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the channel modeling method in the embodiment of the present disclosure, and the processor 102 executes the computer program stored in the memory 104 by running the Various functional applications and service chain address pool slicing processing realize the above-mentioned method.
  • the memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory.
  • the memory 104 may further include a memory that is remotely located relative to the processor 102, and these remote memories may be connected to the mobile terminal through a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • the transmission device 106 is used to receive or transmit data via a network.
  • the specific example of the above network may include a wireless network provided by the communication provider of the mobile terminal.
  • the transmission device 106 includes a network interface controller (NIC for short), which can be connected to other network devices through a base station so as to communicate with the Internet.
  • the transmission device 106 may be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet in a wireless manner.
  • RF Radio Frequency
  • FIG. 2 is a flowchart of a channel modeling method according to an embodiment of the disclosure. As shown in FIG. 2 , the process includes the following step:
  • Step S202 acquiring the external field measurement data when the external field communication devices at multiple sampling points receive signals, and obtaining multiple external field measurement data;
  • Step S204 respectively performing feature extraction on the plurality of external field measurement data to obtain channel characteristics of the plurality of sampling points;
  • Step S206 performing model regression processing according to the channel characteristics of the plurality of sampling points.
  • the above step S206 may specifically include: performing statistical model regression or network model regression according to the channel characteristics of the multiple sampling points, and outputting the channel characteristics of all sampling points; using the channel characteristics of all the sampling points as the channel The parameters are input into the multipath channel model, and the analog channel at any point output by the multipath channel model is obtained.
  • step S204 may specifically include:
  • the current external field measurement data is called the current external field measurement data:
  • the multi-paths are clustered to obtain a plurality of target paths, specifically, the correlation of the channels between the multi-paths is calculated; according to the correlation to the The multi-paths are clustered to obtain a multi-cluster set, wherein each cluster set includes at least one path; respectively selecting the target path with the largest average power in the multi-cluster set to obtain the multiple target paths;
  • the polarization characteristics are correlations between channels in different polarization directions, wherein the channel characteristics include the multipath The time delay of the multipath, the average power of the multipath, the angle of arrival of the multipath, and the polarization characteristics of the multipath.
  • step S202 may specifically include:
  • step S2021 may specifically include: performing channel estimation when the field communication device receives a signal to obtain an estimated wireless channel value; obtaining quality label data of the estimated wireless channel value and network label of the estimated wireless channel value data, wherein the measurement data includes the wireless channel estimation value, the quality label data of the wireless channel estimation value, and the grid label data of the wireless channel estimation value.
  • step S2022 may specifically include:
  • the quality label data of the wireless channel estimation value it is judged whether the measurement data of the plurality of sampling points meet the preset conditions.
  • the quality label data of the wireless channel estimation value includes the signal interference noise of the current received signal.
  • the SINR of the measurement data of the plurality of sampling points is greater than a first preset threshold, and whether the Doppler frequency shift is is less than the second preset threshold and whether the estimated motion speed is less than the third preset threshold; if the judgment result is yes, it is determined that the preset condition is met; if the judgment result is no, it is determined that it is not satisfied the pre-conditions;
  • Fig. 3 is a block diagram of the channel construction system according to this embodiment. As shown in Fig. 3, it includes three subsystems: raw data acquisition, channel feature extraction, and model regression.
  • the raw data acquisition subsystem is mainly used to obtain the sampling points of the external field channel that can be used by the channel feature extraction subsystem and the model regression subsystem, where:
  • the field channel measurement acquisition is to obtain the field measurement data from the wireless equipment running in the field.
  • the measurement data includes the wireless channel estimation value extracted from the external field received signal, the quality label data of the external field wireless channel estimation value, and the grid label data of the external field wireless channel estimation value.
  • the estimation method can be a current general method or other methods, and its output value includes: wireless channel estimation value, quality label data of the external field wireless channel estimation value, and grid label data of the external field wireless channel estimation value.
  • the wireless channel estimation value can be the time-frequency or frequency domain estimation value estimated by the receiver on the pilot frequency of various signals, and other channel estimation values obtained by various methods;
  • the quality label data can be the signal interference of the current received signal Noise ratio, Doppler frequency shift, user motion speed estimation (ie, the above motion speed estimation value), data reception accuracy rate and other measurements reflecting the quality of data reception and channel estimation;
  • grid label data which can be other transmitters And receiver's latitude and longitude coordinates, map coordinates, sampling time stamp, or other data reflecting geographic location and time information. Mass label data and grid label data are optional data.
  • Channel data screening is mainly from the field measurement data.
  • the quality label data is available, the corresponding data that does not meet the quality requirements are screened out according to the quality label data, and only the data that meets the requirements are processed subsequently.
  • the quality label data is compared with it according to the pre-specified criteria, and the data that does not meet the requirements are deleted.
  • the criterion can be a single condition or a combination of multiple conditions.
  • These conditions can be: the received SINR is greater than a certain threshold; the Doppler frequency shift is less than a certain threshold; the user's motion speed is less than a certain threshold; the relevant available channel The number of estimated values exceeds the threshold; and other conditions that can determine that the channel sampling is obtained under the conditions of low mobility, high communication quality, and large amount of related data.
  • This criterion is to improve the accuracy of channel feature estimation for this location point, and it does not mean that this method can only be applied to low-speed channel modeling. By combining motion and trajectory modeling, the method can be applied to channel modeling at arbitrary speeds.
  • the channel feature extraction subsystem is mainly used to extract the multipath delay, angle of arrival and polarization correlation characteristics of the channel from the data meeting the quality requirements, where:
  • the estimation method can use the Estimating Signal Parameter via Rotational Invariance techniques (referred to as ESPRIT) or the multiple signal classification (Multiple Signal Classification Algorithm, referred to as MUSIC) algorithm and other high-precision spectral estimation algorithms and their Variants.
  • ESPRIT Estimating Signal Parameter via Rotational Invariance techniques
  • MUSIC Multiple Signal Classification Algorithm
  • the detection principle of the number of paths may be a criterion that the energy is greater than a predetermined threshold, and then it is determined as one path, or a criterion that the total energy of each path finally determined is greater than the total energy percentage threshold.
  • Multipath delay estimation estimates the delay of each path according to the output multipath number. Estimation methods can use ESPRIT or MUSIC algorithms and other high-precision spectral estimation algorithms and their variants.
  • the time-domain response estimation of each path calculates the time-domain response of each path from the wireless channel data according to the estimated time delay, including the amplitude and phase of each receiving antenna, and estimates the average power value of each path.
  • the least squares method can be used for estimation, and signal estimation algorithms known in the industry can also be used, such as maximum likelihood, maximum prior probability algorithm, and so on.
  • the channels of each path are clustered, and the paths with high correlation are clustered and replaced by a channel to maintain the low correlation between the paths after processing.
  • Correlation can be calculated according to the output of time-delay response estimation of each path.
  • one path in the cluster can be selected, or a path representing the whole cluster can be obtained by means of weighted average of each path in the cluster. And update the estimated channel path number.
  • the horizontal angle of arrival estimation of each path estimates the information of the angle of arrival in the horizontal direction for each path channel separated by the clustering of each path channel.
  • the algorithm for estimating the angle of arrival may also be a multi-step iterative search algorithm, or a spectral estimation algorithm, such as ESPRIT and MUSIC algorithms, and their variants.
  • the algorithm for estimating the angle of arrival may also be a multi-step iterative search algorithm, or a spectral estimation algorithm, such as ESPRIT and MUSIC algorithms, and their variants.
  • Polarization direction correlation estimation for each path channel separated by clustering of each path channel, the correlation between channels with different polarization directions is estimated.
  • the model regression subsystem uses the channel features extracted by the channel feature extraction subsystem to generate different models suitable for different scenarios, among which:
  • Point model generation the channel path number, time delay, power of each path, vertical and horizontal angle of arrival, and polarization correlation output by the channel feature extraction subsystem are input into the multipath channel model as parameters to generate channels; these parameters are used as The overall sampling points of a channel are saved and output to statistical model regression and grid model regression.
  • the channel model used here can be any model that can calculate channel coefficients based on multipath parameters, such as the multipath channel model defined in 3GPP TR38.901, or a spatial channel model (Spatial Channel Model, referred to as SCM) , expand the channel model such as the Spatial Channel Model Extension (SCME for short), and use the parameters obtained in the previous steps to replace the pre-fitted parameters according to the classic scene in its Cluster Delay Line (CDL) parameters, a channel model that is in line with a specific external field environment and fits well with the real channel of the external field can be obtained.
  • CDL Cluster Delay Line
  • Statistical model regression which performs statistical model regression on multiple overall channel sampling points output by the point model to obtain the probability distribution of channel parameters, such as delay, power, and spatial angle distribution.
  • the probability distribution is used to generate channel parameters and channel models for locations in the external field that are not measured.
  • the probability distribution used can be the current classical channel model, or a priori external field parameter distribution model improved by any other means.
  • Grid model regression perform grid model regression on the overall sampling points of multiple channels output by the point model, and combine and store the extracted channel parameters and the grid channel using the point. Using grid information and measured data, the predicted channel parameters and channel models are output for networks without measured data.
  • the field communication equipment can be any kind of multi-antenna transceiver equipment with wireless transceiver function.
  • the equipment includes but is not limited to: 5G base station, 4G base station, WiFi node, etc.
  • the processing flow includes the following steps:
  • the outfield communication equipment When the outfield communication equipment receives signals, it needs to perform channel estimation, signal quality measurement, and transmitter receiver position estimation and other measurement processes. And output the measured estimated value according to the specified format.
  • the reference signal for detection can be a Sounding Reference Signal (SRS for short), or a demodulation reference signal (Demodulation Reference Signal, DMRS for short) of a data and control channel, and other known and available Reference signal sequence for channel estimation.
  • the output value includes: estimated value of wireless channel, measured value of SINR, Doppler frequency shift of transmitter, geographic latitude and longitude coordinates of transmitter, time stamp of estimated data of wireless channel.
  • Wireless channel estimation value the result of the frequency domain channel estimation obtained by using the frequency domain reference signal for the transmitter u is H u (k, t, m, n, p), where k is OFDM (Orthogonal Frequency Division, referred to as orthogonal frequency division multiplexing), t is the moment of time-domain sampling, m is the number of antenna elements in the horizontal direction in the receiver antenna panel, and n is the number of antenna elements in the vertical direction in the receiver antenna panel, p is the polarization number of antenna dipoles in different polarization directions in the receiver antenna panel. The value range of each number is from 1 to the number represented by its corresponding uppercase symbol.
  • T is the total number of sampling moments in the time domain
  • M is the total number of antenna elements in the horizontal direction in the receiver antenna panel
  • N is the total number of antenna elements in the vertical direction in the receiver antenna panel
  • P is the total number of antenna elements in the receiver antenna panel. number of orientations.
  • Data channel screening to filter out inputs that do not meet the pre-defined requirements filter out data samples with a signal-to-interference-noise ratio lower than 10dB, a Doppler frequency shift higher than 300Hz, and samples with a sampling frequency less than 20 times in time.
  • the above numbers and thresholds can be adjusted based on empirical data or other guidelines.
  • the extraction of channel multipath number can adopt ESPRIT algorithm. If the input channel estimate is a time-domain channel, it can be transformed into a frequency-domain channel value by Fourier transform, including:
  • H u, f (t, m, n, p) [H u (1, t, m, n, p)...H u (K, t, m, n, p)] T ;
  • Multipath delay estimation can use spectral estimation to obtain accurate and independent delays that are less than the sampling rate of the communication system.
  • I is a unit diagonal matrix of size (K-1) ⁇ (K-1), and 0 is a (K-1) ⁇ 1 vector of all zeros.
  • ⁇ is a vector containing L eigenvalues, where the lth eigenvalue is ⁇ l , then the time delay estimate of the lth path is obtained as:
  • f ⁇ , H is the frequency domain interval of two channel sampling points in the original data
  • arg() is the operation for calculating the complex argument.
  • the time-domain response estimation of each path uses the time-delay information of each path obtained in the previous steps and the original frequency-domain channel estimation to obtain the response value and power value of each path on the time delay.
  • the algorithm using least squares estimation is as follows:
  • a matrix of size K ⁇ L is estimated as follows:
  • the least squares estimation of each path channel in the time domain is:
  • the power distribution of each path can be obtained as:
  • the channels of each path are clustered, and the correlation value of the channels between the paths is calculated, and the paths with high correlation are clustered and replaced by a single-path channel.
  • the steps are as follows, put each path into an original collection, and perform the following steps:
  • Step 1 calculate the correlation R(i, j) between each path
