CN112019251A - Communication method, communication device, and storage medium - Google Patents

Communication method, communication device, and storage medium Download PDF

Info

Publication number
CN112019251A
CN112019251A CN202010706632.5A CN202010706632A CN112019251A CN 112019251 A CN112019251 A CN 112019251A CN 202010706632 A CN202010706632 A CN 202010706632A CN 112019251 A CN112019251 A CN 112019251A
Authority
CN
China
Prior art keywords
channel
information
communication method
communication
tracking
Prior art date
Legal status (The legal status 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 status listed.)
Pending
Application number
CN202010706632.5A
Other languages
Chinese (zh)
Inventor
高跃
于佳冬
刘小兰
沈学民
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Yueyueyuebaohe Technology Co Ltd
Original Assignee
Shenzhen Yueyueyuebaohe Technology Co Ltd
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 Shenzhen Yueyueyuebaohe Technology Co Ltd filed Critical Shenzhen Yueyueyuebaohe Technology Co Ltd
Priority to CN202010706632.5A priority Critical patent/CN112019251A/en
Publication of CN112019251A publication Critical patent/CN112019251A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/14—Relay systems
    • H04B7/15—Active relay systems
    • H04B7/185—Space-based or airborne stations; Stations for satellite systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Astronomy & Astrophysics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Radio Relay Systems (AREA)

Abstract

The invention relates to a communication method, a communication device and a storage medium, wherein the method comprises the following steps: in a communication network, a signal is channel tracked. Further, the channel tracking is channel tracking in a three-dimensional space. Further, the method includes recursively tracking the dynamic channel according to a three-dimensional space two-dimensional markov model based on a three-dimensional dynamic Turbo approximation message passing algorithm. The invention can effectively improve the accuracy of channel estimation.

