Fusion enhanced positioning method based on 5G and Beidou

Abstract

The invention discloses a fusion enhanced positioning method based on 5G and Beidou, and relates to the technical field of satellite navigation and communication, and the method comprises the steps: collecting multi-system satellite signals, carrying out the signal analysis, and fusing the analyzed signal features into a comprehensive feature vector; calculating the quality level of a satellite signal by using a deep learning model, and selecting a high-quality signal according to the signal quality level; a multi-system fusion positioning algorithm is adopted, high-quality signals are fused, and a preliminary positioning result is calculated; extracting the space-time characteristics of the high-quality signal through signal analysis, and fusing the space-time characteristics with the comprehensive feature vector to form an extended feature vector; and enhancing the preliminary positioning result by using the extended feature vector, sending the enhanced positioning result back to the user terminal, and performing data correction and service optimization. According to the method, through key technologies such as fusion of multi-system signals, intelligent evaluation of signal quality and fusion of time-space characteristics, the positioning precision and robustness are remarkably improved, and particularly, the performance is outstanding in a complex environment.

Description

Technical Field

[0001] The present invention relates to the field of satellite navigation and communication technology, and in particular to a fusion enhanced positioning method based on 5G and Beidou. Background Art

[0002] With the rapid development of mobile communication technology and the widespread application of satellite positioning systems, fusion positioning technology has become one of the key technologies to improve positioning accuracy and reliability. Traditional positioning methods usually rely on a single satellite navigation system, such as GPS or Beidou system, but in complex urban environments, due to the obstruction of high-rise buildings and the influence of multipath effects, traditional positioning methods often cannot meet the needs of high-precision positioning. In recent years, with the popularization of 5G communication technology and the continuous improvement of the Beidou navigation system, fusion positioning technology based on 5G and Beidou has developed rapidly. The 5G network, with its high speed and low latency, provides powerful data transmission capabilities for fusion positioning; the Beidou system, with its high precision and wide coverage, provides a reliable space-time reference for positioning.

[0003] Although the fusion positioning technology based on 5G and Beidou has made certain progress, there are still some challenges and shortcomings. On the one hand, traditional fusion positioning methods usually adopt a fixed weight distribution strategy, which is difficult to adapt to dynamically changing environmental conditions, resulting in limited positioning accuracy. On the other hand, due to the large differences in signal characteristics between different satellite systems, how to effectively fuse multiple satellite signals has become a major problem. In addition, existing technologies often ignore the impact of signal quality on positioning results, resulting in the inability to effectively guarantee the accuracy of positioning results in complex environments. Especially in environments with significant multipath effects such as urban canyons, signal quality evaluation and selection become particularly important, but existing methods often lack effective mechanisms to deal with these problems. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a fusion enhanced positioning method based on 5G and Beidou to solve the problems of inaccurate signal quality assessment and insufficient intelligence of multi-system signal fusion strategy in the prior art.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a fusion enhanced positioning method based on 5G and Beidou, which includes collecting multi-system satellite signals and performing signal analysis, fusing the analyzed signal features into a comprehensive feature vector; using a deep learning model to calculate the quality level of the satellite signal, and selecting a high-quality signal according to the signal quality level; using a multi-system fusion positioning algorithm to fuse the high-quality signals and calculate the preliminary positioning results; extracting the spatiotemporal characteristics of the high-quality signal through signal analysis, fusing the spatiotemporal characteristics with the comprehensive feature vector to form an extended feature vector; using the extended feature vector to enhance the preliminary positioning results, and sending the enhanced positioning results back to the user terminal for data correction and service optimization.

[0008] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the multi-system satellite signals are collected and the signals are analyzed, and the analyzed signal features are integrated into a comprehensive feature vector. The specific steps are as follows:

[0009] Use high-precision 5G base stations and BeiDou-compatible receivers to simultaneously receive BeiDou, GPS, Galileo and GLONASS satellite signals;

[0010] Perform preliminary analysis on the received signal and extract the characteristic vectors of pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio;

[0011] A convolutional neural network is used to fuse pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio into a comprehensive feature vector X.

