Method for processing map data of digital map

Abstract

The invention discloses a method for processing map data of a digital map, and relates to the technical field of vehicle system control, and the method comprises the steps: S1, data collection and preprocessing; s2, data fusion and analysis; s3, updating and maintaining in real time; s4, performing real-time decision making and control; and S5, performing feedback and optimization. Compared with data acquisition completely depending on network transmission, the method for processing the map data of the digital map is less affected by network conditions, can transmit the data in real time in a vehicle system in cooperation with an advanced data transmission protocol and a network technology, and effectively improves the real-time performance of the map data. Meanwhile, the map data of the digital map can be rapidly, accurately and efficiently processed, on the basis of the efficient processing flow, the incremental updating means is matched, the method is particularly suitable for the conditions of large data size and high real-time performance requirement, the real-time performance of the map data is further improved, and therefore the map data processing effect of the digital map is improved.

Description

Technical Field

[0001] The invention relates to the technical field of vehicle system control, in particular to a method for processing map data of a digital map. Background Art

[0002] Vehicle system refers to the general term for all subsystems inside a car, which work together to enable the car to operate normally. The vehicle system covers many aspects, including but not limited to the power system, transmission system, driving system, braking system, steering system, electrical system, cab system, frame system and auxiliary driving system, and each subsystem has its specific functions and tasks.

[0003] In modern automotive technology, the relationship between vehicle systems and digital map data is becoming increasingly close. Digital map data provides vehicles with precise geographic location information, allowing them to accurately determine their own positions and plan the best driving routes. At the same time, it can also be combined with positioning systems such as GPS to achieve real-time positioning and navigation functions for vehicles. During navigation, the vehicle system will continuously adjust and optimize the driving route based on map data and real-time traffic information to avoid congested sections or construction areas, and improve driving efficiency and safety. In addition, the vehicle system can also monitor and remind the driver's driving behavior based on the traffic rules information in the map data to ensure that the driving behavior complies with traffic rules, and identify potential dangerous factors (such as sharp turns, steep slopes, intersections, etc.) through in-depth analysis and processing of digital map data, and issue warnings or take preventive measures in advance to reduce the occurrence of traffic accidents and improve driving safety.

[0004] However, in actual use, the vehicle system control has high real-time requirements for map data, while the transmission of map data requires a certain amount of time, especially under poor network conditions. Data transmission delays will be further aggravated, and it also takes time to process and analyze map data. Especially when the data volume is large and the processing algorithm is complex, the processing time will be longer. This makes the existing methods often have delays when processing map data, making it difficult to meet the real-time requirements of vehicle system control.

[0005] Therefore, it is urgent to improve this shortcoming. The present invention studies and improves the existing technology and its shortcomings, and provides a method for processing map data of a digital map. Summary of the invention

[0006] The object of the present invention is to provide a method for processing map data of a digital map to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a method for processing map data of a digital map, comprising the following steps:

[0008] S1. Data collection and preprocessing: Using sensors (such as GPS, LiDAR, cameras, etc.) and real-time traffic data sources, key information such as road geometry, obstacle locations, traffic signs, and traffic flow are acquired in real time. The collected raw data is cleaned, denoised, and filtered to remove invalid and duplicate data, and to repair missing values ​​to improve data quality and accuracy.

[0009] S2. Data fusion and analysis: Fusion of map data from different sources, using geographic information system (GIS) technology to process (including coordinate conversion, data format conversion, data compression, etc.), analyze (spatial relationship analysis, statistical analysis and modeling of map data) and display (intuitively display the analysis results in the form of geographic charts, statistical charts, dynamic demonstrations, etc.) geographic information data, real-time analysis of vehicle location, matching vehicle positioning data with map data, and determining the exact location of the vehicle on the map;

[0010] S3. Real-time update and maintenance: Establish a real-time update mechanism to automatically obtain new data from the data source and perform incremental updates to ensure the timeliness and accuracy of map data. At the same time, regularly verify and repair map data to ensure data integrity and reliability.

