Human-vehicle positioning method and system driven by grid coding

Abstract

The present invention discloses a method and system for positioning people and vehicles driven by grid coding, which relates to the field of satellite navigation and location service technology. The method includes: using radar and GIS to obtain topographic information, combining the Beidou coordinate system to perform multi-level spatial grid division, and obtaining a grid set with coding and fuzzy bands. The real-time position data of people and vehicles are collected by Beidou terminals, and fuzzy band intrusion analysis is performed after matching the grids to generate grid codes with normal / abnormal identification. When abnormal, the code is retroactively corrected, and when normal, it is directly used to achieve accurate positioning. The present invention solves the technical problems of poor adaptability of traditional grid coding, imperfect boundary processing and dynamic tracking mechanisms, easy grid jumps, and insufficient accuracy in the process of realizing human and vehicle positioning based on radio wave positioning, and achieves the technical effect of optimizing coding adaptation, boundaries and tracking mechanisms to achieve accurate and continuous positioning of people and vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite navigation and location services, and in particular to a method and system for positioning people and vehicles driven by grid coding. Background Art

[0002] In scenarios such as smart transportation and public safety, accurate positioning of people and vehicles is crucial for management and decision-making, and radio wave positioning technology is a key enabler for achieving this. Existing radio wave-based positioning solutions for people and vehicles struggle with accurate and continuous positioning in complex geographical environments and dynamic scenarios due to poor grid coding adaptability, imperfect boundary processing, and incomplete dynamic tracking mechanisms. Traditional technologies fail to fully integrate topography and landforms to optimize grid division, and lack effective deviation correction strategies. This leads to problems such as grid skipping and insufficient accuracy in positioning data, making it unable to meet the needs of police, transportation, and other fields for full-area, high-precision positioning of people and vehicles. Innovative positioning methods and systems are urgently needed. Summary of the Invention

[0003] This application provides a human-vehicle positioning method and system driven by grid coding, which is used to solve the technical problems of poor adaptability, imperfect boundary processing and dynamic tracking mechanism of traditional grid coding, easy positioning jump and insufficient accuracy in the process of realizing human-vehicle positioning based on radio wave positioning.

[0004] In a first aspect, the present application provides a method for positioning a person and a vehicle driven by a grid code, the method comprising: detecting a target area using a radar, and acquiring topographic information of the target area in combination with a GIS system; dividing the target area into a multi-level spatial grid based on the topographic information and the geographic coordinate system of the Beidou satellite navigation system to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band; collecting the real-time position of a person and a vehicle using a terminal device equipped with a Beidou positioning module to obtain real-time position data of the target person and vehicle; matching the multi-level spatial grid set based on the real-time position data of the target person and vehicle to obtain a matched multi-level spatial grid; performing adaptive fuzzy band intrusion analysis on the multi-level spatial grid based on the real-time position data of the target person and vehicle to obtain an identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification; when the identification grid code is an abnormal identification, performing positioning time backtracking, correcting the identification grid code to obtain a corrected grid code, and using the corrected grid code as the target person and vehicle positioning result; when the identification grid code is a normal identification, using the identification grid code as the target person and vehicle positioning result.

[0005] The second aspect of the present application provides a human-vehicle positioning system driven by grid coding, the system comprising: a terrain information acquisition module, which uses radar to detect the target area and combines the GIS system to obtain the terrain information of the target area; a spatial grid set acquisition module, which is used to combine the terrain information and the geographic coordinate system of the Beidou satellite navigation system to divide the target area into multi-level spatial grids to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band; a human-vehicle real-time position acquisition module, which uses a terminal device equipped with a Beidou positioning module to collect the real-time position of the human and vehicle and obtain the real-time position data of the target human and vehicle; a matching spatial grid acquisition module, which is used to match the real-time position of the target human and vehicle based on the real-time position of the target human and vehicle. The position data is matched with the multi-level spatial grid set to obtain a matched multi-level spatial grid; an identification grid code acquisition module is used to perform adaptive fuzzy band intrusion analysis of the multi-level spatial grid in combination with the real-time position data of the target person and vehicle, and obtain an identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification; a correction grid code acquisition module is used to perform positioning time backtracking when the identification grid code is an abnormal identification, correct the identification grid code, obtain a corrected grid code, and use the corrected grid code as the target person and vehicle positioning result; a person and vehicle positioning result acquisition module is used to use the identification grid code as the target person and vehicle positioning result when the identification grid code is a normal identification.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] This application uses radar, GIS system and Beidou satellite navigation system to divide the target area into multi-level spatial grids, collects human and vehicle position data through terminal equipment, matches grids, analyzes fuzzy band intrusions, and combines historical positioning time series backtracking correction to accurately obtain the dynamic position of people and vehicles, adapt to multi-level spatial positioning needs, and provide reliable data for traffic management and safety control, achieving the technical effect of optimizing coding adaptation, boundaries and tracking mechanisms, and realizing accurate and continuous positioning of people and vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 This is a flow chart of a human-vehicle positioning method driven by grid coding provided in an embodiment of the present application.

[0010] Figure 2 This is a structural diagram of a human-vehicle positioning system driven by grid coding provided in an embodiment of the present application.

[0011] Explanation of the accompanying symbols: terrain information acquisition module 1, spatial grid set acquisition module 2, human and vehicle real-time position acquisition module 3, matching spatial grid acquisition module 4, identification grid code acquisition module 5, correction grid code acquisition module 6, human and vehicle positioning result acquisition module 7. DETAILED DESCRIPTION

[0012] This application provides a human-vehicle positioning method and system driven by grid coding, which is used to solve the technical problems of poor adaptability, imperfect boundary processing and dynamic tracking mechanism of traditional grid coding, easy positioning jump and insufficient accuracy in the process of realizing human-vehicle positioning based on radio wave positioning.

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0014] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0015] Example 1, as Figure 1 As shown, a method for positioning a person and a vehicle driven by grid coding, wherein the method includes:

[0016] Step A100: Use radar to detect the target area and use the GIS system to obtain the topographic information of the target area.

[0017] In the embodiment of the present application, the GIS system is a geographic information system used to obtain topographic information of the target area.

[0018] Specifically, first, the target area is detected by radar, and the radio wave reflection characteristics of the radar are used to obtain the spatial structure data of the target area, such as surface contours, obstacle distribution, preliminary detection results of target positions, etc., to provide basic perception information for spatial modeling of terrain and landforms.

[0019] At the same time, the GIS system is used to obtain topographic and geomorphic information for the target area. This includes extracting elevation data (such as terrain undulations at different altitudes), slope information (the degree of surface inclination), land cover types (such as buildings, vegetation, and water bodies), and geographic boundary data (administrative boundaries, natural geographic boundaries, etc.). This multi-dimensional geographic information is integrated through the spatial analysis capabilities of the GIS system to form a structured topographic and geomorphic dataset.

[0020] Subsequently, the radar data was integrated with topographic information acquired through GIS, and a projection transformation was performed within the geographic coordinate system of the BeiDou satellite navigation system, converting the three-dimensional terrain features into a two-dimensional projection spatial map of the terrain. This process required coordinate system alignment and data registration to ensure precise spatial alignment between the radar and GIS data, resulting in a unified visualization model that encompasses both terrain details and spatial structure.