[0012] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the quality level of satellite signals is calculated using a deep learning model, and the specific steps are as follows:

[0013] The synthesized feature vector is input into the deep learning model to evaluate the signal quality, and the expression is:

[0014]

[0015] in, is the signal quality score predicted by the model, σ is the sigmoid function, O is the length of the convolutional layer output vector, and g(X) i is the eigenvalue after convolution and activation function processing, f is the nonlinear transformation function, b is the output layer bias term, θ is the parameter set of the model, and i is the index variable of the signal;

[0016] Use historical data to train the deep learning model and adjust the deep learning model parameters to minimize the loss function;

[0017] The gradient descent method is used to calculate the gradient of the loss function with respect to the model parameters θ, and the model parameters are updated along the negative direction of the gradient.

[0018] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the high-quality signal is selected according to the signal quality level, and the specific steps are as follows:

[0019] Define a signal standard threshold τ based on application requirements;

[0020] when When , the signal is considered to be a high-quality signal;

[0021] when When , the signal is considered as a low-quality signal;

[0022] High-quality signals are selected from all signals and combined into a high-quality signal set S.

[0023] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the multi-system fusion positioning algorithm is adopted to fuse high-quality signals and calculate the preliminary positioning results. The specific steps are as follows:

[0024] The high-quality signal set S is used to construct the high-quality signal measurement matrix As. In the measurement matrix As, each row corresponds to the observation equation of each signal in s.

[0025] The weight matrix W of high-quality signals is constructed using the reliability function of each signal;

[0026] Based on the high-quality signal observation matrix As and the high-quality signal weight matrix W, a multi-system fusion positioning algorithm is used to calculate the preliminary positioning result, which is expressed as:

[0027]

[0028] Among them, P is the preliminary positioning result, As represents the observation matrix of the observation equation corresponding to each high-quality signal in s in each row, W is the weight matrix of the high-quality signal, and z is the observation value of each high-quality signal.

[0029] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the spatiotemporal characteristics of high-quality signals are extracted by signal analysis, and the spatiotemporal characteristics are integrated with the comprehensive feature vector to form an extended feature vector. The specific steps are as follows:

[0030] The signal analysis method is used to extract the spatiotemporal characteristics from the high-quality signal set S, including the arrival time difference of the signal, the trend of signal strength change, and the Doppler frequency shift change of the signal;

[0031] The spatiotemporal characteristics are fused with the comprehensive feature vector X into an extended feature vector, which is expressed as:

[0032]

[0033] Among them, X is the comprehensive feature vector, T is the spatiotemporal feature vector, X t is the extended eigenvector.

[0034] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the extended feature vector is used to enhance the preliminary positioning result and the final positioning result is calculated. The specific steps are as follows:

[0035] Use a shallow neural network model to expand the feature vector X t Perform nonlinear transformation to enhance the preliminary positioning result P, the expression is:

[0036] P ′ =W o ·σ(W h X t +b h )+b o ;

[0037] Among them, W h is the weight matrix of the hidden layer of the shallow neural network model, b h is the bias vector of the hidden layer of the shallow neural network model, h is the activation vector of the hidden layer of the shallow neural network model, and W o is the weight matrix of the output layer of the shallow neural network model, b o is the bias vector of the output layer of the shallow neural network model, P ′ It is the enhanced positioning result.

[0038] As a preferred solution of the fusion enhanced positioning method based on 5G and Beidou described in the present invention, wherein: the enhanced positioning result is sent to the user terminal for data correction and service optimization, and the specific steps are as follows:

[0039] The enhanced positioning result P ′ Sent back to the user terminal via the 5G network;

[0040] The user terminal receives the positioning result P ′ , and compared with the historical positioning data stored locally, the Kalman filter is used to correct the positioning error. The expression is:

[0041] P cr =KF(P ′ ,P ht );

[0042] Among them, P crrepresents the corrected positioning result, KF represents the Kalman filter, P ht Represents historical positioning data;

[0043] The user terminal will correct the positioning result P cr Send it back to the positioning server to optimize the multi-system fusion positioning algorithm;

[0044] Based on the corrected positioning results, services are optimized for user terminals, including updating map display and route planning.