[0011] S4, Real-time decision-making and control: Based on the real-time updated map data and vehicle status information, the vehicle system control module makes real-time decisions, such as path planning, lane keeping, obstacle avoidance, etc., and converts the decision results into specific control instructions, and controls the vehicle's motion state through the actuator to ensure that the vehicle can maintain a stable motion state under various conditions;

[0012] S5. Feedback and optimization: Collect feedback information during vehicle operation, such as actual driving trajectory, changes in road conditions, etc., compare and analyze it with map data, and optimize the map data processing method based on the feedback results to improve processing efficiency and accuracy.

[0013] Furthermore, in step S1, the sensor includes GPS, LiDAR, camera, and laser scanner;

[0014] The GPS receives signals from satellites to determine the position (longitude, latitude and altitude) and time information of the receiver, which is used for vehicle positioning and provides the vehicle's geographic location information in real time;

[0015] The LiDAR determines the distance and position of the target by emitting a laser beam and measuring the time it takes for it to reflect back, and can generate a three-dimensional point cloud map of the surrounding environment, providing high-precision spatial information. LiDAR is used for environmental perception, including measurement of road geometry, precise detection of obstacle locations, and identification of vehicles and pedestrians.

[0016] The camera generates images or videos by capturing light, and captures different ranges of visual information according to different resolutions and focal lengths, including recognition of traffic signs and traffic signals, lane line detection, tracking of vehicles and pedestrians, etc. It cooperates with image processing and computer vision technology to extract rich visual information, providing important perception capabilities for autonomous driving and intelligent transportation systems;

[0017] The laser scanner measures the geometric shape of the road surface, including the flatness, slope, curvature, etc. of the road surface, by emitting a laser beam and receiving its reflected signal.

[0018] Furthermore, in step S1, the processing operation of the original data includes:

[0019] 1) Data cleaning:

[0020] Missing value processing: For missing sensor data or blank information in the data source, adopt a strategy of filling (such as using the mean, median, mode, etc.) or deleting (if the missing data has little impact on the overall analysis);

[0021] Outlier detection and processing: Use statistical methods (such as Z-score, IQR, etc.) or machine learning algorithms to identify outliers, and choose to retain, modify or delete these outliers according to the actual situation (in traffic data, outliers include sudden speed changes, unreasonable location information, etc.);

[0022] Data standardization and normalization: standardize or normalize data from different sensors and data sources to facilitate comparison and fusion in subsequent analysis;

[0023] 2) Denoising: Use digital filtering (such as low-pass filtering, high-pass filtering, band-pass filtering, etc.) and wavelet transform denoising methods to reduce or eliminate the interference of noise (such as electromagnetic interference, mechanical vibration, etc.) on sensor data during the acquisition process;

[0024] 3) Filtering: Improve signal quality by retaining useful components in the signal and suppressing or removing noise components. In real-time traffic data processing, the Kalman filter algorithm is used to smooth the position jitter problem in GPS data.

[0025] Furthermore, in step S2, the preprocessed map data is transmitted to the vehicle system in real time through data transmission protocols (such as TCP / IP, UDP, etc.) and network technologies (such as 5G, Wi-Fi 6, etc.).

[0026] Furthermore, in step S2, the specific operation of multi-source data fusion is: according to the characteristics of the data and the fusion target, a fusion method is selected, and through feature extraction and matching algorithms, feature points from different data sources are associated to establish geometric and semantic connections between the data for subsequent fusion processing.

[0027] Furthermore, the fusion method includes model-based fusion, weight-based fusion and feature-based fusion;

[0028] The model-based fusion integrates map data from different sources by building a mathematical model. The specific operation steps are as follows: according to the characteristics of the data and the fusion target, a mathematical model (including a statistical model, a machine learning model, etc.) is selected for construction to describe the internal relationship and law between the data, and then the model is trained using part of the data, and the model is optimized according to the training results. Finally, the trained model is used to fuse the map data from different sources to obtain the fused map data;

[0029] The weight-based fusion is achieved by assigning different weights to map data from different sources. The specific operation steps are: by analyzing the accuracy, completeness, timeliness and other aspects of the data, the map data from different sources are evaluated to determine their importance and reliability in the fusion process, and then according to the data evaluation results, corresponding weights are assigned to the data from different sources, and finally, the map data from different sources are weightedly fused according to the assigned weights to obtain the fused map data;

[0030] The feature-based fusion extracts feature information of map data from different sources and performs fusion processing on the feature level. The specific operation steps are as follows: using a feature extraction algorithm to extract key feature information from map data from different sources, including geographic location, road information, building information, etc., and matching the extracted feature information to find feature points or feature areas with similarities between different data sources, and then fusion processing is performed on the matched feature information on the feature level through feature weighting, feature selection or feature fusion algorithm, and finally reconstructing the original map data according to the fused feature information to obtain fused map data.