  • Step 2 select the path k with the highest power among the paths in the current set
  • Step 3 classify each path j of R(k, j)>Threshold into one cluster, and delete it from the original set;
  • Step 4 repeat steps 2 and 3 until there is no path in the set;
  • Step 5 each cluster selects the path with the strongest power, and deletes the rest.
  • Threshold can be selected as 0.5 or adjusted to any value according to experience and other criteria.
  • the original estimated number of paths is updated to the clustered value.
  • h l, hor (t, n, p) [h l (t, 1, n, p)...h l (t, M, n, p)] T ;
  • rough estimation it is divided into multiple iterations of rough estimation and fine estimation.
  • rough estimation according to the given initial estimation step size to recompute the finer power spectrum, and then use a finer search to get a finer estimate.
  • ⁇ is a matrix containing the beam vectors of all sampling points. A more precise estimate can be obtained by recalculating the AoA and ASA according to the same formula as the rough estimate.
  • the above fine estimate can be iterated multiple times.
  • the vertical arrival angle estimation of each path is the same as the previous step, and the direction is changed from horizontal to vertical.
  • Polarization direction correlation estimation calculates the correlation value between two sets of channel estimation values with different polarizations, and finally obtains the correlation value.
  • Point model generation according to the path information, power, time delay, vertical angle of arrival, horizontal angle of arrival, and correlation information of each path generated in the previous steps, input to the corresponding channel generation model to generate channel coefficients.
  • the channel model used is the multipath channel model defined in 3GPP TR38.901.
  • CDL Cluster Delay Line
  • the parameters obtained in the previous steps are used to replace the parameters fitted in advance according to the classic scene , to obtain a channel model that is facing a specific external field environment and has an excellent fit with the real channel of the external field.
  • the result of the channel characteristic parameters and the data label information are stored as a sampling point for subsequent model regression.
  • the multipath time delay spread, horizontal and vertical spatial angle power spectrums are calculated. This calculation is performed according to the mathematical definition of the values. Then assume that the calculated value is X, and according to the model in 3GPP TR38.901, it is considered to be in line with the lognormal distribution in the classic scene, and the subscript u represents different sampling points, then the mean and variance estimates of the distribution can be obtained, respectively for:
  • any number of random channel sampling eigenvalues conforming to the distribution can be generated, multipath parameters can be generated, and input into the multipath channel generation model, such as the channel coefficient generated in the 3GPP TR38.901 multipath channel generation model.
  • the channel features output by the point model are stored together with the grid label information.
  • the method in step 10 is used to output the channel coefficients.
  • the tag data is used to generate a space-time two-dimensional grid index
  • the transmitter's latitude and longitude is used to generate a spatial grid dimension
  • the time stamp is used to generate a time grid dimension.
  • the grid granularity is set according to the actual data volume and demand.
  • the field communication equipment can be any kind of multi-antenna transceiver equipment with wireless transceiver function.
  • the equipment includes but is not limited to: 5G base station, 4G base station, WiFi node, etc.
  • the processing flow includes the following steps:
  • the outfield communication equipment When the outfield communication equipment receives signals, it needs to perform channel estimation, signal quality measurement, and transmitter receiver position estimation and other measurement processes. And output the measured estimated value according to the specified format.
  • the reference signal for detection can be a Sounding Reference Signal (SRS for short), or a demodulation reference signal (Demodulation Reference Signal, DMRS for short) of a data and control channel, and other known and available Reference signal sequence for channel estimation.
  • the output value includes: estimated value of wireless channel, received reference signal strength, moving speed of transmitter, geographical longitude and latitude coordinates of transmitter, time stamp of estimated data of wireless channel.
  • Wireless channel estimation value the result of the frequency domain channel estimation obtained by using the frequency domain reference signal for the transmitter u is H u (k, t, m, n, p), where k is the number of the frequency domain subcarrier in OFDM, t is the time domain sampling moment, m is the number of antenna elements in the horizontal direction in the receiver antenna panel, n is the number of antenna elements in the vertical direction in the receiver antenna panel, and p is the number of antenna elements in different polarization directions in the receiver antenna panel Polarization number.
  • the value range of each number is from 1 to the number represented by its corresponding uppercase symbol.
  • T is the total number of sampling moments in the time domain
  • M is the total number of antenna elements in the horizontal direction in the receiver antenna panel
  • N is the total number of antenna elements in the vertical direction in the receiver antenna panel
  • P is the total number of antenna elements in the receiver antenna panel. number of orientations.
  • Data channel screening to filter out inputs that do not meet the implementation-defined requirements filter out data samples whose reference signal strength is lower than -110dB, whose motion speed is greater than 10 kilometers per hour, and samples whose sampling times are less than 20 times in time.
  • the above numbers and thresholds can be adjusted based on empirical data or other guidelines.
  • Multipath delay estimation can use spectral estimation to obtain accurate and independent delays that are less than the sampling rate of the communication system.
  • ESPRIT method it is estimated as follows:
  • H u, f (k, t, m, n, p) [H u (1, t, m, n, p)...H u (K, t, m, n, p)] T ;
  • I is a unit diagonal matrix of size (K-1) ⁇ (K-1), and 0 is a (K-1) ⁇ 1 vector of all zeros.
  • ⁇ is a vector containing L eigenvalues, where the lth eigenvalue is ⁇ l , then the time delay estimate of the lth path is obtained as:
  • H is the frequency domain interval of two channel sampling points in the original data.
  • a matrix of size K ⁇ L is estimated as follows:
  • the least squares estimation of each path channel in the time domain is:
  • the power distribution of each path can be obtained as:
  • each path is clustered, and the correlation value of the channel between the paths is calculated, and the paths with high correlation are clustered, replaced by a single-path channel, and each path is put into an original set, and the following steps are performed:
  • Step 1 calculate the correlation R(i, j) between each path
  • Step 2 select the path k with the highest power among the paths in the current set
  • Step 3 classify each path j of R(k, j)>Threshold into one cluster, and delete it from the original set;
  • Step 4 repeat steps 2 and 3 until there is no path in the set;
  • step 5 the time of each cluster is averaged by the power weighting of each path in each cluster to obtain the final time delay of the representative path, and the power is obtained by using the average power of each path to obtain the final power of the representative path.
  • Threshold can be selected as 0.5 or adjusted to any value according to experience and other criteria. Update the original estimated number of paths to the clustered value.
  • h l, hor (t, n, p) [h l (t, 1, n, p)...h l (t, M, n, p)] T ;
  • the vertical arrival angle estimation of each path is the same as the previous step, and the direction is changed from horizontal to vertical.
  • Polarization direction correlation estimation calculates the correlation value between two sets of channel estimation values with different polarizations, and finally obtains the correlation value.
  • Point model generation according to the path information, power, time delay, vertical angle of arrival, horizontal angle of arrival, and correlation information of each path generated in the previous steps, input to the corresponding channel generation model to generate channel coefficients.
  • the channel model used is input into the SCME channel model, and the parameters obtained in the previous steps are used to replace the parameters fitted in advance according to the classic scene, so that a specific external field environment and an excellent fit to the real channel of the external field can be obtained.
  • channel model At the same time, the result of the channel characteristic parameters and the data label information are stored as a sampling point for subsequent model regression.
  • the multipath time delay spread, horizontal and vertical spatial angle power spectrums are calculated. This calculation is performed according to the mathematical definition of the values. Then assuming that the calculated value is X, and according to the 3GPP TR38.901 model, it is considered to be in line with the lognormal distribution in the classic scene, and the subscript u represents different sampling points, then the mean and variance estimates of the distribution can be obtained as :
  • any number of random channel sampling eigenvalues conforming to the distribution can be generated to generate multipath parameters, which can be input into the SCME multipath channel generation model, and other arbitrary multipath channel generation models to generate channel coefficients.
  • the channel features output by the point model are stored together with the grid label information.
  • the method in step 10 is used to output the channel coefficients.
  • the tag data is used to generate a space-time two-dimensional grid index
  • the transmitter's latitude and longitude is used to generate a spatial grid dimension
  • the time stamp is used to generate a time grid dimension.
  • the grid granularity is set according to the actual data volume and demand.
  • the wireless channel data collected on site is used for data analysis and then the channel model is generated, which effectively avoids the mismatch and error of the generalized model based on the classic scene for a specific scene, and is conducive to the system and algorithm on this basis. and precise research on the problem.
  • the time-domain spectrum estimation plus channel classification clustering can effectively eliminate or suppress the expansion introduced by the filter and other processing modules in the communication equipment, accurately separate the irrelevant paths in the channel, and approach the real external field channel.
  • the point model, statistical model, and grid model used can establish appropriate models under different sampled data volumes, different data integrity, and different application scenarios, and support modeling and simulation from the link to the system level.
  • FIG. 4 is a block diagram of a channel modeling device according to this embodiment. As shown in FIG. 4 , it includes:
  • the acquisition module 42 is configured to obtain the external field measurement data when the external field communication equipment of multiple sampling points receives the signal, and obtain multiple external field measurement data;
  • the feature extraction module 44 is configured to perform feature extraction on the plurality of external field measurement data respectively, to obtain channel characteristics of the plurality of sampling points;
  • the processing module 46 is configured to perform model regression processing according to the channel characteristics of the plurality of sampling points.
  • Fig. 5 is a block diagram one of the channel modeling device according to this preferred embodiment, as shown in Fig. 5, the processing module 46 includes:
  • the output submodule 52 is configured to perform statistical model regression or network model regression according to the channel characteristics of the plurality of sampling points, and output the channel characteristics of all sampling points;
  • the input sub-module 54 is configured to input the channel characteristics of all the sampling points into the multipath channel model as channel parameters, and obtain the analog channel at any point output by the multipath channel model.
  • Fig. 6 is a block diagram two of the channel modeling device according to this preferred embodiment, as shown in Fig. 6, the feature extraction module 44 includes:
  • the execution sub-module 62 is configured to perform the following operations on each of the plurality of external field measurement data to obtain the channel characteristics of the plurality of sampling points, wherein the external field measurement data being executed is called the current external field Measurement data:
  • the polarization characteristics are correlations between channels in different polarization directions, wherein the channel characteristics include the multipath The time delay of the multipath, the average power of the multipath, the angle of arrival of the multipath, and the polarization characteristics of the multipath.
  • the execution sub-module 62 is also set to
  • FIG. 7 is a third block diagram of a channel modeling device according to this preferred embodiment. As shown in FIG. 7, the acquisition module 42 includes:
  • the acquisition sub-module 72 is configured to acquire the measurement data of the external field channel when the external field communication equipment of the plurality of sampling points receives signals;
  • the filtering sub-module 74 is configured to filter out the plurality of external field measurement data satisfying preset conditions from the measurement data of the plurality of sampling points respectively.
  • the acquiring submodule 72 is also set to
  • the measurement data includes the estimated wireless channel value, the quality label data of the estimated wireless channel value, and the Grid label data for wireless channel estimates.
  • the screening submodule 74 includes:
  • the judging unit is configured to judge whether the measurement data of the plurality of sampling points meet a preset condition according to the quality label data of the estimated wireless channel value;
  • the deleting unit is configured to respectively delete data that does not meet the preset condition from the measurement data of the plurality of sampling points, so as to obtain the plurality of external field measurement data that meet the preset condition.
  • the judging unit is further configured to
  • the signal of the measurement data of the plurality of sampling points is judged respectively. Whether the interference-to-noise ratio is greater than a first preset threshold, whether the Doppler frequency shift is less than a second preset threshold and whether the estimated motion speed is less than a third preset threshold;
  • Embodiments of the present disclosure also provide a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps in any one of the above method embodiments when running.
  • the above-mentioned computer-readable storage medium may include but not limited to: U disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM) , mobile hard disk, magnetic disk or optical disk and other media that can store computer programs.
  • ROM read-only memory
  • RAM random access memory
  • mobile hard disk magnetic disk or optical disk and other media that can store computer programs.
  • Embodiments of the present disclosure also provide an electronic device, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.
  • the electronic device may further include a transmission device and an input and output device, wherein the transmission device is connected to the processor, and the input and output device is connected to the processor.
  • each module or each step of the above-mentioned disclosure can be realized by a general-purpose computing device, and they can be concentrated on a single computing device, or distributed in a network composed of multiple computing devices In fact, they can be implemented in program code executable by a computing device, and thus, they can be stored in a storage device to be executed by a computing device, and in some cases, can be executed in an order different from that shown here. Or described steps, or they are fabricated into individual integrated circuit modules, or multiple modules or steps among them are fabricated into a single integrated circuit module for implementation. As such, the present disclosure is not limited to any specific combination of hardware and software.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Monitoring And Testing Of Transmission In General (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本公开实施例提供了一种信道建模方法、装置、存储介质及电子装置，该方法包括：获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；根据所述多个采样点的信道特征进行模型回归处理，可以解决相关技术中要模拟某个特定具体场景中的信道，就会因为具体信道和经典场景差异大，或者经典场景内部的方差，而产生较大偏差的问题，避免使用基于经验数据的通用信道模型出现的和不同外场实测信道的偏差，同时可以直接使用通信系统工作中自动生成的测量数据，避免了额外的测量和开销。