[0045] In a second aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the fusion enhanced positioning method based on 5G and Beidou as described in the first aspect of the present invention is implemented.

[0046] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein: when the computer program is executed by a processor, any step of the fusion enhanced positioning method based on 5G and Beidou as described in the first aspect of the present invention is implemented.

[0047] The beneficial effects of the present invention are as follows: by collecting multi-system satellite signals and performing signal analysis, the analyzed signal features are fused into a comprehensive feature vector, a deep learning model is used to calculate the quality level of satellite signals and select high-quality signals, a multi-system fusion positioning algorithm is used to fuse high-quality signals to calculate preliminary positioning results, and the temporal and spatial characteristics of high-quality signals are further extracted through signal analysis and fused with the comprehensive feature vector to form an extended feature vector, and finally the extended feature vector is used to enhance the preliminary positioning results. This series of steps effectively improves the accuracy and robustness of positioning, especially in complex environments, and provides users with more accurate and reliable positioning services. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0049] Figure 1 This is a flow chart of the fusion enhanced positioning method based on 5G and Beidou in Example 1. DETAILED DESCRIPTION

[0050] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0053] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and provides a fusion enhanced positioning method based on 5G and Beidou, comprising the following steps:

[0054] S1. Collect multi-system satellite signals and perform signal analysis, and fuse the analyzed signal features into a comprehensive feature vector.

[0055] Going a step further, using high-precision 5G base stations and BeiDou-compatible receivers, it can simultaneously receive BeiDou, GPS, Galileo and GLONASS satellite signals;

[0056] Perform preliminary analysis on the received signal and extract the characteristic vectors of pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio;

[0057] The convolutional neural network is used to fuse pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio into a comprehensive feature vector X, which is expressed as:

[0058]

[0059] Among them, p is pseudorange, d is Doppler frequency shift, s is signal strength, m is multipath effect index, and snr is signal-to-noise ratio;

[0060] It should be noted that pseudorange = propagation time × speed of light + receiver clock error;

[0061] The expression of Doppler frequency shift is: Where v is the relative speed between the receiver and the satellite, c is the speed of light, and f0 is the transmission frequency;

[0062] Signal strength can usually be expressed using the Received Signal Strength Indicator (RSSI);

[0063] The multipath effect index is usually used to quantify the impact of multipath propagation on the signal, and the expression is: Among them, Po m is the main path power, Po t is the total power of all paths;

[0064] SNR is the signal-to-noise ratio, which is the ratio of signal power to noise power, and the expression is: Among them, Po s is the power of the signal, Po n is the power of the noise;

[0065] It should also be noted that these signal features are processed by a convolutional neural network (CNN), first converting each signal feature (pseudorange p, Doppler shift d, signal strength s, multipath effect index m, and signal-to-noise ratio snr) into a corresponding feature map. These feature maps are then convolved through a convolutional layer to capture the local correlation between different features. Next, after being processed by an activation function, these feature maps are further combined and passed through a fully connected layer, and finally fused into a comprehensive feature vector X. This process effectively integrates multiple signal features together, providing rich information for subsequent signal quality assessment and positioning.

[0066] S2. Use the deep learning model to calculate the quality level of the satellite signal and select high-quality signals based on the signal quality level.

[0067] Furthermore, the integrated feature vector is input into the deep learning model to evaluate the signal quality, which is expressed as:

[0068]

[0069] in, is the signal quality score predicted by the model, σ is the sigmoid function, which is used to map the output to 1, indicating the quality of the signal, O is the length of the convolutional layer output vector, and g(X) i is the feature value after convolution and activation function processing, f is a nonlinear transformation function, which is used to perform further nonlinear transformation on the output of the convolution layer to enhance the expression ability of the model, b is the output layer bias term, which is used to adjust the baseline value of the signal quality score, θ is the parameter set of the model, and i is the index variable of the signal;

[0070] Use historical data to train the deep learning model and adjust the deep learning model parameters to minimize the loss function, which is expressed as:

[0071]

[0072] in, is the actual signal quality score, is the signal quality score predicted by the model, N represents the number of samples, and i is the index variable of the signal;

[0073] The gradient descent method is used to calculate the gradient of the loss function with respect to the model parameter θ, and the model parameters are updated along the negative direction of the gradient. The expression is:

[0074]

[0075] Among them, θ (l) is the model parameter at the lth iteration, η is the learning rate, is the loss function L(θ (l) ) about the parameter θ (l) The gradient of θ (l+1) is the model parameter at the l+1th iteration;

[0076] When the gradient approaches zero or the change in parameters in several consecutive iterations is less than a certain threshold, the model is considered to have converged and the iteration stops;

[0077] Furthermore, a signal standard threshold τ is defined based on application requirements, and the signal quality score is compared and the size of the signal standard threshold τ to evaluate the quality of the signal;

[0078] when When , the signal is considered as a high-quality signal and will be selected as a high-quality signal;

[0079] when When , the signal is considered as a low-quality signal and will not be selected;

[0080] High-quality signals are selected from all signals and combined into a high-quality signal set S.

[0081] It should also be noted that application requirements usually refer to the specific usage scenarios and goals of the positioning system, such as the required positioning accuracy, response speed, stability, etc. By setting τ reasonably, it can be ensured that the system only uses high-quality signals for positioning calculations, thereby improving the accuracy and reliability of positioning.

[0082] S3, using multi-system fusion positioning algorithm, integrating high-quality signals, and calculating preliminary positioning results.

[0083] Furthermore, the high-quality signal set S is used to construct the high-quality signal measurement matrix A, in which each row corresponds to the observation equation of each signal in S;

[0084] The expression of the observation matrix A is:

[0085] Where n is the number of observed signals, E is the number of unknown parameters, and a ij represents the contribution of the i-th signal to the j-th unknown parameter;

[0086] The expression of the observation equation is: i =ρ i +cΔt+e i ; Among them, p i is the pseudorange observation value of the ith signal, ρ i is the true distance, c is the speed of light, Δt is the receiver clock error, e i is the observation error of each signal;

[0087] The weight matrix W of high-quality signals is constructed by using the reliability of each high-quality signal in the high-quality signal set S. This matrix is ​​used to represent the contribution of each high-quality signal to the positioning result. The expression is:

[0088]

[0089] Where n is the number of high-quality observation signals, w i is the weight value of the i-th signal;

[0090] Based on the high-quality signal observation matrix A and the high-quality signal weight matrix W, a multi-system fusion positioning algorithm is used to calculate the preliminary positioning result, which is expressed as:

[0091]

[0092] Among them, P is the preliminary positioning result, A represents the observation matrix of the observation equation of each high-quality signal corresponding to S in each row, W is the weight matrix of the high-quality signal, and z is the observation value of each high-quality signal.

[0093] S4. Extract the spatiotemporal characteristics of high-quality signals through signal analysis, and fuse the spatiotemporal characteristics with the comprehensive feature vector to form an extended feature vector.

[0094] Furthermore, the signal analysis method is used to extract the spatiotemporal characteristics from the high-quality signal set S, including the arrival time difference of the signal, the trend of signal strength change, and the Doppler frequency shift change of the signal;

[0095] The spatiotemporal characteristics are fused with the comprehensive feature vector X into an extended feature vector, which is expressed as:

[0096]

[0097] Among them, X is the comprehensive feature vector, T is the spatiotemporal feature vector, X t is the extended eigenvector.

[0098] It should be noted that the signal analysis method extracts spatiotemporal characteristics from the high-quality signal set S, including the arrival time difference of the signal, the trend of signal strength change, and the Doppler frequency shift change of the signal. These characteristics reflect the changes in the spatial and temporal dimensions of the signal, which are crucial to improving positioning accuracy. These spatiotemporal characteristics are fused with the original comprehensive feature vector X to form an extended feature vector X t ,By fusing the spatiotemporal characteristics in this way, the characteristics of the ,signal can be more comprehensively characterized, thus providing richer information ,for subsequent positioning result enhancement.

[0099] S5. Use the extended feature vector to enhance the preliminary positioning result, and send the enhanced positioning result to the user terminal for data correction and service optimization.

[0100] Furthermore, the enhanced positioning result P ′ Sent back to the user terminal via the 5G network;

[0101] The user terminal receives the positioning result P ′ , and compared with the historical positioning data stored locally, the Kalman filter is used to correct the positioning error. The expression is:

[0102] P cr =KF(P ′ ,P ht );

[0103] Among them, P cr represents the corrected positioning result, KF represents the Kalman filter, P ht Represents historical positioning data;

[0104] The user terminal will correct the positioning result P cr Send it back to the positioning server to optimize the multi-system fusion positioning algorithm;

[0105] It should be noted that the user terminal will correct the positioning result P cr The purpose of this process is to use the corrected data to optimize the multi-system fusion positioning algorithm. By feeding back these more accurate positioning results to the server, the positioning algorithm can be adjusted and improved to improve the accuracy of future positioning results. This feedback mechanism helps to continuously iterate and optimize the positioning algorithm, ensuring that it can adapt to different environmental conditions and continuously improve positioning accuracy over time.

[0106] Based on the corrected positioning results, services are optimized for user terminals, including updating map display and route planning.

[0107] It should also be noted that according to the corrected positioning result P cr, the user terminal can perform a series of service optimization measures, such as updating map display, improving route planning, etc. By using more accurate positioning information, the map application can more accurately display the user's location and improve the user's navigation experience. In addition, the route planning function can also use these corrected positioning results to recommend better routes, avoid congested or unsuitable roads, thereby improving the user's travel efficiency and satisfaction. These service optimization measures are designed to enhance the user experience and ensure that users can obtain timely, accurate and practical information.

[0108] This embodiment also provides a computer device, which is suitable for the case of a fusion enhanced positioning method based on 5G and Beidou, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the fusion enhanced positioning method based on 5G and Beidou proposed in the above embodiment.

[0109] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0110] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by the processor, the method for implementing the fusion enhanced positioning method based on 5G and Beidou as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0111] In summary, the present invention collects multi-system satellite signals and performs signal analysis, fuses the analyzed signal features into a comprehensive feature vector, uses a deep learning model to calculate the quality level of satellite signals and selects high-quality signals, uses a multi-system fusion positioning algorithm to fuse high-quality signals to calculate preliminary positioning results, further extracts the spatiotemporal characteristics of high-quality signals through signal analysis and fuses them with the comprehensive feature vector to form an extended feature vector, and finally uses the extended feature vector to enhance the preliminary positioning results. This series of steps effectively improves the accuracy and robustness of positioning, especially in complex environments, and provides users with more accurate and reliable positioning services.

[0112] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A fusion enhanced positioning method based on 5G and Beidou, characterized in that: include, Collect multi-system satellite signals and perform signal analysis, and fuse the analyzed signal features into a comprehensive feature vector; Use deep learning models to calculate the quality level of satellite signals and select high-quality signals based on the signal quality level; Adopt multi-system fusion positioning algorithm, fuse high-quality signals, and calculate preliminary positioning results; Extract the spatiotemporal characteristics of high-quality signals through signal analysis, and fuse the spatiotemporal characteristics with the comprehensive feature vector to form an extended feature vector; The initial positioning result is enhanced using the extended feature vector and the enhanced positioning result is sent back to the user terminal for data correction and service optimization.

2. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 1, characterized in that: The steps of collecting multi-system satellite signals and performing signal analysis, and fusing the analyzed signal features into a comprehensive feature vector are as follows: Use high-precision 5G base stations and BeiDou-compatible receivers to simultaneously receive BeiDou, GPS, Galileo and GLONASS satellite signals; Perform preliminary analysis on the received signal and extract the characteristic vectors of pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio; A convolutional neural network is used to fuse pseudorange, Doppler frequency shift, signal strength, multipath effect index and signal-to-noise ratio into a comprehensive feature vector X.

3. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 2, characterized in that: The specific steps of using the deep learning model to calculate the quality level of satellite signals are as follows: The synthesized feature vector is input into the deep learning model to evaluate the signal quality, and the expression is: in, is the signal quality score predicted by the model, σ is the sigmoid function, O is the length of the convolutional layer output vector, g(X) i is the eigenvalue after convolution and activation function processing, f is the nonlinear transformation function, b is the output layer bias term, θ is the parameter set of the model, and i is the index variable of the signal; Use historical data to train the deep learning model and adjust the deep learning model parameters to minimize the loss function; The gradient descent method is used to calculate the gradient of the loss function with respect to the model parameters θ, and the model parameters are updated along the negative direction of the gradient.

4. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 3 is characterized in that: The steps of selecting a high-quality signal according to the signal quality level are as follows: Define a signal standard threshold τ based on application requirements; when When , the signal is considered to be a high-quality signal; when When , the signal is considered as a low-quality signal; High-quality signals are selected from all signals and combined into a high-quality signal set S.

5. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 4, characterized in that: The multi-system fusion positioning algorithm is used to fuse high-quality signals and calculate the preliminary positioning results. The specific steps are as follows: The high-quality signal set S is used to construct the high-quality signal measurement matrix As. In the measurement matrix As, each row corresponds to the observation equation of each signal in s. The weight matrix W of high-quality signals is constructed using the reliability function of each signal; Based on the high-quality signal observation matrix As and the high-quality signal weight matrix W, a multi-system fusion positioning algorithm is used to calculate the preliminary positioning result, which is expressed as: Among them, P is the preliminary positioning result, As represents the observation matrix of the observation equation corresponding to each high-quality signal in s in each row, W is the weight matrix of the high-quality signal, and z is the observation value of each high-quality signal.

6. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 5, characterized in that: The method of extracting the spatiotemporal characteristics of high-quality signals through signal analysis and fusing the spatiotemporal characteristics with the comprehensive feature vector to form an extended feature vector is as follows: The signal analysis method is used to extract the spatiotemporal characteristics from the high-quality signal set S, including the arrival time difference of the signal, the trend of signal strength change, and the Doppler frequency shift change of the signal; The spatiotemporal characteristics are fused with the comprehensive feature vector X into an extended feature vector, which is expressed as: Among them, X is the comprehensive feature vector, T is the spatiotemporal feature vector, and X t is the extended eigenvector.

7. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 6, characterized in that: The use of the extended feature vector to enhance the preliminary positioning result and calculate the final positioning result is performed in the following specific steps: Use a shallow neural network model to expand the feature vector X t Perform nonlinear transformation to enhance the preliminary positioning result P, the expression is: P ′ =W o ·σ(W h X t +b h )+b o ; Among them, W h is the weight matrix of the hidden layer of the shallow neural network model, b h is the bias vector of the hidden layer of the shallow neural network model, h is the activation vector of the hidden layer of the shallow neural network model, and W o is the weight matrix of the output layer of the shallow neural network model, b o is the bias vector of the output layer of the shallow neural network model, P ′ It is the enhanced positioning result.

8. The fusion enhanced positioning method based on 5G and BeiDou as claimed in claim 7, characterized in that: The enhanced positioning result is sent to the user terminal for data correction and service optimization. The specific steps are as follows: The enhanced positioning result P ′ Sent back to the user terminal via the 5G network; The user terminal receives the positioning result P ′ , and compared with the historical positioning data stored locally, the Kalman filter is used to correct the positioning error. The expression is: P cr =KF(P ′ ,P ht ); Among them, P cr represents the corrected positioning result, KF represents the Kalman filter, P ht Represents historical positioning data; The user terminal will correct the positioning result P cr Send it back to the positioning server to optimize the multi-system fusion positioning algorithm; Based on the corrected positioning results, services are optimized for user terminals, including updating map display and route planning.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the fusion enhanced positioning method based on 5G and Beidou as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fusion enhanced positioning method based on 5G and Beidou as described in any one of claims 1 to 7 are implemented.

Images