[0031] Furthermore, in step S2, the specific operations of vehicle position resolution and map matching are as follows: obtaining the real-time position information of the vehicle through GPS or other positioning technologies, and converting it into coordinates in the map coordinate system, then matching the vehicle position data with the map data, and determining the exact position of the vehicle on the map through the steps of error area selection, road selection set creation, angle filtering and distance calculation;

[0032] Among them, the specific steps of matching vehicle positioning data with map data are:

[0033] Selection of error region: Since GPS signals or other sensor data may have errors, a reasonable error region must be defined in order to search for possible matching roads within this region;

[0034] Creation of road selection set: extract all candidate roads in the area from the map database to form a road selection set;

[0035] Angle filtering: By calculating and comparing the angle between the vehicle's travel direction and the candidate road direction, roads with too large angles are excluded, thus further narrowing the range of the road selection set.

[0036] Distance calculation: further calculate the distance between the vehicle position and the remaining candidate roads to determine the final matching road;

[0037] Matching data update: Use the matching results to update the vehicle location data to ensure that the vehicle displayed on the map is consistent with the actual location.

[0038] Furthermore, in step S3, when performing the data update operation, the incremental update only processes the data that has changed since the last update, and does not repeatedly update the data that has not changed or has been updated. The specific operation of the incremental update is:

[0039] Log tracking: Use database logs (such as MySQL binlog), event sources (such as Kafka), or data version control mechanisms to track data changes;

[0040] Difference detection: Compare the data snapshot from the last synchronization with the current data to determine the changed parts;

[0041] Update only changed data: Only the changed data is written to the target database or data warehouse.

[0042] Furthermore, in step S4, the real-time decision algorithm specifically includes:

[0043] Path planning algorithm: responsible for generating an optimal driving path based on the vehicle's current position and target position, combined with map information and traffic rules. The path planning algorithm includes global path planning (under the condition of known map information, combined with local environmental information and real-time vehicle feedback information, the drivable area and reference driving path are generated, and the global path planning algorithm is usually based on graph search, heuristic search (such as A algorithm) or sampling method (such as RRT algorithm)), local path planning (based on global path planning, considering real-time environmental changes and vehicle dynamics characteristics, local path adjustment; local path planning algorithms such as artificial potential field method, dynamic window method and model predictive control, etc.);

[0044] Lane keeping algorithm: used to ensure that the vehicle can drive stably in its lane and avoid deviating from the lane. The lane keeping algorithm includes: vision-based lane keeping (using a camera to capture lane line information, identifying lane lines through image processing technology, and calculating the steering wheel angle based on the relative position between the vehicle and the lane line to achieve lane keeping), model-prediction-based lane keeping (using a vehicle dynamics model to predict the vehicle's future driving trajectory, and combining lane line information to adjust control parameters to ensure that the vehicle stays in the lane);

[0045] Obstacle avoidance algorithm: responsible for making timely decisions when obstacles are detected to avoid collisions with obstacles. The obstacle avoidance algorithm uses sensors such as radar, laser radar (LiDAR), and cameras to detect obstacles around the vehicle, and assesses the risk of collision based on information such as the position, speed, and size of the obstacles. Then, based on the risk assessment results, it formulates obstacle avoidance strategies such as deceleration, lane changing, and emergency braking. After the decision is made, local path planning is performed to generate a driving path that avoids obstacles.

[0046] Furthermore, in step S5, the map data processing method is optimized by:

[0047] Data update and correction: timely update map data based on feedback, including road information, traffic information, traffic facilities, etc., and correct any data errors or omissions found;

[0048] Algorithm optimization: Optimize the map data processing algorithm based on the problems found during data analysis. For example, improve the path planning algorithm to improve accuracy; optimize the traffic prediction algorithm to more accurately reflect the real-time traffic conditions, and introduce machine learning and deep learning technologies to improve the intelligence level of map data processing and traffic prediction;

[0049] Quality control and evaluation: Establish a quality control system to monitor and evaluate the processing of map data, regularly evaluate the accuracy and availability of map data, and ensure that the quality of map data meets user needs;

[0050] User participation and feedback mechanism: Establish a user participation mechanism to encourage users to actively feedback traffic information and map issues, and improve the feedback processing process to ensure that user feedback can be responded to and processed in a timely manner.