[0021] Through the collaborative work of radar and GIS systems, multi-dimensional data collection and spatial modeling of the target area's terrain and landforms are achieved, providing accurate geographic feature basis for the subsequent adaptive division of multi-level spatial grids, enabling grid coding to more realistically reflect geographic spatial attributes, and laying the foundation for subsequent improvement in the accuracy of human and vehicle positioning.

[0022] Step A200: Combining the topographic information and the geographic coordinate system of the BeiDou satellite navigation system, the target area is divided into multi-level spatial grids to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band.

[0023] In the embodiment of the present application, the adaptive fuzzy band is due to geographical factors, interference from surrounding equipment, etc., which may cause positioning errors to occur at the edges of some grids. It is necessary to identify the size of the edge area that is prone to occur and use it as an adaptive fuzzy band.

[0024] Optionally, combined with the elevation data, slope information, land cover type and geographic boundary data in the above-mentioned terrain information, a projection transformation is performed in the geographic coordinate system of the Beidou satellite navigation system to obtain a terrain projection space map, and based on this map, spatial grid division template matching is performed to obtain a multi-level matching template set, and the set is used to perform multi-level division of the map to obtain a grid set and encode it, and at the same time, the edge fuzzy band of the grid set is identified to obtain a unique grid code and adaptive fuzzy band corresponding to each grid. The specific steps are described in detail in A210-A240.

[0025] Step A300: Use a terminal device equipped with a Beidou positioning module to collect the real-time location data of people and vehicles to obtain the real-time location data of the target people and vehicles.

[0026] In one embodiment of the present application, first, the Beidou positioning module of a terminal device (such as a smartphone or an in-vehicle terminal) continuously scans for visible satellites, performs carrier phase measurement and pseudorange calculations after capturing signals, and solves the initial position using the least squares method. Subsequently, the device's built-in Kalman filter is used to perform noise reduction processing on the raw data, filtering out high-frequency noise and outliers to improve data smoothness. At the same time, the pre-processed real-time location data of people and vehicles is transmitted to the back-end server via a network communication module (such as 4G / 5G). The server stores the data in chronological order and generates a target trajectory sequence. For example, in urban scenarios, the Beidou module uses a multipath suppression algorithm to reduce errors caused by building reflections, ensuring that stable latitude and longitude data can still be output in areas with dense high-rise buildings.

[0027] This method leverages the BeiDou system's global short message communication capabilities to enable reliable transmission of location data in areas without terrestrial network coverage (such as mountainous areas and deserts), addressing the limitations of traditional GPS, which relies on base stations. Furthermore, through the unique identification of device numbers, it enables the parallel tracking of multiple targets (such as multiple vehicles or pedestrians), supporting real-time monitoring of large groups of people and vehicles.

[0028] Through terminal equipment equipped with Beidou positioning modules, high-frequency and high-precision collection of human and vehicle location data is achieved, providing a reliable data source for subsequent matching analysis based on multi-level spatial grids and adaptive fuzzy band intrusion detection, solving the problems of insufficient accuracy and poor real-time performance of traditional positioning methods in complex environments, and improving the accuracy, continuity and system robustness of human and vehicle positioning.

[0029] Step A400: matching the multi-level spatial grid set based on the real-time position data of the target person or vehicle to obtain a matching multi-level spatial grid.

[0030] Specifically, the longitude and latitude coordinates collected in real time, such as those in the WGS84 coordinate system with meter-level accuracy, are first converted into corresponding grid codes using the BeiDou Grid Location Code standard. This standard, based on a geographic coordinate system, divides the spatial map of topographic projections into multiple levels of grids (e.g., city-level, meter-level, etc.), with each grid corresponding to a unique code. For example, the longitude and latitude coordinates of a vehicle are mapped to a meter-level grid code using the standard, which contains administrative division and spatial location information.

[0031] Secondly, timestamps are used to perform time-series filtering and alignment on location data. Real-time location data is sorted by timestamp (accurate to the millisecond), and valid data segments are filtered using a preset time window (e.g., 5 seconds). This eliminates outdated data due to transmission delays, such as inertial delay. For example, if the same device reports multiple locations within a short period of time (e.g., 2 seconds), only the most recent or most densely populated points are retained for matching after sorting by timestamp, avoiding errors caused by duplicate matching.

[0032] Next, multi-target parallel matching is achieved through device identification. Each terminal device's unique identification number (such as the IMEI) serves as an identification key, separating the location data sequence of the same device from other targets. This ensures that there is no confusion when matching multiple targets (such as multiple vehicles on a road) simultaneously. For example, in urban traffic monitoring scenarios, the location data of thousands of devices can be processed simultaneously, with each device corresponding to an independent matching thread, improving the system's concurrent processing capabilities.

[0033] During the matching process, adaptive fuzzy bands within multi-level grids are used to address positioning errors. When longitude and latitude coordinates fall within the fuzzy band of a particular grid (e.g., ±5 meters around the grid edge), the system triggers a secondary verification process. The system uses historical location sequences (using timestamps to extract the most recent N points, N=5) to determine the target's movement trend. If three consecutive points move in the same direction, the target is prioritized for matching to the adjacent grid. If the position fluctuation is minor, the error is considered normal within the current grid, and the match is maintained on the original grid.

[0034] Through the above steps, accurate mapping from latitude and longitude coordinates, multi-level grid coding and multi-target matching is achieved, solving the problems of multi-target confusion, boundary error sensitivity and lack of real-time performance in traditional methods. The real-time position data of people and vehicles are efficiently mapped to a multi-level grid set, providing a reliable spatial reference for subsequent positioning analysis (such as fuzzy band intrusion detection and time series backtracking correction), and improving the overall accuracy and robustness of the positioning system.

[0035] Step A500: In combination with the real-time position data of the target person or vehicle, perform adaptive fuzzy band intrusion analysis of the multi-level spatial grid to obtain an identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification.

[0036] Specifically, the real-time latitude and longitude coordinates are first mapped to a multi-level spatial grid set. Using the Beidou grid location code standard, the coordinates are converted to a grid code of the corresponding level (e.g., a meter-level grid code). The adaptive fuzzy band for that grid is then determined: the area between the edge of the first spatial grid of interest and the initial edge, with a width determined by the density of historical anomaly locations.

[0037] Next, the real-time location is determined to have entered the fuzzy zone: the spatial relationship between the coordinates and the grid edge and the fuzzy zone boundary is calculated, such as the distance from the point to the line segment. If the distance from the coordinate to the grid edge is less than or equal to the fuzzy zone width (e.g., ≤5 meters), it is considered to have entered the fuzzy zone and a grid code with an abnormality flag is generated. If the distance is greater than the fuzzy zone width, it is marked as normal. This process requires filtering out invalid data based on timestamps, for example, discarding locations with delays exceeding 5 seconds, to ensure that only real-time, valid data is analyzed.

[0038] In multi-target scenarios, the system groups each target by device number, ensuring independent analysis of each target's position sequence. For example, when tracking 100 vehicles simultaneously, the system creates a separate thread for each device number and concurrently determines whether its position intrudes into the corresponding grid's fuzzy zone, preventing data aliasing.

