CN118945593A - A positioning method based on 5G base station - Google Patents

Abstract

The invention discloses a positioning method based on 5G base stations, which belongs to the technical field of positioning of 5G base stations, and comprises the steps of constructing a signal characteristic matrix and a signal difference coefficient matrix according to the signal characteristic of each 5G base station, wherein the signal characteristic matrix reflects the signal characteristic of each 5G base station, the signal difference coefficient matrix reflects the difference condition of each signal characteristic sequence which is sequentially achieved, the difference of signal characteristics at adjacent moments, the characteristic distinction degree between the positions of different 5G base stations is enhanced, the signal characteristic matrix and the signal difference coefficient matrix are processed by adopting a multi-fusion positioning network to obtain an initial position, so that the multi-fusion positioning network combines static characteristics and dynamic change information, the terminal position is accurately estimated, the influence of instantaneous noise and abnormal values is reduced, the stability of a positioning result is improved, and the initial position is corrected by adopting the terminal position at a plurality of moments of histories, so that the positioning precision is further improved.

Description

Positioning method based on 5G base station

Technical Field

The invention relates to the technical field of positioning of 5G base stations, in particular to a positioning method based on a 5G base station.

Background

With the rapid development and wide deployment of 5G technology, high-precision positioning services have become a research area of great interest. Traditional positioning methods such as GPS often have difficulty providing accurate positioning that meets the needs of modern applications in indoor environments or urban dense areas. The advent of 5G networks has provided new possibilities for solving this problem, especially in scenarios requiring high accuracy, low latency positioning, such as intelligent transportation, augmented reality, and industrial automation. In 5G networks, terminals are typically able to receive signals from multiple base stations, and how to coordinate and fuse such information to improve positioning accuracy is an important issue. The existing positioning technology based on the 5G base station mainly comprises the following steps: 1. the position is determined by measuring the angle at which the signal arrives at the terminal, 2, estimating the distance based on the received signal strength, and further determining the position. However, in reality, more obstacles exist to cause signal reflection and scattering, so that multipath effect and non-line-of-sight propagation are caused, and positioning accuracy is affected.

Disclosure of Invention

Aiming at the defects in the prior art, the positioning method based on the 5G base station solves the problem of low positioning precision in the prior art.

In order to achieve the aim of the invention, the invention adopts the following technical scheme: a positioning method based on a 5G base station comprises the following steps:

S1, constructing a signal characteristic sequence according to the signal characteristics of each 5G base station;

s2, sequencing each signal characteristic sequence according to the signal arrival time of each 5G base station, and constructing a signal gap coefficient matrix and a signal characteristic matrix;

s3, processing the signal characteristic matrix and the signal gap coefficient matrix by adopting a multi-fusion positioning network to obtain an initial position;

S4, correcting the initial position according to the terminal positions at a plurality of historical moments to obtain the terminal position at the current moment.

The beneficial effects of the invention are as follows: according to the invention, a signal characteristic sequence is constructed according to the signal characteristics of each 5G base station, the signal characteristic sequences are ordered based on the arrival time of signals, a signal difference coefficient matrix is constructed, the signal characteristic matrix reflects the signal characteristics of each 5G base station, the signal difference coefficient matrix reflects the difference condition of each signal characteristic sequence which is sequentially achieved, the difference of the signal characteristics at adjacent moments is enhanced, the characteristic distinction degree between the positions of different 5G base stations is enhanced, the signal characteristic matrix and the signal difference coefficient matrix are processed by adopting a multi-fusion positioning network to obtain an initial position, the multi-fusion positioning network is enabled to accurately estimate the position of a terminal by combining static characteristics and dynamic change information, the influence of instantaneous noise and abnormal values is reduced, the stability of a positioning result is improved, the initial position is corrected by adopting the terminal position at a plurality of moments of histories, and the positioning precision is further improved.

Further, the elements in the signal feature sequence in S1 include: received signal power, signal time of arrival, angle of arrival, and channel state information.

Further, the step S2 includes the following sub-steps:

S21, sequencing each signal characteristic sequence according to the signal arrival time of each 5G base station;

S22, taking the ordered signal characteristic sequence as a row vector to form a signal characteristic matrix;

S23, subtracting two signal feature sequences at adjacent moments to obtain a signal feature gap sequence, wherein D=X_t- X_t-1, D is the signal feature gap sequence, X_t is the signal feature sequence of the t arrival time, X_t-1 is the signal feature sequence of the t-1 arrival time, and t is the number of the arrival time;

s24, carrying out normalization processing on elements in each signal characteristic gap sequence to obtain a signal gap coefficient sequence;

s25, taking each signal gap coefficient sequence as a row vector to form a signal gap coefficient matrix.

