High-positioning-precision vehicle adaptive navigation method and vehicle-mounted navigator

Abstract

The invention relates to the technical field of navigation, in particular to a high-positioning-precision vehicle self-adaptive navigation method and a vehicle-mounted navigator, and the method comprises the following steps: obtaining real-time vehicle positioning data, real-time event information and weather data, carrying out association matching with geographical road sections, and establishing a data record for each road section. According to the invention, quantitative evaluation is carried out on each road segment factor based on the preset reference, and a road segment navigation risk score reflecting a real-time risk level is generated in combination with a dynamic weight distribution mechanism of events and weather, so that the response sensitivity to emergencies is enhanced. By analyzing the vehicle track speed deviation value and correcting the comprehensive traffic cost of the road section, the historical traffic flow density mean value is introduced to be compared with real-time data, and it is ensured that cost calculation gives consideration to both stability and timeliness. In the path planning stage, traffic cost and risk scores are synchronously fused, a screening model of a product of geometric average risk indexes and total cost is adopted, a dual-objective optimization function is constructed, and efficiency and safety requirements are effectively balanced.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation, and particularly to a vehicle adaptive navigation method with high positioning accuracy and a vehicle navigator. Background Art

[0002] The technical field of navigation encompasses a technical system for determining the spatial position of an object or vehicle and guiding its movement through multi-source information fusion and dynamic path planning.

[0003] Existing navigation technologies rely on static or low-frequency updated road network data, and there is a lag in perceiving real-time events and weather changes, resulting in difficulties in path planning to avoid sudden risks in a timely manner. At the same time, they focus on historical average travel time or distance, ignoring the dynamic impact of real-time speed deviation on travel efficiency, resulting in cost estimates deviating from the actual road conditions. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a vehicle adaptive navigation method with high positioning accuracy and a vehicle navigator.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions. A vehicle adaptive navigation method with high positioning accuracy includes the following steps: Obtain real-time vehicle positioning data, real-time event information, and weather data, associate and match them with geographical road segments, establish data records for each road segment, and establish a set of real-time road segment states; Based on the set of real-time road segment states, extract the real-time vehicle positioning data, real-time event information, and weather data of each road segment, compare them with preset benchmarks respectively to obtain road segment factor quantization values, and based on the road segment factor quantization values, set weight coefficients for event impacts and weather impacts to establish a road segment navigation risk score; Retrieve the current vehicle trajectory data, calculate the passing speed of the vehicle trajectory data on each road segment, compare the passing speed with the reference speed of the corresponding road segment provided by the map data to generate a road segment speed deviation amount, determine an adjustment coefficient according to the road segment speed deviation amount, perform correction, and generate a road segment comprehensive passing cost; Based on the road segment comprehensive passing cost, use the current vehicle positioning data to set a navigation starting point, determine a navigation destination in combination with user input, search for all paths connecting the starting point and the destination in the road network map, decompose each path into a road segment sequence, summarize the road segment comprehensive passing costs of each path sequence package, obtain a candidate path cost list, and based on the candidate path cost list and the road segment navigation risk score, perform comparison and screening to obtain an optimal adaptive navigation path.

[0006] Preferably, the step of obtaining the set of real-time road segment states is as follows: Obtain real-time vehicle positioning data, real-time event information, and weather data, collect vehicle longitude and latitude coordinates, receive accident and construction event description texts, obtain temperature and precipitation intensity values through a meteorological interface, and obtain unmatched vehicle positioning data, event information data, and weather data; Based on the unmatched vehicle positioning data, event information data, and weather data, parse the set of polygon vertex coordinates of geographical sections, traverse the spatial relationship between vehicle longitude and latitude coordinates and section polygon vertices. If the vehicle coordinates are within a certain section polygon, bind the corresponding event description text and weather values to the unique identifier of that section, and obtain the associated section-vehicle positioning data, section-event information, and section-weather information; Based on the associated section-vehicle positioning data, section-event information, and section-weather information, append the UTC timestamp of the acquisition moment to each data record according to the timestamp field, and integrate the timestamp field, vehicle positioning coordinate field, event description text field, and meteorological value field under the same section unique identifier to form a section real-time status set.