Description

Communication method, communication device, and storage medium
Technical Field
The present invention relates to the field of communications technologies, and in particular, to a communication method, a communication apparatus, and a storage medium.
Background
The statements herein merely provide background information related to the present disclosure and may not necessarily constitute prior art.
"space-air-ground" integrated network (also called "space-air-ground" integrated network, abbreviated as "SAGIN") attracts much attention because it can provide seamless wide area connection for 5G and above communications, high throughput and strong reduction capability. SAGIN multidimensional networks mainly comprise three parts: the space portion of the satellite network, the air portion of the aviation network, and the ground portion of the ground network. Since the "satellite-ground" link is very sensitive to ground interference, it is necessary to utilize unmanned aerial vehicles (abbreviated as UAVs) as relay stations to improve the communication capability of the coverage area.
Compared with the traditional ground or satellite network, the auxiliary communication of the unmanned aerial vehicle is used as an important component of SAGIN, is not only suitable for the condition that ground facilities cannot be accessed, but also is suitable for emergency communication in crowded or disaster areas. In addition, as the unmanned aerial vehicle is deployed in the middle position of the ground space communication, the performance of a short-range line of sight (LoS for short) link can be remarkably improved in the case of a long-distance link.
For the drone-satellite link, in order to alleviate the spectrum congestion problem, the millimeter wave (mmWave for short) Ka band (26.5-36GHz) provides an extra frequency range and has a higher data rate. Although the Ka band has a high weather dependency due to its high path loss, an antenna array with a large number of antenna elements may provide a higher transmission gain to compensate for propagation loss.
In the above scenario, due to satellite orbits and drone three-dimensional trajectories, accurate channel estimation with small pilot overhead is performed for dynamic wireless channels of drone satellite links, such as near earth orbit satellites (also called low orbit satellites, abbreviated LEO satellites) and medium earth orbit satellites (also called medium orbit satellites, abbreviated MEO satellites) and air-to-ground links, which is a serious challenge.
Disclosure of Invention
The invention aims to provide a novel communication method, a communication device and a storage medium.
The purpose of the invention is realized by adopting the following technical scheme. The communication method provided by the invention comprises the following steps: in a communication network, a signal is channel tracked.
The object of the invention can be further achieved by the following technical measures.
In the foregoing communication method, the receiving end and/or the sending end of the signal move in a three-dimensional space, and have a horizontal displacement and a vertical displacement, and performing channel tracking on the signal includes: and carrying out three-dimensional space channel tracking on the signals.
In the foregoing communication method, the performing channel tracking on the signal includes: determining channel information from information of the received signal, and a priori information
In the foregoing communication method, the performing channel tracking on the signal includes: the channel tracking is performed according to a three-dimensional space two-dimensional markov model.
In the foregoing communication method, the performing channel tracking on the signal includes: representing the channel information by using a channel gain and an antenna response matrix, representing the channel gain by using angle information of the channel gain and magnitude information of the channel gain, and representing the angle information of the channel gain by using azimuth information of the channel gain and elevation information of the channel gain.
In the foregoing communication method, the performing channel tracking on the signal includes: from the received signal y during the period t(t)Estimating a dynamic channel vector gtElement g in (1)t，mIs expected to
Figure BDA0002595005450000021
Wherein the expectation is based on a probability p (g)t，m|y(t)) Obtained, the probability p (g)t，m|y(t)) Obtained according to the following formula
Figure BDA0002595005450000022
Wherein the content of the first and second substances,
Figure BDA0002595005450000023
represents a vector g(t)In the middle withoutWith the element gt，mVector of (a), b(t)The angle information representing the channel gain is determined,
Figure BDA0002595005450000024
amplitude information representing the channel gain; to track the dynamic channel vector g at time tt。
In the foregoing communication method, the performing channel tracking on the signal includes: the dynamic channel is recursively tracked based on a three-dimensional dynamic Turbo approximate message passing algorithm.
The communication method comprises the steps that the communication network comprises a space-air-ground integrated network; the transmitting end of the signal comprises one of a satellite and an unmanned aerial vehicle, and the receiving end of the signal comprises the other one of the satellite and the unmanned aerial vehicle.
The object of the present invention is also achieved by the following technical means. According to the invention, a communication device is provided for implementing any one of the possible communication methods.
The object of the present invention is also achieved by the following technical means. A computer storage medium according to the present invention includes computer instructions that, when executed on a device, cause the device to perform any one of the possible communication methods described above.
Compared with the prior art, the invention has obvious advantages and beneficial effects. By the technical scheme, the communication method, the communication device and the storage medium provided by the invention at least have the following advantages and beneficial effects:
(1) the invention proposes a new communication method based on a new dynamic channel model, called three-dimensional space two-dimensional markov model (3D-2D-MM for short). The model captures the structural sparsity and probabilistic temporal correlations of the azimuth and elevation spatial domains. The model establishes a deeper probability relationship of the hidden value vector and the joint hidden support vector. The invention proposes to use the fading channel as a dynamic probability model. The method is particularly suitable for unmanned aerial vehicle-satellite communication. By utilizing the method provided by the invention, a more real three-dimensional channel environment is considered, and the accuracy of channel estimation can be effectively improved.
(2) In order to model the satellite communication channel of the unmanned aerial vehicle in reality, a dynamic space domain transfer probability is designed to replace a complex continuous relative displacement expression between a three-dimensional unmanned aerial vehicle track and an orbit satellite.
(3) The invention provides a new communication method, which is based on a new three-dimensional dynamic Turbo approximate message passing (3D-DTAMP) algorithm, and utilizes newly added 3D-2D-MM prior information to recursively track a dynamic channel. The specific algorithm recursively tracks the three-dimensional dynamic channel according to the relation supported by the azimuth angle and the elevation angle, and the accuracy of channel estimation is effectively improved by exploring the structure of the dynamic channel. Thus the method has lower pilot overhead, equivalent complexity and better recovery performance.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical means of the present invention more clearly understood, the present invention may be implemented in accordance with the content of the description, and in order to make the above and other objects, features, and advantages of the present invention more clearly understandable, the following preferred embodiments are described in detail with reference to the accompanying drawings.
Drawings
FIG. 1a is a schematic diagram of a space-air-ground integrated network provided by an embodiment of the invention;
FIG. 1b is a schematic diagram of a UAV satellite system model provided by one embodiment of the present invention;
FIG. 2 is a factor graph of a 3D-2D-MM for a channel provided by one embodiment of the invention;
FIG. 3a is a schematic diagram of a hidden support matrix provided by an embodiment of the invention;
FIG. 3b is a factor graph of 2D-MM for the azimuth support vector provided by one embodiment of the invention;
FIG. 3c is a factor graph of a 3D-2D-MM joint latent support vector provided by an embodiment of the invention;
FIG. 3d is a diagram illustrating specific factors provided by an embodiment of the present invention;
FIG. 4a is a schematic diagram of a single display implicit value vector according to an embodiment of the present invention;
FIG. 4b is a factor graph of a Gaussian Markov model for hidden value vectors provided by one embodiment of the present invention;
FIG. 5 is an overall factor graph of a 3D-2D-MM provided by an embodiment of the invention;
FIG. 6 is a schematic diagram of the delivery of messages over a sequential path of time provided by one embodiment of the present invention;
FIG. 7 is a diagram illustrating the relationship between the operation time and the number of antennas between the method provided by one embodiment of the present invention and the conventional method;
FIG. 8 is a graphical illustration of TNMSE versus total time for a method provided by one embodiment of the present invention versus a prior art method;
FIG. 9 is a schematic diagram of the TNMSE versus signal-to-noise ratio of a method provided by one embodiment of the present invention and a prior art method;