Description

一种信道建模方法、装置、存储介质及电子装置
相关申请的交叉引用
本公开基于2021年07月26日提交的发明名称为“一种信道建模方法、装置、存储介质及电子装置”的中国专利申请CN202110845872.8，并且要求该专利申请的优先权，通过引用将其所公开的内容全部并入本公开。
技术领域
本公开实施例涉及通信领域，具体而言，涉及一种信道建模方法、装置、存储介质及电子装置。
背景技术
目前无线信道建模常用的建模方式为从典型场景测量数据中导出多径散射模型的方式。如5G系统中常用的信道模型TR38.901。先在一些典型的场景用专用的无线信号测量和采集无线信道特征，经过数据分析后，将其归纳为不同的多径时域，空域分布特征，计算出对应的参数。使用信道模型时，先选定需要的场景，然后采用先前计算好的参数来生成对应场景下的模型。其缺点是只能对某类经典场景做泛化研究使用，一旦要模拟某个特定具体场景中的信道，就会因为具体信道和经典场景差异大，或者经典场景内部的方差，而产生较大偏差。
针对相关技术中要模拟某个特定具体场景中的信道，就会因为具体信道和经典场景差异大，或者经典场景内部的方差，而产生较大偏差的问题，尚未提出解决方案。
发明内容
本公开实施例提供了一种信道建模方法、装置、存储介质及电子装置，以至少解决相关技术中要模拟某个特定具体场景中的信道，就会因为具体信道和经典场景差异大，或者经典场景内部的方差，而产生较大偏差的问题。
根据本公开的一个实施例，提供了一种信道建模方法，包括：
获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
根据所述多个采样点的信道特征进行模型回归处理。
根据本公开的另一个实施例，还提供了一种信道建模装置，包括：
获取模块，设置为获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
特征提取模块，设置为分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
处理模块，设置为根据所述多个采样点的信道特征进行模型回归处理。
根据本公开的又一个实施例，还提供了一种计算机可读的存储介质，所述存储介质中存储有计算机程序，其中，所述计算机程序被设置为运行时执行上述任一项方法实施例中的步骤。
根据本公开的又一个实施例，还提供了一种电子装置，包括存储器和处理器，所述存储器中存储有计算机程序，所述处理器被设置为运行所述计算机程序以执行上述任一项方法实施例中的步骤。
附图说明
图1是本公开实施例的信道建模方法的移动终端的硬件结构框图；
图2是根据本公开实施例的信道建模方法的流程图；
图3是根据本实施例的信道构建系统的框图；
图4是根据本实施例的信道建模装置的框图；
图5是根据本优选实施例的信道建模装置的框图一；
图6是根据本优选实施例的信道建模装置的框图二；
图7是根据本优选实施例的信道建模装置的框图三。
具体实施方式
下文中将参考附图并结合实施例来详细说明本公开的实施例。
需要说明的是，本公开的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。
本公开实施例中所提供的方法实施例可以在移动终端、计算机终端或者类似的运算装置中执行。以运行在移动终端上为例，图1是本公开实施例的信道建模方法的移动终端的硬件结构框图，如图1所示，移动终端可以包括一个或多个(图1中仅示出一个)处理器102(处理器102可以包括但不限于微处理器MCU或可编程逻辑器件FPGA等的处理装置)和用于存储数据的存储器104，其中，上述移动终端还可以包括用于通信功能的传输设备106以及输入输出设备108。本领域普通技术人员可以理解，图1所示的结构仅为示意，其并不对上述移动终端的结构造成限定。例如，移动终端还可包括比图1中所示更多或者更少的组件，或者具有与图1所示不同的配置。
存储器104可用于存储计算机程序，例如，应用软件的软件程序以及模块，如本公开实施例中的信道建模方法对应的计算机程序，处理器102通过运行存储在存储器104内的计算机程序，从而执行各种功能应用以及业务链地址池切片处理，即实现上述的方法。存储器104可包括高速随机存储器，还可包括非易失性存储器，如一个或者多个磁性存储装置、闪存、或者其他非易失性固态存储器。在一些实例中，存储器104可进一步包括相对于处理器102远程设置的存储器，这些远程存储器可以通过网络连接至移动终端。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
传输装置106用于经由一个网络接收或者发送数据。上述的网络具体实例可包括移动终端的通信供应商提供的无线网络。在一个实例中，传输装置106包括一个网络适配器(Network Interface Controller，简称为NIC)，其可通过基站与其他网络设备相连从而可与互联网进行通讯。在一个实例中，传输装置106可以为射频(Radio Frequency，简称为RF)模块，其用于通过无线方式与互联网进行通讯。
在本实施例中提供了一种运行于上述移动终端或网络架构的信道建模方法，图2是根据 本公开实施例的信道建模方法的流程图，如图2所示，该流程包括如下步骤：
步骤S202，获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
步骤S204，分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
步骤S206，根据所述多个采样点的信道特征进行模型回归处理。
本实施例中，上述步骤S206具体可以包括：根据所述多个采样点的信道特征进行统计模型回归或网络模型回归，输出所有采样点的信道特征；将所述所有采样点的信道特征作为信道参数输入到多径信道模型中，得到所述多径信道模型输出的任意点的模拟信道。
通过上述步骤S202至S206，可以解决相关技术中要模拟某个特定具体场景中的信道，就会因为具体信道和经典场景差异大，或者经典场景内部的方差，而产生较大偏差的问题，避免使用基于经验数据的通用信道模型出现的和不同外场实测信道的偏差，同时可以直接使用通信系统工作中自动生成的测量数据，避免了额外的测量和开销。
本实施例中，上述步骤S204具体可以包括：
对所述多个外场测量数据中的每个外场测量数据执行以下操作，得到所述多个采样点的信道特征，其中，对于正在执行的外场测量数据称为当前外场测量数据：
根据所述当前外场测量数据确定信道中多径的径数；
根据所述信道的径数预估所述多径的时延；
根据所述多径的时延从所述当前外场测量数据中计算各径的时域响应，根据所述各径的时域响应预估所述多径的平均功率；
根据信道中所述多径之间的相关性将所述多径进行聚类归簇，得到多个目标径，具体的，计算所述多径之间信道的相关性；根据所述相关性对所多径进行归簇，得到多簇集合，其中，每簇集合中包括至少一径；分别在所述多簇集合中选取所述平均功率最大的目标径，得到所述多个目标径；
确定所述多个目标径的到达角度以及所述多径的极化特征，其中，所述极化特征为不同极化方向上信道间的相关性，其中，所述信道特征包括所述多径的时延、所述多径的平均功率、所述多径的到达角度以及所述多径的极化特征。
本实施例中，上述步骤S202具体可以包括：
S2021，获取所述多个采样点的外场通信设备接收信号时外场信道的测量数据；
进一步的，上述步骤S2021具体可以包括：当所述外场通信设备接收信号时进行信道估计，得到无线信道估计值；获取所述无线信道估计值的质量标签数据与所述无线信道估计值的网络标签数据，其中，所述测量数据包括所述无线信道估计值、所述无线信道估计值的质量标签数据以及所述无线信道估计值的网格标签数据。
S2022，分别从所述多个采样点的测量数据中筛选出满足预设条件的所述多个外场测量数据。
进一步的，上述步骤S2022具体可以包括：
分别根据所述无线信道估计值的质量标签数据判断所述多个采样点的测量数据是否满足预设条件，具体的，在所述无线信道估计值的质量标签数据包括当前接收信号的信干噪比、多普勒频移以及运动速度估计值的情况下，分别判断所述多个采样点的测量数据的所述信干 噪比是否大于第一预设阈值，所述多普勒频移是否小于第二预设阈值以及所述运动速度估计值是否小于第三预设阈值；在判断结果为是的情况下，确定满足所述预设条件；在判断结果为否的情况下，确定不满足所述预设条件；
分别从所述多个采样点的测量数据中删除不满足所述预设条件的数据，得到满足所述预设条件的所述多个外场测量数据。
图3是根据本实施例的信道构建系统的框图，如图3所示，包括原始数据获取，信道特征提取，模型回归三个子系统。
原始数据获取子系统主要用于获取被信道特征提取子系统和模型回归子系统可以使用的外场信道采样点，其中：