[0007] Preferably, the steps for obtaining the section factor quantization value are as follows: Based on the section real-time status set, traverse the data records corresponding to each section unique identifier in the set, and separately isolate the set of vehicle positioning coordinates associated with the timestamp field, the set of event description text fields, and the set of meteorological value fields, to obtain the real-time vehicle positioning data set, real-time event information data set, and weather data set for each section; Based on the real-time vehicle positioning data set, real-time event information data set, and weather data set, call the standard coordinate set, standard event level set, and standard meteorological threshold set in the preset benchmark, calculate the Euclidean distance difference between the vehicle positioning coordinates and the standard coordinates in the preset benchmark, match the grade difference value between the event description text and the corresponding event type in the standard event level set, and compare the deviation degree of the meteorological value from the temperature and precipitation intensity in the standard meteorological threshold set, to obtain the section factor quantization value.

[0008] Preferably, the steps for obtaining the section navigation risk score are as follows: Based on the section factor quantization value, parse the factor quantization value data field of each section, isolate the sub-factor quantization value representing the event impact intensity and the sub-factor quantization value representing the weather impact degree, to obtain the section event impact factor quantization value set and the section weather impact factor quantization value set; Based on the section event impact factor quantization value set and the section weather impact factor quantization value set, according to the association frequency distribution of event types and meteorological conditions in historical accident data, allocate the weight ratios of event impact and weather impact, and generate the event impact dynamic weight parameter set and the weather impact dynamic weight parameter set; Calculate the road section navigation risk score for each road section based on the event impact dynamic weight parameter set and the weather impact dynamic weight parameter set.

[0009] Preferably, the steps for obtaining the road section speed deviation amount are as follows: Retrieve the current vehicle trajectory data, obtain the original trajectory data containing the timestamp field and the longitude and latitude coordinate fields from the in-vehicle recorder, parse the timestamp field and the longitude and latitude coordinate fields, and generate a vehicle trajectory coordinate sequence with timestamps. Based on the vehicle trajectory coordinate sequence with timestamps, traverse the trajectory coordinate points corresponding to each road section unique identifier, calculate the displacement distance and time difference between adjacent coordinate points, divide the total displacement distance by the total time difference to obtain the average passing speed of the road section, and generate a set of actual passing speeds for each road section. Based on the set of actual passing speeds for each road section, retrieve the legal speed limit or historical average speed value corresponding to the road section unique identifier stored in the map data, calculate the absolute difference between the actual passing speed and the map reference speed for each road section, and generate the road section speed deviation amount.

[0010] Preferably, the steps for obtaining the road section comprehensive passing cost are as follows: Based on the road section speed deviation amount, retrieve the road section length field and the historical average passing time field corresponding to the road section unique identifier in the map data, and combine with the real-time traffic flow density to generate a set of basic parameters. Based on the set of basic parameters, calculate the adjustment coefficient for each road section. Based on the adjustment coefficient, superimpose the preset benchmark passing cost parameter in the map data to generate the road section comprehensive passing cost.

[0011] Preferably, the steps for obtaining the candidate path cost list are as follows: Based on the road section comprehensive passing cost, obtain the current vehicle longitude and latitude coordinate data in real time, receive the destination name or coordinate input, parse and convert it into the standard geographic coordinate format, and generate the navigation starting point coordinate and the navigation destination coordinate. Based on the navigation starting point coordinate and the navigation destination coordinate, call the topological structure data of the road network map, traverse all feasible paths connecting the starting point and the destination, disassemble each path into an ordered permutation and combination of road section unique identifiers, and generate a set of path road section sequences. Based on the set of path road section sequences, traverse the road section unique identifier sequence of each path, extract the road section comprehensive passing cost corresponding to the road section unique identifier, accumulate the road section comprehensive passing costs of all road sections for each path one by one, and generate the candidate path cost list.

[0012] Preferably, the steps for obtaining the optimal adaptive navigation path are as follows: Based on the candidate path cost list and the road segment navigation risk score, traverse each candidate path in the candidate path cost list, extract the total path cost and the road segment navigation risk score set corresponding to the unique identification codes of all road segments included in the path, and generate a total path cost-road segment risk score association set; Based on the path total cost-section risk score association set, calculate the comprehensive screening index for each candidate path; Based on the comprehensive screening index, the comprehensive screening indexes of all candidate paths are sorted in ascending order, and the path with the smallest comprehensive screening index is selected as the optimal adaptive navigation path with the lowest total cost and the lowest risk.

