Fusion positioning method and system for autonomous vehicle

Abstract

The invention relates to the technical field of modern traffic and intelligent navigation, and discloses a fusion positioning method and system for an autonomous vehicle, and the method comprises the steps: carrying out the format conversion and consistency verification of a satellite signal, sensor data and motion track calculation data through a pre-established multi-source data standardization model; and generating a preliminary direction estimation value by adopting a weighted average fusion algorithm, and carrying out credibility evaluation and correction in combination with a Bayesian probability model. Then, the direction information is smoothed and predicted by using a Kalman filtering algorithm, local updating is performed according to real-time external environment feedback data, and abnormal fluctuation is detected and corrected through multi-time window trend analysis. Finally, the optimized direction information and position information are jointly optimized, a final fusion positioning result is generated, model parameters are periodically updated through a machine learning method, and high-precision, high-reliability and high-adaptability positioning in a complex scene is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of modern transportation and intelligent navigation, and in particular to a fusion positioning method and system for an autonomous driving vehicle. Background Art

[0002] In modern transportation and intelligent navigation, precise positioning technology is the cornerstone for ensuring safe travel and improving efficiency. Its importance is self-evident, especially in scenarios such as autonomous driving and drone navigation, where the accuracy and real-time nature of positioning are directly related to system reliability and user safety. However, current mainstream positioning methods often have significant flaws. Single positioning methods, such as the Global Positioning System (GPS), are susceptible to signal interference or obstruction in complex environments, resulting in reduced accuracy. Furthermore, the lack of unified standards for integrating multi-source data makes it difficult to achieve stable output.

[0003] In existing technologies, the core challenges facing this field mainly focus on how to effectively integrate multi-source data and ensure high credibility of the information. First, data from different sources, such as satellite signals, vehicle sensor information, and motion trajectory estimation data, vary significantly in format and accuracy. This heterogeneity makes data fusion extremely complex. A greater challenge arises from the lack of a unified standard for processing and encoding this data, especially the lack of standardized processing of directional information, which makes it difficult for the system to accurately judge directional changes in dynamic environments. Further deduction shows that due to the imperfect credibility assessment mechanism of directional information, the system is often unable to update and correct erroneous data in a timely manner when faced with complex scenarios, which ultimately affects the reliability of the positioning results.

[0004] Therefore, how to build a unified data processing framework to standardize the encoding of multi-source information and evaluate the credibility of directional data in real time, while ensuring the timely update of positioning information in a dynamic environment, has become a key issue that needs to be urgently addressed in this study. Summary of the Invention

[0005] The present invention provides a fusion positioning method and system for an autonomous driving vehicle to achieve the reliability of positioning results.

[0006] In a first aspect, in order to solve the above technical problems, the present invention provides a fusion positioning method for an autonomous driving vehicle, comprising:

[0007] Converting the format of the original data using a pre-established multi-source data standardization model to generate a first data set;

[0008] Filtering the heterogeneous data in the first data set, if the deviation value of one of the data exceeds a preset deviation threshold, removing the corresponding record to obtain a second data set;

[0009] performing weighted calculation on direction information of different data in the second data set to generate a first direction estimation value;

[0010] Performing a credibility evaluation on the first direction estimation value, and if the credibility is lower than a preset evaluation threshold, triggering a data correction process to obtain a second direction estimation value;

[0011] performing smoothing and prediction on the direction information in the second direction estimate to generate a third direction estimate;

[0012] Acquire real-time feedback data of the external environment, and generate a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold;

[0013] detecting short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, performing smoothing correction based on historical data to obtain a fifth direction estimate;

[0014] The positioning coordinate data obtained in real time is used to jointly optimize the direction information and position information in the fifth direction estimate to generate a fused positioning result;

[0015] The parameters and weight distribution rules of the multi-source data standardization model are regularly updated according to the fusion positioning results to generate a new model parameter configuration for the next round of multi-source data processing cycle.

[0016] Preferably, the filtering of the heterogeneous data in the first data set, if the deviation value of one of the data exceeds a preset deviation threshold, then excluding the corresponding record to obtain the second data set, includes:

[0017] Using a data verification tool to compare the heterogeneous data in the first data set item by item, calculating the deviation value of each data with a preset deviation threshold, and deleting the relevant records of the data if the deviation exceeds the threshold range, thereby obtaining a preliminary filtered data set;

[0018] Cleaning the preliminarily screened data set using a data cleaning tool to generate a cleaned data set;

[0019] Using a data integration tool, the cleaned data set is reorganized according to a unified structure. If a field is missing, it is filled in using pre-established field completion rules to determine the reorganized data set;

[0020] The reorganized data sets are aggregated and formatted into a standard structure to generate the second data set. Preferably, the step of performing weighted calculation on the direction information of different data in the second data set to generate the first direction estimate includes:

[0021] Acquire direction information and real-time data from different data sources from the second data set, assign initial weights to the direction information using a pre-established weight assignment rule, and obtain a preliminarily weighted direction information set;

[0022] Determine the change in signal strength in the real-time data. If the signal strength is higher than a preset change threshold range, increase the weight ratio of the corresponding data source; otherwise, decrease the weight ratio to obtain a dynamically adjusted weight combination;

[0023] performing a weighted average calculation on the initially weighted direction information set according to the dynamically adjusted weight combination to obtain a fused direction information value to obtain a first direction estimation value;

[0024] The method then includes: associating and storing the first direction estimation value with the data source and weight distribution record through a data storage tool to generate a structured direction estimation data record.

[0025] Preferably, the performing of a credibility evaluation on the first direction estimation value, and triggering a data correction process to obtain a second direction estimation value if the credibility is lower than a preset evaluation threshold, includes:

[0026] Obtaining a signal interference level and sensor data records from a dynamic environment based on the first direction estimate to obtain a collated data set;

[0027] According to a pre-established evaluation standard, if the credibility value in the collated data set is lower than a preset credibility threshold, determining a revised data range;

[0028] Reacquiring sensor data from the dynamic environment according to the corrected data range and combining the signal interference level to obtain an adjusted direction reference value;

[0029] The adjusted direction reference value is associated with the reliability value and saved to obtain the second direction estimation value.

[0030] Preferably, the smoothing and predicting of the direction information in the second direction estimate to generate a third direction estimate includes:

[0031] According to the second direction estimation value, obtaining vehicle speed change and road curvature information from the dynamic environment to obtain a sorted parameter set;

[0032] Based on the organized parameter set and in combination with pre-established dynamic environment perception rules, a comprehensive evaluation is performed on the impact of the vehicle speed and road curvature to determine fused direction reference data;

[0033] Continuously adjusting the directional information of the fused directional reference data, and if the fluctuation of the adjusted data exceeds a preset fluctuation threshold, reacquiring real-time parameters in the dynamic environment to obtain adjusted directional information;

[0034] According to the adjusted direction information, the deviation between the prediction result and the actual environment perception data is corrected, the corrected direction estimation value is saved, and the third direction estimation value is generated.

