Mobile beidou high-precision positioning and tracking system for vehicle intelligent monitoring

Abstract

The application discloses a mobile Beidou high-precision positioning tracking system for vehicle intelligent monitoring and relates to the technical field of vehicle intelligent control.The system comprises a multi-source sensing collection module, a vehicle-level hardware is used to collect various data, and the processed data is transmitted to a multi-dimensional fusion modeling module; the module constructs a three-dimensional continuous field basic model and generates a unified feature set, which is transmitted to an error trend prediction module; the module transmits the predicted error trend to a trajectory precision correction module; after the trajectory correction, the data is transmitted to a high-precision data application module, and finally stored and uploaded to a vehicle intelligent monitoring platform; the application constructs a three-dimensional continuous field model by integrating multi-source sensing data, realizes multi-dimensional data coupling and cooperation, and improves positioning precision; the trajectory is accurately corrected by pre-determining the positioning error trend and combining a composite correction algorithm; hierarchical encryption storage and multi-mode communication are adopted to guarantee data safe transmission, improve vehicle driving safety and intelligence, and broaden system application scenarios.

Description

Technical Field

[0001] This invention relates to the field of vehicle intelligent control technology, specifically a mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring. Background Technology

[0002] With the rapid development of vehicle intelligence technology, high-precision positioning and multi-dimensional collaborative control have become the core support for improving vehicle driving safety and intelligence. The integrated application of visual perception technology and the concept of multi-agent collaboration has made it possible for vehicles to obtain comprehensive driving-related data. Currently, during vehicle operation, it is necessary to integrate information from multiple aspects such as positioning, attitude, dynamics, and road environment to cope with complex and ever-changing road conditions and environmental interference. The gradual maturity of technologies such as Beidou positioning, inertial navigation, and environmental perception has laid the foundation for multi-source data collection. However, how to effectively integrate scattered multi-source data, construct a unified model that reflects the relationship between roads, vehicles, and the environment, and achieve deep data collaboration and efficient utilization has become a key issue that needs to be addressed in the industry. Against this backdrop, the development of a mobile Beidou high-precision positioning and tracking system for intelligent vehicle monitoring can provide accurate data support and decision-making basis for intelligent vehicle control, which is in line with the overall trend of intelligent transportation development.

[0003] Traditional vehicle positioning and control systems often rely on a single data source or simple data overlay processing, making it difficult to fully consider the interplay of road geometry, vehicle driving status, and environmental interference factors. This results in positioning accuracy being easily constrained by external conditions. While some technologies attempt to integrate multi-source data, they lack systematic modeling of the coupling errors among these three sources, failing to accurately quantify the comprehensive impact of various factors on the positioning results. Consequently, the models suffer from insufficient stability and adaptability. In terms of error handling, traditional technologies tend to focus on post-error correction, lacking the ability to anticipate error evolution trends. This leads to delayed correction actions, making it difficult to cope with dynamically changing driving scenarios. Furthermore, traditional correction methods often employ fixed strategies, failing to adapt to differences in vehicle type and driving scenarios, thus failing to meet the high-precision positioning requirements of various usage needs. In addition, traditional systems lack standardized processing for data storage and transmission, making it difficult to guarantee data security and integrity, which affects the reliability of subsequent applications. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring. This system integrates multi-dimensional data from positioning, inertial navigation, dynamics, road conditions, and the environment through a multi-source sensing and acquisition module. After standardized processing, the data is transmitted to a multi-dimensional fusion modeling module to construct a three-dimensional continuous field basic model. An error trend prediction module predicts the evolution trend of positioning errors and iteratively optimizes the model. A feedforward-feedback composite correction algorithm then accurately corrects trajectory and inertial navigation drift. Finally, a high-precision data application module ensures secure data storage and efficient transmission. The system achieves deep coupling and dynamic optimization of multi-source data, adapting to different vehicle types and complex driving scenarios, significantly improving positioning accuracy and stability. This provides reliable technical support for intelligent vehicle monitoring and intelligent driving, helping to enhance vehicle driving safety and intelligence.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring, the system comprising:

[0006] Multi-source perception and acquisition module: Equipped with automotive-grade acquisition hardware, it simultaneously acquires BeiDou positioning trajectory, inertial navigation drift, vehicle dynamics, road real scene images, environmental perception, and high-precision map road geometry data. After timestamp calibration, anomaly filtering, intelligent completion, and standardization processing, it is transmitted to the multi-dimensional fusion modeling module.

[0007] Multidimensional fusion modeling module: Receives standardized data, extracts road geometric features, quantifies environmental interference factors, associates vehicle dynamics data with road geometric features, constructs a three-dimensional continuous field basic model of road-vehicle-environment through a three-dimensional coupled error modeling algorithm, fuses and generates a unified feature set, and transmits it to the error trend prediction module;

[0008] Error trend prediction module: Receives modeling data and unified feature set, converts them into corresponding point clouds with BeiDou positioning trajectory data, performs ICP precise registration and extracts registration features, iteratively optimizes the three-dimensional continuous field basic model, uses error evolution trend prediction algorithm to calculate the current positioning matching deviation, predicts the positioning error evolution trend in the next 3-5 seconds, integrates the calculation results and transmits them to the trajectory precise correction module;

[0009] The trajectory precision correction module receives the transmitted calculation and prediction results, sets the positioning error threshold and performs dual correction trigger judgment, extracts the original coordinates, modeling data and prediction data of Beidou positioning, corrects the Beidou positioning trajectory and inertial navigation drift through the feedforward-feedback composite correction algorithm, and transmits the verification data, correction data and correction log to the high-precision data application module after verification.