FIG. 10 is a diagram illustrating the TNMSE versus pilot number of the method provided by one embodiment of the present invention and the existing method;
FIG. 11 is a message on module provided by one embodiment of the present invention
Figure BDA0002595005450000041
Schematic diagram of the transfer of sequential paths.
Detailed Description
To further explain the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given of the embodiments, structures, features and effects of the communication method, the communication device and the storage medium according to the present invention with reference to the accompanying drawings and the preferred embodiments.
Brief introduction to three-dimensional channel tracking for drone-satellite communications in a "space-air-ground" integrated network.
For conventional channel estimation methods, such as Least Squares (LS) and Minimum Mean Square Error (MMSE), the number of training pilots must be at least the number of antenna elements, which is impractical for large antenna array systems, such as millimeter-wave large MIMO systems with thousands of antenna elements. A great deal of research is to exploit the limited scatterers in the propagation environment, and different compressed sensing (CS for short) algorithms have been developed to reduce the training overhead by exploring the static and sparse properties of the channel, such as matching pursuit (MP for short) and orthogonal matching pursuit (OMP for short).
Besides using traditional compressed sensing algorithms to solve the channel estimation problem, several structured sparse channel estimation algorithms have been developed by learning spatial correlation. The spatial correlation is a characteristic in which non-zero elements in the angular domain channel are concentrated in multiple bursts.
Although the channel estimation or tracking method can be directly applied to the air-to-ground link in consideration of the characteristics of the ground environment. However, as another part of SAGIN, drone satellite communication has the following characteristics: high non-stationary channel conditions and dominant line-of-sight rays. Due to the finite nature of scatterers in the terrestrial propagation environment, the angular domain sparsity of the channel. This behavior is similar to drone satellite communications with one concentrated ray.
However, in practice, it is very difficult to acquire channel information in real time due to satellite orbits and three-dimensional drone trajectory motion, which presents a huge challenge to modeling and tracking of dynamic channels.
In order to obtain more real satellite dynamic channel information of the unmanned aerial vehicle, the invention provides a method for taking a fading channel (fading channel) as a dynamic probability model, researching the space correlation of a dynamic angular domain and researching a three-dimensional dynamic angular domain sparse structure along with time.
The invention provides an unmanned aerial vehicle-satellite communication system of an unmanned aerial vehicle adopting a Uniform Plane Array (UPA), and provides a method for recursively tracking a three-dimensional dynamic channel, wherein the method has the advantages of low pilot frequency overhead, considerable complexity and better recovery performance.
The main contributions are summarized below:
(1) the invention provides a three-dimensional channel statistical model suitable for unmanned aerial vehicle-satellite communication, which is called a three-dimensional space two-dimensional Markov model (3D-2D-MM for short). The 3D-2D-MM captures structural sparsity and probabilistic temporal correlation in the azimuth and elevation spatial domains.
(2) In order to model the satellite communication channel of the unmanned aerial vehicle in reality, a dynamic space domain transfer probability is designed to replace a complex continuous relative displacement expression between a three-dimensional unmanned aerial vehicle track and an orbit satellite.
(3) The invention provides a new three-dimensional dynamic Turbo approximate message passing (3D dynamic Turbo approximate message passing, also called 3D dynamic enhanced approximate message passing, 3D-DTAMP for short) algorithm, which recursively tracks a three-dimensional dynamic channel according to the relation supported by an azimuth angle and an elevation angle.
(4) The 3D-DTAMP algorithm provided by the invention obtains better reconstruction performance with lower pilot frequency overhead and equivalent complexity.
The following common symbols are used in this specification. For plural use
Figure BDA0002595005450000051
To indicate. Transpose and conjugate transpose are denoted (·)TAnd (·)H. I is the identity matrix.
Figure BDA0002595005450000052
(mean, covariance) represents the complex gaussian random vector with defined mean and covariance. Is l | | · | |2And (4) norm.
Second, system model
Figure 1 shows SAGIN and the proposed drone satellite model of the present invention. Wherein, fig. 1(a) is a schematic diagram of a space-air-ground integrated network. Fig. 1(b) is a schematic diagram of a drone satellite system model.
Table 1 summarizes the main symbols, as follows.
TABLE 1 Main symbol table
Figure BDA0002595005450000053
Figure BDA0002595005450000061
Figure BDA0002595005450000071
The invention proposes a channel model for orthogonal frequency division multiplexing (OFDM for short) transmission via the downlink of a drone satellite communication. Optionally, the channel model is a narrowband channel model.
Array antennas have been widely used in commercial satellites to improve spectral efficiency by generating directional beams. The model proposed by the invention does not assume: low earth orbit satellites are equipped with uniform planar arrays (referred to as UPA for short) M ═ Nx×NyAn antenna.
In addition, a single omnidirectional antenna (single omnidirectional antenna) is arranged on the unmanned aerial vehicle.
For the training process, symbols
Figure BDA0002595005450000072
Is transmitted. Alternatively, S may be considered a signal transmitted by a satellite.
Thus, the signal received by the drone may be denoted as
Figure BDA0002595005450000073
And is
Figure BDA0002595005450000074
Is additive white gaussian noise.
One purpose of the method proposed by the invention is to solve the aforementioned channel HtI.e. H in equation (1)tIs to be solved.
For a Ka-band fixed satellite communication system, the line-of-sight ray is the dominant path. Thus HtThe matrix representing the frequency domain channel with one line-of-sight path is given by the following equation
Ht＝αtAH(θt，φt) Equation (2).
Wherein alpha istIs the channel gain, optionally it is the complex channel gain (complex channel gain);
Figure BDA0002595005450000075
represents transmit array response vectors (transmit array response vectors), which are also antenna response matrices, where θtAnd phitThe azimuth and elevation directions as the departure angle AoD at time t. Alternatively, the azimuth angle represents the case of the horizontal direction, and the elevation angle represents the case of the vertical direction.
With a uniform planar array UPA in mind, a steering vector (steering vector) can be written as
Figure BDA0002595005450000076
Where λ is the signal wavelength and the size of the antenna array is Nx×NyAnd 0. ltoreq.nx＜NxAnd 0. ltoreq. ny＜NyIs the index of the antenna element.
The invention considers the case of a discretized departure angle pair (AoD pair), namely thetat∈{0，θrange/Nx，...，θrange(Nx-1)/NxAnd phit∈{0，φrange/Ny，...，φrange(Ny-1)/NyThis means that the starting angle pair is from Nx×NySelected from a uniform grid.
Note that because the drone is located a significant distance from the satellite, the range of motion in both azimuth and elevation is limited. In addition, the present invention contemplates discretized departure angles AoDs.
By arranging a large number of antenna units on the low-orbit satellite, the spatial resolution of the angle base is improved. Therefore, the channel is sparsely processed, and the channel model can be used as a virtual model, which is not called a virtual channel model, and then processed by a compressed sensing technology. The angular domain channel in the virtual channel model is represented as
Figure BDA0002595005450000081
Wherein the content of the first and second substances,
Figure BDA0002595005450000082
is a sparse matrix containing one non-zero element (which means there is only one main path),
Figure BDA0002595005450000083
is an array response vector dictionary matrix with a spatial grid M.
In some examples, g in equation (4)tAnd alpha in the above formula (2)tCorrespondingly, also used to represent channel gain; in equation (4)
Figure BDA0002595005450000084
And A (θ) in equation (2)t，φt) Correspondingly, it is also used to represent the antenna response matrix.
Note that in some examples, the antenna response matrix in equation (2) or equation (4) is known and it is the channel gain that is to be solved for. For example, g in equation (4)tIs to be solved. Therefore, the channel gain g is not settReferred to as channel vectors.
Channel model and problem equation
In this section, a 3D-2D-MM channel model is presented and the problem equations for dynamic channel tracking are described.
3.1, 3D-2D-MM channel model
The dynamic angular domain channel is modeled as a probabilistic signal model. In particular, the channel gain g(T)＝{g1，...，gTIt can be modeled as a probabilistic signal model with two hidden stochastic processes, b(T)＝{b1，...，bTAnd
Figure BDA0002595005450000085
without changing the channel gain g(T)Also known as channel vectors.
Wherein the content of the first and second substances,
Figure BDA0002595005450000091
a joint hidden support vector (joint hidden support vector) representing time t, where b represents the channel sparsityt，mE {1, 0} is equal to 0 or 1.
Note that m is m ═ n (n)x-1)×Nx+ny。
Figure BDA0002595005450000092
Is a hidden value vector (also called hidden value vector) used to represent the time correlation of the channel coefficients.
The dynamic channel element in the mth angle of departure (AoD) direction at time t can be written as