[0013] The present invention provides a vehicle-mounted navigator, comprising: The real-time data fusion module obtains real-time vehicle positioning data, real-time event information and weather data, associates and matches them with geographical road sections, creates data records for each road section, and establishes a real-time status set for the road section; The risk assessment modeling module extracts the real-time vehicle positioning data, real-time event information and weather data of each road section, and compares them with the preset benchmarks to obtain the quantitative value of the road section factor. Based on the quantitative value of the road section factor, weight coefficients are set for the event impact and weather impact to establish the road section navigation risk score; The dynamic cost optimization module retrieves the current vehicle trajectory data, calculates the speed of the vehicle trajectory data on each road section, and compares the speed with the corresponding road section reference speed provided by the map data to generate a road section speed deviation, determines the adjustment coefficient according to the road section speed deviation, performs correction, and generates a comprehensive road section travel cost; The intelligent path generation module sets the navigation starting point based on the comprehensive travel cost of the road section, uses the current vehicle positioning data, determines the navigation destination in combination with user input, searches for all paths connecting the starting point and the destination in the road network map, and decomposes each path into a road section sequence, summarizes the comprehensive travel cost of the road section of each path sequence package, obtains a candidate path cost list, and compares and screens the candidate path cost list and the road section navigation risk score based on the candidate path cost list and the road section navigation risk score to obtain the preferred adaptive navigation path.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: The present invention quantifies and evaluates each road segment factor based on a preset benchmark, combines a dynamic weight allocation mechanism for events and weather, generates a road segment navigation risk score reflecting the real-time risk level, and enhances the response sensitivity to emergencies. By analyzing the vehicle trajectory speed deviation and correcting the comprehensive road segment passing cost, and introducing the comparison between the historical traffic flow density mean value and real-time data, the stability and timeliness of cost calculation are ensured. In the path planning stage, the passing cost and risk score are synchronously fused, and a screening model of the product of the geometric mean risk index and the total cost is adopted to construct a bi-objective optimization function, effectively balancing the efficiency and safety requirements. The navigation path can still maintain the low-latency update ability under sudden road conditions, and at the same time reduce the local sub-optimal problem caused by the optimization of a single index. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] Please refer to Figure 1 , the present invention provides a technical solution, a vehicle adaptive navigation method with high positioning accuracy, including the following steps: Obtain real-time vehicle positioning data, real-time event information and weather data, associate and match them with road segments on the geography, establish a data record for each road segment, and establish a road segment real-time state set; Based on the road segment real-time state set, extract the real-time vehicle positioning data, real-time event information and weather data of each road segment, and compare them with the preset benchmark respectively to obtain the road segment factor quantization value. Based on the road segment factor quantization value, set weight coefficients for the event impact and weather impact, and establish a road segment navigation risk score; Retrieve the current vehicle trajectory data, calculate the passing speed of the vehicle trajectory data on each road segment, compare the passing speed with the reference speed of the corresponding road segment provided by the map data, generate a road segment speed deviation amount, determine an adjustment coefficient according to the road segment speed deviation amount, and perform correction to generate a road segment comprehensive passing cost; Based on the road segment comprehensive passing cost, use the current vehicle positioning data to set the navigation starting point, combine the user input to determine the navigation destination, search for all paths connecting the starting point and the destination in the road network map, decompose each path into a road segment sequence, summarize the road segment comprehensive passing costs of each path sequence package, obtain a candidate path cost list, and perform comparison and screening based on the candidate path cost list and the road segment navigation risk score to obtain an optimal adaptive navigation path.

[0018] The steps for obtaining the set of real-time road segment status are as follows: Obtain real-time vehicle positioning data, real-time event information, and weather data, collect vehicle longitude and latitude coordinates, receive accident and construction event description texts, obtain temperature and precipitation intensity values through the meteorological interface, and obtain unmatched vehicle positioning data, event information data, and weather data; Based on the unmatched vehicle positioning data, event information data, and weather data, parse the set of geographic road segment polygon vertex coordinates, traverse the spatial relationship between vehicle longitude and latitude coordinates and road segment polygon vertices. If the vehicle coordinates are within a certain road segment polygon, bind the corresponding event description text and weather values to the unique identifier of this road segment, and obtain the associated road segment - vehicle positioning data, road segment - event information, and road segment - weather information; Based on the associated road segment - vehicle positioning data, road segment - event information, and road segment - weather information, append the UTC timestamp of the collection moment to each data record according to the timestamp field, and integrate the timestamp field, vehicle positioning coordinate field, event description text field, and meteorological value field under the same road segment unique identifier to form the set of real-time road segment status.