[0010] High-precision data application module: Receives verification data and correction logs, integrates data from the entire system process to form a complete dataset, stores it in a hierarchical and encrypted manner using automotive-grade storage units, and uploads trajectory data to the vehicle intelligent monitoring platform for storage via a multi-mode wireless communication unit.

[0011] Furthermore, the multi-source sensing and acquisition module completes data acquisition through its automotive-grade acquisition hardware. Specifically, the BeiDou positioning module acquires BeiDou positioning trajectory data, the inertial navigation module acquires inertial navigation drift data, the CAN bus acquisition unit acquires vehicle dynamics data, the camera acquires real-world road images, the environmental sensor acquires environmental perception data, and the high-precision map offline retrieval unit retrieves high-precision map road geometry data. During data processing, the timestamps of all acquired data are first calibrated based on the timestamps of the BeiDou positioning module. Abnormal data caused by signal obstruction and sensor failure are filtered out. Intelligent completion is achieved through adjacent valid data. Finally, all data is uniformly converted into the common encoding format of vehicle-mounted equipment to complete the standardization process.

[0012] Furthermore, the road geometric features extracted by the multi-dimensional fusion modeling module include: lane lines, road edge lines, overpass outlines, tunnel entrance / exit outlines, guardrail orientation, and road markings. During the extraction process, the spatial coordinates, morphological orientation, and size parameters of each feature are acquired simultaneously. When quantifying environmental interference factors, fog and rain intensity, road surface condition, and road surface material are classified into different levels. Fog and rain intensity is divided into 5 levels based on visibility, road surface condition is divided into 4 levels based on slipperiness, and road surface material is divided into 3 levels based on friction coefficient. Each level corresponds to a quantitative value of 1-5, 1-4, and 1-3, respectively. When associating vehicle dynamics data with road geometric features, the vehicle dynamics data is spatiotemporally matched with the road geometric features of the corresponding driving segment by combining the vehicle's real-time driving speed and steering angle.

[0013] Furthermore, when the multi-dimensional fusion modeling module constructs a three-dimensional continuous field basic model of road-vehicle-environment using the three-dimensional coupling error modeling algorithm, it first uses the extracted road geometric features, the quantitatively processed environmental interference factors, the correlated vehicle dynamics data, and the road geometric feature data as input data for the three-dimensional coupling error modeling algorithm. Then, it sets the parameters of the three-dimensional coupling error modeling algorithm, where the dynamic weight coefficients of road, vehicle, and environmental factors all range from 0 to 1, and the sum of the three is 1. The three-dimensional factor coupling coefficient ranges from 0.01 to 0.05. The three-dimensional coupling operator is used to realize the nonlinear correlation calculation of the errors among the three. Subsequently, through the three-dimensional coupling error modeling algorithm, the road geometric feature data, the quantitative environmental interference factors, and the correlated vehicle dynamics data are used to construct road dimension models, environmental dimension models, and vehicle dimension models, respectively. Then, the three-dimensional coupling operator couples the three-dimensional models, integrating the correlation parameters and data of the three dimensions to form a three-dimensional continuous field basic model of road-vehicle-environment. After the model is constructed, it is initialized and calibrated, with the calibration standard being that the deviation between the model input data and the actual collected data is ≤0.05m.

[0014] Furthermore, the mathematical expression of the three-dimensional coupling error modeling algorithm used in the multi-dimensional fusion modeling module is: ;in, The error in the three-dimensional continuous field coupling modeling at time t is... , , These are the dynamic weight coefficients of road, vehicle, and environmental factors at time t, respectively, and the sum of the three is 1; Errors induced by road geometric features For vehicle dynamics induced error, Errors induced by environmental interference; The three-dimensional factor coupling coefficient at time t; It is a three-dimensional coupling operator used to realize the nonlinear correlation calculation of errors among roads, vehicles, and the environment.

[0015] Furthermore, in the error trend prediction module, when converting modeling data, unified feature sets, and BeiDou positioning trajectory data into corresponding point clouds, the module first extracts the spatial coordinates of road geometric features, quantitative values ​​of environmental interference, and spatial coordinate information corresponding to vehicle dynamic parameters from the modeling data and unified feature sets. Feature points are then selected at a fixed sampling step size of 0.1m to generate a feature point cloud, with a point density of 10 points per square meter. Each feature point contains corresponding spatial coordinates and numerical parameter information. Next, the latitude, longitude, azimuth, and altitude positioning coordinate information from the BeiDou positioning trajectory data are extracted. Positioning points are selected at a sampling step size of 0.1s / point to generate a trajectory point cloud, with each positioning point containing corresponding positioning coordinates and acquisition time information. Both types of point clouds are sampled... Using the common point cloud data format for vehicle-mounted equipment; before performing precise ICP registration, noise reduction is performed on both types of point clouds to remove abnormal points that exceed the normal value range; during registration, the initial alignment pose of the point clouds is initialized first, and then corresponding feature points in the two types of point clouds are selected for matching one by one, and the point cloud alignment parameters are iteratively adjusted. The number of iterations is set to 10-20 times, and the iteration convergence threshold is set to 0.01m. The iteration stops when the point cloud alignment deviation meets the preset requirements, and the precise ICP registration is completed; when extracting registration features, the point cloud alignment deviation value, the number of feature point matching pairs, the coordinate deviation of each pair of matching points, and the iteration convergence time are extracted during the registration process. All registration features are recorded in numerical form and organized in timestamp order to form a registration feature dataset.