外场信道测量获取，从外场运行的无线设备中获取外场测量数据。测量数据包括从外场接受信号中提取的无线信道估计值，外场无线信道估计值的质量标签数据，以及外场无线信道估计值的网格标签数据。估计方法可以是目前通用的方法以及其他方法，其输出值包括：无线信道估计值，外场无线信道估计值的质量标签数据，以及外场无线信道估计值的网格标签数据。无线信道估计值可以是接收机在各种信号的导频上估计得到的时频或者频域估计值，以及其他用各种方法获得的信道估计值；质量标签数据可以是当前接收信号的信干噪比，多普勒频移，用户运动速度估计(即上述的运动速度估计值)，数据接收正确率及其他反应数据接收和信道估计质量的测量值；网格标签数据，可以是其他发射机和接收机的经纬度坐标，地图坐标，采样时间戳，或者是其他反映地理位置和时间信息的数据。质量标签数据和网格标签数据为可选的数据。
信道数据筛选，主要从外场测量数据中，当质量标签数据具备时，根据其质量标签数据，筛除相应的不符合质量需求的数据，只将符合需求的数据进行后续处理。根据事先规定的准则，将质量标签数据与其作比较，删除不符合要求的数据。准则可以是单个条件和多个条件的组合，这些条件可以是：接收的信干噪比大于一定的门限；多普勒频移小于一定的门限；用户运动速度小于一定的门限；相关可用的信道估计值数量超过门限；以及其他的，能够判断该信道采样是在移动性不强，通信质量较高，相关数据量大的条件下取得的条件。此准则时为了提高对该位置点的信道特征估计的精度，并非本方法仅可应用于低速信道建模。通过对运动和轨迹建模结合，本方法可以应用与任意速度的信道建模。
信道特征提取子系统主要用于从符合质量需求的数据中提取信道的多径时延，到达角度和极化相关性特征，其中：
信道多径径数估计，估计当前分析的无线信道数据中可以辨析的多径的径数。估计方法可以采用基于旋转不变技术的信号参数估计(Estimating Signal Parameter via Rotational Invariance techniques，简称为ESPRIT)或多信号分类(Multiple Signal Classification Algorithm，简称为MUSIC)算法以及其他高精度谱估计算法及其变体。径数的检测原则可以是能量大于预定门限则确定为一径的准则，或者是最后确定的各径总体能量大于总能量百分比门限的准则。
多径时延估计，根据输出的多径数来估计各径的时延。估计方法可以采用ESPRIT或MUSIC算法以及其他高精度谱估计算法及其变体。
各径时域响应估计，根据估计出的时延来从无线信道数据中计算出各径的时域响应，包含各个接收天线上的幅度和相位，并估计各径的平均功率值。可采用最小二乘法估计，也可 采用目前业界已知的信号估计算法，如最大似然，最大先验概率算法等等。
各径信道归簇，将高相关性的各径进行聚类归簇，用一径信道来替代，维持处理后的各径之间的低相关性。相关性可以根据各径时延响应估计的输出来计算，在按照相关性聚簇之后，可以选择簇中的一径，或者簇中各径加权平均的方式得到最后代表整簇的一径。并更新信道径数估计值。
各径水平到达角估计，对各径信道归簇分离出的各径信道，估计其水平方向的达到角的信息。以上在得到相关矩阵后，估计达到角度的算法也可以是多步迭代的搜索算法，可以是谱估计算法，如ESPRIT和MUSIC算法，以及它们的变体。
各径垂直到达角估计，对各径信道归簇分离出的各径信道，估计其垂直方向的达到角的信息。以上在得到相关矩阵后，估计达到角度的算法也可以是多步迭代的搜索算法，可以是谱估计算法，如ESPRIT和MUSIC算法，以及它们的变体。
极化方向相关性估计，对各径信道归簇分离出的各径信道，估计不同极化方向信道间的相关性。
模型回归子系统利用信道特征提取子系统提取的信道特征，生成不同的模型，适用于不同的场景，其中：
点模型生成，将信道特征提取子系统输出的信道径数，时延，各径功率，垂直水平到达角以及极化相关性等特征作为参数输入到多径信道模型，生成信道；将这些参数作为一个信道整体采样点保存，并输出到统计模型回归和网格模型回归。此处所用的信道模型可以是任何一个可以根据多径参数来计算信道系数的模型，比如3GPP TR38.901中定义的多径信道模型，也可以是空间信道模型(Spatial Channel Model，简称为SCM)，扩展空间信道模型(Spatial Channel Model Extension，简称为SCME)等信道模型，用前述步骤获取的参数在其簇时延模型(Cluster Delay Line，简称为CDL)中代替其预先根据经典场景拟合的参数，即可获得正对一个具体外场环境的，和外场真实信道拟合性极好的信道模型。同时存储此次信道特征参数的结果，以及数据标签信息作为一个采样点，为后续模型回归使用。
统计模型回归，将点模型输出的多个信道整体采样点进行统计模型回归，得到信道参数，如时延，功率，空间角度分布等的概率分布。利用该概率分布生成外场中没有被测量的位置的信道参数和信道模型。所使用的概率分布可以是目前经典信道模型中，或者是任意其他手段提高的先验的外场参数分布模型。
网格模型回归，将点模型输出的多个信道整体采样点进行网格模型回归，将提取的信道参数和采用点的网格信道合并存储。利用网格的信息和已经测量的数据，对没有测量数据的网络输出预测信道参数和信道模型。
外场通信设备可以使任意一种具有无线收发功能的多天线收发设备。该设备包括但不限于：5G基站，4G基站，WiFi节点等等。处理流程包括以下步骤：
外场信道测量获取，外场通信设备进行接收信号时，需要进行信道估计，信号质量测量，以及发射机接收机位置估计等测量过程。并将其测量估计值按照规定格式进行输出。检测用的参考信号可以是探测参考信号(Sounding Reference Signal，简称为SRS)，或者是数据和控制信道的解调参考信号(Demodulation Reference Signal，简称为DMRS)，以及其他接收端已知的，可用于信道估计的参考信号序列。输出值包括：无线信道估计值，信干噪比测量值，发射机的多普勒频移，发射机的地理经纬度坐标，无线信道估计数据的时间戳。
无线信道估计值：对发射机u利用频域参考信号获得的频域信道估计的结果为H u(k，t，m，n，p)，其中k为OFDM(Orthogonal Frequency Division，简称为正交频分复用)中的频域子载波的编号，t为时域采样的时刻，m为接收机天线面板中水平方向的天线振子编号，n为接收机天线面板中垂直方向的天线振子编号，p为接收机天线面板中不同极化方向的天线振子极化编号。各个编号的取值范围为1到其对应大写符号代表的数字。其中，T为总时域采样时刻数，M为接收机天线面板中水平方向的总天线振子数，N为接收机天线面板中垂直方向的总天线振子数，P接收机天线面板中总同极化方向数。如果输入信道估计值是时域信道，则可通过傅里叶变换将其变换为频域信道值。
数据信道筛选，滤除不符合事先定义要求的输入：将信干噪比低于10dB，多普勒频移动高于300Hz的数据样本，以及时间上采样次数少于20次的样本滤除。以上数字和门限可以根据经验数据或其他准则调整。
信道多径数的提取，可采用ESPRIT算法。如果输入信道估计值是时域信道，则可通过傅里叶变换将其变换为频域信道值，包括：
首先估计一个原始信道样本中频域相关矩阵，将不同采样时刻，不同天线单元上的信道按其子载波的频率位置排列为列矢量：
H u，f(t，m，n，p)＝[H u(1，t，m，n，p)…H u(K，t，m，n，p)] T；
进而估计频域信道的互相关矩阵：
Figure PCTCN2022106543-appb-000001
采用Forward-Back方案进一步稳定统计结果，压低噪声：
Figure PCTCN2022106543-appb-000002
其中J为一个K维的反对角矩阵，conj(·)为一个逐元素取共轭的操作。并对其做SVD分解：
[U u，f，Λ u，f，V u，f]＝svd(R u，f，FB)；
其中svd()函数为对矩阵奇异值分解运算，Λ u，f＝diag{λ 1 … λ K}包含了矩阵R u，f，FB的K个奇异值。根据这些奇异值，计算当径数为i时的最小表示长度：
Figure PCTCN2022106543-appb-000003
则最后得到的径数估计为：
Figure PCTCN2022106543-appb-000004
多径时延估计，多径时延估计可以采用谱估计的方式获得小于通信系统采样率的，精确且独立的时延。
当采用ESPRIT方法估计如下：
令矩阵U L为矩阵U u，f中包含前L个特征向量的，大小为K×L的矩阵。在此基础上，产 生两个新的矩阵：
U 1＝[I，0]U L；
U 2＝[0，I]U L；
其中I是大小为(K-1)×(K-1)的单位对角阵，0为(K-1)×1的全0矢量。按照下面公式对以下矩阵进行特征值分解：
Figure PCTCN2022106543-appb-000005
其中eig()为求矩阵的特征值分解。
Π是一个包含L个特征值的向量，其中第l个特征值为γ l，则得到第l径的时延估计为：