[0019] Specifically, to obtain real-time vehicle positioning data, real-time event information, and weather data, the vehicle's current longitude and latitude coordinates are continuously collected through an in-vehicle global positioning system (GPS) receiver at a preset frequency, such as once per second. At the same time, the system accesses the data interface of an external traffic information service provider. For example, by subscribing to a real-time traffic event push service based on the TPEG (Transport Protocol Experts Group) or DATEX II standard, or by polling (e.g., once every 5 minutes) the HTTPS API provided by a specific vendor at regular intervals to receive event reports on traffic accidents and road construction. These reports contain event types, location descriptions or coordinates, and detailed event description texts. The system only filters the text descriptions of accident and construction events and performs preliminary parsing to extract key information. In addition, the system also needs to call the meteorological service application programming interface (API), such as using the API provided by a commercial meteorological service, to query the current temperature (in degrees Celsius) and precipitation intensity (in millimeters per hour) values of the area at set time intervals, such as once every 15 minutes, based on the vehicle's current location or the planned route area. These independently collected sequences of vehicle coordinates, lists of event messages, and sets of weather readings, each with its own timestamp but not yet associated with specific road segments on the map road network, together constitute the initial unmatched vehicle positioning data, event information data, and weather data.

[0020] Based on the unmatched vehicle location data, event information data, and weather data obtained in the previous stage, first load the digital map database locally or remotely. This database contains the road network topology, where each road segment is identified by a unique identifier (SegmentID) and is accompanied by a geometric description of its geographical boundary, usually represented as a polygon composed of a sequence of vertices defined by latitude and longitude coordinates. The system then parses out the set of vertex coordinates of these road segment polygons. Next, for each unmatched vehicle location data record (containing latitude and longitude coordinates ), the system traverses the road segment polygons in the map database and applies a spatial geometry judgment algorithm, such as the RayCastingAlgorithm, to determine whether the vehicle coordinate point falls within the interior region of a certain road segment polygon. This algorithm emits a virtual ray from the vehicle coordinate point in an arbitrary fixed direction (e.g., due east) and calculates the number of intersection points of this ray with the boundary edges of the polygon. If the number of intersection points is odd, it is determined that the point is inside the polygon; if it is even or the ray exactly passes through a vertex or coincides with an edge (requiring special handling), it is outside or on the boundary. Once it is determined that the vehicle coordinate is within a certain road segment polygon, the system binds the vehicle location data to the unique identifier of this road segment. Similarly, for the event information data containing location information (which may be coordinates or text descriptions that can be geocoded into coordinates ), the same ray projection algorithm is executed for spatial matching, binding the event description text to the unique identifier of the corresponding road segment. For the weather data, it is bound to the unique identifier of the road segment where the vehicle is located during the same time period. If the weather data is regional, it may be bound to all road segments within that region. Through this series of matching and binding operations, the original, scattered data is effectively associated with specific geographical road segments, obtaining the associated road segment - vehicle location data, road segment - event information, and road segment - weather information.

[0021] Based on the associated road segment-vehicle location data, road segment-event information, and road segment-weather information generated in the previous step, the system needs to perform standardized and structured integration on these associated records. First, it checks whether each associated record contains an accurate acquisition timestamp. If the record itself lacks a timestamp or the timestamp format is inconsistent, the system will supplement or convert it to generate a standard Coordinated Universal Time (UTC) timestamp based on the time recorded during the original data acquisition or the time point when the associated data is received. For example, a vehicle location association record may be processed into a data structure that includes the unique road segment identification code, UTC timestamp, data type (such as 'location'), and specific longitude and latitude coordinates. Next, the system uses an aggregation logic to organize these associated records with UTC timestamps, usually using a data structure such as a hash table (HashMap) or a similar key-value pair, with the unique road segment identification code as the primary key (Key), and organizes all data records belonging to the same road segment identification code together. Specifically, it traverses all associated records, and according to the unique road segment identification code in the record, appends the timestamp field, vehicle location coordinate fields (longitude, latitude), event description text field, and meteorological value fields (temperature, precipitation intensity) to the data set corresponding to the road segment identification code. This set is usually a list or sequence sorted by time, retaining the dynamic information of various data changes over time. For example, for road segment 'SegmentX', its corresponding value is a list containing multiple objects [{timestamp: T1, type: 'location', data: {lon, lat}}, {timestamp: T2, type: 'event', data: {desc}}, {timestamp: T3, type: 'weather', data: {temp, precip}},...]. Finally, through such integration processing of the data for all road segments, a structured road segment real-time status set containing detailed time series status information for each road segment is formed.