[0016] Furthermore, in the error trend prediction module, the specific steps for iteratively optimizing the three-dimensional continuous field basic model are as follows: First, the registration features extracted after precise ICP registration, real-time vehicle dynamics data, and factor data after quantitative processing of environmental interference are input into the three-dimensional continuous field basic model; then, three standards for model iterative optimization are preset, namely, the deviation threshold between the model input data and the actual collected data is ≤0.05m, the overlap of feature matching within the model is ≥98%, and the convergence speed of model parameter adjustment is ≤50ms; subsequently, based on the feature fitting status reflected by the registration features, the correlation parameters of road, vehicle, and environment dimensions in the model are adjusted successively according to the preset standards, with the parameter adjustment range controlled between 0.01 and 0.05; finally, the fit between the model output results and the preset standards is continuously verified until all output results meet all preset standards, thus completing the single iterative optimization of the three-dimensional continuous field basic model.

[0017] Furthermore, the error evolution trend prediction algorithm used by the error trend prediction module is mathematically expressed as follows: ;in, The location error at time t+Δt is the predicted time interval; The error in the three-dimensional continuous field coupling modeling at time t is denoted by ; k is the error evolution coefficient. The residual gradient of ICP registration at time t; This is the time weighting coefficient; The rate of change of the modeling error at time t.

[0018] Furthermore, the positioning error threshold in the trajectory precision correction module The threshold is set to 0.8m and can be adjusted according to vehicle type: 0.8-1.0m for freight vehicles, 0.6-0.8m for passenger vehicles and city buses, and 0.5-0.7m for special operation vehicles. The threshold adjustment is completed by remotely issuing instructions through the vehicle intelligent monitoring platform. When correcting inertial navigation drift, a drift zeroing operation is triggered when the inertial navigation drift amount is ≥0.05m / s. The lateral drift correction coefficient is 0.8-0.9 in wet and slippery curve scenarios, and the longitudinal drift correction coefficient is 0.8-0.9 in high-speed straight driving scenarios. After correction, the inertial navigation drift amount is ≤0.02m / s.

[0019] Furthermore, the mathematical expression of the feedforward-feedback composite correction algorithm used in the trajectory precision correction module is as follows: ;in, These are the corrected BeiDou high-precision positioning coordinates at time t. Let t be the original BeiDou positioning coordinates at time t; α is the feedback correction coefficient, with a value range of 0.7-0.9; The error in the three-dimensional continuous field coupling modeling at time t is denoted by β; β is the feedforward correction coefficient, with a value ranging from 0.6 to 0.8. The location error at the predicted time t+Δt; This is the positioning error threshold; For a sign function, when Output +1 when Output 0.

[0020] Compared with existing technologies, this mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring has the following advantages:

[0021] I. This invention integrates multi-source sensing data and constructs a three-dimensional continuous field basic model of road-vehicle-environment after standardization processing. It transforms scattered positioning, inertial navigation, dynamics, road and environmental information into a unified feature set, realizing deep coupling and collaboration of multi-dimensional data. By quantifying environmental interference factors and associating vehicle driving status with road geometric features, it effectively eliminates the limitations of single data sources, reduces the impact of environmental changes and road conditions on positioning results, and improves the integrity and consistency of positioning data. Through iterative model optimization, it continuously calibrates data deviations to ensure the stability of the positioning foundation, providing high-precision data support for subsequent error prediction and trajectory correction. It significantly reduces positioning deviations caused by various interference factors, ensures reliable output of positioning results, and meets the core requirement of high-precision positioning for vehicle intelligent control.

[0022] II. This invention achieves precise trajectory correction and dynamic optimization by predicting the evolution trend of positioning errors in advance and combining a dual correction triggering judgment mechanism. Relying on a feedforward-feedback composite correction algorithm, it integrates real-time modeling errors and future error prediction results to specifically adjust positioning coordinates and inertial navigation drift, effectively compensating for the lag of traditional correction methods. The correction strategy is adjusted according to different vehicle types and driving scenarios to adapt to diverse usage needs, ensuring stable positioning accuracy even in complex road conditions and environments. Through hierarchical encrypted storage and multi-mode wireless communication, it achieves secure data storage and efficient transmission, providing complete and accurate trajectory data for intelligent vehicle monitoring, helping to improve the safety and intelligence level of vehicle driving, adapting to the actual operating needs of various vehicles, and broadening the application scenarios and practical value of the system.

[0023] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0025] Figure 1 A flowchart illustrating the workflow of a mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring.

[0026] Figure 2 Diagram of a mobile BeiDou high-precision positioning and tracking system module and data transmission framework for intelligent vehicle monitoring;

[0027] Figure 3 This is a flowchart illustrating the workflow of the trajectory precision correction module in a mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring. Detailed Implementation

[0028] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0029] Example 1:

[0030] Urban bus morning rush hour road driving scene

[0031] During the morning rush hour, city buses travel along fixed routes, passing through intersections, bus lanes, crosswalks, and some construction sections. There are localized slippery areas on the road, but visibility is good. After the system starts, the multi-source sensing and acquisition module, equipped with automotive-grade hardware, simultaneously acquires BeiDou positioning trajectory inertial navigation drift vehicle dynamics road scene images, environmental perception, and high-precision map road geometry data. This comprehensively captures various key information during the bus's operation. After acquisition, the timestamp of all data is calibrated based on the BeiDou positioning module's timestamp, ensuring consistent time sequence across different data sources. Abnormal data caused by temporary obstructions from construction sections is filtered out, invalid information is removed, and intelligent completion is performed using adjacent valid data to fill data gaps. Finally, all data is uniformly converted to a common encoding format for onboard equipment, completing standardization processing and ensuring data format compatibility with subsequent module processing requirements. The data is then transmitted to the multi-dimensional fusion modeling module, such as... Figure 1 As shown.

[0032] After receiving standardized data, the multi-dimensional fusion modeling module extracts road geometric features such as lane lines, road edge lines, zebra crossings, construction area enclosure outlines, etc., and simultaneously acquires the spatial coordinates, shape, direction, and size parameters of each feature. It clarifies the boundaries and reference markers for bus travel, quantifies environmental interference factors, assigns a quantitative value of 1 to fog and rain intensity (level 1 based on visibility), a quantitative value of 2 to road surface condition (level 2 based on slipperiness), and a quantitative value of 2 to road surface material (level 2 based on friction coefficient), making the degree of environmental interference clearly controllable. Combined with the vehicle's real-time driving speed (30-40 km / h and steering angle), the module performs spatiotemporal matching between vehicle dynamics data and the corresponding road geometric features, ensuring a precise correspondence between vehicle operating status and the road environment. Subsequently, the module uses the extracted road geometric feature data, along with the quantified environmental interference factors and the resulting vehicle dynamics data, as input. It sets the total dynamic weight coefficient of road vehicle environmental factors to 1, and the three-dimensional factor coupling coefficient to 0.03. A three-dimensional coupling error modeling algorithm is used to construct road dimension model, environmental dimension model, and vehicle dimension model, respectively. The mathematical expression of the three-dimensional coupling error modeling algorithm is: ;in, The error in the three-dimensional continuous field coupling modeling at time t is... , , These are the dynamic weight coefficients of road, vehicle, and environmental factors at time t, respectively, and the sum of the three is 1; Errors induced by road geometric features For vehicle dynamics induced error, Errors induced by environmental interference; The three-dimensional factor coupling coefficient at time t; It is a three-dimensional coupling operator used to realize the nonlinear correlation calculation of the errors of road, vehicle and environment. The three-dimensional coupling operator is then used to couple the road-vehicle-environment three-dimensional continuous field basic model. Multi-dimensional information is integrated to ensure data integrity and correlation. After initialization and calibration, it is ensured that the deviation between the model input data and the actual collected data is ≤0.05m, so that the model is highly consistent with the actual driving scenario. The unified feature set is generated and transmitted to the error trend prediction module.

[0033] After receiving the modeling data and unified feature set, the error trend prediction module extracts the spatial coordinates of road geometric features, quantitative values ​​of environmental interference, and spatial coordinate information corresponding to vehicle dynamic parameters. It then selects feature points at a fixed sampling step size of 0.1m to generate a feature point cloud with a point density of 10 points per square meter. Simultaneously, it extracts latitude, longitude, azimuth, and altitude from the BeiDou positioning trajectory data, selecting positioning points at a sampling step size of 0.1s / point to generate a trajectory point cloud. Both types of point clouds use the common point cloud data format for vehicle-mounted equipment, making the abstract data more concrete and facilitating subsequent matching operations. Noise reduction processing is applied to both types of point clouds to reduce the influence of irrelevant interference factors. The initial alignment posture of the point clouds is initialized, and corresponding feature points are selected sequentially for precise ICP registration. The iteration count is set to 15 iterations with a convergence threshold of 0.01m. Iteration stops when the point cloud alignment deviation meets the requirements, ensuring precise alignment between the feature point cloud and the trajectory point cloud. Registration features such as the point cloud alignment deviation value and the number of matching feature points are extracted during the registration process and compiled into a dataset, providing crucial information for model optimization. Subsequently, the real-time vehicle dynamics data and environmental interference quantification factor data of the registration features are input into the three-dimensional continuous field basic model. The convergence speed is adjusted to ≤50ms according to the preset standard deviation threshold ≤0.05m and feature matching overlap ≥98%. The correlation parameters of the road vehicle environment dimension in the model are successively adjusted by an adjustment increment of 0.02, allowing the model parameters to continuously adapt to the real-time driving scenario until all output results meet the preset standards, completing the model iterative optimization. An error evolution trend prediction algorithm is used to calculate the current positioning matching deviation and predict the positioning error evolution trend in the next 3-5 seconds. The mathematical expression of the error evolution trend prediction algorithm is: ;in, The location error at time t+Δt is the predicted time interval; The error in the three-dimensional continuous field coupling modeling at time t is denoted by ; k is the error evolution coefficient. The residual gradient of ICP registration at time t; This is the time weighting coefficient; The rate of change of error is modeled for time t, so as to understand the error change pattern in advance and avoid the accumulation of deviation. The calculation results are then integrated and transmitted to the trajectory precision correction module.

[0034] After receiving the data, the trajectory precision correction module sets the positioning error threshold to 0.7m according to passenger vehicle standards, adapting to the driving accuracy requirements of urban buses. It extracts the original coordinate modeling data and prediction data from the BeiDou positioning system and uses a feedforward-feedback composite correction algorithm to specifically adjust the BeiDou positioning trajectory. The mathematical expression of the feedforward-feedback composite correction algorithm is as follows: ;in, These are the corrected BeiDou high-precision positioning coordinates at time t. Let t be the original BeiDou positioning coordinates at time t; α is the feedback correction coefficient, with a value range of 0.7-0.9; The error in the three-dimensional continuous field coupling modeling at time t is denoted by β; β is the feedforward correction coefficient, with a value ranging from 0.6 to 0.8. The location error at the predicted time t+Δt; This is the positioning error threshold; For a sign function, when Output +1 when The system outputs 0, effectively reducing positioning deviation. Simultaneously, it detects an inertial navigation drift of 0.06 m / s, triggering a drift zeroing operation. When navigating intersection curves, a lateral drift correction factor of 0.85 is applied to correct the inertial navigation drift, adapting to the characteristics of slippery curves. After correction, the inertial navigation drift is reduced to 0.015 m / s ≤ 0.02 m / s, ensuring stable output from the inertial navigation system. The verification process rigorously checks the corrected data to ensure data reliability. After confirming compliance, the verification data, corrected data, and correction log are transmitted to the high-precision data application module. Figure 3 As shown.

[0035] After receiving the data, the high-precision data application module integrates the data from the entire system process to form a complete dataset. The dataset is then stored in a vehicle-grade storage unit with hierarchical encryption to ensure data security during storage. Simultaneously, the trajectory data is uploaded to the vehicle intelligent monitoring platform for storage via a multi-mode wireless communication unit, facilitating subsequent tracking and management of the bus's driving trajectory.

[0036] In summary, during the morning rush hour of urban public transport, the mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring comprehensively captures various key data through a multi-source sensing and acquisition module, and ensures data validity and compatibility through standardized processing. The multi-dimensional fusion modeling module extracts road geometric features and quantifies environmental interference factors, constructing a precise three-dimensional continuous field basic model using a three-dimensional coupled error modeling algorithm. The error trend prediction module predicts error changes through point cloud conversion, precise ICP registration, and model iterative optimization, combined with an error evolution trend prediction algorithm. The trajectory precision correction module sets positioning error thresholds according to passenger vehicle standards, uses a feedforward-feedback composite correction algorithm to correct positioning trajectory and inertial navigation drift, and finally, the high-precision data application module achieves encrypted data storage and uploading, adapting to complex road conditions during the morning rush hour and ensuring accurate bus positioning and stable operation.

[0037] Example 2:

[0038] Long-distance freight transport on highways.

[0039] The freight vehicle was traveling a long distance on the highway, passing through long straight sections of viaducts and tunnels. Some sections exhibited minor road surface wear, and the weather was clear with high visibility. After system startup, the multi-source perception and acquisition module, using automotive-grade hardware, simultaneously acquired BeiDou positioning trajectory, inertial navigation drift, vehicle dynamics, road scene images, environmental perception, and high-precision map road geometry data, covering various key information related to long-distance highway travel. After acquisition, the timestamp of all data was calibrated based on the BeiDou positioning module's timestamp to ensure consistent data timing across different time periods during the long journey. Abnormal data caused by brief signal interference within tunnels was filtered out to ensure data validity. Intelligent supplementation using adjacent valid data filled potential data gaps during long-distance travel. All data was uniformly converted to a common encoding format for onboard equipment, completing standardization processing and providing high-quality data support for subsequent continuous modeling. The data was then transmitted to the multi-dimensional fusion modeling module, such as... Figure 2 As shown.

[0040] After receiving standardized data, the multi-dimensional fusion modeling module extracts road geometric features such as lane lines, road edges, viaduct outlines, tunnel entrance / exit outlines, and guardrail orientation. Simultaneously, it acquires the spatial coordinates, shape, orientation, and size parameters of each feature to meet the road recognition requirements of highway sections. It quantifies environmental interference factors, with fog and rain intensity corresponding to a quantified value of 1, road surface condition to a quantified value of 1, and road surface material to a quantified value of 2, clearly presenting the interference situation on dry, sunny roads. Combined with the vehicle's real-time driving speed (80-90 km / h) and steering angle (close to 0 on long straight sections), it performs spatiotemporal matching of vehicle dynamics data with the corresponding road geometric features, ensuring precise alignment between the operating data of heavy-duty freight vehicles and the highway environment. Using the extracted road geometric feature data and the quantified environmental interference factors-correlated vehicle dynamics data as input, the module sets the total dynamic weight coefficient of road vehicle environmental factors to 1, and the three-dimensional factor coupling coefficient to 0.04. A three-dimensional coupling error modeling algorithm is used to construct a three-dimensional model of the road environment and vehicles. The mathematical expression of the three-dimensional coupling error modeling algorithm is: ;in, The error in the three-dimensional continuous field coupling modeling at time t is... , , These are the dynamic weight coefficients of road, vehicle, and environmental factors at time t, respectively, and the sum of the three is 1; Errors induced by road geometric features For vehicle dynamics induced error, Errors induced by environmental interference; The three-dimensional factor coupling coefficient at time t; It is a three-dimensional coupling operator used to realize the nonlinear correlation calculation of errors among roads, vehicles, and the environment. The three-dimensional coupling operator couples to form a three-dimensional continuous field basic model of road-vehicle-environment, which is adapted to the dynamic changes of high-speed long-distance driving. The initial calibration ensures that the deviation between the model input data and the actual collected data is ≤0.05m, ensuring the accuracy of the model in the initial stage of long-distance driving. The unified feature set is generated and transmitted to the error trend prediction module.

[0041] After receiving data, the error trend prediction module selects feature points at a fixed sampling step size of 0.1m to generate a feature point cloud, and selects positioning points at a sampling step size of 0.1s / point to generate a trajectory point cloud. This ensures the continuity and integrity of data sampling during high-speed driving. Noise reduction is performed to reduce interference factors in high-speed scenarios. ICP precise registration is performed with 18 iterations and a convergence threshold of 0.01m to ensure the accuracy of point cloud matching during high-speed driving. Registration features are extracted and organized into a dataset, providing a valid basis for model optimization. The registration features, real-time vehicle dynamics data, and quantitative environmental interference factors are input into the 3D continuous field basic model. The convergence speed is adjusted to ≤50ms according to the preset standard deviation threshold ≤0.05m and feature matching overlap ≥98%. The model correlation parameters are adjusted by an increment of 0.03 to allow the model to continuously adapt to road condition changes during long-distance driving, completing iterative model optimization. An error evolution trend prediction algorithm is used to calculate the current positioning matching deviation and predict the positioning error evolution trend in the next 3-5 seconds. The mathematical expression for the error evolution trend prediction algorithm is: ;in, The location error at time t+Δt is the predicted time interval; The error in the three-dimensional continuous field coupling modeling at time t is denoted by ; k is the error evolution coefficient. The residual gradient of ICP registration at time t; This is the time weighting coefficient; The rate of change of the error at time t is modeled to avoid the accumulation of errors during long-distance travel affecting positioning accuracy, and the integrated results are transmitted to the trajectory precision correction module.

[0042] The trajectory precision correction module sets the positioning error threshold to 0.9m according to freight vehicle standards, adapting to the driving accuracy requirements of heavy-duty freight vehicles. Relevant data is extracted and a feedforward-feedback composite correction algorithm is used to specifically correct BeiDou positioning trajectory deviations. The mathematical expression of the feedforward-feedback composite correction algorithm is: ;in, These are the corrected BeiDou high-precision positioning coordinates at time t. Let t be the original BeiDou positioning coordinates at time t; α is the feedback correction coefficient, with a value range of 0.7-0.9; The error in the three-dimensional continuous field coupling modeling at time t is denoted by β; β is the feedforward correction coefficient, with a value ranging from 0.6 to 0.8. The location error at the predicted time t+Δt; This is the positioning error threshold; For a sign function, when Output +1 when The output was 0, indicating an inertial navigation drift of 0.04 m / s, which did not trigger a drift zeroing operation. When driving on long straight sections, a longitudinal drift correction coefficient of 0.88 was used for inertial navigation drift correction to meet the drift control requirements of high-speed straight-line driving scenarios. After correction, the inertial navigation drift was reduced to 0.018 m / s, ensuring the positioning stability of heavy-duty trucks during long-distance travel. The verification process rigorously checked the corrected data to ensure it met the requirements of long-distance transportation. After passing the verification, the corrected data and correction logs were transmitted to the high-precision data application module.

[0043] The high-precision data application module integrates data from the entire system process to form a complete dataset, which is then stored in a vehicle-grade storage unit with hierarchical encryption to protect the security of long-distance transportation data. The trajectory data is uploaded to the vehicle intelligent monitoring platform for storage through a multi-mode wireless communication unit, enabling effective supervision and subsequent traceability of the entire trajectory of freight vehicles.