[0035] Preferably, the acquiring of real-time feedback data of the external environment, and generating a fourth direction estimate if a deviation between the feedback data and the third direction estimate exceeds a preset dynamic threshold, includes:

[0036] Obtain satellite signal strength and road feature point data from external environment feedback to obtain a collated environmental data set;

[0037] Performing a deviation analysis on the third direction estimate and the environmental data set, and if the deviation exceeds a preset dynamic threshold, initiating a local data update process to determine a revised direction data range;

[0038] Acquire the latest external environment feedback information, adjust the revised direction data range, and obtain a preliminary revised direction reference value;

[0039] The preliminary corrected direction reference value is finally compared with the real-time data to generate the fourth direction estimation value.

[0040] Preferably, detecting short-term fluctuations in the direction information in the fourth direction estimate and, if abnormal fluctuations are detected, performing smoothing correction based on historical data to obtain a fifth direction estimate includes:

[0041] Obtaining historical records of the fourth direction estimation value, and using a time series decomposition tool to segment the distribution of the historical records in multiple time windows to obtain separate data sets of short-term fluctuations and long-term changes;

[0042] Using a data comparison tool to detect the short-term fluctuations, if the short-term fluctuations exceed a preset fluctuation threshold, the abnormal data is marked as a range to be adjusted;

[0043] Extracting corresponding direction estimation records from a historical data storage unit according to the range to be adjusted, and performing weighted processing on the range to be adjusted using a data smoothing tool to obtain a preliminary revised direction data set;

[0044] According to the preliminary corrected direction data set, a comparison and adjustment is performed using a data verification tool in combination with the long-term change to determine the fifth direction estimate.

[0045] Preferably, the real-time acquired positioning coordinate data jointly optimizes the direction information and position information in the fifth direction estimation value to generate a fused positioning result, including:

[0046] Performing preliminary registration of the fifth direction estimate and the positioning coordinates using a spatial geometric mapping tool, performing calibration processing using a weighted average tool, and determining a preliminary integrated spatial information set;

[0047] If it is detected, based on the preliminary integrated spatial information set, that a deviation between the fifth direction estimate and the position of the positioning coordinate exceeds a preset deviation threshold, a dynamic adjustment is performed using a deviation correction tool to obtain an adjusted data set;

[0048] The direction estimation value and the positioning coordinates in the adjusted data set are finally integrated, and smoothed in combination with historical records to obtain a fused positioning result.

[0049] Preferably, the step of regularly updating the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning results to generate a new model parameter configuration for the next multi-source data processing cycle includes:

[0050] Acquire historical data and real-time feedback data of the fused positioning result, and synchronize them using a time alignment tool to obtain a data set under a unified time reference;

[0051] Classify and summarize the historical data and the real-time feedback data using a data integration tool based on the data set under the unified time base;

[0052] For the classified features, the weighted average tool is used to adjust the weight distribution and determine the weight combination suitable for the current cycle processing;

[0053] Based on the weight combination, if the detected weight distribution deviates from the preset weight threshold, the weight is fine-tuned through a dynamic optimization tool, and the characteristics of the multi-source data are combined to determine the adjusted weight set that meets the cyclic processing requirements;

[0054] The parameters of the multi-source data standardization model configuration are updated using the adjusted weight set and an information fusion tool. The configuration parameters suitable for the next round of multi-source data processing are obtained in combination with the training results.

[0055] In a second aspect, the present invention provides a fusion positioning system for an autonomous driving vehicle, comprising:

[0056] A first acquisition module is used to convert the format of the original data using a pre-established multi-source data standardization model to generate a first data set;

[0057] A second acquisition module is used to filter the heterogeneous data in the first data set, and if the deviation value of one of the data exceeds a preset deviation threshold, the corresponding record is removed to obtain a second data set;

[0058] a third acquisition module, configured to perform weighted calculation on direction information of different data in the second data set to generate a first direction estimation value;

[0059] a fourth acquisition module, configured to perform a credibility evaluation on the first direction estimation value, and trigger a data correction process to obtain a second direction estimation value if the credibility is lower than a preset evaluation threshold;

[0060] a fifth acquisition module, configured to perform smoothing processing and prediction on the direction information in the second direction estimation value to generate a third direction estimation value;

[0061] a sixth acquisition module, configured to acquire real-time feedback data of the external environment, and generate a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold;

[0062] a seventh acquisition module, configured to detect short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, perform smoothing correction based on historical data to obtain a fifth direction estimate;

[0063] an eighth acquisition module, configured to acquire positioning coordinate data in real time, jointly optimize the direction information and position information in the fifth direction estimation value, and generate a fused positioning result;

[0064] A configuration module is used to regularly update the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning results, and generate a new model parameter configuration for the next round of multi-source data processing cycle.

[0065] Compared with the existing technology, the present invention discloses a fusion positioning method for autonomous driving vehicles, which converts the format and performs consistency verification on the raw data from different sources through a pre-established multi-source data standardization model, uses a weighted average fusion algorithm to generate a first direction estimate, and combines it with a Bayesian probability model for credibility assessment and correction. Subsequently, the present invention uses a Kalman filter algorithm to smooth and predict the direction information, and performs local updates based on real-time external environment feedback data, and detects and corrects abnormal fluctuations through multi-time window trend analysis. Finally, the present invention jointly optimizes the optimized direction information and position information to generate the final fusion positioning result, and regularly updates the model parameters through machine learning methods, thereby achieving high-precision, high-reliability and high-adaptability positioning in complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1This is a flow chart of a fusion positioning method for an autonomous driving vehicle provided by an embodiment of the present invention;

[0067] Figure 2 It is a schematic diagram of the structure of a fusion positioning system for an autonomous driving vehicle provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0069] Reference Figure 1 The first embodiment of the present invention provides a flow chart of a fusion positioning method for an autonomous driving vehicle, including the following steps:

[0070] S11, converting the format of the original data using a pre-established multi-source data standardization model to generate a first data set;

[0071] S12, screening the heterogeneous data in the first data set, and if the deviation value of one of the data exceeds a preset deviation threshold, removing the corresponding record to obtain a second data set;

[0072] S13, performing weighted calculation on direction information of different data in the second data set to generate a first direction estimation value;

[0073] S14, performing a credibility assessment on the first direction estimate using a Bayesian probability model. If the credibility is lower than a preset assessment threshold, a data correction process is triggered to obtain a second direction estimate.

[0074] S15, using a Kalman filter algorithm, smoothing and predicting the direction information in the second direction estimate to generate a third direction estimate;

[0075] S16, obtaining real-time feedback data of the external environment, and generating a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold;

[0076] S17, detecting short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, performing smoothing correction based on historical data to obtain a fifth direction estimate;

[0077] S18, using the real-time acquired positioning coordinate data, jointly optimizing the direction information and position information in the fifth direction estimation value to generate a fused positioning result;

[0078] S19, regularly updating the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning result, generating a new model parameter configuration for the next round of multi-source data processing cycle.

[0079] In step S11, the original data is formatted using a pre-established multi-source data standardization model to generate a first data set, including:

[0080] Raw data is obtained from multiple data sources. Satellite signals, sensor data, and motion trajectory estimation data are extracted using pre-established acquisition tools to obtain an initial data set. This initial data set is processed using a format conversion tool. If data fields are inconsistent, adjustments are made using a field mapping table to establish a unified data structure. Based on this unified data structure, the processed data is encoded using standardized encoding rules. Data conforming to a standard format is generated using a pre-set encoding template to obtain a standardized data set. This standardized data set is aggregated using a data integration tool to generate a first data set, which is stored in a designated database for subsequent processing.