[0039] Through the above steps, dynamic error detection of the real-time position of people and vehicles is achieved, and the spatial relationship between position data and grid fuzzy bands is converted into a clear identification result (normal / abnormal). This improves the positioning system's ability to perceive and respond to boundary errors, providing a key basis for subsequent positioning corrections.

[0040] Step A600: When the identification grid code is an abnormal identification, perform positioning time sequence backtracking, correct the identification grid code, obtain a corrected grid code, and use the corrected grid code as the target person and vehicle positioning result.

[0041] In an embodiment of the present application, positioning timing backtracking is a process of obtaining a preset positioning timing backtracking window when the identification grid code is an abnormal identification, extracting a historical human and vehicle positioning result sequence based on the window, and then correcting the identification grid code based on this sequence to obtain the corrected grid code and use it as the target human and vehicle positioning result.

[0042] Specifically, when the identification grid code is an abnormal identification, a preset positioning time sequence backtracking window is obtained, from which a sequence of historical human and vehicle positioning results is extracted, and the identification grid code is corrected based on the sequence to obtain a corrected grid code as the target human and vehicle positioning result. The specific steps are described in detail in A610-A630.

[0043] Step A700: When the identification grid code is a normal identification, the identification grid code is used as the target person and vehicle positioning result.

[0044] Specifically, first, real-time location data (including latitude and longitude coordinates, timestamp, and device ID) is collected from a terminal device equipped with a Beidou positioning module. These coordinates are then calculated using the Beidou satellite navigation system, achieving accuracy ranging from meters to centimeters. Timestamps are used to ensure data timeliness. Subsequently, the real-time longitude and latitude coordinates are mapped to corresponding grids in a multi-level spatial grid set to obtain a unique grid code.

[0045] Next, an adaptive fuzzy band intrusion analysis is performed: Based on the fuzzy band range pre-determined by the historical location anomaly data for that grid (e.g., the edge indentation area determined by an iterative shrinkage algorithm), the distance from the real-time location to the grid edge is calculated. If the distance is greater than the fuzzy band width (e.g., a preset 5 meters), the location is considered normal and the current grid code is directly output as the location result without triggering the correction mechanism.

[0046] Through the above process, the system directly outputs the grid code when it is marked as normal. Relying on the high-precision data of Beidou positioning, the precise modeling of dynamic fuzzy bands and the multi-dimensional verification mechanism, it ensures the accuracy and stability of the positioning results, and achieves the efficiency and reliability of the human and vehicle positioning results.

[0047] Furthermore, step A300 in the method provided in the embodiment of the present application includes:

[0048] A310: The real-time location data of each person or vehicle in the real-time location data sequence includes latitude and longitude coordinates, timestamp, and device number.

[0049] Specifically, in the grid-coding-driven human-vehicle positioning method, the collection of real-time location data of humans and vehicles relies on terminal devices equipped with Beidou positioning modules, such as smartphones and vehicle-mounted terminals. Its core data elements include latitude and longitude coordinates, timestamps, and device numbers. The acquisition and processing process of each element is as follows:

[0050] Acquiring Longitude and Latitude Coordinates: The Beidou positioning module receives signals from Beidou satellites in the B1 and B2 frequency bands and uses pseudorange measurement and carrier phase differential technology to calculate the target position. The module first captures signals from at least four satellites and calculates the signal transmission delay to determine the pseudorange. Then, using the least squares method to solve geometric equations, it obtains the target's three-dimensional latitude and longitude coordinates with accuracy ranging from meters to centimeters. To enhance anti-interference capabilities, the module incorporates a built-in multipath mitigation algorithm to reduce errors caused by building reflections and terrain obstruction, ensuring stable coordinate output even in complex environments (such as urban canyons and mountainous areas).

[0051] Timestamp Generation: The Beidou system inherently features high-precision timing. The positioning module synchronizes its internal clock with satellite time, automatically recording the current moment upon completion of each positioning solution and generating a timestamp accurate to the millisecond. This timestamp not only marks the moment the position data was collected but also facilitates timing calibration during subsequent trajectory analysis. For example, when performing backtracking for positioning timing, the timestamp is used to align historical data to ensure temporal continuity of the trajectory.

[0052] Device ID: The device number is a unique identifier for a terminal device, such as the International Mobile Equipment Identity (IMEI) or the vehicle terminal ID. It is read and associated by the system during device initialization. This number is transmitted along with location data to the backend server and used to distinguish between different targets, such as multiple vehicles or pedestrians, enabling multi-target parallel tracking. For example, in smart city management scenarios, the device number can be used to quickly filter the travel trajectory of a specific vehicle and, combined with grid coding, perform path analysis.

[0053] Through the above technical means, the accurate collection and unique identification of human and vehicle location data are achieved, solving problems such as multi-target confusion and time synchronization errors in traditional positioning, and improving the real-time, accuracy and multi-target management capabilities of the positioning system.

[0054] Furthermore, step A200 in the method provided in the embodiment of the present application includes:

[0055] A210: Extracting elevation data, slope information, land feature coverage type and geographic boundary data from the topographic information, performing projection transformation in the geographic coordinate system, and obtaining a topographic projection space map.

[0056] A220: Perform spatial grid division template matching based on the topographic projection spatial map to obtain a multi-level matching spatial grid division template set.

[0057] A230: Perform multi-level division of the topographic projection spatial map using the multi-level matching spatial grid division template set to obtain a multi-level spatial grid set, and encode the multi-level spatial grid set to obtain a corresponding grid code.

[0058] A240: Perform edge fuzzy band identification on the multi-level spatial grid set to obtain an adaptive fuzzy band corresponding to each multi-level spatial grid.

[0059] In the embodiment of the present application, elevation data is numerical information used to describe the altitude of each point on the surface, reflecting the vertical undulation characteristics of the terrain, such as the altitude differences of mountains, valleys, and plains.

[0060] Optionally, first extract the elevation, slope, land cover type, and geographic boundary data from the terrain information and project it into the Beidou geographic coordinate system to generate a projected spatial map of the terrain. This step converts three-dimensional terrain features into a two-dimensional spatial model. For example, using DEM data from a GIS system to obtain elevation and combining it with slope analysis tools to generate slope information, accurately aligning the terrain data with geographic coordinates and providing a spatial foundation for gridding.

[0061] Next, a multi-level feature extractor is used to extract multi-level features from the terrain projection spatial map, which are matched with the template features in the preset template library. The template with the highest matching degree is included in the multi-level matching spatial grid division template set. The specific steps are described in detail in A221-A222.

[0062] Then, the terrain projection spatial map is divided into multiple levels using the matching spatial division template, as shown in Table 1: First, the target area (such as the urban area) is coarsely divided into city-level grids (approximately 10 km × 10 km) to cover the entire area; within the city-level grid, the corresponding template is matched according to the terrain (such as mountainous areas, plains, and densely built-up areas) and target density (such as commercial areas with heavy pedestrian and vehicle traffic, using street / community-level grids of approximately 100 m × 100 m; suburban areas with sparse pedestrian and vehicle traffic, using district / county-level grids of approximately 1 km × 1 km), (for example, in mountainous areas with complex terrain, meter-level templates are prioritized to improve positioning accuracy), and the grids are further divided into district / county-level and street / community-level grids; for areas requiring high-precision positioning (such as streets and intersections in the urban core area), meter-level precision grids (10 m × 10 m or 1 m × 1 m) are further subdivided to achieve fine coverage of local areas. During the division process, the nested structure is strictly followed to ensure that the upper and lower grids are completely contained in space (for example, a 1km×1km district / county-level grid contains multiple 100m×100m street-level grids, and each street-level grid contains multiple meter-level grids), forming a pyramid-like multi-level spatial grid collection.