Figure BDA0002595005450000093
Wherein, bt，mIndicates whether there is an active path (i.e., b)t，m1 means that the mth departure angle path is activated),
Figure BDA0002595005450000094
a complex path gain (complex path gain) is indicated.
In some alternative embodiments, the channel gain comprises two parts: some are the aforementioned joint implicit support vectors, which represent the angular splitting of the channel gain (or, the splitting of the channel gain in angular space); some are the aforementioned implicit value vectors, which represent a split of the magnitude of the channel gain. And multiplying the splitting of the angle of the channel gain and the splitting of the amplitude of the channel gain to obtain the channel gain.
To illustrate a more realistic dynamic channel (i.e., the inactive AoD path at the current time slot t may be activated at time slot t + 1), a probabilistic channel model with a prior distribution of channels may be built, as follows
Figure BDA0002595005450000095
Wherein the channel vector condition is a priori
Figure BDA0002595005450000096
Where (·) is a dirac delta function. The factor graph of the combined channel is shown in FIG. 2, where πt，mIs a conditional prior
Figure BDA0002595005450000101
Optionally, each time slot represents a separate estimated transmission frame.
The joint implicit support vector and implicit value vector for the azimuth and elevation directions are described as follows:
FIG. 2 is a graph showing the relationship between N andx8 and NyFactor graph of 3D-2D-MM for the channel at 8. Joint implicit support vector btVector of sum hidden values
Figure BDA0002595005450000102
The detailed models of (a) are shown in fig. 3 (fig. 3a, 3b, 3c) and fig. 4 (fig. 4a, 4b), respectively. It should be noted that the factor graphs such as fig. 2, fig. 3, and fig. 4 may represent specific operation processes, or each of the factor graphs represents a specific operation corresponding to a specific operation.
3.1.1 Joint latent support vector for azimuth and elevation
Fig. 3a is a latent support matrix diagram. FIG. 3b shows the case when T is 2 and NxThe factor graph of 2D-MM for the azimuth support vector at 8. FIG. 3c shows that when T is equal to 1, Nx＝8，Ny3D-2D-MM joint implicit support vector b when equal to 8tFactor graph of (c). Note that M ═ N according to the foregoing descriptionx×NyYielding M-64 in this example. FIG. 3d is a factorExamples are given.
Azimuthal support a(T)And elevation support e(T)The relationship of (c) can be found in fig. 3 a. Where only one active path is considered. I.e. the one marked with grey in figure 3 a. Optionally, the active path corresponds to an active channel.
Note that in FIG. 3a, N may also be takenx＝8，NyThis is the fact that 8 is divided into 8 parts for azimuth and 8 parts for elevation. Taking the azimuth angle as an example, 360 degrees in total, the azimuth angle is divided into 8 parts by 45 degrees. In fact, Nx、NyIs similar to resolution, and for greater accuracy, N may be setx、NyThe value of (a) is set sufficiently large.
In addition, as a specific example, fig. 3d shows an exemplary hidden support element b in the hidden support vector1，30And an exemplary one of the overall factor structures forming the hidden support element comprises a1，4And e1，6Schematic representation of the factor (c). Wherein, a1，4Is an element in the azimuth support vector, e1，6Is an element in the elevation support vector, b1，30、a1，4、e1，6The foregoing relationship is satisfied between the subscripts (also referred to as indexes) of the three: m ═ nx-1)×Nx+ny(4-1) × 8+6 ═ 30. Note that in fact, as in fig. 3c, the overall factor structure forming the implicit support element will typically include multiple factors, with fig. 3d showing only one factor.
In the model proposed by the present invention, the known low orbit satellite orbits and uncertain drone movement are considered to be relative angular changes, not a certain position. Specifically, due to the three-dimensional trajectory of the drone, coupled with the fixed orbit navigation of the low orbit satellite, the relative displacement of the drone and the low orbit satellite is continuous. This means that for the time-varying case, the azimuth and elevation angle will only vary with some probability to adjacent degrees (degree) in the angular domain. That is, as shown in the right diagram of fig. 3a, in the next time slot, the active channels represented by the gray grids will change to the adjacent grids with a certain probability, such as moving to the left and right grids in the right diagram of fig. 3a due to the change of azimuth angle, or moving to the upper and lower grids in the right diagram of fig. 3a due to the change of elevation angle. This probability can be seen as an expression of the movement of both the drone and the satellite. In practice, based on the expectation-maximization algorithm, the parameters may be pre-trained and stored for real-time channel tracking.
In other words, the spatial sparse pattern will be highly correlated with the a priori neighboring patterns. For example,
Figure BDA0002595005450000111
with a higher probability of relying on neighbors in the last slot
Figure BDA0002595005450000112
This also verifies that the support changes slowly over time, indicating that
Figure BDA0002595005450000113
With a higher probability of relying on
Figure BDA0002595005450000114
Spatial sparsity, the probabilistic relationship passing over time, is referred to as spatial correlation and temporal correlation, respectively. The reason for naming the azimuthal support vector model as 2D-MM is that it combines both spatial and temporal correlation.
In particular, support in the azimuthal direction a(T)The time-lapse situation can be obtained using the following equation:
Figure BDA0002595005450000115
Figure BDA0002595005450000121
wherein the content of the first and second substances,
Figure BDA0002595005450000122
and
Figure BDA0002595005450000123
is the transition probability, which is equal to the factor node k in fig. 3b and 3c for the azimuth direction1，1、
Figure BDA0002595005450000124
kt，1And
Figure BDA0002595005450000125
note that azimuth support
Figure BDA0002595005450000126
Namely, the value is 0 or 1, which respectively represents the activation condition of non-activation and activation.
The factor graph of FIG. 3b shows a graphical representation of the 2D-MM. When the t is equal to 1, the first step is carried out,
Figure BDA0002595005450000127
adjacent to the activation situation of
Figure BDA0002595005450000128
And
Figure BDA0002595005450000129
correlation, transition probabilities are respectively
Figure BDA00025950054500001210
And
Figure BDA00025950054500001211
in an alternative example, when t is 2,
Figure BDA00025950054500001212
adjacent to the activation situation of
Figure BDA00025950054500001213
And
Figure BDA00025950054500001214
associated with, and of itself of the preceding time slot
Figure BDA00025950054500001215
And (4) correlating. In another alternative, when t is 2,
Figure BDA00025950054500001216
active case of and previous slot itself
Figure BDA00025950054500001217
Related and adjacent to the previous time slot
Figure BDA00025950054500001218
And
Figure BDA00025950054500001219
and (4) correlating.
Similarly, support e for elevation direction of departure angle AoD(T)It is also modeled in a similar manner, using an equation similar to equation (8) to derive its propagation over time. Only the azimuth angle in equation (8) needs to be replaced by the elevation angle.
Thus, the total departure angle AoD supported 3D-2D-MM is shown in fig. 3c, which includes a combination of support in both azimuth and elevation directions. Note that the implicit support vector btEach hidden support element b int，mThe method is characterized in that: n thxAn azimuth support vector and an nthyThe combination of the elevation support vectors. The joint conditional prior of the channel support probability is given by
Figure BDA00025950054500001220
3.1.2 Gaussian Markov model of hidden value vector
FIG. 4a is a diagram showing a hidden value vector at a single time. Fig. 4b is a factor graph of a gaussian markov model for hidden value vectors, where T is 2 and M is 64.
Hidden value vectors are shown in fig. 4a and 4b, respectively
Figure BDA00025950054500001221
And a factor graph of a gaussian markov model.
Based on the path gain varying slowly over time, the hidden value vector can be represented as a Gaussian Markov process
Figure BDA0002595005450000131
Wherein, beta is ∈ [0, 1 ]]，
Figure BDA0002595005450000132
Figure BDA0002595005450000133
Is the average value of the process.
When beta is 0, then
Figure BDA0002595005450000134
This means
Figure BDA0002595005450000135
And remain unchanged over time.
When β is 1, then
Figure BDA0002595005450000136
This means that over time the hidden value vector represents a gaussian distribution with an independent iso-distribution (i.i.d.) of the mean value μ.
If 0 < β < 1, the conditional probability can be recorded as
Figure BDA0002595005450000137
The steady state profile of the process is
Figure BDA0002595005450000138
And the number of the first and second electrodes,
Figure BDA0002595005450000139
can be expressed as
Figure BDA00025950054500001310
Wherein the content of the first and second substances,
Figure BDA00025950054500001311
and
Figure BDA00025950054500001312
is the transition probability, which is equal to the factor node f in FIG. 4b1，mAnd ft，m。
3.2, dynamic sparse channel tracking problem equation
By considering the angular domain channel, the received signal can be written as
Figure BDA00025950054500001313
Wherein the content of the first and second substances,
Figure BDA00025950054500001314
the received signal can then be rewritten to a standard compressed sensing model (CS model for short)
yt＝Φgt+ntIn the formula (14),
wherein the content of the first and second substances,
Figure BDA0002595005450000141
and a perception matrix
Figure BDA0002595005450000142
It is noted that the transmitted pilot is defined as
Figure BDA0002595005450000143