Further, the multi-converged positioning network in S3 includes: the system comprises a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit, a fourth scale feature extraction unit, a first feature fusion unit, a second feature fusion unit, a first full-connection layer, a second full-connection layer, an adder A1 and an output layer;

the input end of the first scale feature extraction unit is used for inputting a signal feature matrix;

The input end of the second scale feature extraction unit is used for inputting a signal feature matrix;

The input end of the third scale feature extraction unit is used for inputting a signal difference coefficient matrix;

the input end of the fourth scale feature extraction unit is used for inputting a signal difference coefficient matrix;

The input end of the first feature fusion unit is respectively connected with the output end of the first scale feature extraction unit and the output end of the third scale feature extraction unit, and the output end of the first feature fusion unit is connected with the input end of the first full-connection layer;

The input end of the second feature fusion unit is respectively connected with the output end of the second scale feature extraction unit and the output end of the fourth scale feature extraction unit, and the output end of the second feature fusion unit is connected with the input end of the second full-connection layer;

The first input end of the adder A1 is connected with the output end of the first full-connection layer, the second input end of the adder A1 is connected with the output end of the second full-connection layer, and the output end of the adder A1 is connected with the input end of the output layer;

and the output end of the output layer is used as the output end of the multi-fusion positioning network.

The beneficial effects of the above further scheme are: the invention processes the same signal feature matrix through the first scale feature extraction unit and the second scale feature extraction unit respectively, extracts the features of different scales, processes the same signal gap coefficient matrix through the third scale feature extraction unit and the fourth scale feature extraction unit, extracts the features of different scales, fuses the output features of the first scale feature extraction unit and the third scale feature extraction unit at the first feature fusion unit, fuses the output features of the second scale feature extraction unit and the fourth scale feature extraction unit at the second feature fusion unit, realizes that the features of the same scale are fused in one feature fusion unit, processes the fused features through the full-connection layer, each neuron in the full-connection layer is connected with all neurons of the previous layer, and can automatically determine the importance of the different features by learning the weights of the connection, optimize the combination mode of the features and output the comprehensive adder A1 of the output layer to obtain the initial position.

Further, the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit, and the fourth scale feature extraction unit each include: a first convolution block, an SPP layer, a second convolution block, a Sigmoid layer, a multiplier M1 and an adder A2;

The input end of the first convolution block is used as the input end of a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit or a fourth scale feature extraction unit; the input end of the SPP layer is connected with the output end of the first convolution block, and the output end of the SPP layer is respectively connected with the input end of the second convolution block and the first input end of the adder A2; the output end of the second convolution block is respectively connected with the input end of the Sigmoid layer and the first input end of the multiplier M1; the second input end of the multiplier M1 is connected with the output end of the Sigmoid layer, and the output end of the multiplier M1 is connected with the second input end of the adder A2; the output end of the adder A2 is used as the output end of the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit or the fourth scale feature extraction unit;

The size of the convolution kernel in the first convolution block in the first scale feature extraction unit and the third scale feature extraction unit is；

The size of the convolution kernel in the first convolution block in the second scale feature extraction unit and the fourth scale feature extraction unit is。

The beneficial effects of the above further scheme are: according to the invention, the first scale feature extraction unit and the third scale feature extraction unit are adopted to extract features of the same scale, the second scale feature extraction unit and the fourth scale feature extraction unit are adopted to extract features of the same scale, the SPP layer is adopted to enhance the adaptability of the network to different scale inputs, the robustness of feature extraction is improved, the network can adaptively pay attention to important features through the Sigmoid layer and the multiplier M1, irrelevant information is ignored, the efficiency and the precision of feature extraction are improved, the adder A2 realizes residual connection, the output of the SPP layer is added with the features after attention processing, the gradient vanishing problem of the deep network is relieved, and the training of the deeper network is promoted.