[0063] Furthermore, multi-level spatial grid sets are encoded. Based on the Beidou grid location code standard, each grid is assigned a unique code. For example, the coding of meter-level precision grids (1m×1m) is accurate to the smallest granularity. Through the nested structure coding rules, it is ensured that the code contains the hierarchical information of all upper-level grids (for example, the meter-level grid code can be parsed to the street / community, district / county, and city-level grids), achieving one-code positioning of the entire domain. The coding rules ensure traceability (for example, the spatial scope and hierarchical affiliation of the grid can be queried through the code) and uniqueness (each grid code is globally unique), supporting high-precision human and vehicle positioning scenarios, such as the precise positioning and trajectory tracking of vehicles within a 1m×1m grid.

[0064] Finally, the first multi-level spatial grid is extracted, its historical positioning abnormal position is mined and the edge is identified to obtain the first adaptive fuzzy band, and then all grids are traversed to obtain the adaptive fuzzy band corresponding to each multi-level spatial grid. The specific steps are described in detail in A241-A243.

[0065] Through the above steps, a terrain spatial map is first constructed, a template is adapted to grid division and encoding, and finally, edge fuzzy bands are identified, achieving multi-level, terrain-aware grid division. This process combines geographic features with the BeiDou coordinate system to optimize the spatial adaptability of the grid encoding, providing a foundation for subsequent positioning data processing and improving the accuracy and stability of human and vehicle positioning in complex scenarios.

[0066] Table 1: Multi-level spatial grid division information table

[0067] Grid Level Division basis Grid size Application Scenario Nested relationships City-level grid Administrative level and full coverage requirements About 10km×10km Coarse-grained division of urban areas Includes district / county level grids District / county-level grid Topography (mountainous areas, suburban areas, etc.), target density (areas with few people and vehicles) About 1km×1km The city-level grid is further divided into suburban areas, such as Contains street / neighborhood level grids Street / community-level grid Topography (densely built areas, etc.), target density (areas with high pedestrian and vehicle traffic, such as commercial areas) About 100m×100m Within district / county-level grids, such as urban commercial areas Contains meter-level precision grid Meter-level precision grid High-precision positioning requirements (streets and intersections in urban core areas, etc.) 10m×10m, 1m×1m Fine positioning of urban core areas No sub-level nesting

[0068] Furthermore, step A220 in the method provided in the embodiment of the present application includes:

[0069] A221: Perform multi-level feature extraction on the topographic projection spatial map using a multi-level feature extractor set to obtain a multi-level spatial map feature set.

[0070] A222: Matching the multi-level spatial graph feature set with the template spatial graph features in the preset template library respectively, and adding the template corresponding to the maximum matching degree into the multi-level matching spatial grid division template set.

[0071] Specifically, first, a multi-level feature extractor set (neural network model) is constructed. The model adopts a multi-branch architecture and takes the terrain projection spatial map (including rasterized geographic data such as elevation, slope, land cover type, geographic boundary data, etc.) as input. It is first converted into a four-dimensional tensor such as 256×256×4 (corresponding to elevation, slope, land cover and geographic boundary respectively) and normalized; then the basic features are extracted through the initial layer of the convolutional neural network (CNN) (such as 4 ResNet modules) to capture the global structural information of the terrain; after the shared feature layer, multiple parallel branches corresponding to the city level, district level, street level, meter level, etc. are set. Each branch contains spatial pyramid pooling (SPP), which captures terrain features of different scales through pooling kernels of different sizes (such as 16×16 pooling for the city-level branch and 1×1 pooling for the meter-level branch), and the SE (Squeeze-and-Excitation) module is used to weight the features of each channel to highlight the key features. Features (such as areas with drastic elevation changes) are then mapped into 256-dimensional fixed-length feature vectors through a fully connected layer to form hierarchical feature encoding. During training, the aggregated labeled samples containing hierarchical labels and grid division results in step A222-3 are used to design a multi-task loss function, including feature reconstruction loss to ensure that the hierarchical features can accurately restore the original terrain data, hierarchical classification loss to constrain the model to correctly distinguish terrain features at different levels, and template matching loss to guide the model to extract the most relevant feature dimensions of the preset template library. The Adam optimizer is used to train samples in batches of 32, and the network parameters are adjusted for 100 epochs to converge the loss function. Finally, multi-level spatial graph feature vectors are output, such as city-level feature vectors representing the overall terrain complexity of the region and meter-level feature vectors representing the local terrain curvature, providing a quantitative basis for subsequent template matching. Through multi-branch and multi-task design, the model automatically learns the inherent laws of terrain features at different levels, improving the accuracy and adaptability of template matching.

[0072] Next, a spatial map is projected onto the current topography, and the trained model is used to extract multi-level features. For example, given terrain data for a mountainous area (including high slopes and complex landforms), the model outputs city-level features (e.g., if the area is mountainous, the overall grid must fit within a 10km×10km template) and meter-level features (e.g., if a local area requires a fine 10m×10m grid to cover the terrain), forming a multi-level spatial map feature set.

[0073] Finally, these features are matched against a pre-set template library (the specific construction process is detailed in A222-1-A222-5, which contains a mapping between terrain features at each level and corresponding grid templates). The similarity of the feature vectors is calculated (e.g., using Euclidean distance or cosine similarity). The template with the highest matching score is selected for each level (e.g., city level, meter level). For example, a 10 km × 10 km template is used for the city level, and a 10 m × 10 m template is used for the meter level. This creates a multi-level matching spatial grid partitioning template set.

[0074] By building a multi-level neural network feature extractor and performing similarity matching with a preset template library, the problem of insufficient scene adaptation of traditional template matching is solved, accurate multi-level template support is provided for grid division, and the grid division accuracy and system adaptability of human and vehicle positioning in complex terrain are improved.

[0075] Furthermore, step A222 in the method provided in the embodiment of the present application includes:

[0076] A222-1: Obtain multiple sample topography projection space maps and corresponding multiple sample space grid division results, wherein each sample topography projection space map has a hierarchical identifier.

[0077] A222-2: Traverse the plurality of sample topography projection space maps to perform feature extraction and obtain a plurality of sample space map features.

[0078] A222-3: In combination with the hierarchical identification, the multiple sample terrain and landform projection spatial maps are aggregated from the two dimensions of spatial map feature similarity and the same level to obtain multiple aggregated sample terrain and landform projection spatial map sets, and the multiple sample space grid division results are mapped and aggregated to obtain multiple aggregated sample space grid division result sets.

[0079] A222-4: Calculate the mean of the multiple aggregated sample spatial graph feature sets corresponding to the multiple aggregated sample terrain projection spatial graph sets to obtain multiple template spatial graph features.

[0080] A222-5: Perform mean processing on the plurality of aggregated sample spatial grid division result sets to obtain a plurality of spatial grid division templates, map and associate the plurality of template spatial graph features with the plurality of spatial grid division templates to obtain a preset template library.