[0081] For example, when processing satellite signals, sensor data, and motion trajectory estimation data, a specific scenario can be used to understand how this process is implemented. Consider an intelligent traffic monitoring system that requires vehicle operation data from multiple sources. Satellite signals might come from GPS devices, providing the vehicle's real-time location; sensor data from onboard equipment records speed and acceleration; and motion trajectory estimation data uses algorithms to predict the vehicle's likely path. These data sources come in various formats, with varying field names and units. Therefore, pre-established collection tools are needed to extract the raw data and form an initial data set.

[0082] To convert the format of the initial data set, a format conversion tool can be used to unify data from different sources. Specifically, a pre-established multi-source data standardization model defines unified processing rules for three primary raw data sources: satellite signals, sensor data, and dead reckoning data. This model includes specific format conversion templates and consistency verification logic. During the conversion process, the model first identifies the source type of the input data (e.g., GPS signals, IMU data, wheel speed meter data, visual positioning data, etc.). It then applies pre-defined parsing rules for that type, deconstructing the raw format (e.g., NMEA sentences, CAN bus messages, custom binary streams, image coordinates, etc.) and extracting key information fields. The model then uses a built-in field mapping table to map fields with the same meaning but inconsistent naming or units across different data sources (e.g., "timestamp" and "time_record," latitude and longitude degrees, minutes, seconds and decimal degrees, speed m / s and km / h) to pre-defined standard field names and unified units.

[0083] For example, the time field in GPS data might be "timestamp," while sensor data might be "time_record." A field mapping table unifies both into "time." If GPS location data is in degrees, while sensor data is in meters, the mapping table adjusts the units to ensure consistency. This unified data structure lays the foundation for subsequent processing, avoiding data conflicts caused by format differences and improving data processing accuracy.

[0084] It should be noted that the process of establishing a multi-source data standardization model: the construction of this model begins with an in-depth analysis of multi-source heterogeneous data in autonomous driving scenarios, covering the physical characteristics and logical structure of data sources such as satellite positioning (such as GPS / Beidou's NMEA protocol), inertial sensors (IMU's angular velocity / acceleration raw values), wheel speed meter pulse signals, visual positioning coordinates, and high-precision map feature points. Based on this, a three-layer standardized architecture was designed. The bottom parsing layer customizes binary / text parsing rules for each data source (for example, breaking down the GPGGA field in NMEA sentences into latitude, longitude, altitude, and satellite number; parsing CAN bus ID 0x0A0 messages into yaw rate). The middle mapping layer establishes a dynamic field mapping table to resolve semantic conflicts (for example, renaming the "heading" field in different systems uniformly as "heading" and forcing units to 0-360 degrees; standardizing "timestamps" to nanosecond UTC time in ISO 8601 format). The top validation layer embeds physical logic constraints (for example, a speed range of 0-200 km / h and a road curvature radius threshold of 500 meters) and performs adaptive infill (missing GPS elevation data is supplemented by Kalman filtering and IMU Z-axis acceleration estimation). The model is trained and validated with historical multi-source data, and its parameters are solidified, ultimately forming a multi-source data standardization model that includes data pattern recognition, outlier filtering, unit conversion, and spatial coordinate system conversion (for example, converting WGS84 to the local ENU coordinate system).

[0085] In one possible implementation, the application of standardized encoding rules can further optimize data quality. For example, if processed data needs to conform to specific industry standards, a pre-set encoding template can be used to standardize the time format to "YYYY-MM-DDHH:MM:SS" and the location data to latitude and longitude format, such as "latitude:39.9042, longitude:116.4074." This encoding method generates a standardized data set, ensuring seamless data transfer between different systems and reducing errors caused by inconsistent formats.

[0086] For example, using data integration tools can aggregate standardized data sets into a first dataset. For example, suppose the GPS locations, speeds, and predicted trajectories of all vehicles are integrated into a single table. Each record contains fields such as vehicle ID, time, location, and speed. This first dataset is ultimately generated and stored in a designated database. This integration facilitates subsequent analysis, such as traffic flow forecasting or congestion warnings, significantly improving data utilization efficiency.

[0087] It's important to note that the storage process in a designated database also requires attention to data security and access speed. Storing the first dataset in a distributed database ensures efficient large-scale data queries, while also implementing permission controls to protect data privacy. This approach not only improves data processing reliability but also provides solid support for subsequent steps, such as real-time monitoring or historical data analysis. Through this process, each step—from data collection to standardization and integration—is seamlessly integrated, ensuring data consistency and availability. Ultimately, this provides a high-quality data foundation for decision-making within intelligent transportation systems, delivering enhanced business value and operational efficiency.

[0088] In step S12, the heterogeneous data in the first data set is screened. If the deviation value of one of the data exceeds a preset deviation threshold, the corresponding record is removed to obtain a second data set, including:

[0089] Using a data verification tool to compare the heterogeneous data in the first data set item by item, calculating the deviation value of each data with a preset deviation threshold, and deleting the relevant records of the data if the deviation exceeds the threshold range, thereby obtaining a preliminary filtered data set;

[0090] Cleaning the preliminarily screened data set using a data cleaning tool to generate a cleaned data set;

[0091] Using a data integration tool, the cleaned data set is reorganized according to a unified structure. If a field is missing, it is filled in using pre-established field completion rules to determine the reorganized data set;

[0092] The reorganized data sets are aggregated and formatted into a standard structure to generate the second data set.

[0093] For example, in the scenario of an intelligent traffic monitoring system, for the processing flow of the first data set, the specific implementation methods of each technical topic can be explored in depth from multiple perspectives and analyzed in detail in combination with the business background.

[0094] For example, the core of using data verification tools lies in the item-by-item comparison and deviation calculation of heterogeneous data. Assuming that the first data set contains location data from a GPS device and speed data from an on-board sensor, the verification tool will compare each vehicle's record one by one to check whether the location data is within a reasonable range, such as whether the longitude and latitude exceed the city boundaries, and whether the speed data conforms to a conventional range, such as 0 to 120 kilometers per hour. If a record shows a speed of 200 kilometers per hour, which clearly exceeds the preset threshold range of 80 to 150 kilometers per hour, its deviation value is calculated and marked as an item to be eliminated. This method can effectively identify abnormal data and provide a more reliable foundation for subsequent processing.

[0095] For example, when using data cleaning tools, removing flagged records from a pre-screened data set is a key step. For example, suppose one of the multiple records for a particular vehicle is flagged for abnormal speed. Using the cleaning tool, this record is removed while retaining other normal records, such as those whose location and time fields meet thresholds. This streamlined data set ensures that subsequent analysis is not affected by outliers.

[0096] For example, the data integration tool's restructuring process focuses on unified structure and field completion. For example, suppose some records in a cleaned data set lack acceleration fields. The integration tool will use pre-established rules to infer missing values based on speed trends. For example, if a car's speed changes from 40 kilometers per hour to 50 kilometers per hour in two seconds, the acceleration can be inferred to a certain value and filled in the field. This completion method makes the reorganized data set more complete and facilitates unified analysis.