Further, the expression of the first feature fusion unit is: Wherein H₁ is the output of the first feature fusion unit, Y₁ is the output of the first scale feature extraction unit, Y₃ is the output of the third scale feature extraction unit,Is Hadamard product;

the expression of the second feature fusion unit is as follows: Wherein H₂ is the output of the second feature fusion unit, Y₂ is the output of the second scale feature extraction unit, and Y₄ is the output of the fourth scale feature extraction unit.

The beneficial effects of the above further scheme are: the invention multiplies the characteristics extracted from the signal characteristic matrix and the characteristics extracted from the signal gap coefficient matrix by elements, and can more accurately describe the change characteristics of the signal in space time by combining static and dynamic characteristics.

Further, the expression of the output layer is:， where long₀ is the longitude of the initial position, lat₀ is the latitude of the initial position, x_i is the ith eigenvalue output by adder A1, ω_1,i is the longitude weight of the ith eigenvalue, b_1,i is the longitude bias of the ith eigenvalue, ω_2,i is the latitude weight of the ith eigenvalue, b_2,i is the latitude bias of the ith eigenvalue, N is the number of eigenvalues output by adder A1, and i is a positive integer.

Further, the step S4 includes the steps of:

s41, fitting a plurality of longitudes by using a quadratic polynomial according to terminal positions at a plurality of moments of history to obtain a terminal movement longitude track, and fitting a plurality of dimensions by using a quadratic polynomial to obtain a terminal movement latitude track;

s42, predicting longitude of the next moment by using the terminal moving longitude track, and predicting latitude of the next moment by using the terminal moving latitude track to obtain a predicted position;

S43, calculating a deviation coefficient of the initial position and the predicted position;

s44, weighting the initial position and the predicted position according to the deviation coefficient of the initial position and the predicted position to obtain the terminal position at the current moment.

Further, the formula for calculating the bias coefficient between the initial position and the predicted position in S43 is as follows: Where γ is a deviation coefficient, arctan is an arctangent function, long₀ is the longitude of the initial position, lat₀ is the latitude of the initial position,In order to predict the longitude of the location,Is the latitude of the predicted location.

Further, the weighting formula in S44 is:， where x is the longitude of the terminal position at the current time and y is the latitude of the terminal position at the current time.

The beneficial effects of the above further scheme are: according to the method, the historical position data are fitted by using the quadratic polynomial, and the movement track models of the longitude and the latitude are respectively established. The position at the next moment is then predicted based on these trajectory models, which provides a powerful a priori information for the position estimation. By calculating the deviation coefficient of the initial position and the predicted position, the influence of the historical track information and the current measurement data can be dynamically balanced. And finally, fusing the initial position and the predicted position by adopting a weighting processing method, thereby not only considering the historical movement trend, but also keeping the accuracy of the current measurement. The method not only can smooth the positioning result and reduce the influence of instantaneous errors, but also can improve the continuity and stability of positioning. Particularly, under the condition of unstable signals or noise in measurement, the method can remarkably improve the positioning accuracy.

Drawings

FIG. 1 is a flow chart of a positioning method based on a 5G base station;

FIG. 2 is a schematic diagram of a multi-converged positioning network;

fig. 3 is a schematic structural diagram of a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit, and a fourth scale feature extraction unit.

Detailed Description

The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.

As shown in fig. 1, a positioning method based on a 5G base station includes the following steps:

S1, constructing a signal characteristic sequence according to the signal characteristics of each 5G base station;

s2, sequencing each signal characteristic sequence according to the signal arrival time of each 5G base station, and constructing a signal gap coefficient matrix and a signal characteristic matrix;

s3, processing the signal characteristic matrix and the signal gap coefficient matrix by adopting a multi-fusion positioning network to obtain an initial position;

S4, correcting the initial position according to the terminal positions at a plurality of historical moments to obtain the terminal position at the current moment.

The elements in the signal characteristic sequence in the S1 comprise: received signal power, signal time of arrival, angle of arrival, and channel state information.

The invention comprehensively utilizes multidimensional features such as received signal power, signal arrival time, arrival angle, channel state information and the like to construct a comprehensive signal feature sequence, and fully captures the rich characteristics of 5G signals.