[0081] Specifically, we first obtain multiple sample topographic projection spatial maps and the corresponding sample spatial grid division results. Each sample is annotated with a hierarchical identifier (such as city level, district / county level, street / community level, meter level, etc.). For example, we collect terrain data at different levels, such as urban central areas (meter level), mountainous areas (10m level), and suburban areas (100m level), to form a sample set covering multi-scale terrain features.

[0082] Next, the projected spatial maps of all sample terrain features are traversed, and a feature extractor is used to extract terrain features (such as elevation, slope, feature density, and geographic boundary curvature) to generate corresponding sample spatial map features. For example, high slopes and complex boundaries are extracted for mountainous samples, while low slopes and regular boundaries are extracted for plain samples, forming a multi-dimensional feature vector.

[0083] Then, combined with the hierarchical identifier, samples are aggregated based on two dimensions: spatial map feature similarity and the same hierarchy. Samples at the same hierarchy with feature similarity above a threshold (e.g., cosine similarity > 0.8) are grouped together to form a set of aggregated sample topographic projection spatial maps. At the same time, the corresponding sample spatial grid division results are mapped and aggregated, such as by taking the intersection or average of the grid division parameters of the same group of samples to obtain multiple sets of aggregated sample spatial grid division results. For example, all meter-level mountain samples are aggregated into one group, and their corresponding grid division templates all use a fine-grained (10m×10m) size.

[0084] Afterwards, each aggregated sample set is statistically processed: first, based on the hierarchical identification and spatial map feature similarity, the sample topography projection spatial maps with the same hierarchical level and similar features are aggregated into several sets, such as the city-level mountain sample set and the meter-level plain sample set. At the same time, the corresponding sample space grid division results are mapped and aggregated. Subsequently, for each aggregated terrain feature set, such as the elevation, slope, and feature coverage data of all mountain samples in a certain aggregate set, its mean data is calculated, specifically including the average elevation, average slope, average feature density, average geographic boundary curvature, etc., to form a template spatial map feature vector that can characterize this type of terrain feature. For example, high slope + low feature density corresponds to the mountain template feature.

[0085] Finally, the spatial gridding results from multiple aggregated samples are averaged, such as by calculating average grid edge length and hierarchical nesting rules. This generates a spatial gridding template that maps one-to-one with the template spatial graph features, such as a meter-level fine grid template for mountainous areas. Finally, the features and templates are associated and stored to construct a preset template library. This process abstracts and standardizes terrain features through statistical modeling, providing a quantitative basis for template matching during subsequent gridding and improving the accuracy of gridding template adaptation for complex terrain.

[0086] Through the above steps, based on multi-level aggregation and statistical analysis of terrain features, a library of preset templates containing multiple scales and terrain sensitivity was constructed. This provided support for quickly matching the optimal template during subsequent grid division, and improved the adaptability and positioning accuracy of the human-vehicle positioning system in complex terrain.

[0087] Furthermore, step A240 in the method provided in the embodiment of the present application includes:

[0088] A241: Extracting a first multi-level spatial grid from the multi-level spatial grid set, performing historical positioning anomaly position mining on the first multi-level spatial grid to obtain a first historical positioning anomaly position set.

[0089] A242: Perform abnormal position edge identification on the first historical positioning abnormal position set to obtain a first adaptive fuzzy band.

[0090] A243: Traverse the multi-level spatial grid set to identify edge fuzzy bands, and obtain an adaptive fuzzy band corresponding to each multi-level spatial grid.

[0091] In an embodiment of the present application, the abnormal historical positioning position refers to a set of coordinate points in the historical positioning data where the actual position of the person or vehicle does not match the grid code, fluctuates frequently, or triggers a grid jumping phenomenon due to factors such as grid boundary ambiguity, positioning system errors, or transmission delays.

[0092] Specifically, first, the first multi-level spatial grid is extracted in sequence from the multi-level spatial grid set (for example, 100m is processed first). A density clustering algorithm (such as DBSCAN) is used to mine outlier locations based on the historical positioning data of that grid, including latitude and longitude coordinates and timestamps. The algorithm first sets a neighborhood radius (e.g., 5 meters) and a minimum number of samples (e.g., 10 points). The neighborhood density of each data point is calculated, and points with a density below the threshold are marked as outliers. Coordinate points with position fluctuations exceeding a preset threshold (e.g., two consecutive positionings span multiple grid levels) or triggering frequent grid code switching (e.g., switching ≥3 times in a short period of time) are then screened and overlaid with the density clustering results to form the first set of historical positioning outlier locations. These outliers are primarily distributed within the grid edges, with a significantly higher spatial distribution density than within the grid interior. This reflects areas of concentrated positioning deviations due to grid boundary ambiguity, providing data support for the subsequent dynamic construction of edge fuzzy bands.

[0093] Next, the first spatial grid edge of the first multi-level spatial grid is extracted, and the first abnormal edge neighborhood is constructed according to the preset bandwidth combined with the first historical positioning abnormal position set. By iteratively limiting the edge and comparing the neighborhood density, the area between the target edge and the initial edge is used as the first adaptive fuzzy band. The specific steps are described in detail in A242-1-A242-5.

[0094] Furthermore, for other grids in the multi-level spatial grid set, the same principle is applied. The historical positioning anomaly location set for each grid is extracted in sequence, and the edge reduction and neighborhood density analysis process is repeated to determine the adaptive fuzzy band for each grid. This process leverages the statistical properties of historical positioning data to dynamically identify the boundary error ranges of different grids due to factors such as terrain and signal quality, thus avoiding the blindness of traditional fixed buffers.

[0095] Through the edge fuzzy band identification mechanism driven by historical positioning anomaly data, dynamic quantification and buffering of grid boundary errors are achieved, so that the adaptive fuzzy band of each grid accurately matches the actual positioning error range, enhancing the stability and reliability of the human-vehicle positioning system in complex environments.

[0096] Furthermore, step A242 in the method provided in the embodiment of the present application includes:

[0097] A242-1: Extract the first spatial grid edge of the first multi-level spatial grids.

[0098] A242-2: According to a preset bandwidth, in the first multi-level spatial grid, combined with the first historical positioning abnormal position set, construct a first abnormal edge neighborhood of the edge of the first spatial grid.

[0099] A242-3: Narrow the first spatial grid edge toward the inside of the grid according to a preset bandwidth to obtain an iterative first spatial grid edge, and construct an iterative abnormal edge neighborhood of the iterative spatial grid edge.

[0100] A242-4: Determine whether the neighborhood density of the iterative abnormal edge neighborhood is greater than or equal to the neighborhood density of the first abnormal edge neighborhood. If so, continue to shrink the iterative first spatial grid edge toward the inside of the grid according to the preset bandwidth until the preset number of iterations is met to obtain the target first spatial grid edge.

[0101] A242-5: Use the area from the edge of the target first spatial grid to the edge of the first spatial grid as the first adaptive fuzzy band.

[0102] In this embodiment of the present application, the preset bandwidth is a pre-set distance range used to construct an outlier edge neighborhood. A neighborhood is an area defined by the preset bandwidth, based on the grid edge. Neighborhood density is a statistical measure of the area of ​​the neighborhood.