Figure PCTCN2022106543-appb-000006
其中f Δ，H为原始数据中两个信道采样点的频域间隔，arg()为求复数辐角的运算。
各径时域响应估计，利用前面步骤获得的各径时延信息和原先的频域信道估计，获得各径在时延上的响应值和功率值。采用最小二乘估计的算法如下：
根据上一步得到参数估计一个大小为K×L的矩阵如下：
Figure PCTCN2022106543-appb-000007
时域各径信道的最小二乘估计为：
Figure PCTCN2022106543-appb-000008
从而可得各径的功率分布为：
Figure PCTCN2022106543-appb-000009
各径信道归簇，计算各径之间信道的相关值，进行将相关性高的各径归簇，用一条单径信道代替。步骤如下，将各径放入一个原始集合，按下列步骤执行：
步骤1，计算各径之间的相关性R(i，j)；
步骤2，在当前集合各径中选择功率最大的径k；
步骤3，将R(k，j)＞Threshold的各径j都归为一簇，从原始集合中删去；
步骤4，重复步骤2，3，直到集合中没有径；
步骤5，各簇选择其功率最强的一径，删除其余径。
以上门限值Threshold可以选择为0.5或根据经验和其他准则调整为任意值。原先估计的径数更新为归簇后的值。
各径水平达到角估计，利用多步搜索法，获取水平空间域的角度功率，以下计算对归簇后的独立各径进行：
首先，计算第l径的信道的在水平方向的相关矩阵，将水平方向的第l个径的信道排列为矢量：
h l，hor(t，n，p)＝[h l(t，1，n，p)…h l(t，M，n，p)] T；
获得相关矩阵估计：
Figure PCTCN2022106543-appb-000010
分成粗估计精估计多次迭代，粗估计中，按照给定的初始估计步长
Figure PCTCN2022106543-appb-000011
去重新计算精细的功率谱，然后再用更精细的搜索来获取更精细的估计值。粗估计获得的角度功率谱按下
Figure PCTCN2022106543-appb-000012
其中
Figure PCTCN2022106543-appb-000013
为包含间隔为
Figure PCTCN2022106543-appb-000014
方向矢量的矩阵。
然后根据每个点所代表的角度信息，得到其功率最大的点：
Figure PCTCN2022106543-appb-000015
作方向估计值。
在[AoA l，coarse-ASA l，coarse，AoA l，coarse+ASA l，coarse]的范围内，按照给定的步长
Figure PCTCN2022106543-appb-000016
去重新计算精细的功率谱：
Figure PCTCN2022106543-appb-000017
Φ是包含所有采样点波束矢量的矩阵。按照和粗估计相同的公式重新计算AoA和ASA，可以获得更精确的估计。
以上精估计可以进行多次迭代。
各径垂直达到角估计，处理过程同上一步，方向从水平改变为垂直。
极化方向相关性估计，在极化不同的两组信道估计值之间计算其相关值，最后得到相关值。
点模型生成，根据前述步骤生成的各径信息，功率，时延，垂直到达角，水平到达角，以及各径相关性信息，输入到对应的信道生成模型，产生信道系数。所用的信道模型为3GPP TR38.901中定义的多径信道模型，在其簇时延模型(Cluster Delay Line，简称为CDL)中，用前述步骤获取的参数代替其预先根据经典场景拟合的参数，即可获得正对一个具体外场环境的，和外场真实信道拟合性极好的信道模型。同时存储此次信道特征参数的结果，以及数据标签信息作为一个采样点，为后续模型回归使用。
统计模型回归，当点模型中的采样点积累到一定数量，通过估计这些信道特征的概率分布，预测网络中没有测量的点的可能值。
首先根据各采样点的估计值，计算多径时延扩展，水平和垂直空间角度功率谱。该计算根据各值的数学定义进行。然后假设计算后的值为X，并根据3GPP TR38.901中模型认为其符合经典场景中的对数正态分布，下标u代表不同的采样点，则可得到该分布的均值和方差估计分别为：
Figure PCTCN2022106543-appb-000018
Figure PCTCN2022106543-appb-000019
利用获得的该分布，可以生成任意个符合该分布的随机信道采样特征值，生成多径参数，输入多径信道生成模型，如3GPP TR38.901多径信道生成模型中产生信道系数。
网格模型回归，将点模型输出的信道特征和网格标签信息一起存储。当需要输出本网格的信道时，采用步骤10的方法输出信道系数。利用标签数据生成时空两维的网格索引，利用发射机的经纬度生成空间的网格维度，利用时间戳生成的时间的网格维度。网格粒度根据实际的数据量和需求设定。当需要输出任意一个点的信道参数时，首先寻找所有网格中有数据，空间位置上最接近的网格集合，再在此基础上，在该网格集合上寻找时间上最接近网格。输出该网格的点模型参数到多径信道模型，按照步骤10的办法生成信道系数。
外场通信设备可以使任意一种具有无线收发功能的多天线收发设备。该设备包括但不限于：5G基站，4G基站，WiFi节点等等。处理流程包括以下步骤：
外场信道测量获取，外场通信设备进行接收信号时，需要进行信道估计，信号质量测量，以及发射机接收机位置估计等测量过程。并将其测量估计值按照规定格式进行输出。检测用的参考信号可以是探测参考信号(Sounding Reference Signal，简称为SRS)，或者是数据和控制信道的解调参考信号(Demodulation Reference Signal，简称为DMRS)，以及其他接收端已知的，可用于信道估计的参考信号序列。输出值包括：无线信道估计值，接收的参考信号强度，发射机的运动速度，发射机的地理经纬度坐标，无线信道估计数据的时间戳。
无线信道估计值：对发射机u利用频域参考信号获得的频域信道估计的结果为H u(k，t，m，n，p)，其中k为OFDM中的频域子载波的编号，t为时域采样的时刻，m为接收机天线面板中水平方向的天线振子编号，n为接收机天线面板中垂直方向的天线振子编号，p为接收机天线面板中不同极化方向的天线振子极化编号。各个编号的取值范围为1到其对应大写符号代表的数字。其中，T为总时域采样时刻数，M为接收机天线面板中水平方向的总天线振子数，N为接收机天线面板中垂直方向的总天线振子数，P接收机天线面板中总同极化方向数。如果输入信道估计值是时域信道，则可通过傅里叶变换将其变换为频域信道值。
数据信道筛选，滤除不符合实现定义要求的输入：将参考信号强度低于-110dB，运动速度大于10公里每小时的数据样本，以及时间上采样次数少于20次的样本滤除。以上数字和门限可以根据经验数据或其他准则调整。
信道多径数的提取，按照经验值，固定信道最多径数为L＝20。
多径时延估计，多径时延估计可以采用谱估计的方式获得小于通信系统采样率的，精确且独立的时延。当采用ESPRIT方法估计如下：
首先估计一个原始信道样本中频域相关矩阵，将不同采用时刻，不同天线单元上的信道按其子载波的频率位置排列为列矢量：
H u，f(k，t，m，n，p)＝[H u(1，t，m，n，p)…H u(K，t，m，n，p)] T；
进而估计频域信道的互相关矩阵：
Figure PCTCN2022106543-appb-000020
采用Forward-Back方案进一步稳定统计结果，压低噪声：
Figure PCTCN2022106543-appb-000021
其中J为一个K维的反对角矩阵，conj(·)为一个逐元素取共轭的操作。并对其做SVD分解：
[U u，f，Λ u，f，V u，f]＝svd(R u，f，FB)；
令矩阵U L为矩阵U u，f中包含前L＝20个特征向量的，大小为K×L的矩阵。在此基础上，产生两个新的矩阵：
U 1＝[I，0]U L；
U 2＝[0，I]U L；
其中I是大小为(K-1)×(K-1)的单位对角阵，0为(K-1)×1的全0矢量。按照下面公式对以下矩阵进行特征值分解：
Figure PCTCN2022106543-appb-000022
Π是一个包含L个特征值的向量，其中第l个特征值为γ l，则得到第l径的时延估计为：
Figure PCTCN2022106543-appb-000023
其中f Δ，H为原始数据中两个信道采样点的频域间隔。
以上eig()，svd()，arg()的含义同上述实施例。
各径时域响应估计。利用前面步骤获得的各径时延信息和原先的频域信道估计，获得各径在时延上的响应值和功率值。采用最小二乘估计的算法如下：
根据上一步得到参数估计一个大小为K×L的矩阵如下：
Figure PCTCN2022106543-appb-000024
时域各径信道的最小二乘估计为：
Figure PCTCN2022106543-appb-000025
从而可得各径的功率分布为：
Figure PCTCN2022106543-appb-000026