The step S2 comprises the following sub-steps:

S21, sequencing each signal characteristic sequence according to the signal arrival time of each 5G base station;

S22, taking the ordered signal characteristic sequence as a row vector to form a signal characteristic matrix;

In this embodiment, during sorting, the first row of the signal feature matrix is arranged in ascending order according to the arrival time, so that the arrival time of the first row is the earliest and the arrival time of the last row is the latest;

S23, subtracting two signal feature sequences at adjacent moments to obtain a signal feature gap sequence, wherein D=X_t- X_t-1, D is the signal feature gap sequence, X_t is the signal feature sequence of the t arrival time, X_t-1 is the signal feature sequence of the t-1 arrival time, and t is the number of the arrival time;

s24, carrying out normalization processing on elements in each signal characteristic gap sequence to obtain a signal gap coefficient sequence;

s25, taking each signal gap coefficient sequence as a row vector to form a signal gap coefficient matrix.

As shown in fig. 2, the multi-converged positioning network in S3 includes: the system comprises a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit, a fourth scale feature extraction unit, a first feature fusion unit, a second feature fusion unit, a first full-connection layer, a second full-connection layer, an adder A1 and an output layer;

the input end of the first scale feature extraction unit is used for inputting a signal feature matrix;

The input end of the second scale feature extraction unit is used for inputting a signal feature matrix;

The input end of the third scale feature extraction unit is used for inputting a signal difference coefficient matrix;

the input end of the fourth scale feature extraction unit is used for inputting a signal difference coefficient matrix;

The input end of the first feature fusion unit is respectively connected with the output end of the first scale feature extraction unit and the output end of the third scale feature extraction unit, and the output end of the first feature fusion unit is connected with the input end of the first full-connection layer;

The input end of the second feature fusion unit is respectively connected with the output end of the second scale feature extraction unit and the output end of the fourth scale feature extraction unit, and the output end of the second feature fusion unit is connected with the input end of the second full-connection layer;

The first input end of the adder A1 is connected with the output end of the first full-connection layer, the second input end of the adder A1 is connected with the output end of the second full-connection layer, and the output end of the adder A1 is connected with the input end of the output layer;

and the output end of the output layer is used as the output end of the multi-fusion positioning network.

The invention processes the same signal feature matrix through the first scale feature extraction unit and the second scale feature extraction unit respectively, extracts the features of different scales, processes the same signal gap coefficient matrix through the third scale feature extraction unit and the fourth scale feature extraction unit, extracts the features of different scales, fuses the output features of the first scale feature extraction unit and the third scale feature extraction unit at the first feature fusion unit, fuses the output features of the second scale feature extraction unit and the fourth scale feature extraction unit at the second feature fusion unit, realizes that the features of the same scale are fused in one feature fusion unit, processes the fused features through the full-connection layer, each neuron in the full-connection layer is connected with all neurons of the previous layer, and can automatically determine the importance of the different features by learning the weights of the connection, optimize the combination mode of the features and output the comprehensive adder A1 of the output layer to obtain the initial position.

As shown in fig. 3, the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit, and the fourth scale feature extraction unit each include: a first convolution block, an SPP layer, a second convolution block, a Sigmoid layer, a multiplier M1 and an adder A2;

The input end of the first convolution block is used as the input end of a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit or a fourth scale feature extraction unit; the input end of the SPP layer is connected with the output end of the first convolution block, and the output end of the SPP layer is respectively connected with the input end of the second convolution block and the first input end of the adder A2; the output end of the second convolution block is respectively connected with the input end of the Sigmoid layer and the first input end of the multiplier M1; the second input end of the multiplier M1 is connected with the output end of the Sigmoid layer, and the output end of the multiplier M1 is connected with the second input end of the adder A2; the output end of the adder A2 is used as the output end of the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit or the fourth scale feature extraction unit;

The size of the convolution kernel in the first convolution block in the first scale feature extraction unit and the third scale feature extraction unit is；

The size of the convolution kernel in the first convolution block in the second scale feature extraction unit and the fourth scale feature extraction unit is。

The SPP layer is SPATIAL PYRAMID Pooling layers, and the convolution block comprises: convolution layer, reLU layer and BN layer.

In this embodiment, the multi-fusion positioning network is trained by gradient descent.

According to the invention, the first scale feature extraction unit and the third scale feature extraction unit are adopted to extract features of the same scale, the second scale feature extraction unit and the fourth scale feature extraction unit are adopted to extract features of the same scale, the SPP layer is adopted to enhance the adaptability of the network to different scale inputs, the robustness of feature extraction is improved, the network can adaptively pay attention to important features through the Sigmoid layer and the multiplier M1, irrelevant information is ignored, the efficiency and the precision of feature extraction are improved, the adder A2 realizes residual connection, the output of the SPP layer is added with the features after attention processing, the gradient vanishing problem of the deep network is relieved, and the training of the deeper network is promoted.