[0044] In summary, in the scenario of long-distance highway transportation for freight vehicles, the system's multi-source perception and acquisition module covers all dimensions of high-speed driving data, providing high-quality data support through calibration, filtering, and standardization. The multi-dimensional fusion modeling module specifically extracts the geometric features of highways and quantifies the interference of dry, sunny environments, constructing a three-dimensional continuous field basic model adapted for long-distance driving through a three-dimensional coupled error modeling algorithm. The error trend prediction module completes point cloud generation, accurate ICP registration, and model optimization, relying on error evolution trend prediction algorithms to avoid error accumulation during long-distance driving. The trajectory precision correction module sets positioning error thresholds according to freight vehicle standards, ensuring stable positioning for heavy-load vehicles through a feedforward-feedback composite correction algorithm. The high-precision data application module achieves encrypted storage and traceability of data throughout the entire process, adapting to the needs of high-speed long-distance transportation and ensuring positioning accuracy and data security.

[0045] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring, characterized in that: The system includes: Multi-source perception and acquisition module: Equipped with automotive-grade acquisition hardware, it simultaneously acquires BeiDou positioning trajectory, inertial navigation drift, vehicle dynamics, road real scene images, environmental perception, and high-precision map road geometry data. After timestamp calibration, anomaly filtering, intelligent completion, and standardization processing, it is transmitted to the multi-dimensional fusion modeling module. Multidimensional fusion modeling module: Receives standardized data, extracts road geometric features, quantifies environmental interference factors, associates vehicle dynamics data with road geometric features, constructs a three-dimensional continuous field basic model of road-vehicle-environment through a three-dimensional coupled error modeling algorithm, fuses and generates a unified feature set, and transmits it to the error trend prediction module; Error trend prediction module: Receives modeling data and unified feature set, converts them into corresponding point clouds with BeiDou positioning trajectory data, performs ICP precise registration and extracts registration features, iteratively optimizes the three-dimensional continuous field basic model, uses error evolution trend prediction algorithm to calculate the current positioning matching deviation, predicts the positioning error evolution trend in the next 3-5 seconds, integrates the calculation results and transmits them to the trajectory precise correction module; The trajectory precision correction module receives the transmitted calculation and prediction results, sets the positioning error threshold and performs dual correction trigger judgment, extracts the original coordinates, modeling data and prediction data of Beidou positioning, corrects the Beidou positioning trajectory and inertial navigation drift through the feedforward-feedback composite correction algorithm, and transmits the verification data, correction data and correction log to the high-precision data application module after verification. High-precision data application module: Receives verification data and correction logs, integrates data from the entire system process to form a complete dataset, stores it in a hierarchical and encrypted manner using automotive-grade storage units, and uploads trajectory data to the vehicle intelligent monitoring platform for storage via a multi-mode wireless communication unit.

2. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The multi-source sensing and acquisition module completes data acquisition through automotive-grade acquisition hardware, including Beidou positioning module acquiring Beidou positioning trajectory data, inertial navigation module acquiring inertial navigation drift data, CAN bus acquisition unit acquiring vehicle dynamics data, camera acquiring real-world road images, environmental sensor acquiring environmental perception data, and high-precision map offline retrieval unit retrieving high-precision map road geometry data. During data processing, the timestamps of all collected data are first calibrated based on the timestamps of the Beidou positioning module. Abnormal data caused by signal obstruction and sensor failure are filtered out. Intelligent completion is achieved by using adjacent valid data. Finally, all data is uniformly converted into the common encoding format of vehicle equipment to complete the standardization process.

3. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The road geometric features extracted by the multi-dimensional fusion modeling module include: lane lines, road edge lines, viaduct outlines, tunnel entrance / exit outlines, guardrail orientation, and road markings. During the extraction process, the spatial coordinates, morphological orientation, and size parameters of each feature are acquired simultaneously. When quantifying environmental interference factors, fog and rain intensity, road surface condition, and road surface material are classified into different levels. Fog and rain intensity is divided into 5 levels based on visibility, road surface condition is divided into 4 levels based on slipperiness, and road surface material is divided into 3 levels based on friction coefficient. Each level corresponds to a quantitative value of 1-5, 1-4, and 1-3, respectively. When associating vehicle dynamics data with road geometric features, the vehicle dynamics data is spatiotemporally matched with the road geometric features of the corresponding driving segment by combining the vehicle's real-time driving speed and steering angle.

4. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, When the multidimensional fusion modeling module constructs a three-dimensional continuous field basic model of road-vehicle-environment through the three-dimensional coupling error modeling algorithm, it first uses the extracted road geometric features, the quantitatively processed environmental interference factors, the correlated vehicle dynamics data and road geometric feature data as input data for the three-dimensional coupling error modeling algorithm. Next, the parameters of the three-dimensional coupling error modeling algorithm are set, where the dynamic weight coefficients of road, vehicle, and environmental factors are all in the range of 0-1 and the sum of the three factors is 1. The three-dimensional factor coupling coefficient is in the range of 0.01-0.

05. The three-dimensional coupling operator is used to realize the nonlinear correlation calculation of the errors of the three factors. Then, through the three-dimensional coupling error modeling algorithm, the road geometric feature data, quantitative environmental interference factors, and correlated vehicle dynamics data are respectively used to construct road dimension model, environmental dimension model, and vehicle dimension model. Then, the three-dimensional coupling operator is used to couple the three-dimensional models, integrate the correlation parameters and data of the three dimensions, and form a basic model of road-vehicle-environment three-dimensional continuous field. After the model is built, the model is initialized and calibrated. The calibration standard is that the deviation between the model input data and the actual collected data is ≤0.05m.

5. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The mathematical expression for the three-dimensional coupling error modeling algorithm used in the multi-dimensional fusion modeling module is: ;in, The error in the three-dimensional continuous field coupling modeling at time t is... , , These are the dynamic weight coefficients of road, vehicle, and environmental factors at time t, respectively, and the sum of the three is 1; Errors induced by road geometric features For vehicle dynamics induced error, Errors induced by environmental interference; The three-dimensional factor coupling coefficient at time t; It is a three-dimensional coupling operator used to realize the nonlinear correlation calculation of errors among roads, vehicles, and the environment.

6. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, In the error trend prediction module, when converting modeling data, unified feature sets, and BeiDou positioning trajectory data into corresponding point clouds, the module first extracts the spatial coordinates of road geometric features, quantitative values ​​of environmental interference, and spatial coordinate information corresponding to vehicle dynamic parameters from the modeling data and unified feature sets. Feature points are then selected at a fixed sampling step size of 0.1m to generate a feature point cloud, with a point density of 10 points per square meter. Each feature point contains corresponding spatial coordinates and numerical parameter information. Next, the module extracts latitude, longitude, azimuth, and altitude positioning coordinate information from the BeiDou positioning trajectory data. Positioning points are then selected at a sampling step size of 0.1s / point to generate a trajectory point cloud, with each positioning point containing corresponding positioning coordinates and acquisition time information. Both types of point clouds are generated using a vehicle-mounted system. The equipment uses a common point cloud data format. Before performing precise ICP registration, noise reduction is performed on both types of point clouds to remove abnormal points that exceed the normal value range. During registration, the initial alignment posture of the point cloud is initialized first, and then corresponding feature points in the two types of point clouds are selected for matching one by one. The point cloud alignment parameters are iteratively adjusted, with the number of iterations set to 10-20 times and the iteration convergence threshold set to 0.01m. The iteration stops when the point cloud alignment deviation meets the preset requirements, thus completing the precise ICP registration. When extracting registration features, the point cloud alignment deviation value, the number of feature point matching pairs, the coordinate deviation of each pair of matching points, and the iteration convergence time are extracted. All registration features are recorded in numerical form and organized in timestamp order to form a registration feature dataset.

7. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, In the error trend prediction module, the specific steps for iteratively optimizing the three-dimensional continuous field basic model are as follows: first, input the registration features extracted after ICP accurate registration, real-time vehicle dynamics data, and factor data after quantitative processing of environmental disturbances into the three-dimensional continuous field basic model. Three pre-defined criteria for model iteration optimization are: a deviation threshold between the model input data and the actual collected data ≤ 0.05m, an overlap of feature matching within the model ≥ 98%, and a convergence speed of model parameter adjustment ≤ 50ms. Subsequently, based on the feature fitting status reflected by the registration features, the correlation parameters of the road, vehicle, and environment dimensions in the model are adjusted successively according to the pre-defined criteria, with the parameter adjustment range controlled between 0.01 and 0.

05. Finally, the fit between the model output results and the pre-defined criteria is continuously verified until all output results meet all pre-defined criteria, completing the single-iteration optimization of the three-dimensional continuous field basic model.

8. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The error evolution trend prediction algorithm used in the error trend prediction module is mathematically expressed as follows: ;in, The location error at time t+Δt is the predicted time interval; The error in the three-dimensional continuous field coupling modeling at time t is denoted by ; k is the error evolution coefficient. The residual gradient of ICP registration at time t; This is the time weighting coefficient; The rate of change of the modeling error at time t.

9. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The positioning error threshold in the trajectory accuracy correction module The threshold is set to 0.8m and can be adjusted according to vehicle type: 0.8-1.0m for freight vehicles, 0.6-0.8m for passenger vehicles and city buses, and 0.5-0.7m for special operation vehicles. The threshold adjustment is completed by remotely issuing instructions through the vehicle intelligent monitoring platform. When correcting inertial navigation drift, a drift zeroing operation is triggered when the inertial navigation drift amount is ≥0.05m / s. The lateral drift correction coefficient is 0.8-0.9 in wet and slippery curve scenarios, and the longitudinal drift correction coefficient is 0.8-0.9 in high-speed straight driving scenarios. After correction, the inertial navigation drift amount is ≤0.02m / s.

10. The mobile BeiDou high-precision positioning and tracking system for intelligent vehicle monitoring according to claim 1, characterized in that, The mathematical expression for the feedforward-feedback composite correction algorithm used in the trajectory precision correction module is: ;in, These are the corrected BeiDou high-precision positioning coordinates at time t. Let t be the original BeiDou positioning coordinates at time t; α is the feedback correction coefficient, with a value range of 0.7-0.9; The error in the three-dimensional continuous field coupling modeling at time t is denoted by β; β is the feedforward correction coefficient, with a value ranging from 0.6 to 0.

8. The location error at the predicted time t+Δt; This is the positioning error threshold; For a sign function, when Output +1 when Output 0.

Images