各径信道归簇，计算各径之间信道的相关值，进行将相关性高的各径归簇，用一条单径信道代替，将各径放入一个原始集合，按下列步骤执行：
步骤1，计算各径之间的相关性R(i，j)；
步骤2，在当前集合各径中选择功率最大的径k；
步骤3，将R(k，j)＞Threshold的各径j都归为一簇，从原始集合中删去；
步骤4，重复步骤2，3，直到集合中没有径；
步骤5，各簇的时间通过各簇中各径的功率加权进行平均，得到所代表径的最后时延，功率用各径的平均功率得到代表径的最终功率。
以上门限值Threshold可以选择为0.5或根据经验和其他准则调整为任意值。将原先估计的径数更新为归簇后的值。
各径水平达到角估计，利用多步搜索法，获取水平空间域的角度功率，以下计算对归簇后的独立各径进行：
首先，计算第l径的信道的在水平方向的相关矩阵，将水平方向的第l个径的信道排列为矢量：
h l，hor(t，n，p)＝[h l(t，1，n，p)…h l(t，M，n，p)] T；
获得相关矩阵估计
Figure PCTCN2022106543-appb-000027
以上在得到相关矩阵后，在其上运行谱估计算法，如ESPRIT和MUSIC算法，以及它们的变体，来获得各径的角度估计。
各径垂直达到角估计，处理过程同上一步，方向从水平改变为垂直。
极化方向相关性估计，在极化不同的两组信道估计值之间计算其相关值，最后得到相关值。
点模型生成，根据前述步骤生成的各径信息，功率，时延，垂直到达角，水平到达角，以及各径相关性信息，输入到对应的信道生成模型，产生信道系数。所用的信道模型，输入到SCME信道模型中，用前述步骤获取的参数代替其预先根据经典场景拟合的参数，即可获得正对一个具体外场环境的，和外场真实信道拟合性极好的信道模型。同时存储此次信道特征参数的结果，以及数据标签信息作为一个采样点，为后续模型回归使用。
统计模型回归。当点模型中的采样点积累到一定数量，通过估计这些信道特征的概率分布，预测网络中没有测量的点的可能值。
首先根据各采样点的估计值，计算多径时延扩展，水平和垂直空间角度功率谱。该计算根据各值的数学定义进行。然后假设计算后的值为X，并根3GPP TR38.901模型认为其符合经典场景中的对数正态分布，下标u代表不同的采样点，则可得到该分布的均值和方差估计分别为：
Figure PCTCN2022106543-appb-000028
Figure PCTCN2022106543-appb-000029
利用获得的该分布，可以生成任意个符合该分布的随机信道采样特征值，生成多径参数， 输入SCME多径信道生成模型，等其他任意多径信道生成模型中产生信道系数。
网格模型回归，将点模型输出的信道特征和网格标签信息一起存储。当需要输出本网格的信道时，采用步骤10的方法输出信道系数。利用标签数据生成时空两维的网格索引，利用发射机的经纬度生成空间的网格维度，利用时间戳生成的时间的网格维度。网格粒度根据实际的数据量和需求设定。当需要输出任意一个点的信道参数时，首先寻找所有网格中有数据，空间位置上最接近的网格集合，再在此基础上，在该网格集合上寻找时间上最接近网格。将距离最近的3个网格之间的参数进行平均后，得到输出的信道参数，将该参数输出该网格的点模型参数到多径信道模型，生成信道系数。
本实施例采用现场采集的无线信道数据进行数据分析而后生成信道模型，有效避免了一般基于经典场景的泛化模型对于某个具体场景的不匹配和误差，有利于在此基础上进行系统，算法和问题的精确研究。采用的时域谱估计加信道分类归簇可以有效排除或者抑制通信设备中滤波器及其其他处理模块引入的扩展，准确的分离出信道中不相干的各径，逼近真正的外场信道。采用的点模型，统计模型，网格模型可以在不同的采样数据量，不同的数据完整度，不同的应用场景下，都建立合适的模型，支持从链路到系统级别的建模和仿真。
根据本公开的另一个实施例，还提供了一种信道建模装置，图4是根据本实施例的信道建模装置的框图，如图4所示，包括：
获取模块42，设置为获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
特征提取模块44，设置为分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
处理模块46，设置为根据所述多个采样点的信道特征进行模型回归处理。
图5是根据本优选实施例的信道建模装置的框图一，如图5所示，所述处理模块46包括：
输出子模块52，设置为根据所述多个采样点的信道特征进行统计模型回归或网络模型回归，输出所有采样点的信道特征；
输入子模块54，设置为将所述所有采样点的信道特征作为信道参数输入到多径信道模型中，得到所述多径信道模型输出的任意点的模拟信道。
图6是根据本优选实施例的信道建模装置的框图二，如图6所示，所述特征提取模块44包括：
执行子模块62，设置为对所述多个外场测量数据中的每个外场测量数据执行以下操作，得到所述多个采样点的信道特征，其中，对于正在执行的外场测量数据称为当前外场测量数据：
根据所述当前外场测量数据确定信道中多径的径数；
根据所述信道的径数预估所述多径的时延；
根据所述多径的时延从所述当前外场测量数据中计算各径的时域响应，根据所述各径的时域响应预估所述多径的平均功率；
根据信道中所述多径之间的相关性将所述多径进行聚类归簇，得到多个目标径；
确定所述多个目标径的到达角度以及所述多径的极化特征，其中，所述极化特征为不同极化方向上信道间的相关性，其中，所述信道特征包括所述多径的时延、所述多径的平均功 率、所述多径的到达角度以及所述多径的极化特征。
在一示例性实施例中，所述执行子模块62，还设置为
计算所述多径之间信道的相关性；
根据所述相关性对所多径进行归簇，得到多簇集合，其中，每簇集合中包括至少一径；
分别在所述多簇集合中选取所述平均功率最大的目标径，得到所述多个目标径。
图7是根据本优选实施例的信道建模装置的框图三，如图7所示，所述获取模块42包括：
获取子模块72，设置为获取所述多个采样点的外场通信设备接收信号时外场信道的测量数据；
筛选子模块74，设置为分别从所述多个采样点的测量数据中筛选出满足预设条件的所述多个外场测量数据。
在一示例性实施例中，所述获取子模块72，还设置为
当所述外场通信设备接收信号时进行信道估计，得到无线信道估计值；
获取所述无线信道估计值的质量标签数据与所述无线信道估计值的网络标签数据，其中，所述测量数据包括所述无线信道估计值、所述无线信道估计值的质量标签数据以及所述无线信道估计值的网格标签数据。
在一示例性实施例中，所述筛选子模块74包括：
判断单元，设置为分别根据所述无线信道估计值的质量标签数据判断所述多个采样点的测量数据是否满足预设条件；
删除单元，设置为分别从所述多个采样点的测量数据中删除不满足所述预设条件的数据，得到满足所述预设条件的所述多个外场测量数据。
在一示例性实施例中，所述判断单元，还设置为
在所述无线信道估计值的质量标签数据包括当前接收信号的信干噪比、多普勒频移以及运动速度估计值的情况下，分别判断所述多个采样点的测量数据的所述信干噪比是否大于第一预设阈值，所述多普勒频移是否小于第二预设阈值以及所述运动速度估计值是否小于第三预设阈值；
在判断结果为是的情况下，确定满足所述预设条件；
在判断结果为否的情况下，确定不满足所述预设条件。
本公开的实施例还提供了一种计算机可读存储介质，该计算机可读存储介质中存储有计算机程序，其中，该计算机程序被设置为运行时执行上述任一项方法实施例中的步骤。
在一个示例性实施例中，上述计算机可读存储介质可以包括但不限于：U盘、只读存储器(Read-Only Memory，简称为ROM)、随机存取存储器(Random Access Memory，简称为RAM)、移动硬盘、磁碟或者光盘等各种可以存储计算机程序的介质。
本公开的实施例还提供了一种电子装置，包括存储器和处理器，该存储器中存储有计算机程序，该处理器被设置为运行计算机程序以执行上述任一项方法实施例中的步骤。
在一个示例性实施例中，上述电子装置还可以包括传输设备以及输入输出设备，其中，该传输设备和上述处理器连接，该输入输出设备和上述处理器连接。
本实施例中的具体示例可以参考上述实施例及示例性实施方式中所描述的示例，本实施 例在此不再赘述。
显然，本领域的技术人员应该明白，上述的本公开的各模块或各步骤可以用通用的计算装置来实现，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，并且在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本公开不限制于任何特定的硬件和软件结合。
以上所述仅为本公开的优选实施例而已，并不用于限制本公开，对于本领域的技术人员来说，本公开可以有各种更改和变化。凡在本公开的原则之内，所作的任何修改、等同替换、改进等，均应包含在本公开的保护范围之内。

