CN-121984625-A - Modeling method and device for non-stationary cluster delay line channel based on cluster evolution

CN121984625ACN 121984625 ACN121984625 ACN 121984625ACN-121984625-A

Abstract

The application relates to the technical field of channel modeling, and provides a non-stationary cluster delay line channel modeling method and device based on cluster evolution. The present application describes the non-stationary nature of urban multi-antenna channels by characterizing the process of extinction of each cluster using a markov process. Meanwhile, in the process of generating and extinguishing each cluster, in the current generating and extinguishing interval, cluster power is smoothed by combining cluster power with sampling time by using a cluster power evolution function, and the problem of step jump of cluster power in a traditional model is solved by introducing double time parameters of generating and extinguishing time and sampling time to carry out cluster power smoothing, so that the power attenuation of channel impulse response is more fit with the physical rule of actual propagation, and modeling errors caused by power mutation are reduced.

Inventors

LI WEI
ZHANG XIAOYING
MU ZHIHAO
WEI JIBO
ZHAO HAITAO
XIONG JUN
LIU XIAORAN
MA DONGTANG

Assignees

中国人民解放军国防科技大学

Dates

Publication Date: 20260505
Application Date: 20260330

Claims (10)

1. The modeling method of the non-stationary cluster delay line channel based on cluster evolution is characterized by comprising the following steps of: Collecting channel data and GPS data of a target area, and extracting time delay, power and angle information of channel multipath; based on the extracted time delay, power and angle information of the channel multipath, channel parameters including intra-cluster parameters and inter-cluster parameters are obtained through a clustering algorithm and a multipath tracking algorithm, and the occurrence and extinction process of each cluster is represented by a Markov process to obtain the occurrence and extinction probability of the cluster; fitting the cluster power change condition of each cluster through a primary function to obtain a cluster power evolution function corresponding to each cluster, and smoothing the cluster power by combining the cluster power evolution function with sampling time in the current extinguishing interval in the extinguishing process of each cluster; And constructing the channel impulse response based on the channel parameters, the cluster extinction probability and the cluster power after the smoothing.
2. The method for modeling a non-stationary cluster delay line channel based on cluster evolution of claim 1, wherein the channel impulse response is: ; Wherein, the L is the cluster number under the current snapshot; Is the central delay of the first cluster; K is the number of multipaths; is the time delay of the kth path in the ith cluster; complex amplitude for the kth path in the kth cluster; A random phase for the kth path in the kth cluster; The cluster center arrival angle of the first cluster; The arrival angle of the kth path in the kth cluster; is a motion velocity direction vector; A direction vector of a kth path in a kth cluster; A position vector for the nth antenna; Is the signal wavelength; The probability of being on/off for the first cluster.
3. The method for modeling a non-stationary cluster delay line channel based on cluster evolution according to claim 1, wherein the intra-cluster parameters include an intra-cluster multipath number, an intra-cluster delay offset, an intra-cluster angle offset, an intra-cluster K factor; the inter-cluster parameters comprise cluster time delay, cluster power and cluster angle.
4. The method for modeling a non-stationary cluster delay line channel based on cluster evolution of claim 3, wherein the step of characterizing a process of occurrence of each cluster using a markov process to obtain a probability of occurrence of the cluster, comprises: in the Markov process, the existing state of the cluster is regarded as a1 state, the disappearing state of the cluster is regarded as a 0 state, and the disappearing state of each cluster is judged based on the cluster power; And respectively calculating a transition matrix and a steady-state matrix of the on-off state of each cluster to obtain the cluster on-off probability.
5. The method of modeling a non-stationary cluster delay line channel based on cluster evolution of claim 4, wherein the transfer matrix is calculated according to the following formula: ; Wherein, the Is a transfer matrix; to be a probability of transitioning from the a-state to the b-state, ; The steady state matrix is calculated according to the following formula: ; Wherein, the Is a steady state matrix; is the steady state probability of the o-state, 。
6. The modeling method of a non-stationary cluster delay line channel based on cluster evolution according to claim 1, wherein obtaining channel parameters by a clustering algorithm and a multipath tracking algorithm based on extracted delay, power and angle information of channel multipath comprises: Calculating the kernel density of the multipath target point based on the extracted time delay and angle information of the multipath of the channel; calculating the relative density based on the kernel density of the multipath target point, and taking a set formed by multipath points with the relative density of 1 as an initial cluster center; The multipath point x-bit arc tail of any non-cluster center is used, the distance x is nearest and the density is larger than that of the multipath Obtaining arcs corresponding to multipath points of each non-cluster center for the arc heads to form a global pointing relation table, and carrying out preliminary clustering based on the global pointing relation table; if a communication path exists between the centers of the two clusters and the relative density of all multipath components on the path is higher than a preset threshold value, combining the two clusters to finish final clustering; tracking the same cluster among a plurality of snapshots by adopting a tracking algorithm based on multipath component distances, judging the existence time of each cluster, and extracting channel parameters.
7. The method for modeling a non-stationary cluster delay line channel based on cluster evolution of claim 3, wherein the cluster angle and cluster delay are calculated according to the following formula: ; The cluster power is calculated according to the following formula: ; Wherein, the The cluster angle or cluster time delay is used, and N is the number of multipath; The power of the ith multipath in the cluster; delay or angle of ith multipath in the cluster; Is cluster power.
8. The method for modeling a non-stationary cluster delay line channel based on cluster evolution according to claim 6, wherein tracking the same cluster among a plurality of snapshots by using a tracking algorithm based on a multipath component distance, judging the existence time of each cluster, and extracting channel parameters, comprises: detecting any two multipaths for obtaining a plurality of snapshots, and extracting channel parameters to obtain a joint parameter representation corresponding to the multipaths; Calculating the multipath component distances of two multipaths in an angle domain and a time delay domain respectively based on the joint parameter representation, and further calculating to obtain the multipath component distances of the two multipath totalities; If the distance between the multipath components of the two multipath populations is smaller than a preset threshold, judging that the cluster corresponding to the two multipath components is the continuation of the same cluster, and if the distance is larger than the preset threshold, judging that a new cluster is generated; Traversing all the snapshots, and judging the existence time of each cluster.
9. The method of modeling a non-stationary cluster delay line channel based on cluster evolution of claim 6, wherein the kernel density is calculated according to the following formula: ; Wherein, the Is the kernel density of the multipath point x, y is the index of the multipath points around the multipath point x, ; The number of multipath components nearest to x; Is the time delay of the ith multipath; An angle of the ith multipath; is the standard deviation of signal distribution in the time delay domain; is the standard deviation of the signal distribution in the angle domain; is the amplitude of the ith multipath.
10. A non-stationary cluster delay line channel modeling apparatus based on cluster evolution, comprising: The data acquisition module is used for acquiring channel data and GPS data of a target area and extracting time delay, power and angle information of channel multipath; The parameter acquisition module is used for acquiring channel parameters, including intra-cluster parameters and inter-cluster parameters, through a clustering algorithm and a multipath tracking algorithm based on the extracted time delay, power and angle information of the channel multipath, and representing the occurrence and extinction process of each cluster by utilizing a Markov process to obtain the occurrence and extinction probability of the cluster; The power smoothing module is used for fitting the cluster power change condition of each cluster through a primary function to obtain a cluster power evolution function corresponding to each cluster, and smoothing the cluster power by combining the cluster power with the sampling time in the current extinguishing interval in the extinguishing process of each cluster by using the cluster power evolution function; and the model construction module is used for constructing channel impulse response based on the channel parameters, the cluster generation and extinction probability and the cluster power after the smoothing processing.