[0097] For example, when using data storage tools, the final step is to format the reorganized data set into a standard structure and generate a secondary data set. Suppose the data is formatted as a table containing fields such as vehicle ID, time, location, and speed. The time for each record is standardized, such as 2023-10-01 08:00:00, and location data is stored as latitude and longitude. This standard structure facilitates subsequent system access and processing, ensuring efficient data flow between different modules. Through the detailed processing of each of the above steps, from verification to cleaning, integration, and storage, each step is closely aligned with the needs of intelligent traffic monitoring, ensuring data accuracy and consistency, laying a solid foundation for subsequent business analysis.

[0098] In step S13, performing weighted calculation on the direction information of different data in the second data set to generate a first direction estimation value includes:

[0099] Acquire direction information and real-time data from different data sources from the second data set, assign initial weights to the direction information using a pre-established weight assignment rule, and obtain a preliminarily weighted direction information set;

[0100] Determine the change in signal strength in the real-time data. If the signal strength is higher than a preset change threshold range, increase the weight ratio of the corresponding data source; otherwise, decrease the weight ratio to obtain a dynamically adjusted weight combination;

[0101] performing a weighted average calculation on the initially weighted direction information set according to the dynamically adjusted weight combination to obtain a fused direction information value to obtain a first direction estimation value;

[0102] The method then includes: associating and saving the first direction estimation value with the data source and weight distribution record through a data storage tool to generate a structured direction estimation data record.

[0103] For example, in the context of an intelligent traffic monitoring system, the specific implementation methods for processing the directional information of the second data set can be explored from multiple perspectives and analyzed in detail in combination with the business context. The process of obtaining directional information and real-time data from different data sources can be explored.

[0104] It's understandable that data sources may include roadside cameras, onboard navigation devices, and positioning modules in mobile devices. For example, cameras provide vehicle heading angle data, onboard navigation devices provide direction estimates based on historical trajectories, and mobile devices provide real-time heading information using gyroscopes. Data from these sources is accompanied by historical error rates and signal strength records. For example, a camera might have an error rate of 5% due to lighting conditions, while an onboard device might have an error rate as high as 15% in tunnels with weak signals.

[0105] For example, for links assigned initial weights using a pre-established weighting rule, assuming the weighting rule is based on a comprehensive evaluation of historical error rates and signal strength, a camera might be assigned an initial weight of 0.5 due to its low error rate, an onboard device 0.3, and a mobile terminal 0.2. This initial weighting reflects the differences in the credibility of the data sources, forming a preliminary weighted set of directional information.

[0106] It should be noted that the design of weight distribution rules needs to take into account the particularity of the business scenario. For example, in densely populated urban areas, camera data may be more reliable, while in suburban areas it may rely on vehicle-mounted equipment.

[0107] For example, when using data processing tools to dynamically adjust weights, incorporating changes in signal strength in real-time data is crucial. For example, if the camera's signal strength drops below a preset threshold due to weather conditions, its weight might drop from 0.5 to 0.3. However, if the vehicle's signal strength is above the threshold, its weight might increase from 0.3 to 0.4. This dynamic adjustment better adapts to real-time environmental changes and ensures accurate directional information.

[0108] For example, in the data fusion tool's weighted average calculation, assuming the direction information from three data sources is 10 degrees, 15 degrees, and 20 degrees east of north, respectively, and the dynamically adjusted weights are 0.3, 0.4, and 0.3, the weighted average yields a fused direction value of approximately 14.5 degrees, which serves as the first direction estimate. This fusion method combines the advantages of multiple data sources and improves the stability of direction estimation.

[0109] For example, in the step of using a data storage tool to associate and save data, assume that the first direction estimate of 14.5 degrees is stored together with the data source and weight distribution records as a structured record, including fields such as timestamp and device ID. This structured storage facilitates subsequent tracing and analysis.

[0110] It should be noted that when determining whether a vehicle meets the preset credibility criteria, it can be compared with historical data. If the estimated value deviates from the actual trajectory by less than 2 degrees, it is considered to meet the criteria. This judgment mechanism helps to screen high-quality direction estimation data, supporting subsequent traffic flow analysis or route planning.

[0111] For example, the implementation of each of the aforementioned steps revolves around intelligent traffic monitoring. From data acquisition to dynamic weight adjustment, integration, and storage, each step is closely aligned with the requirements of vehicle direction estimation, ensuring the practicality and reliability of the data. This refined processing effectively improves the accuracy of direction information, providing more valuable data support for traffic management.

[0112] In step S14, the first direction estimation value is subjected to a credibility evaluation. If the credibility is lower than a preset evaluation threshold, a data correction process is triggered to obtain a second direction estimation value, including:

[0113] Obtaining a signal interference level and sensor data records from a dynamic environment based on the first direction estimate to obtain a collated data set;

[0114] According to a pre-established evaluation standard, if the credibility value in the collated data set is lower than a preset credibility threshold, determining a revised data range;

[0115] Reacquiring sensor data from the dynamic environment according to the corrected data range and combining the signal interference level to obtain an adjusted direction reference value;

[0116] The adjusted direction reference value is associated with the reliability value and saved to obtain the second direction estimation value.

[0117] For example, in the context of intelligent traffic monitoring, subsequent processing of the first directional estimate can be refined from the perspectives of signal interference and data discrepancies. Signal interference can arise from environmental factors, such as weather changes or electromagnetic interference, while sensor data records include directional information collected by different devices. For example, in urban road monitoring, sensor data comes from roadside cameras and onboard equipment. Signal interference can blur camera data due to heavy rain, while onboard equipment can experience data fluctuations due to signal obstruction. Data processing tools can clean and categorize this data, removing significant outliers and forming a consolidated data set that provides a foundation for subsequent evaluation.

[0118] For example, the process of using probability calculation tools to evaluate credibility values.

[0119] It is understandable that pre-established evaluation criteria may be based on a combination of historical data deviations and real-time signal quality. For example, if the collated data set shows that the confidence level of an estimate in a certain direction is 0.6, while the preset threshold is 0.8, then this falls below the standard, triggering subsequent processing.

[0120] It should be noted that this evaluation method can promptly detect potential problems in the data and ensure the reliability of directional information.

[0121] For example, when determining the data range that needs correction and reacquiring data, suppose a data correction tool is used to reacquire sensor data from onboard equipment in a dynamic environment, targeting camera data that is subject to significant interference. This tool also adjusts the direction estimate based on the degree of signal interference, such as signal attenuation caused by heavy rain. For example, if the original estimate was 14.5 degrees east of north, it might become 13.8 degrees east of north after adjustment, which serves as the reference value for the adjusted direction. This approach effectively addresses data deviations caused by environmental changes.

[0122] For example, regarding the step of using a data storage tool to associate and save the adjusted direction reference value and the confidence value.

[0123] It's understood that the stored content may include timestamps, device origins, and adjustment records. Assume the adjusted direction reference value is 13.8 degrees, and the confidence level increases to 0.85. This meets the pre-defined evaluation standard of 0.8 and is ultimately confirmed as the direction reference result. This storage method facilitates subsequent tracing and data analysis.

[0124] For example, the business context for implementing each of the aforementioned steps revolves around intelligent traffic monitoring. From signal interference identification to data collation, probability assessment, data correction, and final storage, each step is closely aligned with the requirements for vehicle direction estimation, ensuring data practicality. This refined processing provides a more reliable directional reference for traffic management, particularly improving the adaptability of data processing in complex environments and laying the foundation for subsequent route planning or traffic analysis.

[0125] In step S15, the direction information in the second direction estimate is smoothed and predicted to generate a third direction estimate, including:

[0126] According to the second direction estimation value, obtaining vehicle speed change and road curvature information from the dynamic environment to obtain a sorted parameter set;

[0127] Based on the organized parameter set and in combination with pre-established dynamic environment perception rules, a comprehensive evaluation is performed on the impact of the vehicle speed and road curvature to determine fused direction reference data;

[0128] Continuously adjusting the directional information of the fused directional reference data, and if the fluctuation of the adjusted data exceeds a preset fluctuation threshold, reacquiring real-time parameters in the dynamic environment to obtain adjusted directional information;

[0129] According to the adjusted direction information, the deviation between the prediction result and the actual environment perception data is corrected, the corrected direction estimation value is saved, and the third direction estimation value is generated.

[0130] For example, in the scenario of intelligent traffic monitoring, the subsequent processing of the second direction estimate can start with the vehicle speed changes and road curvature information in the dynamic environment, and conduct detailed analysis in combination with specific business needs.

[0131] Understandably, vehicle speed changes can be affected by real-time traffic conditions, while road curvature information is related to road design and topography. Assume that in urban road monitoring, vehicle speed data is collected through roadside sensors and onboard equipment. Initially, outliers caused by equipment failures, such as speeds that suddenly drop to zero or exceed a reasonable range, are filtered out to form a consolidated parameter set, providing a reliable foundation for subsequent fusion.

[0132] For example, the application of data fusion tools, combined with pre-established dynamic environment perception rules, can comprehensively assess the impact of vehicle speed and road curvature. For example, if a vehicle's speed decreases from 60 km / h to 40 km / h on a curved road with a moderate curvature, the fusion rules might prioritize the impact of the speed reduction on the directional reference data, thereby deriving a more realistic directional reference value. This approach effectively adapts to changes in the dynamic environment.

[0133] For example, when using the smoothing tool, continuous adjustments to the fused directional reference data can prevent sudden changes in directional information. For example, if the adjusted data fluctuates within a preset range of plus or minus 2 degrees and the actual fluctuation reaches 3 degrees, this exceeds the threshold and requires re-acquiring real-time parameters in the dynamic environment, such as vehicle speed and road curvature, to readjust the directional information. This processing method helps maintain the stability of directional information.

[0134] For example, when using information calibration tools, it's crucial to correct for discrepancies between predicted results and actual environmental perception data. For example, if the predicted direction is 13.8 degrees east of north, while the actual perception data is 14.2 degrees east of north, the calibration tool will fine-tune the direction estimate based on the discrepancy, resulting in a more accurate direction estimate. This calibration improves the accuracy of the direction reference.

[0135] For example, using data storage tools to save corrected direction estimates and determine whether they meet stability requirements can provide a basis for subsequent analysis. For example, if the corrected direction value is 14.0 degrees east of north, and the stability assessment shows that the fluctuation range is within the preset threshold, it is confirmed as the final direction reference. This storage and judgment mechanism facilitates data traceability and long-term monitoring.

[0136] For example, from a broader business perspective, each of the aforementioned steps is closely centered around intelligent traffic monitoring. Data collection, fusion, smoothing, calibration, and storage all aim to provide reliable support for vehicle direction estimation. Especially in complex urban road environments, this multi-step collaborative processing adapts to dynamic changes, ensuring the continuity and stability of direction information, providing valuable insights for traffic management and route planning.

[0137] In step S16, real-time feedback data of the external environment is obtained. If the deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold, a fourth direction estimation value is generated, including:

[0138] Obtain satellite signal strength and road feature point data from external environment feedback to obtain a collated environmental data set;

[0139] Performing a deviation analysis on the third direction estimate and the environmental data set, and if the deviation exceeds a preset dynamic threshold, initiating a local data update process to determine a revised direction data range;

[0140] Acquire the latest external environment feedback information, adjust the revised direction data range, and obtain a preliminary revised direction reference value;

[0141] The preliminary corrected direction reference value is finally compared with the real-time data to generate the fourth direction estimation value.

[0142] For example, in the field of intelligent traffic monitoring, satellite signal strength and road feature point data obtained from external environmental feedback can be initially processed using data collation tools. The primary function of data collation tools is to clean and classify the collected raw information to ensure data reliability for subsequent analysis. For example, in an urban road environment, satellite signal strength is affected by tall buildings, resulting in missing or abnormal data. The data collation tool will prioritize data with signal strength below a certain threshold, such as a signal value below 30%. It will also perform coordinate normalization on the road feature point data to form a collated environmental data set, providing a foundation for subsequent comparisons.

[0143] For example, when using data comparison tools to analyze deviations between third-party direction estimates and environmental data sets, a preset dynamic threshold can be used to determine whether corrections are necessary. For example, suppose the third-party direction estimate is 15.0 degrees east of north, while the reference direction in the environmental data set is 16.5 degrees east of north. For a deviation of 1.5 degrees, if the preset dynamic threshold is 1.0 degrees, this is out of range, requiring a local data update. This comparison method can promptly identify potential problems in direction estimates and ensure data accuracy.

[0144] For example, obtaining the latest external environmental feedback is crucial during local data updates. Combining signal strength fluctuations with road feature matching data, for example, on a major urban road, if signal strength drops to 40% due to temporary construction, while road feature points appear as straight lines, the data fusion tool will reduce the weight of directional corrections based on signal strength fluctuations and rely more on road feature matching data to arrive at a preliminary revised directional reference value, such as 16.2 degrees east of north. This approach dynamically adapts to environmental changes and improves the reliability of directional references.

[0145] For example, when using information verification tools, it's particularly important to perform a final comparison of the initially corrected directional reference value with environmental adaptation rules. For example, if the initial correction value is 16.2 degrees east of north, and the real-time data is updated to 16.0 degrees east of north, the deviation is 0.2 degrees, which meets the preset dynamic threshold of 0.5 degrees. The resulting fourth directional estimate is 16.0 degrees east of north. This verification mechanism further refines the directional data, ensuring that the final result is accurate for the actual environment.

[0146] For example, the implementation of data fusion tools hinges on the comprehensive evaluation of multi-source data. In an urban road scenario, assuming satellite signal strength has recovered to 80% and road feature point data closely matches historical records, the fusion tool will consider the impact of both, appropriately weighting signal strength, and ultimately adjusting the direction reference value. This multi-dimensional fusion effectively addresses the uncertainty of a single data source and improves the stability of direction estimation.