Claims (11)

  1. 一种信道建模方法，包括：
    获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
    分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
    根据所述多个采样点的信道特征进行模型回归处理。
  2. 根据权利要求1所述的方法，其中，根据所述多个采样点的信道特征进行模型回归处理包括：
    根据所述多个采样点的信道特征进行统计模型回归或网络模型回归，输出所有采样点的信道特征；
    将所述所有采样点的信道特征作为信道参数输入到多径信道模型中，得到所述多径信道模型输出的任意点的模拟信道。
  3. 根据权利要求1所述的方法，其中，分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征包括：
    对所述多个外场测量数据中的每个外场测量数据执行以下操作，得到所述多个采样点的信道特征，其中，对于正在执行的外场测量数据称为当前外场测量数据：
    根据所述当前外场测量数据确定信道中多径的径数；
    根据所述信道的径数预估所述多径的时延；
    根据所述多径的时延从所述当前外场测量数据中计算各径的时域响应，根据所述各径的时域响应预估所述多径的平均功率；
    根据信道中所述多径之间的相关性将所述多径进行聚类归簇，得到多个目标径；
    确定所述多个目标径的到达角度以及所述多径的极化特征，其中，所述极化特征为不同极化方向上信道间的相关性，其中，所述信道特征包括所述多径的时延、所述多径的平均功率、所述多径的到达角度以及所述多径的极化特征。
  4. 根据权利要求3所述的方法，其中，根据信道中所述多径之间的相关性将所述多径进行聚类归簇，得到多个目标径包括：
    计算所述多径之间信道的相关性；
    根据所述相关性对所多径进行归簇，得到多簇集合，其中，每簇集合中包括至少一径；
    分别在所述多簇集合中选取所述平均功率最大的目标径，得到所述多个目标径。
  5. 根据权利要求1所述的方法，其中，获取所述多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据包括：
    获取所述多个采样点的外场通信设备接收信号时外场信道的测量数据；
    分别从所述多个采样点的测量数据中筛选出满足预设条件的所述多个外场测量数据。
  6. 根据权利要求5所述的方法，其中，获取所述多个采样点的外场通信设备接收信号时外场信道的测量数据包括：
    当所述外场通信设备接收信号时进行信道估计，得到无线信道估计值；
    获取所述无线信道估计值的质量标签数据与所述无线信道估计值的网络标签数据，其中，所述测量数据包括所述无线信道估计值、所述无线信道估计值的质量标签数据以及所述无线信道估计值的网格标签数据。
  7. 根据权利要求6所述的方法，其中，分别从所述多个采样点的测量数据中筛选出满足预设条件的所述多个外场测量数据包括：
    分别根据所述无线信道估计值的质量标签数据判断所述多个采样点的测量数据是否满足预设条件；
    分别从所述多个采样点的测量数据中删除不满足所述预设条件的数据，得到满足所述预设条件的所述多个外场测量数据。
  8. 根据权利要求7所述的方法，其中，分别根据所述无线信道估计值的质量标签数据判断所述多个采样点的测量数据是否满足预设条件包括：
    在所述无线信道估计值的质量标签数据包括当前接收信号的信干噪比、多普勒频移以及运动速度估计值的情况下，分别判断所述多个采样点的测量数据的所述信干噪比是否大于第一预设阈值，所述多普勒频移是否小于第二预设阈值以及所述运动速度估计值是否小于第三预设阈值；
    在判断结果为是的情况下，确定满足所述预设条件；
    在判断结果为否的情况下，确定不满足所述预设条件。
  9. 一种信道建模装置，包括：
    获取模块，设置为获取多个采样点的外场通信设备接收信号时的外场测量数据，得到多个外场测量数据；
    特征提取模块，设置为分别对所述多个外场测量数据进行特征提取，得到所述多个采样点的信道特征；
    处理模块，设置为根据所述多个采样点的信道特征进行模型回归处理。
  10. 一种计算机可读的存储介质，所述存储介质中存储有计算机程序，其中，所述计算机程序被设置为运行时执行所述权利要求1至8任一项中所述的方法。
  11. 一种电子装置，包括存储器和处理器，所述存储器中存储有计算机程序，所述处理器被设置为运行所述计算机程序以执行所述权利要求1至8任一项中所述的方法。
PCT/CN2022/106543 2021-07-26 2022-07-19 一种信道建模方法、装置、存储介质及电子装置 WO2023005746A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110845872.8A CN115694696A (zh) 2021-07-26 2021-07-26 一种信道建模方法、装置、存储介质及电子装置
CN202110845872.8 2021-07-26