[0051] The present invention provides a method for processing map data of a digital map, which has the following beneficial effects:

[0052] The data acquisition method of the present invention includes not only sensors, but also real-time traffic data sources. Some data comes directly from the devices and systems carried by the vehicle itself. Compared with completely relying on network transmission to obtain data, it is less affected by network conditions. With advanced data transmission protocols and network technologies, the vehicle system that can transmit data in real time effectively improves the real-time performance of map data. At the same time, the present invention can quickly, accurately and efficiently process the map data of digital maps. On the basis of efficient processing procedures, with incremental update means, it is particularly suitable for situations with large amounts of data and high real-time requirements, further improving the real-time performance of map data, thereby improving the map data processing effect of digital maps. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 The present invention is a schematic flow chart of the steps of a method for processing map data of a digital map. DETAILED DESCRIPTION

[0054] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0055] like Figure 1 As shown, a method for processing map data of a digital map comprises the following steps:

[0056] S1. Data collection and preprocessing:

[0057] Utilize sensors (such as GPS, LiDAR, cameras, etc.) and real-time traffic data sources (on-board sensor data, road infrastructure sensor data, public transportation system data, mobile applications and user data, third-party data sources) to obtain key information such as road geometry, obstacle locations, traffic signs, and traffic flow in real time, and clean, denoise, and filter the collected raw data to remove invalid and duplicate data, and repair missing values ​​to improve data quality and accuracy;

[0058] In this embodiment, the sensors include GPS, LiDAR, camera, laser scanner, etc., wherein GPS determines the position (longitude, latitude and altitude) and time information of the receiver by receiving signals from satellites, which is used for vehicle positioning and provides the vehicle's geographic location information in real time; LiDAR determines the distance and position of the target by emitting laser beams and measuring the time for them to be reflected back, and can generate a three-dimensional point cloud map of the surrounding environment, providing high-precision spatial information, and LiDAR is used for environmental perception, including the measurement of road geometry, accurate detection of obstacle positions, and identification of vehicles and pedestrians; the camera generates images or videos by capturing light, and captures different ranges of visual information according to different resolutions and focal lengths, including identification of traffic signs and traffic signals, lane line detection, tracking of vehicles and pedestrians, etc., and cooperates with image processing and computer vision technology to extract rich visual information, providing important perception capabilities for autonomous driving and intelligent transportation systems; the laser scanner measures the geometry of the road surface, including road surface flatness, slope, curvature, etc., by emitting laser beams and receiving their reflected signals;

[0059] In this embodiment, the processing operation of the original data includes:

[0060] 1) Data cleaning:

[0061] Missing value processing: For missing sensor data or blank information in the data source, adopt a strategy of filling (such as using the mean, median, mode, etc.) or deleting (if the missing data has little impact on the overall analysis);

[0062] Outlier detection and processing: Use statistical methods (such as Z-score, IQR, etc.) or machine learning algorithms to identify outliers, and choose to retain, modify or delete these outliers according to the actual situation (in traffic data, outliers include sudden speed changes, unreasonable location information, etc.);

[0063] Data standardization and normalization: standardize or normalize data from different sensors and data sources to facilitate comparison and fusion in subsequent analysis;

[0064] 2) Denoising: Use digital filtering (such as low-pass filtering, high-pass filtering, band-pass filtering, etc.) and wavelet transform denoising methods to reduce or eliminate the interference of noise (such as electromagnetic interference, mechanical vibration, etc.) on sensor data during the acquisition process;

[0065] 3) Filtering: Improve signal quality by retaining useful components in the signal and suppressing or removing noise components. In real-time traffic data processing, the Kalman filter algorithm is used to smooth the position jitter problem in GPS data;

[0066] S2. Data fusion and analysis:

[0067] The preprocessed map data is transmitted to the vehicle system in real time through data transmission protocols (such as TCP / IP, UDP, etc.) and network technologies (such as 5G, Wi-Fi 6, etc.), and the map data from different sources are integrated and processed (including coordinate conversion, data format conversion, data compression, etc.), analyzed (spatial relationship analysis, statistical analysis and modeling of map data, such as calculating the distance between vehicles and roads, analyzing vehicle driving trajectories, predicting traffic flow, etc.) and displayed (the analysis results are intuitively displayed in the form of geographic charts, statistical charts, dynamic demonstrations, etc., such as displaying the vehicle position, driving trajectory and surrounding road information on the map in real time, and displaying traffic flow, congestion, etc. through statistical charts) using geographic information data, real-time analysis of vehicle position, matching vehicle positioning data with map data, and determining the exact position of the vehicle on the map;

[0068] In this embodiment, the specific operation of multi-source data fusion is as follows: according to the characteristics of the data and the fusion target, a fusion method is selected, and feature points from different data sources are associated through feature extraction and matching algorithms to establish geometric and semantic connections between the data for subsequent fusion processing; the fusion methods include model-based fusion, weight-based fusion and feature-based fusion;

[0069] The model-based fusion integrates map data from different sources by building a mathematical model. The method involves an in-depth understanding and analysis of the data so that an accurate mathematical model can be established to describe the relationship between the data. The specific operation steps are as follows: according to the characteristics of the data and the fusion target, a mathematical model (including a statistical model, a machine learning model, etc.) is selected for construction to describe the intrinsic relationship and law between the data, and then the model is trained using part of the data, and the model is optimized according to the training results to improve its accuracy and generalization ability, and finally the trained model is used to fuse the map data from different sources to obtain the fused map data;

[0070] The weight-based fusion is achieved by assigning different weights to map data from different sources. These weights reflect the importance and reliability of different data sources in the fusion process. The specific operation steps are: by analyzing the accuracy, completeness, timeliness and other aspects of the data, the map data from different sources are evaluated to determine their importance and reliability in the fusion process. Then, according to the data evaluation results, corresponding weights are assigned to the data from different sources. The size of the weight reflects the contribution of the data in the fusion result. Finally, the map data from different sources are weightedly fused according to the assigned weights to obtain the fused map data.

[0071] The feature-based fusion extracts feature information of map data from different sources and performs fusion processing at the feature level. The method can make full use of the feature information of the data to improve the accuracy and reliability of the fusion result. The specific operation steps are: using a feature extraction algorithm to extract key feature information from map data from different sources, including geographical location, road information, building information, etc., and matching the extracted feature information to find feature points or feature areas with similarities between different data sources, and then fusing the matched feature information at the feature level through feature weighting, feature selection or feature fusion algorithm, and finally reconstructing the original map data according to the fused feature information to obtain fused map data;

[0072] In this embodiment, the specific operation of vehicle position resolution and map matching is: obtaining the real-time position information of the vehicle through GPS or other positioning technology, and converting it into coordinates in the map coordinate system, and then matching the vehicle position data with the map data to determine the exact position of the vehicle on the map; wherein, the specific steps of matching the vehicle positioning data with the map data are:

[0073] Selection of error region (since GPS signals or other sensor data may contain errors, a reasonable error region must be defined in order to search for possible matching roads within this region): the error region is defined according to the probability criterion, that is, the error region must contain the actual position of the vehicle with a certain probability, and the size of the error region is determined comprehensively based on the accuracy of the GPS signal, the performance of the sensor, and environmental factors (such as urban canyons, high-rise building occlusion, etc.); if the error region is selected too large, the capacity of the road selection set will increase and more computing time will be consumed; if it is selected too small, the correctly matched roads may be missed;

[0074] Creation of road selection set (extracting all candidate roads in the area from the map database to form a road selection set): using the road data in the map database, retrieve all possible roads in the error area, and filter the retrieved roads according to certain rules (such as road type, travel direction, etc.) to form the final road selection set;

[0075] Angle filtering (an effective method to further narrow the range of road selection set by calculating and comparing the angle between the vehicle's current driving direction and the candidate road direction, excluding roads with too large an angle): Calculate the angle between the vehicle's current driving direction and the candidate road direction, set an angle threshold (such as 15°, 30°, etc.), and exclude roads with angles greater than the threshold from the road selection set;

[0076] Distance calculation (further calculating the distance between the vehicle position and the remaining candidate roads to determine the final matching road): considering the characteristics of map data and calculation efficiency, select distance calculation methods such as Euclidean distance and Manhattan distance to calculate the shortest distance between the vehicle position and each candidate road, and select the road with the shortest distance as the final matching road;

[0077] Matching data update (use the matching results to update the vehicle location data to ensure that the vehicle's display on the map is consistent with the actual location): the matched road coordinates are used as the vehicle's current location coordinates. As the vehicle travels and the GPS signal is updated, the above matching process is repeated to achieve dynamic updates of the vehicle's location.

[0078] S3, real-time update and maintenance:

[0079] Establish a real-time update mechanism to automatically obtain new data from the data source and perform incremental updates to ensure the timeliness and accuracy of map data. At the same time, regularly check and repair map data to ensure data integrity and reliability. When performing data update operations, incremental updates only process the data that has changed since the last update, and do not repeat updates of data that has not changed or has been updated. The specific operations of incremental updates are:

[0080] Log tracking: Use database logs (such as MySQL binlog), event sources (such as Kafka), or data version control mechanisms to track data changes;

[0081] Difference detection: Compare the data snapshot from the last synchronization with the current data to determine the changed parts;

[0082] Update only changed data: only write changed data to the target database or data warehouse;

[0083] In this embodiment, the process of establishing a real-time update mechanism is as follows:

[0084] Determine the data source: Identify the types of data that need to be updated in real time, such as traffic conditions, POI information, road changes, etc., and determine a reliable data source. If an API interface is used as a data source, it is necessary to establish a cooperative relationship with the data provider and obtain the necessary API access rights and keys;

[0085] Data collection: Use programming languages ​​(such as Python, Java, etc.) to write automated scripts to obtain data from data sources regularly or in real time. The scripts should be able to process network requests, parse data formats (such as JSON, XML), etc., and use cache technologies (such as Redis, Memcached) to store recently obtained data;

[0086] Data processing and incremental updates: Design incremental detection mechanisms, such as using timestamps, version numbers, or hash values ​​to determine whether data has changed. For data that has changed, perform incremental update operations, add new data to the map database, and update related indexes and caches.

[0087] In this embodiment, the process of map data verification and repair is as follows:

[0088] Establish a regular verification mechanism to conduct a comprehensive check on map data, including but not limited to road network connectivity, POI information accuracy, address matching correctness, etc.;

[0089] Data repair: For problems that can be automatically identified and repaired by algorithms (such as format errors, data redundancy, etc.), develop corresponding repair scripts to handle them; for problems that require manual judgment and repair (such as incorrect POI locations, newly added roads that are not marked, etc.), establish a problem feedback and repair process, and have a dedicated team handle it;

[0090] Set cycle: According to the update frequency and importance of map data, set a reasonable verification and repair cycle; for example, real-time traffic information can be set to be verified once every hour, while static road and POI information can be set to be verified once a day or a week;

[0091] S4. Real-time decision-making and control:

[0092] Based on the real-time updated map data and vehicle status information, the vehicle system control module makes real-time decisions, such as path planning, lane keeping, obstacle avoidance, etc., and converts the decision results into specific control instructions, and controls the vehicle's motion state through actuators to ensure that the vehicle can maintain a stable motion state under various conditions;

[0093] In this embodiment, the real-time decision algorithm specifically includes:

[0094] Path planning algorithm: responsible for generating an optimal driving path based on the vehicle's current position and target position, combined with map information and traffic rules. The path planning algorithm includes global path planning (under the condition of known map information, combined with local environmental information and real-time vehicle feedback information, the drivable area and reference driving path are generated, and the global path planning algorithm is usually based on graph search, heuristic search (such as A algorithm) or sampling method (such as RRT algorithm)), local path planning (based on global path planning, considering real-time environmental changes and vehicle dynamics characteristics, local path adjustment; local path planning algorithms such as artificial potential field method, dynamic window method and model predictive control, etc.);

[0095] Lane keeping algorithm: used to ensure that the vehicle can drive stably in its lane and avoid deviating from the lane. The lane keeping algorithm includes: vision-based lane keeping (using a camera to capture lane line information, identifying lane lines through image processing technology, and calculating the steering wheel angle based on the relative position between the vehicle and the lane line to achieve lane keeping), model-prediction-based lane keeping (using a vehicle dynamics model to predict the vehicle's future driving trajectory, and combining lane line information to adjust control parameters to ensure that the vehicle stays in the lane);

[0096] Obstacle avoidance algorithm: responsible for making timely decisions when obstacles are detected to avoid collisions with obstacles. The obstacle avoidance algorithm uses sensors such as radar, LiDAR, and cameras to detect obstacles around the vehicle, and assesses the risk of collision based on information such as the location, speed, and size of the obstacles. Then, based on the risk assessment results, it formulates obstacle avoidance strategies such as deceleration, lane change, and emergency braking. After the decision is made, local path planning is performed to generate a driving path that avoids obstacles.