[0103] In one embodiment, the first spatial grid edge of the first multi-level spatial grid, i.e., the initial grid boundary, is first extracted as a benchmark for analyzing boundary errors. Based on a preset bandwidth (e.g., 5 meters), historically located outlier locations within this grid whose distance from the first spatial grid edge is less than or equal to the preset bandwidth are screened and incorporated into the neighborhood of the first outlier edge. The density of this neighborhood is then calculated (e.g., the number of outlier locations per unit area, represented by the statistic "number of outlier locations per unit area" / neighborhood area). This density reflects the concentration of errors near the initial boundary.

[0104] Next, the first spatial grid edge is constricted inwards according to a preset bandwidth (e.g., by 5 meters at a time) to generate an iterative first spatial grid edge. A new iterative abnormal edge neighborhood is constructed with the same bandwidth. The density of the iterative neighborhood is recalculated. If this density is greater than or equal to the density of the first abnormal edge neighborhood, the currently constricted edge is still within the error concentration region and the constriction operation needs to be repeated (e.g., setting a preset number of iterations to 5). As the iterations proceed, when the neighborhood density stops increasing or reaches the preset number of iterations, the constriction is stopped and the edge at this point is determined to be the target first spatial grid edge.

[0105] Ultimately, the area between the edge of the target first spatial grid and the edge of the initial first spatial grid (the width of which is the sum of the iterative reduction amounts) is defined as the first adaptive fuzzy band. For example, if the preset bandwidth is 5 meters and the iterative reduction is repeated twice, the fuzzy band width is 10 meters, which encompasses the primary distribution range of boundary errors in historical positioning.

[0106] Through neighborhood density analysis and iterative edge reduction of historical positioning anomalies, the adaptive construction of the grid boundary fuzzy zone is achieved, which improves the continuity and accuracy of human and vehicle positioning in the grid boundary area and reduces the impact of grid jumping on trajectory tracking.

[0107] Furthermore, step A600 in the method provided in the embodiment of the present application includes:

[0108] A610: Get the preset positioning timing backtracking window.

[0109] A620: Extracting a historical human and vehicle positioning result sequence based on the preset positioning time series backtracking window.

[0110] A630: Correct the identification grid code based on the historical human and vehicle positioning result sequence to obtain the corrected grid code.

[0111] In the embodiment of the present application, the positioning time series backtracking window is a preset time interval or data sequence range, which is used to extract the historical human and vehicle positioning result sequence to correct the current abnormal identification grid code.

[0112] Optionally, first, a preset positioning timing lookback window is obtained. Those skilled in the art can set a lookback window of fixed length based on the positioning data update frequency and the allowable delay range, such as the last 10 seconds or a time interval containing 20 positioning points. This window parameter can be determined through statistical analysis of historical data to ensure coverage of time periods that may be affected by delays.

[0113] Next, extract the historical positioning results sequence for people and vehicles. Based on a preset window, retrieve the historical positioning results associated with the current anomaly grid code from the database (which stores historical positioning results sequences for people and vehicles, including longitude and latitude coordinates, corresponding grid codes, and timestamps). This creates a chronological sequence consisting of longitude and latitude coordinates, corresponding grid codes, and timestamps. For example, if the timestamp of the current anomaly location is T, set the window from T-10 seconds to T, and extract the 10 historical positioning points and their grid codes within that interval.

[0114] Finally, the historical grid code sequence is extracted based on the historical human and vehicle positioning result sequence, the main code is traversed and identified, and the consistency is determined with the identification grid code. If it passes, the identification grid code is used as the correction result; if it fails, the main code is used as the correction result. The specific steps are described in detail in A631-A633.

[0115] Through historical data mining, master code identification and consistency correction mechanism of preset time series windows, this method solves the positioning deviation problem caused by inertial delay and boundary error. Finally, through dynamic analysis and statistical modeling of time series data, it achieves the improvement of the accuracy of human and vehicle positioning results and trajectory continuity.

[0116] Furthermore, step A630 in the method provided in the embodiment of the present application includes:

[0117] A631: Extracting a historical grid code sequence based on the historical human and vehicle positioning result sequence.

[0118] A632: Traverse the historical grid code sequence to identify the main code, and perform consistency determination on the identification grid code based on the identified main code. When the consistency determination passes, use the identification grid code as the revised grid code.

[0119] A633: When the consistency determination fails, the main code is used as the modified grid code.

[0120] In one embodiment, a sequence of historical human and vehicle positioning results is first obtained. These sequences contain latitude and longitude coordinates, corresponding grid codes, and timestamps. The length of these sequences is determined by a preset window, such as selecting the last five positioning points or data within the last 10 seconds to form a chronological sequence. For example, based on the timestamp T of the current anomaly location, historical positioning data from T-10 seconds to T is extracted to construct a historical grid code sequence S, which contains the grid code information for each positioning point.

[0121] Next, count the frequency of occurrence of historical grid codes. Traverse the sequence S and count each grid code Count the number of occurrences of , for example, generate a frequency mapping table FreqMap= . For example, in a certain sequence, the code Appeared 3 times, Appeared 1 time, If a word appears once, FreqMap will record the frequency and provide data support for subsequent main code identification.

[0122] Then, identify the dominant grid code (main code). Select the code with the highest frequency from the FreqMap as the main code. If there are codes with the same frequency, the code at the end of the sequence (closest in time) is preferred to enhance the temporal correlation. For example, when When the frequency of occurrence is the highest, That is , represents the dominant grid code in the historical trajectory.

[0123] After that, the abnormal points are removed. Codes with large spatial differences (e.g., distances exceeding two grid spans) are treated as jump points and removed from subsequent corrections to improve code stability. This step, by setting a spatial jump threshold, selects historical codes that match trajectory trends and optimizes the basic data for primary code identification.

[0124] Then determine the consistency between the current identification code and the main code. Check the current abnormal identification grid code Is it equal to If they are consistent, it means that although the current abnormal mark is marked, it is consistent with the historical trajectory trend and can be directly confirmed; if they are inconsistent, it means that the current position deviates from the historical trajectory and needs to be corrected.

[0125] Finally, the modified grid code is generated. When the consistency judgment is passed, As a correction grid code; if it fails, As a modified grid code For example, if and Inconsistency, That is , as the final positioning result at the current moment, ensuring that the positioning code is consistent with the historical trajectory trend.

[0126] Through the above steps, the historical grid code sequence is first extracted. Then, through frequency statistics, master code identification, and consistency determination, the current anomaly identification grid code is corrected. This process leverages the temporal characteristics of historical positioning data to effectively correct the positioning code of instantaneous anomalies, improving the accuracy and stability of human and vehicle positioning, making the positioning results more consistent with historical trajectories, reducing positioning errors caused by local anomalies, achieving historical data-driven positioning code optimization, and enhancing the robustness of the positioning system.

[0127] In summary, the human-vehicle positioning method driven by grid coding provided by the embodiments of the present application has the following technical effects:

[0128] This application uses the Beidou positioning module to collect real-time location data of people and vehicles (including latitude and longitude coordinates, timestamps, and device numbers), obtains identification grid codes through multi-level spatial grid matching and adaptive fuzzy band intrusion analysis, and combines historical positioning result sequences for time series backtracking correction, thereby accurately realizing the positioning of people and vehicles driven by grid codes, making the positioning results of people and vehicles in complex environments more accurate and reliable, and achieving the technical effect of optimizing code adaptation, boundaries and tracking mechanisms, and realizing accurate and continuous positioning of people and vehicles.

[0129] Example 2, as Figure 2 As shown, based on the same inventive concept as the aforementioned embodiment 1, the embodiment of the present application provides a human-vehicle positioning system driven by grid coding, the system comprising:

[0130] The topography and landform information acquisition module 1 uses radar to detect the target area and combines the GIS system to obtain the topography and landform information of the target area.

[0131] The spatial grid set acquisition module 2 is used to combine the terrain information and the geographic coordinate system of the Beidou satellite navigation system to perform multi-level spatial grid division on the target area to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band.

[0132] The real-time position acquisition module 3 of the person and vehicle uses the terminal device equipped with the Beidou positioning module to collect the real-time position of the person and vehicle and obtain the real-time position data of the target person and vehicle.

[0133] The matching spatial grid acquisition module 4 matches the multi-level spatial grid set based on the real-time position data of the target person and vehicle to obtain a matching multi-level spatial grid.

[0134] The identification grid code acquisition module 5 is used to combine the real-time position data of the target person and vehicle, perform adaptive fuzzy band intrusion analysis of the multi-level spatial grid, and obtain the identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification.

[0135] The modified grid code acquisition module 6 is used to perform positioning time sequence backtracking when the identification grid code is an abnormal identification, correct the identification grid code, obtain the modified grid code, and use the modified grid code as the target person and vehicle positioning result.

[0136] The human-vehicle positioning result acquisition module 7 is configured to use the identification grid code as the target human-vehicle positioning result when the identification grid code is a normal identification.

[0137] Furthermore, the real-time location acquisition module 3 is used to perform the following steps:

[0138] The real-time location data of each person or vehicle in the real-time location data sequence includes latitude and longitude coordinates, timestamp, and device number.

[0139] Furthermore, the spatial grid set acquisition module 2 is configured to perform the following steps:

[0140] Extract elevation data, slope information, land feature coverage type and geographic boundary data from the topographic information, perform projection transformation in the geographic coordinate system, and obtain a topographic projection space map; perform spatial grid division template matching based on the topographic projection space map to obtain a multi-level matching spatial grid division template set; use the multi-level matching spatial grid division template set to perform multi-level division on the topographic projection space map to obtain a multi-level spatial grid set, and encode the multi-level spatial grid set to obtain a corresponding grid code; perform edge fuzzy band identification on the multi-level spatial grid set to obtain an adaptive fuzzy band corresponding to each multi-level spatial grid.

[0141] Furthermore, the spatial grid set acquisition module 2 is configured to perform the following steps:

[0142] A multi-level feature extractor set is used to perform multi-level feature extraction on the terrain projection spatial map to obtain a multi-level spatial map feature set; based on the multi-level spatial map feature set, the template spatial map features in the preset template library are matched respectively, and the template corresponding to the maximum matching degree is added to the multi-level matching spatial grid division template set.

[0143] Furthermore, the spatial grid set acquisition module 2 is configured to perform the following steps:

[0144] Acquire multiple sample topography and geomorphology projection space maps and corresponding multiple sample space grid division results, wherein each sample topography and geomorphology projection space map has a hierarchical identifier; traverse the multiple sample topography and geomorphology projection space maps to perform feature extraction and obtain multiple sample space map features; in combination with the hierarchical identifier, aggregate the multiple sample topography and geomorphology projection space maps from two dimensions, namely, spatial map feature similarity and same hierarchy, to obtain multiple aggregated sample topography and geomorphology projection space map sets, and map and aggregate the multiple sample space grid division results to obtain multiple aggregated sample space grid division result sets; perform mean calculation on the multiple aggregated sample space map feature sets corresponding to the multiple aggregated sample topography and geomorphology projection space map sets respectively to obtain multiple template space map features; perform mean processing on the multiple aggregated sample space grid division result sets to obtain multiple space grid division templates, map and associate the multiple template space map features with the multiple space grid division templates to obtain a preset template library.

[0145] Furthermore, the spatial grid set acquisition module 2 is configured to perform the following steps:

[0146] A first multi-level spatial grid is extracted from the multi-level spatial grid set, and historical positioning abnormal position mining is performed on the first multi-level spatial grid to obtain a first historical positioning abnormal position set; abnormal position edge identification is performed on the first historical positioning abnormal position set to obtain a first adaptive fuzzy band; the multi-level spatial grid set is traversed to perform edge fuzzy band identification to obtain an adaptive fuzzy band corresponding to each multi-level spatial grid.

[0147] Furthermore, the spatial grid set acquisition module 2 is configured to perform the following steps:

[0148] Extract the first spatial grid edge of the first multi-level spatial grid; construct a first abnormal edge neighborhood of the first spatial grid edge in combination with the first historical positioning abnormal position set in the first multi-level spatial grid according to a preset bandwidth; constrict the first spatial grid edge toward the inside of the grid according to the preset bandwidth to obtain an iterative first spatial grid edge, and construct an iterative abnormal edge neighborhood of the iterative spatial grid edge; determine whether the neighborhood density of the iterative abnormal edge neighborhood is greater than or equal to the neighborhood density of the first abnormal edge neighborhood, and if so, continue to constrict the iterative first spatial grid edge toward the inside of the grid according to the preset bandwidth until a preset number of iterations is met to obtain a target first spatial grid edge; and use the area from the target first spatial grid edge to the first spatial grid edge as a first adaptive fuzzy band.

[0149] Furthermore, the modified grid code acquisition module 6 is configured to perform the following steps:

[0150] A preset positioning time sequence backtracking window is obtained; a historical human and vehicle positioning result sequence is extracted based on the preset positioning time sequence backtracking window; and the identification grid code is corrected based on the historical human and vehicle positioning result sequence to obtain the corrected grid code.

[0151] Furthermore, the modified grid code acquisition module 6 is configured to perform the following steps:

[0152] A historical grid code sequence is extracted based on the historical human and vehicle positioning result sequence; the historical grid code sequence is traversed to identify the main code, and the identification grid code is subjected to consistency judgment based on the main code obtained by the identification. When the consistency judgment passes, the identification grid code is used as the revised grid code; when the consistency judgment fails, the main code is used as the revised grid code.