The expression of the first feature fusion unit is as follows: Wherein H₁ is the output of the first feature fusion unit, Y₁ is the output of the first scale feature extraction unit, Y₃ is the output of the third scale feature extraction unit,Is Hadamard product;

the expression of the second feature fusion unit is as follows: Wherein H₂ is the output of the second feature fusion unit, Y₂ is the output of the second scale feature extraction unit, and Y₄ is the output of the fourth scale feature extraction unit.

The invention multiplies the characteristics extracted from the signal characteristic matrix and the characteristics extracted from the signal gap coefficient matrix by elements, and can more accurately describe the change characteristics of the signal in space time by combining static and dynamic characteristics.

The expression of the output layer is:， where long₀ is the longitude of the initial position, lat₀ is the latitude of the initial position, x_i is the ith eigenvalue output by adder A1, ω_1,i is the longitude weight of the ith eigenvalue, b_1,i is the longitude bias of the ith eigenvalue, ω_2,i is the latitude weight of the ith eigenvalue, b_2,i is the latitude bias of the ith eigenvalue, N is the number of eigenvalues output by adder A1, and i is a positive integer.

In the invention, the output layer comprises two types of weights and biases, and the longitude and the latitude are respectively predicted.

The step S4 comprises the following steps:

s41, fitting a plurality of longitudes by using a quadratic polynomial according to terminal positions at a plurality of moments of history to obtain a terminal movement longitude track, and fitting a plurality of dimensions by using a quadratic polynomial to obtain a terminal movement latitude track;

In this embodiment, the quadratic polynomial is ：f₁ = a₁t²+b₁t+ c₁,f₂ = a₂t²+b₂t+ c₂,f₁ a terminal movement longitude trace curve, f₂ a terminal movement latitude trace, t is time, a₁、b₁、a₂、b₂ is four multipliers, and c₁、c₂ is two increments;

s42, predicting longitude of the next moment by using the terminal moving longitude track, and predicting latitude of the next moment by using the terminal moving latitude track to obtain a predicted position;

S43, calculating a deviation coefficient of the initial position and the predicted position;

s44, weighting the initial position and the predicted position according to the deviation coefficient of the initial position and the predicted position to obtain the terminal position at the current moment.

The formula for calculating the bias coefficient of the initial position and the predicted position in S43 is as follows: Where γ is a deviation coefficient, arctan is an arctangent function, long₀ is the longitude of the initial position, lat₀ is the latitude of the initial position,In order to predict the longitude of the location,Is the latitude of the predicted location.

The weighting formula in S44 is:， where x is the longitude of the terminal position at the current time and y is the latitude of the terminal position at the current time.

According to the method, the historical position data are fitted by using the quadratic polynomial, and the movement track models of the longitude and the latitude are respectively established. The position at the next moment is then predicted based on these trajectory models, which provides a powerful a priori information for the position estimation. By calculating the deviation coefficient of the initial position and the predicted position, the influence of the historical track information and the current measurement data can be dynamically balanced. And finally, fusing the initial position and the predicted position by adopting a weighting processing method, thereby not only considering the historical movement trend, but also keeping the accuracy of the current measurement. The method not only can smooth the positioning result and reduce the influence of instantaneous errors, but also can improve the continuity and stability of positioning. Particularly, under the condition of unstable signals or noise in measurement, the method can remarkably improve the positioning accuracy.

According to the invention, a signal characteristic sequence is constructed according to the signal characteristics of each 5G base station, the signal characteristic sequences are ordered based on the arrival time of signals, a signal difference coefficient matrix is constructed, the signal characteristic matrix reflects the signal characteristics of each 5G base station, the signal difference coefficient matrix reflects the difference condition of each signal characteristic sequence which is sequentially achieved, the difference of the signal characteristics at adjacent moments is enhanced, the characteristic distinction degree between the positions of different 5G base stations is enhanced, the signal characteristic matrix and the signal difference coefficient matrix are processed by adopting a multi-fusion positioning network to obtain an initial position, the multi-fusion positioning network is enabled to accurately estimate the position of a terminal by combining static characteristics and dynamic change information, the influence of instantaneous noise and abnormal values is reduced, the stability of a positioning result is improved, the initial position is corrected by adopting the terminal position at a plurality of moments of histories, and the positioning precision is further improved.