[0097] In terms of implementing decision-making algorithms, autonomous driving systems usually combine multiple algorithms to cope with complex driving environments, as follows:

[0098] PID control algorithm: used for speed control, position tracking, etc., by adjusting the control parameters to make the vehicle run along the expected trajectory;

[0099] Fuzzy control algorithm: It has advantages in processing fuzzy or uncertain input information and can be used for obstacle avoidance decision-making, driving mode switching, etc.

[0100] Model Predictive Control (MPC): Considers the dynamic characteristics and future state of the vehicle, generates optimal control instructions through prediction and optimization, and is suitable for local path planning and obstacle avoidance control;

[0101] Deep learning algorithms: such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), which can be used for obstacle detection, behavior prediction, and driving decision-making;

[0102] S5. Feedback and Optimization:

[0103] Collect feedback information during vehicle operation, such as actual driving trajectory, road condition changes, etc., compare and analyze it with map data, and optimize the map data processing method based on the feedback results to improve processing efficiency and accuracy;

[0104] In this embodiment, the map data processing method is optimized by:

[0105] Data update and correction: timely update map data based on feedback, including road information, traffic information, traffic facilities, etc., and correct any data errors or omissions found;

[0106] Algorithm optimization: Optimize the map data processing algorithm based on the problems found during data analysis. For example, improve the path planning algorithm to improve accuracy; optimize the traffic prediction algorithm to more accurately reflect the real-time traffic conditions, and introduce machine learning and deep learning technologies to improve the intelligence level of map data processing and traffic prediction;

[0107] Quality control and evaluation: Establish a quality control system to monitor and evaluate the processing of map data, regularly evaluate the accuracy and availability of map data, and ensure that the quality of map data meets user needs;

[0108] User participation and feedback mechanism: Establish a user participation mechanism to encourage users to actively feedback traffic information and map issues, and improve the feedback processing process to ensure that user feedback can be responded to and processed in a timely manner.

[0109] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. A method for processing map data of a digital map, characterized in that The following steps are involved: S1. Data collection and preprocessing: Use sensors and real-time traffic data sources to obtain key road information in real time, and clean, denoise, and filter the collected raw data; S2. Data fusion and analysis: Integrate map data from different sources, use geographic information system technology to process, analyze and display geographic information data, analyze vehicle location in real time, match vehicle positioning data with map data, and determine the exact location of the vehicle on the map; S3. Real-time update and maintenance: Establish a real-time update mechanism to automatically obtain new data from the data source and perform incremental updates, while regularly verifying and repairing map data; S4, real-time decision and control: based on the real-time updated map data and vehicle status information, the vehicle system control module makes real-time decisions and converts the decision results into specific control instructions to control the vehicle's motion state through the actuator; S5. Feedback and optimization: Collect feedback information during vehicle operation, compare and analyze it with map data, and optimize the map data processing method based on the feedback results.

2. A method for processing map data of a digital map according to claim 1, characterized in that: In step S1, the sensors include GPS, LiDAR, camera, and laser scanner; The GPS receives signals from satellites to determine the location and time information of the receiver, which is used for vehicle positioning and provides the vehicle's geographic location information in real time; The LiDAR determines the distance and position of the target by emitting a laser beam and measuring the time it takes for it to reflect back, and can generate a three-dimensional point cloud map of the surrounding environment, providing high-precision spatial information. LiDAR is used for environmental perception; The camera generates images or videos by capturing light, and captures different ranges of visual information according to different resolutions and focal lengths, and cooperates with image processing and computer vision technology to extract rich visual information; The laser scanner measures the geometric shape of the road surface, including the flatness, slope and curvature of the road surface, by emitting laser beams and receiving their reflected signals.

3. A method for processing map data of a digital map according to claim 1, characterized in that: In step S1, the processing operation of the original data includes: 1) Data cleaning: Missing value processing: For missing sensor data or blank information in the data source, adopt a filling or deletion strategy; Outlier detection and processing: Use statistical methods or machine learning algorithms to identify outliers, and choose to retain, modify or delete outliers based on actual conditions; Data standardization and normalization: standardize or normalize data from different sensors and data sources; 2) Denoising: Use digital filtering and wavelet transform to reduce or eliminate noise interference on sensor data during the acquisition process; 3) Filtering: Improve signal quality by retaining useful components in the signal and suppressing or removing noise components. In real-time traffic data processing, the Kalman filter algorithm is used to smooth the position jitter problem in GPS data.

4. A method for processing map data of a digital map according to claim 1, characterized in that: In step S2, the pre-processed map data is transmitted to the vehicle system in real time through data transmission protocol and network technology.

5. A method for processing map data of a digital map according to claim 1, characterized in that: In step S2, the specific operation of multi-source data fusion is: according to the characteristics of the data and the fusion target, a fusion method is selected, and feature points from different data sources are associated through feature extraction and matching algorithms to establish geometric and semantic connections between the data for subsequent fusion processing.

6. A method for processing map data of a digital map according to claim 5, characterized in that: The fusion methods include model-based fusion, weight-based fusion and feature-based fusion, wherein the model-based fusion integrates map data from different sources by constructing a mathematical model, the weight-based fusion achieves fusion by assigning different weights to map data from different sources, and the feature-based fusion extracts feature information from map data from different sources and performs fusion processing at the feature level.

7. A method for processing map data of a digital map according to claim 1, characterized in that: In step S2, the specific operations of vehicle position resolution and map matching are as follows: obtaining the real-time position information of the vehicle through GPS or other positioning technologies, and converting it into coordinates in the map coordinate system, and then matching the vehicle position data with the map data, and determining the exact position of the vehicle on the map through error area selection, road selection set creation, angle filtering, distance calculation and matching data update.

8. A method for processing map data of a digital map according to claim 1, characterized in that: In step S3, when performing the data update operation, the incremental update only processes the data that has changed since the last update, and does not repeatedly update the data that has not changed or has been updated. The specific operation of the incremental update is: Log tracking: Use database logs, event sources, or data version control mechanisms to track data changes; Difference detection: Compare the data snapshot from the last synchronization with the current data to determine the changed parts; Update only changed data: Only the changed data is written to the target database or data warehouse.

9. A method for processing map data of a digital map according to claim 1, characterized in that: In step S4, the real-time decision algorithm specifically includes: Path planning algorithm: responsible for generating an optimal driving path based on the vehicle's current position and target position, combined with map information and traffic rules. The path planning algorithm includes global path planning and local path planning. Lane keeping algorithm: used to ensure that the vehicle can drive stably in its lane and avoid deviating from the lane. The lane keeping algorithm includes: lane keeping based on vision and lane keeping based on model prediction; Obstacle avoidance algorithm: responsible for making timely decisions when obstacles are detected to avoid collisions with obstacles. The obstacle avoidance algorithm uses sensors to detect obstacles around the vehicle and assesses the risk of collision based on information such as the location, speed, and size of the obstacles. Then, based on the risk assessment results, it formulates obstacle avoidance strategies such as deceleration, lane changing, emergency braking, etc. After the decision is made, local path planning is performed to generate a driving path that avoids obstacles.

10. A method for processing map data of a digital map according to claim 1, characterized in that: In step S5, the map data processing method is optimized specifically including: Data update and correction: timely update map data based on feedback results, and correct any data errors or omissions found; Algorithm optimization: Optimize the map data processing algorithm to address problems found during data analysis, introduce machine learning and deep learning technologies, and improve the intelligence level of map data processing and traffic condition prediction; Quality control and evaluation: Establish a quality control system to monitor and evaluate the processing of map data, regularly evaluate the accuracy and availability of map data, and ensure that the quality of map data meets user needs; User participation and feedback mechanism: Establish a user participation mechanism to encourage users to actively feedback traffic information and map issues, and improve the feedback processing process to ensure that user feedback can be responded to and processed in a timely manner.

Images