With randomly selected and reordered N of M identity matricesSAnd (6) rows. In this way, Φ can be treated as an approximate partial orthogonal sensing matrix and provide good performance.
Given a received signal y over a period t(t)The purpose is to track the dynamic channel vector g at time tt. The channel vector can be estimated as an expectation
Figure BDA0002595005450000144
Wherein the expectation is an expectation on a marginal posteriori (margin posterior) as described below
Figure BDA0002595005450000145
Wherein the content of the first and second substances,
Figure BDA0002595005450000146
indicates that there is no element gt，mVector g of(t). Equation (15) is a dynamic sparse channel tracking problem equation. Probability p (g) according to equation (15)t，m|y(t)) A dynamic channel vector g can be obtainedt＝[gt，1，...，gt，M]Expectation of the element (1)
Figure BDA0002595005450000147
In one example, upon receiving signal y(t)Then, the channel vector g can be obtained based on the probability expressed by the above equation (15) or the above expectationtThe expectation is that.
The goal of one embodiment of the method shown in the present invention is to obtain the Minimum Mean Square Error (MMSE) of the estimated g.
Fourthly, the three-dimensional dynamic Turbo approximate message transmission algorithm provided by the invention
FIG. 5 is a factorial representation of the 3D-2D-MM proposed by the present invention. In particular, FIG. 5 shows a 3D-2D-MM total distribution
Figure BDA0002595005450000148
Factor graph of (c). The factor graph contains two subgraphs: joint concealment of time TTibetan support b(T)(on the left side of FIG. 5) and hidden value vector
Figure BDA0002595005450000149
(on the right side of fig. 5).
Total implicit support b(T)Can be further developed into a 3D-2D-MM factor graph comprising an azimuthal support vector a at time T(T)And elevation support vector e(T)Both of which are described below.
The expressions of the function nodes are shown in table 2. Table 2 is a table of factors, distributions and functional forms associated with 3D-2D-MM.
TABLE 2
Figure BDA0002595005450000161
There are two reasons for utilizing the AMP framework (approximate messaging framework) in the model proposed by the present invention: rigorous analysis and extremely fast run time. Conventional AMPs use a static prior. However, for the specially designed factor graph proposed by the present invention, the a priori distributed knowledge of the AMP algorithm will change over time. Since the full-factorial graph structure of 3D-2D-MM is complicated, it is difficult to accurately obtain the edge posteriori in equation (15). Conventional turbo compressed sensing is an iterative algorithm that follows a two-step iterative framework: a Linear Minimum Mean Square Error (LMMSE) estimator, and a Minimum Mean Square Error (MMSE) estimator based on independent equal distribution apriori (i.i.d. prior-based).
In this section, in order to make better use of the dynamic channel structure, details of the algorithm proposed by the present invention based on the turbo iteration framework with the MMSE denoiser of the 3D-2D-MM proposed by the present invention will be described. The 3D-DTAMP algorithm proposed by the present invention is introduced in three parts: module
Figure BDA0002595005450000171
LMMSE estimator module
Figure BDA0002595005450000172
Message passing MMSE denoiser, and message passing over time. Note that what is passed in the message passing process is a distribution, and the received distribution is multiplied by the transition probability and then sent out.
4.1, Module
Figure BDA0002595005450000173
LMMSE estimator
On the module
Figure BDA0002595005450000174
Middle, angular channel vector gtFrom the received signal ytAn estimation is made with a prior distribution of
Figure BDA0002595005450000175
Wherein the external (extrinsic) mean and variance are respectively
Figure BDA0002595005450000176
And
Figure BDA0002595005450000177
gthas an external distribution (external distribution) of
Figure BDA0002595005450000178
Wherein the external mean and variance are respectively
Figure BDA0002595005450000179
Figure BDA0002595005450000181
Figure BDA0002595005450000182
4.2, Module
Figure BDA0002595005450000183
Messaging MMSE denoiser
Figure BDA0002595005450000184
Figure BDA0002595005450000191
On the module
Figure BDA0002595005450000192
In the method, the minimum mean square error MMSE de-noising device for message transmission is realized by utilizing the space sparsity detail of the 3D-2D-MM provided by the invention.
As a core structure of turbo compressed sensing, external mean and variance are sent from a module
Figure BDA0002595005450000193
LMMSE estimator in (1). That is to say that the first and second electrodes,
Figure BDA0002595005450000194
and
Figure BDA0002595005450000195
for message passing algorithms, the basic assumption is that
Figure BDA0002595005450000196
Wherein the content of the first and second substances,
Figure BDA0002595005450000197
independent of gt. The purpose is to calculate
Figure BDA0002595005450000198
An approximate posterior distribution ofThe formula (15) is substituted. A sum-product messaging rule (sum-product messaging rule) is used that follows the messaging structure shown in fig. 5. See the appendix for details. The process is summarized as follows:
first, the hidden value vector in Gaussian is given by equation (38)
Figure BDA00025950054500001910
Message delivery of (2), where mu and sigma2For updating messages
Figure BDA0002595005450000199
Is input.
Next, according to equations (39) to (40), the message passes through the path
Figure BDA0002595005450000201
Figure BDA0002595005450000202
To transmit, wherein the input is
Figure BDA0002595005450000203
And
Figure BDA0002595005450000204
from a module
Figure BDA0002595005450000205
Forward-backward (forward-backward) messaging is then performed on the 3D-2D-MM in three sequential paths: azimuthal support at2D-MM of
Figure BDA0002595005450000206
And elevation support et2D-MM of (1). The details of the 3D-2D-MM azimuth and elevation support non-zero probability estimation are summarized in Algorithm 1.
According to equations (60) to (63), finally, the message passes through the path
Figure BDA0002595005450000207
Figure BDA0002595005450000208
And returning.
Computation-based update messages
Figure BDA0002595005450000209
The posterior distribution can be written as
Figure BDA00025950054500002010
Wherein
Figure BDA00025950054500002011
Then, gtThe posterior mean and variance of each element of (1) can be updated to
Figure BDA00025950054500002012
And
Figure BDA00025950054500002013
wherein the content of the first and second substances,
Figure BDA00025950054500002014
denotes gt,mAbout
Figure BDA00025950054500002015
The conditional variance of (c).
Thus, similar to equations (19) and (20), the external update mean and covariance can be written as
Figure BDA00025950054500002016
And
Figure BDA0002595005450000211
4.3 message delivery over time
For channel tracking, better recovery can be achieved with time correlation based on the a priori information provided. Thus, the present invention also contemplates message delivery over time. There are three main paths for the transfer over time: gaussian hidden value vector
Figure BDA0002595005450000216
Path, azimuth support oftPath of (e) and elevation support etThe path of (2). The sequence paths (sequence paths) can be seen in FIG. 6.
Fig. 6 is a graphical representation of sequential path delivery of messages over time, illustrating the details of equations (27) through (36).
4.3.1 Gaussian hidden value vector
Figure BDA0002595005450000218
Delivery over time
Slave factor node pit,mTo variable nodes
Figure BDA0002595005450000215
Should be the message of
Figure BDA0002595005450000212
It should be noted that if b t,m0 in equation (5), gt,mShould be 0, this results in
Figure BDA0002595005450000217
Not observable. To solve this problem, optionally a threshold value slightly less than 1 is introduced. Thus, the adjusted message is delivered as
Figure BDA0002595005450000213
Wherein the content of the first and second substances,
Figure BDA0002595005450000214
equation (28) and equation (29).
Where e is a small positive value close to 0 and Th is the aforementioned threshold value slightly less than 1.
Then from the factor node ft+1,mTo variable nodes
Figure BDA0002595005450000229
Is a message of
Figure BDA0002595005450000221
Wherein the content of the first and second substances,
Figure BDA0002595005450000222
and is
Figure BDA0002595005450000223
4.3.2 azimuthal support atDelivery over time
Azimuth support prior information passed to next time instant
Figure BDA0002595005450000224
Can be
Figure BDA0002595005450000225
Wherein the content of the first and second substances,
Figure BDA0002595005450000226
4.3.3、elevation support etDelivery over time
Similarly, the elevation support prior information passed to the next time instant may be
Figure BDA0002595005450000227
Wherein the content of the first and second substances,
Figure BDA0002595005450000228
Figure BDA0002595005450000231
finally, algorithm 2 summarizes the 3D-DTAMP proposed by the present invention.
Figure BDA0002595005450000232
Five, numerical results
In this section, the complexity and recovery performance of the algorithm proposed by the present invention is evaluated.
For unmanned aerial vehicle satellite communication systems, the sparsity ratio in azimuth and elevation directions is
Figure BDA0002595005450000233
This indicates that there is only one dominant LoS ray.
The parameters are set as follows:
azimuthal spatial correlation parameter
Figure BDA0002595005450000241
Azimuthal time dependence parameter
Figure BDA0002595005450000242
Azimuth angle related parameter
Figure BDA0002595005450000243
Figure BDA0002595005450000244
Elevation spatial correlation parameter
Figure BDA0002595005450000245
Figure BDA0002595005450000246
Elevation time dependent parameter
Figure BDA0002595005450000247
Elevation angle related parameter
Figure BDA0002595005450000248
Implicit value vector parameters
Figure BDA0002595005450000249
Th＝1-10-2，
Figure BDA00025950054500002410
And e is 10-7。