In reality, due to the existence of more barriers, the signals are reflected and scattered, multipath effect and non-line-of-sight propagation are caused, and positioning accuracy is affected, the invention constructs a signal characteristic matrix according to the signal characteristics of a plurality of 5G base stations, and according to the signal arrival time, and constructing a signal gap coefficient matrix, representing signal difference conditions of different 5G base stations, processing the signal difference conditions and the signal characteristics by adopting a multi-fusion positioning network, predicting initial positioning, effectively reducing accumulation of single-point measurement errors, and providing a more stable and accurate positioning result.

The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

Translated fromChinese

1.一种基于5G基站的定位方法，其特征在于，包括以下步骤：1. A positioning method based on a 5G base station, characterized in that it comprises the following steps:S1、根据每个5G基站的信号特征，构建信号特征序列；S1. Construct a signal feature sequence based on the signal features of each 5G base station;S2、根据各个5G基站的信号到达时间，对各个信号特征序列排序，构建信号差距系数矩阵和信号特征矩阵；S2. Sort each signal feature sequence according to the signal arrival time of each 5G base station, and construct a signal gap coefficient matrix and a signal feature matrix;S3、采用多融合定位网络对信号特征矩阵和信号差距系数矩阵进行处理，得到初始位置；S3, using a multi-fusion positioning network to process the signal feature matrix and the signal gap coefficient matrix to obtain an initial position;S4、根据历史多个时刻的终端位置，对初始位置进行修正，得到当前时刻的终端位置。S4. According to the terminal positions at multiple historical moments, the initial position is corrected to obtain the terminal position at the current moment.2.根据权利要求1所述的基于5G基站的定位方法，其特征在于，所述S1中信号特征序列中元素包括：接收信号功率、信号到达时间、到达角和信道状态信息。2. According to the 5G base station-based positioning method according to claim 1, it is characterized in that the elements in the signal feature sequence in S1 include: received signal power, signal arrival time, arrival angle and channel state information.3.根据权利要求1所述的基于5G基站的定位方法，其特征在于，所述S2包括以下分步骤：3. The positioning method based on a 5G base station according to claim 1, characterized in that S2 comprises the following sub-steps:S21、根据各个5G基站的信号到达时间，对各个信号特征序列排序；S21. Sort each signal feature sequence according to the signal arrival time of each 5G base station;S22、将排序后的信号特征序列作为行向量，构成信号特征矩阵；S22, using the sorted signal feature sequence as a row vector to form a signal feature matrix;S23、将相邻时刻的两个信号特征序列相减，得到信号特征差距序列，D=X_t- X_t-1，其中，D为信号特征差距序列，X_t为第t到达时间的信号特征序列，X_t-1为第t-1到达时间的信号特征序列，t为到达时间的编号；S23, subtracting two signal feature sequences at adjacent moments to obtain a signal feature difference sequence, D=_Xt -Xt_-1 , where D is the signal feature difference sequence,_Xt is the signal feature sequence at the t-th arrival time, Xt_-1 is the signal feature sequence at the t-1th arrival time, and t is the number of the arrival time;S24、对每个信号特征差距序列中元素进行归一化处理，得到信号差距系数序列；S24, normalizing the elements in each signal feature gap sequence to obtain a signal gap coefficient sequence;S25、将每个信号差距系数序列作为行向量，构成信号差距系数矩阵。S25. Take each signal gap coefficient sequence as a row vector to form a signal gap coefficient matrix.4.