[0153] The grid coding driven human and vehicle positioning system provided in the embodiment of the present invention can execute the grid coding driven human and vehicle positioning method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0154] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0155] The above specific embodiments do not constitute a limitation to the scope of protection of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of this application should be included in the scope of protection of this application. In some cases, the actions or steps recorded in this application can be performed in an order different from that in the embodiments and can still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A human-vehicle positioning method driven by grid coding, characterized in that: The method comprises: Use radar to detect the target area and combine it with the GIS system to obtain the topographic information of the target area; Combining the topographic information with the geographic coordinate system of the BeiDou satellite navigation system, the target area is divided into a multi-level spatial grid to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band. The adaptive fuzzy band is due to geographical factors and interference from surrounding equipment, which may cause positioning errors at the edges of some grids. It is necessary to identify the size of the edge area that is prone to positioning errors and use it as the adaptive fuzzy band; Use terminal equipment equipped with Beidou positioning modules to collect real-time location data of people and vehicles, and obtain real-time location data of target people and vehicles; Matching the multi-level spatial grid set based on the real-time position data of the target person and vehicle to obtain a matching multi-level spatial grid; In combination with the real-time position data of the target person or vehicle, performing adaptive fuzzy band intrusion analysis of the multi-level spatial grid to obtain an identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification; When the identification grid code is an abnormal identification, the positioning time sequence is backtracked, the identification grid code is corrected to obtain a corrected grid code, and the corrected grid code is used as the target person and vehicle positioning result; When the identification grid code is a normal identification, the identification grid code is used as the target person and vehicle positioning result; Performing edge fuzzy band identification on the multi-level spatial grid set to obtain an adaptive fuzzy band corresponding to each multi-level spatial grid includes: Extracting a first multi-level spatial grid from the multi-level spatial grid set, performing historical positioning anomaly position mining on the first multi-level spatial grid to obtain a first historical positioning anomaly position set; Performing abnormal position edge identification on the first historical positioning abnormal position set to obtain a first adaptive fuzzy band; The multi-level spatial grid set is traversed to perform edge fuzzy band identification, and an adaptive fuzzy band corresponding to each multi-level spatial grid is obtained.

2. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 1, wherein: The real-time location data of each person or vehicle in the real-time location data sequence includes latitude and longitude coordinates, timestamp, and device number.

3. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 1, wherein: Combining the topographic information with the geographic coordinate system of the BeiDou satellite navigation system, the target area is divided into a multi-level spatial grid to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band, including: Extracting elevation data, slope information, land cover type and geographic boundary data from the topographic information, performing projection transformation in the geographic coordinate system, and obtaining a topographic projection space map; Perform spatial grid division template matching based on the topographic projection spatial map to obtain a multi-level matching spatial grid division template set; Performing multi-level division on the topographic projection spatial map using the multi-level matching spatial grid division template set to obtain a multi-level spatial grid set, and encoding the multi-level spatial grid set to obtain a corresponding grid code; The edge fuzzy band is identified on the multi-level spatial grid set to obtain an adaptive fuzzy band corresponding to each multi-level spatial grid.

4. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 3, wherein: Performing spatial grid division template matching based on the topographic projection spatial map to obtain a multi-level matching spatial grid division template set includes: Performing multi-level feature extraction on the topographic projection spatial map using a multi-level feature extractor set to obtain a multi-level spatial map feature set; Based on the multi-level spatial graph feature set, the template spatial graph features in the preset template library are matched respectively, and the template corresponding to the maximum matching degree is added into the multi-level matching spatial grid division template set.

5. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 4, characterized in that: include: Acquire multiple sample topography projection space maps and corresponding multiple sample space grid division results, wherein each sample topography projection space map has a level identifier; Traversing the plurality of sample topographic projection space maps to perform feature extraction and obtain a plurality of sample space map features; In combination with the hierarchical identifier, the plurality of sample topography and landform projection spatial maps are aggregated from two dimensions: spatial map feature similarity and same hierarchical level, to obtain a plurality of aggregated sample topography and landform projection spatial map sets; and the plurality of sample space grid division results are mapped and aggregated to obtain a plurality of aggregated sample space grid division result sets; performing mean calculation on a plurality of aggregated sample spatial graph feature sets corresponding to the plurality of aggregated sample topography projection spatial graph sets to obtain a plurality of template spatial graph features; Mean processing is performed on the plurality of aggregated sample spatial grid division result sets to obtain a plurality of spatial grid division templates, and the plurality of template spatial graph features and the plurality of spatial grid division templates are mapped and associated to obtain a preset template library.

6. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 1, wherein: Performing abnormal position edge identification on the first historical positioning abnormal position set to obtain a first adaptive fuzzy band includes: Extracting a first spatial grid edge of the first multi-level spatial grid; According to a preset bandwidth, combining the first historical positioning abnormal position set in the first multi-level spatial grid to construct a first abnormal edge neighborhood at the edge of the first spatial grid; The first spatial grid edge is constricted toward the inside of the grid according to a preset bandwidth to obtain an iterative first spatial grid edge, and an iterative abnormal edge neighborhood of the iterative spatial grid edge is constructed; Determine whether the neighborhood density of the iterative abnormal edge neighborhood is greater than or equal to the neighborhood density of the first abnormal edge neighborhood. If so, continue to shrink the iterative first spatial grid edge toward the interior of the grid according to a preset bandwidth until a preset number of iterations is met, thereby obtaining a target first spatial grid edge. The area from the edge of the target first spatial grid to the edge of the first spatial grid is used as a first adaptive fuzzy band.

7. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 1, wherein: When the identification grid code is an abnormal identification, the positioning time sequence is backtracked, the identification grid code is corrected to obtain a corrected grid code, and the corrected grid code is used as the target person and vehicle positioning result, including: Get the preset positioning timing backtracking window; Extracting a historical sequence of human and vehicle positioning results based on the preset positioning time sequence backtracking window; The identification grid code is corrected based on the historical human and vehicle positioning result sequence to obtain the corrected grid code.

8. The method for positioning a person and a vehicle driven by grid coding as claimed in claim 7, characterized in that: Correcting the identification grid code based on the historical human and vehicle positioning result sequence to obtain the corrected grid code includes: Extracting a historical grid code sequence according to the historical human and vehicle positioning result sequence; Traversing the historical grid code sequence to identify the main code, and performing consistency determination on the identification grid code based on the identified main code, and when the consistency determination passes, using the identification grid code as the revised grid code; When the consistency determination fails, the main code is used as the modified grid code.

9. The human-vehicle positioning system driven by grid coding is characterized by: A system for implementing the human-vehicle positioning method driven by grid coding according to any one of claims 1 to 8, comprising: The topography and landform information acquisition module uses radar to detect the target area and combines it with the GIS system to obtain the topography and landform information of the target area; A spatial grid set acquisition module is used to combine the terrain information and the geographic coordinate system of the Beidou satellite navigation system to divide the target area into multi-level spatial grids to obtain a multi-level spatial grid set, wherein each multi-level spatial grid includes a unique grid code and an adaptive fuzzy band. The adaptive fuzzy band is due to geographical factors and interference from surrounding equipment, which may cause positioning errors at the edges of some grids. It is necessary to identify the size of the edge area that is prone to positioning errors and use it as the adaptive fuzzy band; The real-time location acquisition module for people and vehicles uses a terminal device equipped with a Beidou positioning module to collect the real-time location of people and vehicles and obtain the real-time location data of the target people and vehicles; A matching spatial grid acquisition module matches the multi-level spatial grid set based on the real-time position data of the target person and vehicle to obtain a matching multi-level spatial grid; an identification grid code acquisition module, configured to perform adaptive fuzzy band intrusion analysis of the multi-level spatial grid in combination with the real-time position data of the target person or vehicle, and obtain an identification grid code, wherein the identification grid code includes a normal identification and an abnormal identification; A modified grid code acquisition module is used to perform positioning time sequence backtracking when the identification grid code is an abnormal identification, correct the identification grid code, obtain a modified grid code, and use the modified grid code as the target person and vehicle positioning result; The human and vehicle positioning result acquisition module is used to use the identification grid code as the target human and vehicle positioning result when the identification of the identification grid code is a normal identification.

Images