Sparse bayesian learning (abbreviated as SBL) and differential OMP (abbreviated as D-OMP) can be used as the reference algorithm (benchmark algorithms) for the three-dimensional channel model of the present invention.
5.1 complexity analysis
Fig. 7 is a graph of the operating time of various algorithms versus the number of antennas. Wherein, SNR is 15dB, N s10, T50, M16, 64, 256, and Nx＝Ny＝4，8，16。
Module
Figure BDA00025950054500002411
The complexity of the medium LMMSE estimator is
Figure BDA00025950054500002412
This is mainly determined by multiplication of the matrix. Module
Figure BDA00025950054500002413
The complexity of the MMSE denoiser in (1) is a simple sum-product message passing rule, whereby the complexity is
Figure BDA00025950054500002414
To sum up, for each iteration, the total complexity over time T of the algorithm proposed by the present invention is
Figure BDA00025950054500002415
As shown in fig. 7, SBL requires longer operation time, especially for larger antenna elements. This is caused by the high computational complexity required for matrix multiplication and inversion. Compared with the SBL algorithm, the simulation time of the algorithm and the D-OMP algorithm provided by the invention is obviously reduced.
Note that the complexity of the D-OMP is
Figure BDA00025950054500002416
By increasing the number of antennas, the arithmetic time of the algorithm provided by the invention is equivalent to the arithmetic time.
5.2 simulation results
FIG. 8 shows the SNR at 15dB, NsTNMSE is a function of total time at 22.
In fig. 8 (a), M is 64, and N isx＝NyIn fig. 8 (b), M is 256 and N is Nx＝Ny＝16。
FIG. 9 is at NsWhen T is 50, TNMSE is related to the signal-to-noise ratio (SNR). In fig. 9 (a), M is 64, and N isx＝NyIn fig. 9 (b), M is 256 and N is 8x＝Ny＝16。
Fig. 10 shows the relationship between TNMSE and the total number of pilots when SNR is 15dB and T is 50. Where M is 64 and N in fig. 10 (a)x＝NyIn fig. 9 (b), M is 256 and N is 8x＝Ny＝16。
To evaluate the recovery performance of these algorithms, a performance metric called time-averaged normalized mean square error (TNMSE) is proposed, as follows
Figure BDA0002595005450000251
Wherein the content of the first and second substances,
Figure BDA0002595005450000252
is gtThe result of the estimation at time t.
In fig. 8, the value of TNMSE shows a decreasing trend with increasing total amount of time for all methods. This is because the accumulated a priori signal collected from the entire time period can provide additional information for accurate channel tracking. The performance metric trend remains the same for antenna arrays of different sizes. It is easy to see that compared with the reference algorithms SBL and D-OMP, the performance of the algorithm provided by the invention is obviously improved.
In fig. 9, the TNMSE performance versus SNR is further compared for the total number of antenna array elements M64 and 256, respectively. It can be observed that the proposed algorithm has sufficient performance gain compared to the baseline algorithms SBL and D-OMP at different antenna numbers. This demonstrates that the algorithm proposed by the present invention can effectively exploit the refined dynamic azimuth and elevation spatial sparsity of the channel. Meanwhile, compared with other two algorithms, the antenna scale has little influence on the algorithm provided by the invention.
In fig. 10, the effect of the number of pilots is further compared. Obviously, although both SBL and D-OMP have better recovery performance as the number of pilots increases, the algorithm proposed by the present invention always maintains superior performance gain. In addition, unlike the other two algorithms, the proposed algorithm exhibits a near-level performance as the pilot scale increases. This shows that the algorithm proposed by the present invention can achieve stable performance with a smaller pilot overhead. This is because conventionally, increasing the number of pilots can improve the recovery performance based on multiple estimates over time. The algorithm provided by the invention can effectively obtain higher accuracy by exploring the dynamic channel structure.
Sixth, conclusion
The invention proposes a statistical dynamic channel model called 3D-2D-MM, which establishes a deeper probabilistic relationship of hidden value vectors and joint hidden support vectors.
Further, a new 3D-DTAMP algorithm is designed that utilizes the newly added 3D-2D-MM prior information to recursively track the dynamic channel.
The advantages of the algorithm proposed by the present invention include the following three aspects.
First, the messaging rules allow for a more realistic three-dimensional channel environment.
Secondly, the structure of the dynamic channel is considered from the two aspects of time domain and space domain, and the recovery performance is improved.
Finally, it is more advantageous to track the dynamic channel of longer time series with lower pilot overhead.
Analysis and numerical results show that compared with other two reference algorithms, the algorithm provided by the invention has better recovery performance while reducing pilot frequency overhead and time complexity.
Appendix, Module
Figure BDA0002595005450000261
3D-2D-MM messaging in
FIG. 11 is a message at module
Figure BDA0002595005450000262
Showing details of the message delivery sequence path.
(1) Hidden value vector of message in Gauss
Figure BDA0002595005450000269
To be transmitted to
Messages delivered from a previous time slot
Figure BDA0002595005450000263
Is shaped as
Figure BDA0002595005450000264
Slave variable node
Figure BDA00025950054500002610
Transferred to the factor node pit,mCan be represented as
Figure BDA0002595005450000265
Wherein the content of the first and second substances,
Figure BDA0002595005450000266
when t is equal to 1, setting
Figure BDA0002595005450000267
(2) Message is at
Figure BDA0002595005450000268
To be transmitted to
Slave variable node gt,mTo the factor node pit,mIs that
Figure BDA0002595005450000271
Slave factor node pit,mTo variable node bt,mIs that
Figure BDA0002595005450000272
Wherein the content of the first and second substances,
Figure BDA0002595005450000273
slave variable node bt,mTo factor node ut,mOf a message
Figure BDA0002595005450000274
And
Figure BDA0002595005450000275
the same is true. Slave factor node ut,mTo variable node
Figure BDA0002595005450000276
Is that
Figure BDA0002595005450000277
Wherein
Figure BDA0002595005450000278
Wherein
Figure BDA0002595005450000279
And is
Figure BDA00025950054500002710
(3) Message support in azimuth at2D-MM of
First, details of the azimuth support estimation are introduced. Two cases can be distinguished, one when t is 1 and the other when t > 1. Note that the slave factor node ut,mTo variable node
Figure BDA00025950054500002711
Is that
Figure BDA0002595005450000281
This is based on the matched overall gain dimension M, the azimuth supported dimension nxElevation-supported dimension nyAnd derived.
When t is 1, the azimuth forward parameter is
Figure BDA0002595005450000282
Azimuth backward parameter
Figure BDA0002595005450000283
Given by equation (46).
When t is more than 1, azimuth forward parameter
Figure BDA0002595005450000284
Given by equation (47); azimuth backward parameter
Figure BDA0002595005450000285
Wherein gamma is(A)Given by equation (48).
Figure BDA0002595005450000286
Figure BDA0002595005450000287
Figure BDA0002595005450000288
The final azimuth message representing the non-zero probability can be written as
Figure BDA0002595005450000291
(4) Message on path
Figure BDA0002595005450000292
To be transmitted to
Slave factor node
Figure BDA0002595005450000293
To variable node ut,mIs that
Figure BDA0002595005450000294
Wherein
Figure BDA0002595005450000295
It is to be noted that it is preferable that,
Figure BDA0002595005450000296
is shown at nxThe messages passed on the individual azimuth support elements,
Figure BDA0002595005450000297
denotes the n-thxAn azimuth support element and an nthyThe elevation angles support matching relationships between elements. This situation is similar to the elevation support messaging equation.
Slave factor node ut,mTo variable node
Figure BDA0002595005450000298
Is that
Figure BDA0002595005450000299
Wherein the content of the first and second substances,
Figure BDA00025950054500002910
given by equation (53).
Figure BDA00025950054500002911
(5) Message support in elevation et2D-MM of
Derivation of detail and azimuth support atThe estimated messaging process is similar.
Slave factor node ut,mTo variable node
Figure BDA0002595005450000301
The message to be delivered is
Figure BDA0002595005450000302
When t is 1, the elevation forward parameter is
Figure BDA0002595005450000303
Elevation angle backward direction parameter
Figure BDA0002595005450000304
Given by equation (56).
When t > 1, elevation forward parameter
Figure BDA0002595005450000305
Given by equation (57); elevation angle backward direction parameter
Figure BDA0002595005450000306
Wherein gamma is(E)Given by equation (58).
Figure BDA0002595005450000307
Figure BDA0002595005450000308
Figure BDA0002595005450000311
The final elevation non-zero probability output can be written as
Figure BDA0002595005450000312
(6) Message on path
Figure BDA0002595005450000313
To be transmitted to
Slave factor node
Figure BDA0002595005450000314
To variable node ut,mIs that
Figure BDA0002595005450000315
Wherein
Figure BDA0002595005450000316
Slave factor node ut,mTo variable node bt,mIs that
Figure BDA0002595005450000317
Wherein
Figure BDA0002595005450000318
Slave variable node bt,mTo the factor node pit,mOf a message
Figure BDA0002595005450000319
And
Figure BDA00025950054500003110
the same is true.
Slave factor node pit,mReturn to variable node gt,mIs that
Figure BDA0002595005450000321
The invention also provides a communication method, which comprises the communication by using the method and the formula of any one of the embodiments.
In some embodiments of the present invention, the communication method of the present example mainly includes the steps of: in a communication network, a signal is channel tracked. In some embodiments of the invention, the aforementioned channel tracking comprises: and carrying out three-dimensional space channel tracking on the signals.