Description

Modeling method and device for non-stationary cluster delay line channel based on cluster evolution Technical Field The application relates to the technical field of wireless communication and channel modeling of complex urban areas, in particular to a non-stationary cluster delay line channel modeling method and device based on cluster evolution. Background The spatial characteristic is an important characteristic of a complex urban Sub-1GHz wireless communication channel, and macroscopic statistical information (such as path loss, shadow fading, time delay expansion and the like) of the channel in the time domain, the frequency domain and the power domain can only be obtained by adopting single antenna detection, but key characteristics of the channel in the spatial dimension such as arrival angle, departure angle, spatial correlation and the like cannot be revealed. In urban environments with dense buildings and abundant scatterers, the spatial parameters can realize accurate assessment of spatial multiplexing gain, diversity gain and beam forming performance, so that the research on the multi-antenna channel characteristics of complex urban scenes is particularly important. The existing non-stationary cluster delay line model generally models the cluster extinction process as an instantaneous and independent random event by using a Markov process, wherein a cluster appears suddenly at a fixed power or suddenly disappears completely at a certain moment, and the jump process ignores a natural process that the corresponding cluster power gradually increases and decreases when a scatterer in a real channel gradually enters/leaves an effective scattering area. Disclosure of Invention Based on this, it is necessary to provide a modeling method and device for a non-stationary cluster delay line channel based on cluster evolution aiming at the technical problems. A modeling method of a non-stationary cluster delay line channel based on cluster evolution comprises the following steps: Collecting channel data and GPS data of a target area, and extracting time delay, power and angle information of channel multipath; based on the extracted time delay, power and angle information of the channel multipath, channel parameters including intra-cluster parameters and inter-cluster parameters are obtained through a clustering algorithm and a multipath tracking algorithm, and the occurrence and extinction process of each cluster is represented by a Markov process to obtain the occurrence and extinction probability of the cluster; fitting the cluster power change condition of each cluster through a primary function to obtain a cluster power evolution function corresponding to each cluster, and smoothing the cluster power by combining the cluster power evolution function with sampling time in the current extinguishing interval in the extinguishing process of each cluster; And constructing the channel impulse response based on the channel parameters, the cluster extinction probability and the cluster power after the smoothing. In one embodiment, the channel impulse response is: ; Wherein, the L is the cluster number under the current snapshot; Is the central delay of the first cluster; K is the number of multipaths; is the time delay of the kth path in the ith cluster; complex amplitude for the kth path in the kth cluster; A random phase for the kth path in the kth cluster; The cluster center arrival angle of the first cluster; The arrival angle of the kth path in the kth cluster; is a motion velocity direction vector; A direction vector of a kth path in a kth cluster; A position vector for the nth antenna; Is the signal wavelength; The probability of being on/off for the first cluster. In one embodiment, the intra-cluster parameters include intra-cluster multipath number, intra-cluster delay offset, intra-cluster angle offset, intra-cluster K factor; the inter-cluster parameters comprise cluster time delay, cluster power and cluster angle. In one embodiment, the process of generating and extinguishing each cluster is characterized by a Markov process to obtain the probability of generating and extinguishing clusters, which comprises: in the Markov process, the existing state of the cluster is regarded as a1 state, the disappearing state of the cluster is regarded as a 0 state, and the disappearing state of each cluster is judged based on the cluster power; And respectively calculating a transition matrix and a steady-state matrix of the on-off state of each cluster to obtain the cluster on-off probability. In one embodiment, the transfer matrix is calculated according to the following formula: ; Wherein, the Is a transfer matrix; for the probability of transitioning from the a-state to the b-state, a e {0,1}, b e {0,1}; The steady state matrix is calculated according to the following formula: ; Wherein, the Is a steady state matrix; is the steady state probability of the o-state, 。 In one embodiment, obtaining c