根据权利要求1所述的基于5G基站的定位方法，其特征在于，所述S3中多融合定位网络包括：第一尺度特征提取单元、第二尺度特征提取单元、第三尺度特征提取单元、第四尺度特征提取单元、第一特征融合单元、第二特征融合单元、第一全连接层、第二全连接层、加法器A1和输出层；4. The positioning method based on a 5G base station according to claim 1 is characterized in that the multi-fusion positioning network in S3 includes: a first scale feature extraction unit, a second scale feature extraction unit, a third scale feature extraction unit, a fourth scale feature extraction unit, a first feature fusion unit, a second feature fusion unit, a first fully connected layer, a second fully connected layer, an adder A1 and an output layer;所述第一尺度特征提取单元的输入端用于输入信号特征矩阵；The input end of the first scale feature extraction unit is used to input a signal feature matrix;所述第二尺度特征提取单元的输入端用于输入信号特征矩阵；The input end of the second scale feature extraction unit is used for inputting a signal feature matrix;所述第三尺度特征提取单元的输入端用于输入信号差距系数矩阵；The input end of the third scale feature extraction unit is used for inputting a signal gap coefficient matrix;所述第四尺度特征提取单元的输入端用于输入信号差距系数矩阵；The input end of the fourth scale feature extraction unit is used for inputting a signal gap coefficient matrix;所述第一特征融合单元的输入端分别与第一尺度特征提取单元的输出端和第三尺度特征提取单元的输出端连接，其输出端与第一全连接层的输入端连接；The input end of the first feature fusion unit is connected to the output end of the first scale feature extraction unit and the output end of the third scale feature extraction unit respectively, and the output end thereof is connected to the input end of the first fully connected layer;所述第二特征融合单元的输入端分别与第二尺度特征提取单元的输出端和第四尺度特征提取单元的输出端连接，其输出端与第二全连接层的输入端连接；The input end of the second feature fusion unit is connected to the output end of the second scale feature extraction unit and the output end of the fourth scale feature extraction unit respectively, and the output end thereof is connected to the input end of the second fully connected layer;所述加法器A1的第一输入端与第一全连接层的输出端连接，其第二输入端与第二全连接层的输出端连接，其输出端与输出层的输入端连接；The first input terminal of the adder A1 is connected to the output terminal of the first fully connected layer, the second input terminal thereof is connected to the output terminal of the second fully connected layer, and the output terminal thereof is connected to the input terminal of the output layer;所述输出层的输出端作为多融合定位网络的输出端。The output end of the output layer serves as the output end of the multi-fusion positioning network.5.根据权利要求4所述的基于5G基站的定位方法，其特征在于，所述第一尺度特征提取单元、第二尺度特征提取单元、第三尺度特征提取单元和第四尺度特征提取单元均包括：第一卷积块、SPP层、第二卷积块、Sigmoid层、乘法器M1和加法器A2；5. The positioning method based on a 5G base station according to claim 4 is characterized in that the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit and the fourth scale feature extraction unit each include: a first convolution block, an SPP layer, a second convolution block, a Sigmoid layer, a multiplier M1 and an adder A2;所述第一卷积块的输入端作为第一尺度特征提取单元、第二尺度特征提取单元、第三尺度特征提取单元或第四尺度特征提取单元的输入端；所述SPP层的输入端与第一卷积块的输出端连接，其输出端分别与第二卷积块的输入端和加法器A2的第一输入端连接；所述第二卷积块的输出端分别与Sigmoid层的输入端和乘法器M1的第一输入端连接；所述乘法器M1的第二输入端与Sigmoid层的输出端连接，其输出端与加法器A2的第二输入端连接；所述加法器A2的输出端作为第一尺度特征提取单元、第二尺度特征提取单元、第三尺度特征提取单元或第四尺度特征提取单元的输出端；The input end of the first convolution block is used as the input end of the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit or the fourth scale feature extraction unit; the input end of the SPP layer is connected to the output end of the first convolution block, and its output end is respectively connected to the input end of the second convolution block and the first input end of the adder A2; the output end of the second convolution block is respectively connected to the input end of the Sigmoid layer and the first input end of the multiplier M1; the second input end of the multiplier M1 is connected to the output end of the Sigmoid layer, and its output end is connected to the second input end of the adder A2; the output end of the adder A2 is used as the output end of the first scale feature extraction unit, the second scale feature extraction unit, the third scale feature extraction unit or the fourth scale feature extraction unit;在第一尺度特征提取单元和第三尺度特征提取单元中第一卷积块中卷积核的大小为3*3；The size of the convolution kernel in the first convolution block in the first scale feature extraction unit and the third scale feature extraction unit is 3*3;在第二尺度特征提取单元和第四尺度特征提取单元中第一卷积块中卷积核的大小为5*5。