In some embodiments of the present invention, the receiving end of the signal is moved in three dimensions with a horizontal displacement and a vertical displacement.
In some embodiments of the invention, the aforementioned channel tracking comprises: and determining channel information according to the information of the received signal and the prior information. Optionally, the foregoing channel tracking further includes: and determining the received signal or determining the sent signal according to the determined channel.
In some embodiments of the present invention, the aforementioned channel tracking of the signal comprises: channel tracking is performed according to a three-dimensional space two-dimensional markov model.
In some embodiments of the present invention, the aforementioned channel tracking of the signal comprises: the channel information is represented by representing the channel information by using the channel gain and the antenna response matrix, representing the channel gain by using angle information of the channel gain and amplitude information of the channel gain, and representing the angle information of the channel gain by using azimuth information of the channel gain and elevation information of the channel gain.
In some embodiments of the invention, the aforementioned channel tracking of the signal comprises one or more of: determining the current channel gain according to the historical information (prior information) of the channel gain; determining current angle information of the channel gain according to historical information (prior information) of the angle information of the channel gain; determining azimuth information of a current channel gain according to historical information (prior information) of the azimuth information of the channel gain; determining current elevation angle information of the channel gain according to historical information (prior information) of the elevation angle information of the channel gain; the current amplitude information of the channel gain is determined from the history information (a priori information) of the amplitude information of the channel gain.
In some embodiments of the present invention, the aforementioned channel tracking of the signal comprises: from the received signal y during the period t(t)Estimating a dynamic channel vector gtElement g in (1)t,mIs expected to
Figure BDA0002595005450000322
Wherein the expectation is based on the probability p (g)t,m|y(t)) Obtained, probability p (g)t，m|y(t)) Obtained according to the following formula
Figure BDA0002595005450000323
Wherein the content of the first and second substances,
Figure BDA0002595005450000324
represents a vector g(t)Does not have element gt,mVector of (a), b(t)Angle information indicating the gain of the channel is obtained,
Figure BDA0002595005450000325
amplitude information representing a channel gain; to track the dynamic channel vector g at time tt。
In some embodiments of the present invention, the aforementioned channel tracking of the signal comprises: the dynamic channel is recursively tracked based on a three-dimensional dynamic Turbo approximate message passing algorithm.
In some embodiments of the invention, the aforementioned communication network comprises a space-air-ground integrated network; the sending end of the signal comprises one of a satellite and an unmanned aerial vehicle, and the receiving end of the signal comprises the other of the satellite and the unmanned aerial vehicle.
Alternatively, the communication signal used by the aforementioned communication method includes a millimeter wave signal, a microwave signal, and the like.
It is noted that the channel tracking involved in the present invention is not conventional channel estimation. Conventional channel estimation is based only on the received signal and does not take historical information into account. The channel tracking of the invention also utilizes prior information, for example, the information of the last time slot is utilized to process the angle information of azimuth and elevation, amplitude information and the like, thereby effectively improving the accuracy of channel estimation.
The present invention also provides a communication apparatus for implementing the communication method shown in any of the foregoing embodiments. Alternatively, the communication device may be a transmitting end, or may be a receiving end. Alternatively, the communication device may be a space end device located in space (space), such as a satellite, or an air end device located in air (air), such as a drone, or a ground end device located in ground, such as a mobile terminal, a notebook, or a ground base station.
The present invention also proposes a communication system for implementing the communication method shown in any one of the foregoing embodiments. In one embodiment, the communication system includes one or more of a transmitting end device and a receiving end device for implementing the communication method shown in any one of the foregoing embodiments. In another embodiment, the communication system comprises one or more of a space end device, an air end device and a ground end device for implementing the communication method shown in any one of the foregoing embodiments.
Embodiments of the present invention also provide a computer storage medium, where computer instructions are stored, and when the computer instructions are executed on a device, the device executes the above related method steps to implement the communication method in the above embodiments.
Embodiments of the present invention also provide a computer program product, which when run on a computer, causes the computer to execute the above related steps to implement the communication method in the above embodiments.
In addition, the embodiment of the present invention further provides an apparatus, which may specifically be a chip, a component or a module, and the apparatus may include a processor and a memory connected to each other; the memory is used for storing computer execution instructions, and when the device runs, the processor can execute the computer execution instructions stored in the memory, so that the chip can execute the communication method in the above-mentioned method embodiments.
The device, the computer storage medium, the computer program product, or the chip provided by the present invention are all configured to execute the corresponding method provided above, and therefore, the beneficial effects achieved by the device, the computer storage medium, the computer program product, or the chip may refer to the beneficial effects in the corresponding method provided above, and are not described herein again.
Although the present invention has been described with reference to a preferred embodiment, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. A method of communication, the method comprising the steps of:
in a communication network, a signal is channel tracked.
2. The communication method according to claim 1,
the receiving end and/or the emitting end of the signal move in three-dimensional space, and have horizontal displacement and vertical displacement,
the channel tracking of the signal comprises: and carrying out three-dimensional space channel tracking on the signals.
3. The communication method of claim 1, wherein the channel tracking the signal comprises: and determining channel information according to the information of the received signal and the prior information.
4. The communication method of claim 3, wherein the channel tracking the signal comprises: the channel tracking is performed according to a three-dimensional space two-dimensional markov model.
5. The communication method of claim 4, wherein the channel tracking the signal comprises:
representing the channel information by using a channel gain and an antenna response matrix, representing the channel gain by using angle information of the channel gain and magnitude information of the channel gain, and representing the angle information of the channel gain by using azimuth information of the channel gain and elevation information of the channel gain.
6. The communication method of claim 5, wherein the channel tracking the signal comprises:
from the received signal y during the period t(t)Estimating a dynamic channel vector gtElement g in (1)t,mIs expected to
Figure FDA0002595005440000011
Wherein the expectation is based on a probability p (g)t,m|y(t)) Obtained, the probability p (g)t,m|y(t)) Obtained according to the following formula
Figure FDA0002595005440000012
Wherein the content of the first and second substances,
Figure FDA0002595005440000013
represents a vector g(t)Does not have element gt,mVector of (a), b(t)Said angle information, θ, representing said channel gain(t)Amplitude information representing the channel gain; to track the dynamic channel vector g at time tt。
7. The communication method of claim 4, wherein the channel tracking the signal comprises:
the dynamic channel is recursively tracked based on a three-dimensional dynamic Turbo approximate message passing algorithm.
8. The communication method according to claim 1, wherein:
the communication network comprises a space-air-ground integrated network;
the transmitting end of the signal comprises one of a satellite and an unmanned aerial vehicle, and the receiving end of the signal comprises the other one of the satellite and the unmanned aerial vehicle.
9. A communication apparatus for implementing the communication method according to any one of claims 1 to 8.
10. A computer storage medium comprising computer instructions which, when run on a device, cause the device to perform the communication method of any one of claims 1 to 8.
CN202010706632.5A 2020-07-21 2020-07-21 Communication method, communication device, and storage medium Pending CN112019251A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010706632.5A CN112019251A (en) 2020-07-21 2020-07-21 Communication method, communication device, and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010706632.5A CN112019251A (en) 2020-07-21 2020-07-21 Communication method, communication device, and storage medium