[0147] For example, within the overall business context, each of the aforementioned steps is closely centered around intelligent traffic monitoring, forming a complete closed-loop process from data collation to comparison, updating, and verification. Particularly in complex urban environments, this multi-layered data processing approach dynamically adapts to external environmental changes, providing reliable support for vehicle direction estimation and, in turn, providing a crucial reference for traffic management and route planning.

[0148] In step S17, short-term fluctuations of the direction information in the fourth direction estimate are detected. If abnormal fluctuations are detected, smoothing correction is performed based on historical data to obtain a fifth direction estimate, including:

[0149] Obtaining historical records of the fourth direction estimation value, and using a time series decomposition tool to segment the distribution of the historical records in multiple time windows to obtain separate data sets of short-term fluctuations and long-term changes;

[0150] Using a data comparison tool to detect the short-term fluctuations, if the short-term fluctuations exceed a preset fluctuation threshold, the abnormal data is marked as a range to be adjusted;

[0151] Extracting corresponding direction estimation records from a historical data storage unit according to the range to be adjusted, and performing weighted processing on the range to be adjusted using a data smoothing tool to obtain a preliminary revised direction data set;

[0152] According to the preliminary corrected direction data set, a comparison and adjustment is performed using a data verification tool in combination with the long-term change to determine the fifth direction estimate.

[0153] For example, in the field of intelligent traffic monitoring, the historical analysis of fourth-direction estimates can be explored in depth from multiple perspectives. First, regarding the acquisition of historical records from the storage unit, assume that in an urban road scenario, the system stores fourth-direction estimate data generated every minute for the past 24 hours. This data records the changes in vehicle direction on a specific road section, such as fluctuations between 15.0 and 16.5 degrees north-east. When using time series decomposition tools for segmentation processing, the data can be divided into short-term fluctuations and long-term changes. Short-term fluctuations may reflect temporary interference, such as brief jumps in direction values caused by signal obstruction, while long-term changes may indicate gradual changes in road layout or environmental conditions.

[0154] For example, when analyzing short-term fluctuations using a data comparison tool, assume the threshold is set to ±0.5 degrees. If the direction value suddenly jumps to 17.0 degrees east of north within a certain time window, exceeding the threshold, it will be marked as a target for adjustment. This detection method can quickly identify abnormal data and provide a basis for subsequent adjustments.

[0155] It should be noted that the purpose of marking abnormal data is to avoid short-term interference from misleading the overall direction estimation and to ensure the continuity and reliability of the data.

[0156] For example, when weighting the range to be adjusted, suppose the 5-minute records preceding and following an anomaly are extracted from the historical data storage unit and the direction value is found to be stable at approximately 16.0 degrees north-east before the anomaly. Using the data smoothing tool, a lower weight can be assigned to the anomaly and a higher weight to the surrounding stable data, ultimately resulting in a preliminary revised direction data set, such as adjusting it back to 16.1 degrees north-east. This smoothing process effectively reduces the interference of the anomaly data and maintains the stability of the direction data.

[0157] For example, during the final verification and adjustment phase, using data verification tools in conjunction with long-term trends, if the long-term trend shows a direction value gradually shifting toward 16.2 degrees east of north over the past few hours, and the initial correction value is 16.1 degrees east of north, with a deviation within the preset fluctuation range of 0.3 degrees, the fifth direction estimate can be determined to be 16.1 degrees east of north. This verification method, combined with long-term trends, ensures that the final result reflects both the rationality of short-term corrections and conforms to the overall pattern of change.

[0158] For example, from a holistic perspective, the aforementioned steps form a complete data processing chain for urban traffic monitoring. Whether detecting short-term fluctuations or referencing long-term changes, they all aim to improve the stability of direction estimation. This multi-layered processing approach dynamically adapts to various situations, providing strong support for the accuracy of vehicle direction data and, in turn, a reliable basis for traffic management and route optimization.

[0159] In step S18, the real-time acquired positioning coordinate data is used to jointly optimize the direction information and position information in the fifth direction estimation value to generate a fused positioning result, including:

[0160] Performing preliminary registration of the fifth direction estimate and the positioning coordinates using a spatial geometric mapping tool, performing calibration processing using a weighted average tool, and determining a preliminary integrated spatial information set;

[0161] If it is detected, based on the preliminary integrated spatial information set, that a deviation between the fifth direction estimate and the position of the positioning coordinate exceeds a preset deviation threshold, a dynamic adjustment is performed using a deviation correction tool to obtain an adjusted data set;

[0162] The direction estimation value and the positioning coordinates in the adjusted data set are finally integrated, and smoothed in combination with historical records to obtain a fused positioning result.

[0163] For example, in the field of urban traffic monitoring, the specific implementation methods for the integrated processing of vehicle direction and position data can be explored in depth from multiple perspectives. Regarding the step of obtaining the fifth direction estimate and the real-time collected positioning coordinate data from the storage unit, it is assumed that the system stores the direction estimate and the corresponding position coordinate data updated every minute in the past hour. The direction estimate may be 16.1 degrees north-east, while the position coordinate data records the specific latitude and longitude position of the vehicle on a certain road section. When using the data alignment tool for synchronization processing, the system will match the two according to the timestamp to ensure that each set of data is completely consistent in the time dimension, forming a time-consistent direction and position data set.

[0164] For example, regarding the topic of using spatial geometric mapping tools for preliminary alignment, assuming that in an urban road scenario, the system uses geometric mapping to perform a preliminary spatial match between the direction estimate and the position coordinates. If the direction indicates that the vehicle is moving 16.1 degrees north-east, and the position coordinates show that the vehicle is on a straight path on a certain road section, the system will try to align the two. However, due to sensor errors or environmental interference, there may be certain deviations, such as the direction deviating from the actual path by 0.2 degrees. At this time, the weighted average tool is used for calibration. The system will assign different weights based on the reliability of the data, such as assigning a higher weight to the positioning coordinates and a lower weight to the direction estimate, and finally determine the preliminary integrated set of spatial information.

[0165] For example, when detecting deviations beyond a preset threshold and making dynamic adjustments, suppose the preset threshold is 0.3 degrees, and after initial registration, the direction and position deviation is found to be 0.4 degrees, exceeding the threshold. The system will then use a deviation correction tool to dynamically adjust, perhaps interpolating data from surrounding time points to determine an adjusted data set that better reflects the actual scenario, such as reducing the deviation to 0.1 degrees. This approach effectively addresses signal interference or data loss issues common in urban environments.

[0166] For example, regarding the topic of final integration and smoothing with historical records, suppose the adjusted data set shows a direction of 16.0 degrees north-east, and the location coordinates point to a specific point on a certain road section. The system will use information fusion tools to deeply integrate the two, while extracting historical records from the past 10 minutes, and find that the direction value fluctuates between 15.9 degrees and 16.1 degrees north-east. Based on this, the system smoothes the final result, obtaining a fused positioning result of 16.0 degrees north-east. This processing method can reduce data jumps and ensure the continuity of the results, which is particularly important for the stability of vehicle trajectories in urban traffic monitoring.

[0167] For example, from a holistic perspective, the aforementioned steps form a complete data integration chain in urban road scenarios. From time alignment and preliminary registration to deviation correction and final fusion, all aim to improve the matching of direction and position data. This multi-layered processing approach dynamically adapts to the complexities of urban environments, providing reliable support for subsequent traffic management and route planning.