The size of the convolution kernel in the first convolution block in the second scale feature extraction unit and the fourth scale feature extraction unit is 5*5.6.根据权利要求4所述的基于5G基站的定位方法，其特征在于，所述第一特征融合单元的表达式为：，其中，H₁为第一特征融合单元的输出，Y₁为第一尺度特征提取单元的输出，Y₃为第三尺度特征提取单元的输出，为哈达玛积；6. The positioning method based on a 5G base station according to claim 4 is characterized in that the expression of the first feature fusion unit is: , where_H1 is the output of the first feature fusion unit,_Y1 is the output of the first scale feature extraction unit, and_Y3 is the output of the third scale feature extraction unit. For Hadamard;所述第二特征融合单元的表达式为：，其中，H₂为第二特征融合单元的输出，Y₂为第二尺度特征提取单元的输出，Y₄为第四尺度特征提取单元的输出。The expression of the second feature fusion unit is: , where H₂ is the output of the second feature fusion unit, Y₂ is the output of the second scale feature extraction unit, and Y₄ is the output of the fourth scale feature extraction unit.7.根据权利要求4所述的基于5G基站的定位方法，其特征在于，所述输出层的表达式为：7. The positioning method based on 5G base station according to claim 4 is characterized in that the expression of the output layer is:，，其中，long₀为初始位置的经度，lat₀为初始位置的纬度，x_i为加法器A1输出的第i个特征值，ω_1,i为第i个特征值的经度权重，b_1,i为第i个特征值的经度偏置，ω_2,i为第i个特征值的纬度权重，b_2,i为第i个特征值的纬度偏置，N为加法器A1输出的特征值的数量，i为正整数。 , , where long₀ is the longitude of the initial position, lat₀ is the latitude of the initial position,_xi is the i-th eigenvalue output by adder A1,_ω1,i is the longitude weight of the i-th eigenvalue, b1_,i is the longitude bias of the i-th eigenvalue,_ω2,i is the latitude weight of the i-th eigenvalue,_b2,i is the latitude bias of the i-th eigenvalue, N is the number of eigenvalues output by adder A1, and i is a positive integer.8.根据权利要求1所述的基于5G基站的定位方法，其特征在于，所述S4包括以下步骤：8. The 5G base station-based positioning method according to claim 1, wherein S4 comprises the following steps:S41、根据历史多个时刻的终端位置，使用二次多项式对多个经度进行拟合，得到终端移动经度轨迹，使用二次多项式对多个维度进行拟合，得到终端移动维度轨迹；S41. According to the terminal positions at multiple historical moments, a quadratic polynomial is used to fit multiple longitudes to obtain a terminal movement longitude trajectory, and a quadratic polynomial is used to fit multiple dimensions to obtain a terminal movement dimension trajectory;S42、使用终端移动经度轨迹预测下一个时刻的经度，使用终端移动维度轨迹预测下一个时刻的维度，得到预测位置；S42, predicting the longitude at the next moment using the terminal's moving longitude trajectory, and predicting the latitude at the next moment using the terminal's moving latitude trajectory, to obtain a predicted position;S43、计算初始位置与预测位置的偏差系数；S43, calculating the deviation coefficient between the initial position and the predicted position;S44、根据初始位置与预测位置的偏差系数，对初始位置和预测位置进行加权处理，得到当前时刻的终端位置。S44. Perform weighted processing on the initial position and the predicted position according to the deviation coefficient between the initial position and the predicted position to obtain the terminal position at the current moment.9.根据权利要求8所述的基于5G基站的定位方法，其特征在于，所述S43中计算初始位置与预测位置的偏置系数的公式为：9. The positioning method based on a 5G base station according to claim 8, characterized in that the formula for calculating the bias coefficient between the initial position and the predicted position in S43 is:，其中，γ为偏差系数，arctan为反正切函数，long₀为初始位置的经度，lat₀为初始位置的纬度，为预测位置的经度，为预测位置的纬度。 , where γ is the deviation coefficient, arctan is the inverse tangent function, long₀ is the longitude of the initial position, and lat₀ is the latitude of the initial position. is the longitude of the predicted location, is the latitude of the predicted location.10.根据权利要求9所述的基于5G基站的定位方法，其特征在于，所述S44中加权处理公式为：10. The positioning method based on a 5G base station according to claim 9, characterized in that the weighted processing formula in S44 is:，，其中，x为当前时刻的终端位置的经度，y为当前时刻的终端位置的纬度。 , , where x is the longitude of the terminal's location at the current moment, and y is the latitude of the terminal's location at the current moment.