Publications (1)

Publication Number Publication Date
CN112019251A true CN112019251A (en) 2020-12-01

Family

ID=73498846

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010706632.5A Pending CN112019251A (en) 2020-07-21 2020-07-21 Communication method, communication device, and storage medium

Country Status (1)

Country Link
CN (1) CN112019251A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140241239A1 (en) * 2012-09-21 2014-08-28 Donald C.D. Chang Communications Architectures via UAV
CN106855626A (en) * 2016-12-29 2017-06-16 北京奇艺世纪科技有限公司 Vector tracking method and wave filter
CN106953682A (en) * 2017-03-28 2017-07-14 电子科技大学 A kind of method cooperateed with for unmanned plane networking
CN107483131A (en) * 2017-06-30 2017-12-15 西安电子科技大学 The double satellite combined channel Markov state method for generating sequence of high-speed aircraft
CN109743093A (en) * 2018-12-26 2019-05-10 北京邮电大学 A kind of unmanned plane millimetre-wave attenuator beam tracking method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140241239A1 (en) * 2012-09-21 2014-08-28 Donald C.D. Chang Communications Architectures via UAV
CN106855626A (en) * 2016-12-29 2017-06-16 北京奇艺世纪科技有限公司 Vector tracking method and wave filter
CN106953682A (en) * 2017-03-28 2017-07-14 电子科技大学 A kind of method cooperateed with for unmanned plane networking
CN107483131A (en) * 2017-06-30 2017-12-15 西安电子科技大学 The double satellite combined channel Markov state method for generating sequence of high-speed aircraft
CN109743093A (en) * 2018-12-26 2019-05-10 北京邮电大学 A kind of unmanned plane millimetre-wave attenuator beam tracking method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JIADONG YU, ETC.: "3D Channel Tracking for UAV-Satellite Communications in Space-Air-Ground Integrated Networks", 《IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS》 *

Similar Documents

Publication Publication Date Title
Yu et al. 3D channel tracking for UAV-satellite communications in space-air-ground integrated networks
CN110492911B (en) Beam tracking method and system for unmanned aerial vehicle communication
US20220376956A1 (en) Transmission System with Channel Estimation Based on a Neural Network
CN113556300A (en) Joint active terminal detection and channel estimation method based on time domain training sequence
CN113595944A (en) Channel estimation method for millimeter wave MIMO hybrid precoding system
Hu et al. MmWave MIMO communication with semi-passive RIS: A low-complexity channel estimation scheme
CN114172597A (en) Non-iterative parameter joint estimation method based on reconfigurable intelligent surface
Chen et al. Joint initial access and localization in millimeter wave vehicular networks: a hybrid model/data driven approach
CN113965881A (en) Millimeter wave integrated communication and sensing method under shielding effect
Yuan et al. Joint radar-communication-based bayesian predictive beamforming for vehicular networks
CN110086734B (en) Channel tracking and recovery method in time-varying large-scale MIMO system
CN112731273A (en) Low-complexity signal direction-of-arrival estimation method based on sparse Bayes
CN112019251A (en) Communication method, communication device, and storage medium
CN116760439A (en) Robust wave beam forming method based on radar communication integrated system
Xu et al. Joint Scattering Environment Sensing and Channel Estimation Based on Non-stationary Markov Random Field
CN115714612A (en) Perception-based communication beam tracking method
Nayir et al. Hybrid-field channel estimation for massive MIMO systems based on OMP cascaded convolutional autoencoder
CN115549745A (en) RIS phase shift design method, apparatus, computer device and storage medium
Yu et al. 3D on and off-grid dynamic channel tracking for multiple UAVs and satellite communications
Wang et al. Deep learning-based compressive beam alignment in mmWave vehicular systems
Alexandropoulos et al. Joint localization and channel estimation for UAV-assisted millimeter wave communications
Jha et al. Temporally correlated compressed sensing using generative models for channel estimation in unmanned aerial vehicles
Yu et al. On-grid 3D dynamic channel tracking for space-air communications with multiple UAVs
Bellili et al. Massive MIMO mmWave channel estimation using approximate message passing and Laplacian prior
Cheng et al. Gibbs Sampling Based Sparse Bayesian Learning for Direction-of-Arrival Estimation with Impulse Noise Towards 6G

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20201201

WD01 Invention patent application deemed withdrawn after publication