Publications (1)

Publication Number Publication Date
WO2023005746A1 true WO2023005746A1 (zh) 2023-02-02

Family

ID=85043749

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/106543 WO2023005746A1 (zh) 2021-07-26 2022-07-19 一种信道建模方法、装置、存储介质及电子装置

Country Status (2)

Country Link
CN (1) CN115694696A (zh)
WO (1) WO2023005746A1 (zh)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102638290A (zh) * 2012-03-15 2012-08-15 北京邮电大学 一种基于信道测量的多径信号分量提取方法及装置
CN103763719A (zh) * 2014-01-02 2014-04-30 工业和信息化部电信研究院 一种用于td-lte系统的模拟路测的方法
CN107425895A (zh) * 2017-06-21 2017-12-01 西安电子科技大学 一种基于实测的3d mimo统计信道建模方法
CN110113119A (zh) * 2019-04-26 2019-08-09 国家无线电监测中心 一种基于人工智能算法的无线信道建模方法
CN111404847A (zh) * 2020-03-20 2020-07-10 中山大学 一种海洋通信系统的信道估计方法
CN112543471A (zh) * 2020-11-16 2021-03-23 南京邮电大学 一种面向复杂环境的机动5g混合接入链路中断预测方法
WO2021108940A1 (en) * 2019-12-01 2021-06-10 Nokia Shanghai Bell Co., Ltd. Channel state information feedback

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102638290A (zh) * 2012-03-15 2012-08-15 北京邮电大学 一种基于信道测量的多径信号分量提取方法及装置
CN103763719A (zh) * 2014-01-02 2014-04-30 工业和信息化部电信研究院 一种用于td-lte系统的模拟路测的方法
CN107425895A (zh) * 2017-06-21 2017-12-01 西安电子科技大学 一种基于实测的3d mimo统计信道建模方法
CN110113119A (zh) * 2019-04-26 2019-08-09 国家无线电监测中心 一种基于人工智能算法的无线信道建模方法
WO2021108940A1 (en) * 2019-12-01 2021-06-10 Nokia Shanghai Bell Co., Ltd. Channel state information feedback
CN111404847A (zh) * 2020-03-20 2020-07-10 中山大学 一种海洋通信系统的信道估计方法
CN112543471A (zh) * 2020-11-16 2021-03-23 南京邮电大学 一种面向复杂环境的机动5g混合接入链路中断预测方法

Also Published As

Publication number Publication date
CN115694696A (zh) 2023-02-03

Similar Documents

Publication Publication Date Title
KR100930799B1 (ko) 자동화된 클러스터링 방법 및 이를 이용한 이동통신환경에서 다중 경로의 클러스터링 방법 및 장치
Gaber et al. A study of wireless indoor positioning based on joint TDOA and DOA estimation using 2-D matrix pencil algorithms and IEEE 802.11 ac
Salmi et al. Detection and tracking of MIMO propagation path parameters using state-space approach
Kupershtein et al. Single-site emitter localization via multipath fingerprinting
EP2798366B1 (en) Method and system for localization
US10880854B2 (en) Intelligent base station with capability to identify three-dimensional environment, method for determining location thereof and storage medium
CN106909779B (zh) 基于分布式处理的mimo雷达克拉美罗界计算方法
CN108646213B (zh) 一种室内多径环境下直达波aoa判定方法
Zhang et al. An efficient machine learning approach for indoor localization
CN110049549B (zh) 基于WiFi指纹的多融合室内定位方法及其系统
CN112040400B (zh) 一种基于mimo-csi的单站点室内指纹定位方法、存储介质及设备
CN107343315B (zh) 基于无线局域网络多频点联合测量的定位系统及方法
CN109768816B (zh) 一种非高斯噪声3d-mimo系统数据检测方法
CN115486035A (zh) 用于信道估计的nn参数的类别
Vaghefi et al. Cooperative RF pattern matching positioning for LTE cellular systems
Moayyed et al. Clustering algorithms and validation indices for mmWave radio multipath propagation
Barua et al. A survey of direction of arrival estimation techniques and implementation of channel estimation based on SCME
CN111313943A (zh) 深度学习辅助的大规模天线阵列下的三维定位方法及装置
KR102416604B1 (ko) 무선통신 시스템의 정밀 측위 방법 및 장치
Li et al. Static CSI extraction and application in the tomographic channel model
US10901076B2 (en) Temporal analysis for user speed estimation in wireless networks
Tripkovic et al. Cluster density in crowdsourced mobile network measurements
WO2023005746A1 (zh) 一种信道建模方法、装置、存储介质及电子装置
Zhang et al. Single-site positioning method based on high-resolution estimation in VANET localization
Colpaert et al. 3D non-stationary channel measurement and analysis for MaMIMO-UAV communications

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22848351

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE