A method and system for user road-aware positioning in a heterogeneous network

Abstract

The application discloses a user road perception positioning method in a heterogeneous network, and proposes a heterogeneous network road perception positioning method based on road and sub-road segment two-scale feature matching. The method mainly comprises a double-scale signal feature extraction process and a segmented road perception positioning process. In the double-scale signal feature extraction process, the received signal reference power values from the heterogeneous base stations are segmented, and the received signal reference power value features of the road sequence and the sub-segment sequence are extracted respectively. In addition, the curve fitting technology is applied to map the received signal reference power values to the position of each sub-segment. In the segmented road perception positioning process, the UE index of the road sequence and the sub-segment sequence is determined through feature matching. Finally, the position of the UE is obtained through the regression of the fitting curve in the sub-segment.

Description

Technical Field

[0001] This invention belongs to the field of base station signal-based positioning, and particularly relates to a user road perception positioning method in a heterogeneous base station system by analyzing the reference power values ​​of signals received from multiple base stations. Background Technology

[0002] In various industries, including public safety, commercial operations, and network service construction, many applications rely on user location for their functionality. While satellite navigation systems like GPS offer high positioning accuracy in open outdoor spaces, the long propagation distance between satellites and terminal devices, along with obstacles, makes GPS-based positioning impossible in tunnels, overpasses, congested urban areas, and indoor spaces. This makes GPS-independent positioning mechanisms increasingly important. With the advent of the 5G era, heterogeneous network base station systems can provide seamless wired network coverage for user terminals to meet users' demands for high data rates, low latency, and low packet loss rates. Furthermore, the demand for ubiquitous network connectivity from user terminals is growing rapidly. These factors provide favorable conditions for research into user positioning via mobile terminal signal reception.

[0003] Existing base station signal-based positioning methods are mainly divided into fingerprint-based positioning, ID-based positioning, and distance-based positioning. Fingerprint-based positioning can accurately determine the terminal's location by comparing the terminal signal with a predefined database, even in indoor environments with small grid sizes. However, the positioning accuracy of this method is limited by matching accuracy and fingerprint granularity. The accuracy of similarity matching in fingerprint positioning relies excessively on the number of base stations or WiFi access points that the user can detect, while fine-grained fingerprints increase computational complexity but lead to longer positioning times. ID-based positioning technology can locate the user equipment's location using the location of the serving base station, resulting in low computational complexity and short positioning time. However, the positioning accuracy of this method is limited by the base station's coverage radius. To improve positioning accuracy, ranging positioning methods based on distance estimation between at least three base stations are used to obtain the terminal's location. This distance estimation can be achieved by measuring received signal strength, angle of arrival (AOA), or time of arrival (ToA). However, the performance of this positioning method typically depends on the line-of-sight (LoS) link between the base station and the user equipment. This makes distance-based positioning unable to continuously locate the terminal when it leaves an area where multiple base stations overlap. In real-world scenarios, mobile terminals are primarily located on roads in outdoor environments or in corridors in indoor environments. In congested urban environments, existing positioning methods cannot simultaneously achieve both high positioning accuracy and low latency, both of which are crucial for applications such as vehicle navigation. This necessitates the development of a fast and efficient positioning mechanism to accomplish the task of highly reliable and low-latency road perception.

[0004] To address the high-mobility positioning problem for outdoor users on roads, this invention discloses a road-aware positioning method in heterogeneous networks. This method rapidly obtains the user's location on the road and within sub-segments through multi-granularity road segment signal feature matching. Furthermore, after locating the sub-segment, the user's latitude and longitude coordinates within that sub-segment can be obtained through curve regression. This method effectively improves the road positioning speed for outdoor users, eliminates the need for grid fingerprint construction, and reduces fingerprint construction and maintenance costs. It extracts received signal strength information from the signal quality report provided by the user to the base station. The positioning mechanism does not require complex hardware support, making it simple to implement and capable of simultaneously locating multiple data points efficiently. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a mobile user positioning method based on segmented features of user terminal received signals in road positioning scenarios.

[0006] The user road perception and localization method in heterogeneous networks designed in this invention is as follows:

[0007] Step 1, Dual-Scale Data Segmentation. Using the start and end points of each road's geographical location, segmented road-scale sequences are obtained from the initial reference signal received power sampling sequence. Gradient feature segmentation is performed on the drive test signal data from multiple base stations in a heterogeneous network, separating sub-segment sequences from each road sequence. The segmentation points between road segments and sub-segment sequences within a road are called road segmentation points and sub-segment segmentation points, respectively.

[0008] Step 2, Segmented Feature Extraction. Differential features are extracted from the road-scale and sub-segment-scale signal sequences generated in Step 1. For each signal sequence to be represented, a feature set is first calculated, consisting of statistical and trend features related to the sequence data. Then, differential features are selected from the feature set to represent the sequence uniquely. This feature representation method can highlight effective features and suppress useless features. Differential feature selection is achieved using the information gain feature selection method.

[0009] Step 3: Establish a dual-scale feature list. Map the received power of the reference signal in the sub-segment sequence to the location using curve fitting, and store the fitting coefficients in the feature table. The dual-scale feature table is established with the information in the following order: signal features at the road scale extracted in Step 2, signal features at the sub-segment scale, and fitting coefficients.

[0010] Step 4, User Signal Segment Feature Extraction. The reference signal received power reported by the terminal is sampled using equal time windows to obtain the terminal's target reference signal received power sequence. Since the reference signal received power sequence of a user equipment within a time window may contain multiple road or sub-road segment sequences, the user equipment's reference signal received power sequence is first divided into multiple sub-segment sequences based on the road segmentation points and sub-road segmentation points obtained in Step 1. Feature set extraction is performed on each sub-segment sequence;

[0011] Step 5: Dual-scale road feature matching. Different roads and sub-segments within the feature table have different representational features, and these discrepancies are selected from the feature set vector. Based on the features of the signal segment to be located obtained in Step 4, the corresponding discrepancies are calculated and selected from the road and sub-segment features in the feature table. Then, the user equipment features are matched with the features in the data table to find the most similar road segment and sub-segment features.

[0012] Step 6: The index positions of the most similar road matching features and sub-segment matching features obtained in Step 5 within the feature table represent the user's road location and sub-segment location. Then, the latitude and longitude coordinates of the signal points are regressed using the fitted curves within the sub-segments to obtain the latitude and longitude coordinates corresponding to each signal sampling point of the user.

[0013] Based on the same inventive concept, this invention also provides a user road perception and positioning system in a heterogeneous network, which is characterized by including the following modules:

[0014] The dual-scale data segmentation module uses the starting and ending points of each road's geographical location to obtain segmented road scale sequences from the initial reference signal received power sampling sequence. It performs gradient feature segmentation on road test signal data from multiple base stations in a heterogeneous network, and this gradient feature segmentation yields sub-segment scale sequences. The segmentation points between road segments and sub-segment sequences within roads are called road segmentation points and sub-segment segmentation points.

[0015] The segmented feature extraction module extracts differential features from the generated signal sequences at the road scale and sub-segment scale.

[0016] The dual-scale feature list module maps the received power of the reference signal in the segment-scale sequence to the location through curve fitting, and stores the fitting coefficients in the feature table. The information in the feature table includes the extracted signal features at the road scale, the signal features at the segment scale, and the fitting coefficients.

[0017] The user signal segmentation feature extraction module samples the received power of the reference signal reported by the terminal through equal time windows to obtain the received power sequence of the reference signal to be located of the terminal, and performs segmentation and feature extraction according to the dual-scale data segmentation module and the segmentation feature extraction module.

[0018] The dual-scale road feature matching module calculates the difference features between the signal segment to be located in the dual-scale feature list module and the road and sub-segment in the feature table, and then selects the most similar road segment and sub-segment by matching the features of the user equipment with the features in the data table.

[0019] The user positioning module and the dual-scale road feature matching module obtain the index positions of the most similar road matching features and sub-segment matching features in the feature table, which are the user's road position and sub-segment position. Then, the latitude and longitude coordinates of the signal points are obtained by regressing the fitting curves within the sub-segments. The latitude and longitude coordinates of each signal sampling point of the user are then obtained.

[0020] Based on the same inventive concept, the present invention also designs an electronic device, characterized in that it comprises:

[0021] One or more processors;

[0022] Storage device for storing one or more programs;

[0023] When one or more programs are executed by the one or more processors, the one or more processors perform the operations of the user road perception and localization method in a heterogeneous network.

[0024] Based on the same inventive concept, the present invention also designs a computer-readable medium storing a computer program, wherein the program, when executed by a processor, implements the operations performed by the user road perception and positioning method in a heterogeneous network.

[0025] The advantages of this invention are:

[0026] By dividing the original road test signal sequence data into road segments and sub-segments, and extracting differential features from the resulting road segment signal sequences and sub-segment signal sequences, a road location feature list ranging from coarse-grained to fine-grained is established. Road signal sequences are described using features at both the road-scale and sub-segment-scale, improving the accuracy of feature description.

[0027] By using dual-scale feature matching and curve fitting, the location of the user equipment can be quickly determined, including the road, the sub-segments within the road, and the latitude and longitude coordinates within the sub-segments.

[0028] Sampling signals from heterogeneous base stations for positioning can achieve seamless positioning. However, most existing positioning technologies rely on the signal values ​​of neighboring base stations that the user can detect. This results in a limited number of detectable base station signals, especially in situations with strong attenuation due to obstacles in 5G millimeter waves, leading to decreased positioning accuracy.

[0029] This invention discloses a user road-aware positioning method in heterogeneous networks, enabling dual-scale positioning of mobile terminals on roads. With the widespread deployment of heterogeneous network systems in 5G and future 6G network communication systems, heterogeneous base stations can be used to overcome the positioning discontinuity problem caused by signal attenuation in traditional positioning. First, the initial reference signal received power sampling sequence in the heterogeneous base station is segmented and features extracted to establish a dual-scale feature table; then, a polynomial curve is fitted to map the reference signal received power value to the position of each sub-segment sequence. Next, through feature matching and curve regression, indicators such as the road sequence, sub-segment sequence, and latitude and longitude values ​​are obtained sequentially. Because the positioning matching utilizes the different features of the segmented sequence at two scales, rather than the invariant features at a single reference point, the proposed positioning method can accurately locate mobile terminals from coarse-grained to fine-grained levels, with low latency. Attached Figure Description

[0030] Figure 1 This is a diagram illustrating an application scenario of the present invention.

[0031] Figure 2 This is a schematic diagram of the workflow or implementation steps of the present invention.

[0032] Figure 3 This is a schematic diagram of the road and sub-segment scale division of the present invention.

[0033] Figure 4 This is a simulation diagram of the positioning error of the present invention.

[0034] Figure 5 This is a simulation diagram of the positioning time of the present invention.

[0035] Figure 6 This is a schematic diagram of the gradient feature-based segmentation method of the present invention. Detailed Implementation

[0036] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0038] The following disclosure provides numerous different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of the invention. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0039] In complex and congested urban environments, the task of road-aware localization is to accurately locate a person's position on a road or corridor in a short time, whether indoors or outdoors. However, traditional fingerprint and distance-based localization methods struggle to achieve highly reliable and low-latency road localization because the terminal cannot detect enough base station signals to ensure positioning accuracy and cannot effectively reduce computational complexity to save positioning time.

[0040] This invention discloses a user road perception and localization method in heterogeneous networks, and its application scenarios are as follows: Figure 1 As shown, Figure 2 The detailed implementation steps are as follows:

[0041] Step 1, Dual-Scale Data Segmentation. Using the start and end points of each road's location marked on an existing map, a segmented road-scale sequence is obtained from the initial reference signal received power sampling sequence. This road sequence is the road segment. Then, the signal sequence of each road segment is segmented using a hierarchical classification algorithm based on gradient features, thus dividing the road sequence into multiple sub-segment sequences. The gradients of adjacent signal values ​​within these sub-segments are similar, while the gradient values ​​between adjacent sub-segments are different. The dividing points between road segments and their sub-segment sequences are called road dividing points and sub-segment dividing points, respectively.

[0042] Step 2, Segmented Feature Extraction. For the road-scale and sub-segment-scale signal sequences generated in Step 1, differential features are extracted based on information gain in information theory. Differential features are extracted from a feature set based on gradient features to characterize each road segment and sub-segment sequence. This method ensures that the selected differential features are similar within sub-segments but different between road segments, thus increasing the diversity between road segments. For each signal sequence to be characterized, statistical and trend features related to the sequence data are first calculated to form a feature set. Then, differential features are selected from the feature set to perform unique characterization of the sequence. This feature characterization method can highlight effective features and suppress useless features.

[0043] Step 3: Establish a dual-scale feature list. The received power values ​​of the reference signals in the sub-segment sequence are curve-fitted using the least squares method to map the multi-base station signal values ​​to their locations, and the fitting coefficients are stored in the feature table. A dual-scale feature table is established, with the information in the table in the following order: signal features at the road scale extracted in Step 2, signal features at the sub-segment scale, and fitting coefficients.

[0044] Step 4, User Signal Segment Feature Extraction. The reference signal received power reported by a single terminal to be located is sampled using an equal time window sampling method to obtain the reference signal received power sequence for that terminal. Without sampling, the directly obtained user signal values ​​are numerous and contain duplicates, necessitating sampling before location. Since the reference signal received power sequence of a user equipment within a time window may contain multiple road or sub-road segment sequences, the user equipment's reference signal received power sequence is first divided into multiple sub-segment sequences based on the road segmentation points and sub-road segmentation points obtained in Step 1. Feature set extraction is performed on each sub-segment sequence; in this embodiment, the sampling interval is 1 second during simulation, and the time window is at least 4 seconds.

[0045] Step 5: Dual-scale road feature matching. Different roads and sub-segments within the feature table have different representational features, and these discriminatory features are selected from the feature set vector. Based on the features of the signal segment to be located obtained in Step 4, the corresponding discriminatory features of the roads and sub-segments in the feature table are calculated and selected. Then, by calculating the Euclidean distance between the user equipment features and the features in the data table, the corresponding matching probability is obtained to find the most similar road segment and sub-segment features.

[0046] Step 6: The index positions of the most similar road matching features and sub-segment matching features obtained in Step 5 within the feature table represent the user's road location and sub-segment location. Then, the latitude and longitude coordinates of the signal points are regressed using the fitted curves within the sub-segments to obtain the latitude and longitude coordinates corresponding to each signal sampling point of the user.

[0047] Figure 5 This is the cumulative probability distribution curve of the positioning error obtained after implementing the present invention, and the curve is compared with the positioning errors of traditional fingerprint positioning and distance positioning. The comparison results show that the positioning error of the present method is small and the positioning performance is good.

[0048] Figure 4 This is a graph showing the change in positioning time as a function of the positioning window (i.e., the size of the data to be located) after the implementation of this invention. Since distance-based positioning takes a relatively long time, its performance is compared with fingerprint positioning methods. The comparison results show that the positioning time of this invention is shorter and less affected by the size of the positioning window, demonstrating the high efficiency of dual-scale positioning.

[0049] This invention proposes a differential feature selection mechanism, which extracts features from a feature set to accurately represent different signal sequences, compresses the original signal sequence values ​​into feature vectors with obvious differences, reduces localization complexity and storage space.

[0050] This invention proposes a road-aware positioning method using dual-scale feature matching to achieve rapid and accurate location of the user's location on the road and within specific road segments. The location of the identified sub-segments is further determined by performing position regression on the mapped signals within the road using a signal-position mapping curve.

Claims

1. A method for user road awareness positioning in a heterogeneous network, characterized in that, Includes the following steps: Step 1, dual-scale data segmentation: using the starting and ending points of each road's geographical location, obtain the segmented road scale sequence from the initial reference signal received power sampling sequence. Perform gradient feature segmentation on the road test signal data from multiple base stations in the heterogeneous network. This gradient feature segmentation yields the sub-segment scale sequence. The segmentation points between the road segment and the sub-segment sequence within the road are called the road segmentation point and the sub-segment segmentation point. Step 2, segmented feature extraction: extract the differential features from the road-scale and sub-segment-scale signal sequences generated in Step 1; Step 3: Establish a dual-scale feature list. Map the received power of the reference signal in the sub-segment scale sequence to the location through curve fitting, and store the fitting coefficients in the feature table. The information in the feature table is in the order of signal features in the road scale, signal features in the sub-segment scale, and fitting coefficients extracted in Step 2. Step 4, user signal segmentation feature extraction: the received power of the reference signal reported by the terminal is sampled through equal time windows to obtain the received power sequence of the reference signal to be located of the terminal, and segmentation and feature extraction are performed according to steps 1 and 2. Step 5: Dual-scale road feature matching. Based on the features of the signal segment to be located obtained in Step 4, calculate the difference features corresponding to the roads and sub-segments in the feature table, select the most similar road segments and sub-segments by matching the features of the user equipment with the features in the data table. Step 6: The index positions of the most similar road matching features and sub-segment matching features obtained in Step 5 in the feature table are the user's road position and sub-segment position. Then, the latitude and longitude coordinates of the signal points are obtained by regressing the fitting curves within the sub-segments. This yields the latitude and longitude coordinates of each signal sampling point for the user.

2. The user road perception and localization method in a heterogeneous network according to claim 1, characterized in that: In step 2, information gain is used to extract differential features.

3. The method of claim 1, wherein the method further comprises: The specific process of step 2 is as follows: For each signal sequence to be represented, first calculate the statistical and trend features related to the sequence data to form a feature set, and then select the differential features from the feature set to perform unique characterization of the sequence.

4. The method of claim 1, wherein: In step 3, the reference signal received power value in the sub-segment sequence is curve-fitted using the least squares method.

5. The method of claim 1, wherein: In step 5, Euclidean distance is used to calculate the corresponding matching probability and find the most similar road segments and sub-segment features.

6. A user road-aware positioning system in a heterogeneous network, characterized by Includes the following modules: The dual-scale data segmentation module uses the starting and ending points of each road's geographical location to obtain segmented road scale sequences from the initial reference signal received power sampling sequence. It performs gradient feature segmentation on road test signal data from multiple base stations in a heterogeneous network, and this gradient feature segmentation yields sub-segment scale sequences. The segmentation points between road segments and sub-segment sequences within roads are called road segmentation points and sub-segment segmentation points. The segmented feature extraction module extracts differential features from the generated signal sequences at the road scale and sub-segment scale. The dual-scale feature list module maps the received power of the reference signal in the segment-scale sequence to the location through curve fitting, and stores the fitting coefficients in the feature table. The information in the feature table includes the extracted signal features at the road scale, the signal features at the segment scale, and the fitting coefficients. The user signal segmentation feature extraction module samples the received power of the reference signal reported by the terminal through equal time windows to obtain the received power sequence of the reference signal to be located of the terminal, and performs segmentation and feature extraction according to the dual-scale data segmentation module and the segmentation feature extraction module. The dual-scale road feature matching module calculates the difference features between the signal segment to be located in the dual-scale feature list module and the road and sub-segment in the feature table, and then selects the most similar road segment and sub-segment by matching the features of the user equipment with the features in the data table. The user positioning module and the dual-scale road feature matching module obtain the index positions of the most similar road matching features and sub-segment matching features in the feature table, which are the user's road position and sub-segment position. Then, the latitude and longitude coordinates of the signal points are obtained by regressing the fitting curves within the sub-segments. The latitude and longitude coordinates of each signal sampling point of the user are then obtained.

7. The user road-aware positioning system in a heterogeneous network of claim 6, wherein: In step 3, the reference signal received power value in the sub-segment sequence is curve-fitted using the least squares method.

8. The user road-aware positioning system in a heterogeneous network of claim 6, wherein: The dual-scale road feature matching module uses Euclidean distance to calculate the corresponding matching probability and find the most similar road segments and sub-segments.

9. An electronic device, comprising: include: One or more processors; Storage device for storing one or more programs; When one or more programs are executed by the one or more processors, the one or more processors perform the operations performed by the method as described in any one of claims 1-5.

10. A computer readable medium having stored thereon a computer program, characterized in that: When the program is executed by the processor, it performs the operations performed by the method as described in any one of claims 1-5.

Images