[0168] In step S19, the parameters and weight distribution rules of the multi-source data standardization model are regularly updated according to the fusion positioning results to generate a new model parameter configuration for the next round of multi-source data processing cycle, including:

[0169] Acquire historical data and real-time feedback data of the fused positioning result, and synchronize them using a time alignment tool to obtain a data set under a unified time reference;

[0170] Classify and summarize the historical data and the real-time feedback data using a data integration tool based on the data set under the unified time base;

[0171] For the classified features, the weighted average tool is used to adjust the weight distribution and determine the weight combination suitable for the current cycle processing;

[0172] Based on the weight combination, if the detected weight distribution deviates from the preset weight threshold, the weight is fine-tuned through a dynamic optimization tool, and the characteristics of the multi-source data are combined to determine the adjusted weight set that meets the cyclic processing requirements;

[0173] The parameters of the multi-source data standardization model configuration are updated using the adjusted weight set and an information fusion tool. The configuration parameters suitable for the next round of multi-source data processing are obtained in combination with the training results.

[0174] For example, in the field of urban traffic monitoring, the specific implementation methods for processing fused positioning data can be explored from multiple perspectives, particularly in the integration of historical data and real-time feedback data. Regarding the acquisition of fused positioning historical data and real-time feedback data from a storage unit, assume that the system stores historical data updated every 5 minutes over the past 24 hours, including vehicle direction and position information on urban roads, while real-time feedback data is vehicle status information collected by sensors at the current moment. Because the two types of data may have time differences, such as historical data recorded at 2:00 PM and real-time data at 2:02 PM, the system uses a time alignment tool to synchronize the two to a common time base, for example, aligning them to 2:00 PM in minutes, thus forming a unified data set.

[0175] For example, regarding the topic of using data integration tools to categorize and summarize historical data and real-time feedback data, we can first understand the principles. Categorization primarily involves grouping data based on its source and characteristics, such as classifying historical data as a reference for long-term trends and real-time data as a reflection of immediate status. For example, in an urban road scenario, historical data indicates an average vehicle speed of 40 km / h on a certain road section, while real-time data shows a current speed of 35 km / h. The system will classify data based on these characteristics and provide a basis for subsequent weighting.

[0176] For example, when using a weighted average tool to adjust weight distribution, assuming the system assigns weights based on the freshness of the data, real-time data is assigned a weight of 0.7 due to its immediacy, while historical data is assigned a weight of 0.3 due to its stability. This weight combination is suitable for the current cycle processing and can balance the timeliness and reliability of the data. If the weight distribution is detected to deviate from the preset threshold range, such as the weight of real-time data exceeding 0.8, the system will fine-tune it through dynamic optimization tools, taking into account the characteristics of multi-source data, such as the high volatility of real-time data, and appropriately reduce its weight to 0.75 to form an adjusted weight set.

[0177] For example, regarding the topic of updating model configuration parameters using a modified set of weights, suppose the system uses information fusion tools to apply the modified weights of 0.75 and 0.25 to the fusion algorithm. This, combined with training results, such as the model's performance data from the past week, updates the parameters to adapt to the next round of multi-source data processing. This approach ensures continuous model optimization in urban traffic monitoring and adapts to the dynamically changing road environment. The final configuration parameters provide a more realistic reference for subsequent processing, improving data processing adaptability.

[0178] In summary, the present invention discloses a fusion positioning method for autonomous driving vehicles, which performs format conversion and consistency verification on satellite signals, sensor data, and motion trajectory estimation data through a pre-established multi-source data standardization model, uses a weighted average fusion algorithm to generate a preliminary direction estimate, and combines the Bayesian probability model for credibility assessment and correction. Subsequently, the present invention uses a Kalman filter algorithm to smooth and predict the direction information, and performs local updates based on real-time external environment feedback data, and detects and corrects abnormal fluctuations through multi-time window trend analysis. Finally, the present invention jointly optimizes the optimized direction information and position information to generate the final fusion positioning result, and regularly updates the model parameters through machine learning methods, thereby achieving high-precision, high-reliability, and high-adaptability positioning in complex scenarios.

[0179] Reference Figure 2 The second embodiment of the present invention provides a structural diagram of a fusion positioning system for an autonomous driving vehicle, including:

[0180] A first acquisition module 201 is configured to convert the format of the original data using a pre-established multi-source data standardization model to generate a first data set;

[0181] A second acquisition module 202 is configured to filter the heterogeneous data in the first data set, and if the deviation value of one of the data exceeds a preset deviation threshold, the corresponding record is removed to obtain a second data set;

[0182] A third acquisition module 203 is configured to perform weighted calculation on the direction information of different data in the second data set to generate a first direction estimation value;

[0183] A fourth acquisition module 204 is configured to perform a credibility evaluation on the first direction estimation value, and if the credibility is lower than a preset evaluation threshold, trigger a data correction process to obtain a second direction estimation value;

[0184] a fifth acquisition module 205, configured to perform smoothing and prediction on the direction information in the second direction estimation value to generate a third direction estimation value;

[0185] a sixth acquisition module 206 for acquiring real-time feedback data of the external environment, and generating a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold;

[0186] a seventh acquisition module 207 for detecting short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, performing smoothing correction based on historical data to obtain a fifth direction estimate;

[0187] An eighth acquisition module 208 is configured to acquire positioning coordinate data in real time, jointly optimize the direction information and position information in the fifth direction estimation value, and generate a fused positioning result;

[0188] The configuration module 209 is used to regularly update the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning result, and generate a new model parameter configuration for the next round of multi-source data processing cycle.

[0189] It should be noted that the fusion positioning system for an autonomous driving vehicle provided in an embodiment of the present invention is used to execute all the process steps of the fusion positioning method for an autonomous driving vehicle in the above embodiment. The working principles and beneficial effects of the two correspond one to one, so they will not be repeated here.

[0190] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0191] The specific embodiments described above further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A fusion positioning method for an autonomous driving vehicle, characterized in that: include: Converting the format of the original data using a pre-established multi-source data standardization model to generate a first data set; Filtering the heterogeneous data in the first data set, if the deviation value of one of the data exceeds a preset deviation threshold, removing the corresponding record to obtain a second data set; performing weighted calculation on direction information of different data in the second data set to generate a first direction estimation value; Performing a credibility evaluation on the first direction estimation value, and if the credibility is lower than a preset evaluation threshold, triggering a data correction process to obtain a second direction estimation value; performing smoothing and prediction on the direction information in the second direction estimate to generate a third direction estimate; Acquire real-time feedback data of the external environment, and generate a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold; detecting short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, performing smoothing correction based on historical data to obtain a fifth direction estimate; The positioning coordinate data obtained in real time is used to jointly optimize the direction information and position information in the fifth direction estimate to generate a fused positioning result; The parameters and weight distribution rules of the multi-source data standardization model are regularly updated according to the fusion positioning results to generate a new model parameter configuration for the next round of multi-source data processing cycle.

2. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The filtering of the heterogeneous data in the first data set, and if a deviation value of one of the data exceeds a preset deviation threshold, removing the corresponding record to obtain the second data set, includes: Using a data verification tool to compare the heterogeneous data in the first data set item by item, calculating the deviation value of each data with a preset deviation threshold, and deleting the relevant records of the data if the deviation exceeds the threshold range, thereby obtaining a preliminary filtered data set; Cleaning the preliminarily screened data set using a data cleaning tool to generate a cleaned data set; Using a data integration tool, the cleaned data set is reorganized according to a unified structure. If a field is missing, it is filled in using pre-established field completion rules to determine the reorganized data set; The reorganized data sets are aggregated and formatted into a standard structure to generate the second data set.

3. The fusion positioning method for an autonomous driving vehicle according to claim 1 or 2, characterized in that: The performing weighted calculation on the direction information of different data in the second data set to generate a first direction estimation value includes: Acquire direction information and real-time data from different data sources from the second data set, assign initial weights to the direction information using a pre-established weight assignment rule, and obtain a preliminarily weighted direction information set; Determine the change in signal strength in the real-time data. If the signal strength is higher than a preset change threshold range, increase the weight ratio of the corresponding data source; otherwise, decrease the weight ratio to obtain a dynamically adjusted weight combination; According to the dynamically adjusted weight combination, a weighted average calculation is performed on the initially weighted direction information set to obtain a fused direction information value to obtain a first direction estimation value.

4. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The credibility evaluation of the first direction estimation value is performed, and if the credibility is lower than a preset evaluation threshold, a data correction process is triggered to obtain a second direction estimation value, including: Obtaining a signal interference level and sensor data records from a dynamic environment based on the first direction estimate to obtain a collated data set; According to a pre-established evaluation standard, if the credibility value in the collated data set is lower than a preset credibility threshold, determining a revised data range; Reacquiring sensor data from the dynamic environment according to the corrected data range and combining the signal interference level to obtain an adjusted direction reference value; The adjusted direction reference value is associated with the reliability value and saved to obtain the second direction estimation value.

5. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The smoothing and predicting of the direction information in the second direction estimate to generate a third direction estimate includes: According to the second direction estimation value, obtaining vehicle speed change and road curvature information from the dynamic environment to obtain a sorted parameter set; Based on the organized parameter set and in combination with pre-established dynamic environment perception rules, a comprehensive evaluation is performed on the impact of the vehicle speed and road curvature to determine fused direction reference data; Continuously adjusting the directional information of the fused directional reference data, and if the fluctuation of the adjusted data exceeds a preset fluctuation threshold, reacquiring real-time parameters in the dynamic environment to obtain adjusted directional information; According to the adjusted direction information, the deviation between the prediction result and the actual environment perception data is corrected, the corrected direction estimation value is saved, and the third direction estimation value is generated.

6. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The acquiring of real-time feedback data of the external environment, and generating a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold, includes: Obtain satellite signal strength and road feature point data from external environment feedback to obtain a collated environmental data set; Performing a deviation analysis on the third direction estimate and the environmental data set, and if the deviation exceeds a preset dynamic threshold, initiating a local data update process to determine a revised direction data range; Acquire the latest external environment feedback information, adjust the revised direction data range, and obtain a preliminary revised direction reference value; The preliminary corrected direction reference value is finally compared with the real-time data to generate the fourth direction estimation value.

7. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The detecting short-term fluctuation of the direction information in the fourth direction estimate, and if abnormal fluctuation is detected, performing smoothing correction based on historical data to obtain a fifth direction estimate, includes: Obtaining historical records of the fourth direction estimation value, and using a time series decomposition tool to segment the distribution of the historical records in multiple time windows to obtain separate data sets of short-term fluctuations and long-term changes; Using a data comparison tool to detect the short-term fluctuations, if the short-term fluctuations exceed a preset fluctuation threshold, the abnormal data is marked as a range to be adjusted; Extracting corresponding direction estimation records from a historical data storage unit according to the range to be adjusted, and performing weighted processing on the range to be adjusted using a data smoothing tool to obtain a preliminary revised direction data set; According to the preliminary corrected direction data set, a comparison and adjustment is performed using a data verification tool in combination with the long-term change to determine the fifth direction estimate.

8. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The real-time acquired positioning coordinate data, and jointly optimizing the direction information and position information in the fifth direction estimation value to generate a fused positioning result, include: Performing preliminary registration of the fifth direction estimate and the positioning coordinates using a spatial geometric mapping tool, performing calibration processing using a weighted average tool, and determining a preliminary integrated spatial information set; If it is detected, based on the preliminary integrated spatial information set, that a deviation between the fifth direction estimate and the position of the positioning coordinate exceeds a preset deviation threshold, a dynamic adjustment is performed using a deviation correction tool to obtain an adjusted data set; The direction estimation value and the positioning coordinates in the adjusted data set are finally integrated, and smoothed in combination with historical records to obtain a fused positioning result.

9. The fusion positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The method of regularly updating the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning results to generate a new model parameter configuration for the next round of multi-source data processing cycle includes: Acquire historical data and real-time feedback data of the fused positioning result, and synchronize them using a time alignment tool to obtain a data set under a unified time reference; Classify and summarize the historical data and the real-time feedback data using a data integration tool based on the data set under the unified time base; For the classified features, the weighted average tool is used to adjust the weight distribution and determine the weight combination suitable for the current cycle processing; Based on the weight combination, if the detected weight distribution deviates from the preset weight threshold, the weight is fine-tuned through a dynamic optimization tool, and the characteristics of the multi-source data are combined to determine the adjusted weight set that meets the cyclic processing requirements; The parameters of the multi-source data standardization model configuration are updated using the adjusted weight set and an information fusion tool. The configuration parameters suitable for the next round of multi-source data processing are obtained in combination with the training results.

10. A fusion positioning system for an autonomous driving vehicle, characterized in that: include: A first acquisition module is used to convert the format of the original data using a pre-established multi-source data standardization model to generate a first data set; A second acquisition module is used to filter the heterogeneous data in the first data set, and if the deviation value of one of the data exceeds a preset deviation threshold, the corresponding record is removed to obtain a second data set; a third acquisition module, configured to perform weighted calculation on direction information of different data in the second data set to generate a first direction estimation value; a fourth acquisition module, configured to perform a credibility evaluation on the first direction estimation value, and trigger a data correction process to obtain a second direction estimation value if the credibility is lower than a preset evaluation threshold; a fifth acquisition module, configured to perform smoothing processing and prediction on the direction information in the second direction estimation value to generate a third direction estimation value; a sixth acquisition module, configured to acquire real-time feedback data of the external environment, and generate a fourth direction estimation value if a deviation between the feedback data and the third direction estimation value exceeds a preset dynamic threshold; a seventh acquisition module, configured to detect short-term fluctuations in the direction information in the fourth direction estimate, and if abnormal fluctuations are detected, perform smoothing correction based on historical data to obtain a fifth direction estimate; an eighth acquisition module, configured to acquire positioning coordinate data in real time, jointly optimize the direction information and position information in the fifth direction estimation value, and generate a fused positioning result; A configuration module is used to regularly update the parameters and weight distribution rules of the multi-source data standardization model according to the fusion positioning results, and generate a new model parameter configuration for the next round of multi-source data processing cycle.

Images