[0022] The steps for obtaining the quantification value of the road segment factor are as follows: Based on the road segment real-time status set, traverse the data records corresponding to each unique road segment identification code in the set, and separately isolate the vehicle location coordinate set associated with the timestamp field, the event description text field set, and the meteorological value field set, to obtain the real-time vehicle location data set, real-time event information data set, and weather data set for each road segment; Based on the real-time vehicle positioning dataset, real-time event information dataset, and weather dataset, call the standard coordinate set, standard event level set, and standard meteorological threshold set in the preset benchmark, calculate the Euclidean distance difference between the vehicle positioning coordinates and the standard coordinates in the preset benchmark, match the grade difference value between the event description text and the corresponding event type in the standard event level set, and compare the deviation degree of the meteorological value from the temperature and precipitation intensity in the standard meteorological threshold set to obtain the road section factor quantization value.

[0023] Specifically, based on the real-time road section status set, the system starts the parsing and classification processing flow for the data in this set. Specifically, it will traverse each entry in the set, and these entries use the unique road section identification code as the index or grouping basis. For the specific unique road section identification code being processed currently, the system will access all the associated data record lists and check the data type identification field in each record one by one. If the data type of the record is identified as 'vehicle positioning', then extract its corresponding timestamp field and vehicle positioning coordinate fields (including longitude and latitude), and collect these coordinate points and their timestamps to form a time-sorted real-time vehicle position trajectory sequence of this road section. If the data type of the record is identified as 'event information', then extract its timestamp field and event description text field, and store these text descriptions and their timestamps in a set specifically recording the real-time events occurring on this road section. Similarly, if the data type of the record is identified as 'weather data', then extract its timestamp field and meteorological value field containing specific values such as temperature and precipitation intensity, and integrate these meteorological values and their timestamps into the weather condition time series of this road section. By performing the above classification and extraction operations on all the associated data records of the specified road section, three independent datasets are finally generated for this road section, namely, the real-time vehicle positioning dataset containing a series of timestamped coordinate points, the real-time event information dataset containing a series of timestamped event descriptions, and the weather dataset containing a series of timestamped meteorological parameters.

[0024] Based on the real-time vehicle positioning dataset, real-time event information dataset, and weather dataset generated for each road section in the previous step, the system then calls preset benchmark data for comparative quantification. These preset benchmarks include a standard coordinate set, a standard event level set, and a standard meteorological threshold set. Among them, the "standard coordinate set" is derived from a high-precision digital map and defines the geometric coordinate sequence of the ideal driving center line or lane center line for each road section. The "standard event level set" is a pre-constructed mapping table that is formulated based on historical traffic impact analysis (e.g., analyzing the impact of different types of events on the average travel time) and traffic management specifications. It maps various event description texts (such as "minor accident, occupying the emergency lane", "road closed for construction") to predefined numerical severity levels (e.g., from 1 to 5, with level 0 representing no event). For example, through analyzing historical data, it is found that "road closed for construction" on average leads to an increase in travel time of more than 80%, so its level is set to 5, while "minor accident" on average affects the travel time by an increase of 10% - 20%, and the level is set to 2. This mapping table serves as a benchmark. The "standard meteorological threshold set" is set based on the correlation analysis of historical meteorological data with the traffic accident rate and traffic speed to determine the critical meteorological conditions affecting traffic safety. For example, if the analysis shows that when the precipitation intensity exceeds 10 mm / h, the accident rate rises significantly, then 10 mm / h is set as the heavy precipitation threshold. , similarly, set extreme temperature thresholds, such as below 0°C (icing risk threshold ) or above 40°C (high temperature impact threshold ). The quantification process is specifically as follows: For the latest coordinates in the real-time vehicle positioning dataset , calculate the Euclidean distance difference between it and the nearest point on the standard coordinate set (center line) of the corresponding road section . For the latest event description in the real-time event information dataset, query the standard event level set to obtain its corresponding event level . This level value is the level difference value (relative to level 0 of no event). For the latest meteorological values (temperature , precipitation intensity ) in the weather dataset, compare them with the standard meteorological threshold set and calculate the deviation degree. For example, the precipitation deviation degree can be calculated as , and the temperature deviation degree can be calculated as . Combining these calculated distance differences , event level difference values , and various meteorological deviation degrees (such as ), they jointly constitute the road section factor quantification value of this road section, which can be expressed as a structured data containing multiple quantification indicators.

[0025] The steps to obtain the road section navigation risk score are as follows: Based on the quantization values of road section factors, parse the data fields of the quantization values of each road section factor, separate the sub-factor quantization values representing the intensity of event influence and the sub-factor quantization values representing the degree of weather influence, and obtain the set of road section event influence factor quantization values and the set of road section weather influence factor quantization values; Based on the set of road section event influence factor quantization values and the set of road section weather influence factor quantization values, according to the correlation frequency distribution of event types and meteorological conditions in historical accident data, allocate the weight ratios of event influence and weather influence, and generate the set of event influence dynamic weight parameters and the set of weather influence dynamic weight parameters; Based on the set of event influence dynamic weight parameters and the set of weather influence dynamic weight parameters, calculate the road section navigation risk score of each road section. The calculation formula is: ; Wherein, is the road section navigation risk score, is the event influence dynamic weight parameter of the th road section, is the weather influence dynamic weight parameter of the th road section, is the event influence factor quantization value of the th road section, is the event influence benchmark value of the th road section in the preset benchmark, is the weather influence factor quantization value of the th road section, is the average value of the road section weather influence factors in the same historical period.

[0026] Specifically, based on the quantization values of road section factors of each road section obtained in the previous step, this quantization value is a structured data, which includes multiple sub-indicators such as position deviation degree, event level, precipitation deviation degree and temperature deviation degree. Next, the system performs parsing and separation operations. Specifically, for the data record of the quantization value of the road section factor corresponding to each unique road section identification code, the system accesses each data field therein, identifies and extracts the sub-factor quantization value that clearly represents the influence intensity caused by events such as traffic accidents or road construction. This value is directly taken from the event level ( ) in the quantization value of the road section factor. This level itself quantifies the severity of the event and is used as the event influence factor quantization value ( ) of this road section. At the same time, the system also needs to extract the sub-factor quantization value that represents the influence degree brought by the current weather conditions (such as rainfall, temperature anomaly). This involves comprehensively considering multiple sub-indicators related to weather in the quantization value of the road section factor, such as precipitation deviation degree ( ) and temperature deviation degree ( ), the system fuses the quantified values of these weather-related sub-factors into a single quantified value of the weather impact factor according to the preset rules ( ), one fusion method is to use weighted summation, and the calculation formula is , where and are the fixed weights corresponding to the precipitation impact and temperature impact respectively. For example, according to the analysis of historical safety data, the impact of precipitation on driving risk is slightly greater than that of temperature anomalies, and it can be set as and (the sum of the two is 1). These weights are determined based on the statistical regression analysis of a large amount of historical traffic accident data and the corresponding meteorological records, evaluating the relative contribution degrees of different meteorological factors to the accident incidence rate or traffic efficiency. By performing the above parsing, extraction, and fusion calculations on the quantified values of the factors for each road section, a set of quantified values of the road section event impact factors covering all road sections (i.e., the set of for each road section) and a set of quantified values of the road section weather impact factors (i.e., the set of for each road section) are finally obtained.

[0027] Based on the set of quantified values of the road section event impact factors and the set of quantified values of the road section weather impact factors separated in the previous step, the system needs to dynamically determine the proportions of the event impact and weather impact in the final risk assessment for each road section, that is, allocate weight coefficients. This allocation process is based on the in-depth analysis results of the historical accident database, which needs to contain at least several years of accident records, and each record is associated with the unique identification code of the road section at the time of the accident, the specific event type (and its quantified level), and the detailed meteorological conditions (temperature, precipitation intensity, etc.). The core of the analysis is to calculate the conditional probability or correlation frequency of different event levels and different weather conditions (for example, divided into normal, light rain, heavy rain, icing risk, high temperature and other states according to the preset threshold) and the occurrence of accidents on a specific road section . Specifically, it is necessary to statistically calculate, for example, "on the road section , given that the event level is , and the weather condition is normal, the frequency of accidents occurring" , and "on the road section , given that no event occurs, but the weather condition is , the frequency of accidents occurring" . According to the quantified value of the event impact factor reported in real time by the current road section (corresponding to the event level ) and the quantified value of the weather impact factor (corresponding to the weather condition ), the system searches for the frequency data obtained from the above statistics, extracts the historical accident risk correlation corresponding to the current conditions, and sets the event correlation risk degree , the weather correlation risk degree , and then assigns weights according to the relative magnitudes of these two risk degrees. The calculation method is as follows: the dynamic weight parameter of event impact , the dynamic weight parameter of weather impact , where is an extremely small positive number (e.g., ) to prevent the denominator from being zero, which ensures that is close to 1, and the weights are tilted towards the factors with higher historical associated accident frequencies. For example, if there is an event of level 2 on section A and it is raining heavily, and the historical data shows that the accident frequency associated with events of level 2 on this section is 0.005 times per hour, while the accident frequency associated with heavy rain is 0.015 times per hour, then the calculated , , indicating that the risk contribution brought by the current weather is greater. The calculated and values for all sections are respectively stored in the set of dynamic weight parameters of event impact and the set of dynamic weight parameters of weather impact.

[0028] Formula: , the advantage of this formula is that it intelligently adjusts the contributions of event and weather factors to the final risk score through the dynamic weights and , enabling the score to more accurately reflect the source of the dominant risk under the current conditions. It does not simply add up the impacts, but normalizes the event impact relative to the maximum possible impact ( ), and compares the current weather impact with the historical average level ( ), paying attention to its abnormal degree. This design of relative comparison and dynamic weighting makes the risk score not only consider the absolute severity of the current situation, but also consider its deviation from the benchmark and historical normality as well as the relative risk revealed by historical data, thus providing a more comprehensive and instructive risk measure for route navigation; The steps for obtaining the parameters are as follows: The dynamic weight parameter of event impact ( ) for the th section is calculated according to the method described in the previous step. For example, for section A, based on its current event level 2 and historical accident frequency data, the dynamic weight parameter of event impact is calculated; The steps for obtaining the parameters are as follows: The dynamic weight parameter of weather impact ( ) for the )(It) is also calculated according to the method described in the previous step. For example, for road section A, based on its current heavy rain weather and historical accident frequency data, the weather impact dynamic weight parameter is calculated. ; The steps for obtaining the parameter are as follows: The quantification value of the event impact factor for the th road section ( ) is parsed and separated from the quantification values of road section factors. It directly adopts the quantification level of the real-time event currently occurring on the corresponding road section, and this level comes from the preset "standard event level set". For example, if the event currently occurring on road section A is rated as level 2, then ; The steps for obtaining the parameter are as follows: The event impact reference value ( ) for the th road section in the preset benchmark is a reference value for normalizing the event impact. It represents the maximum or standardized impact level that an event may cause. This value is usually set as the highest event level defined in the "standard event level set", which is determined based on domain knowledge and traffic management practices. For example, if the highest event level set is level 5, then the event impact reference values for all road sections are uniformly set as ; The steps for obtaining the parameter are as follows: The quantification value of the weather impact factor for the th road section ( ) is calculated according to the method described in the previous step by comprehensively considering multiple weather-related sub-factors in the current road section factor quantification values. For example, road section A is currently in heavy rain (precipitation intensity 12 mm / h, exceeding the threshold of 10 mm / h, so the precipitation deviation ), the temperature is normal (temperature deviation ), using the weight , the calculated result is ; The steps for obtaining the parameter are as follows: The average value of the weather impact factors of the road section in the same historical period ( ) is a reference value introduced to evaluate the abnormality of the current weather condition. Its acquisition requires querying the historical meteorological database and the corresponding road section information, and calculating the average value of the weather impact factors corresponding to various weather conditions that have occurred in the th road section in the same time period as the current moment (for example, the same season, the same day of the week, similar time periods). For example, the system queries all the weather records of road section A between 8 - 9 am on spring weekdays in the past three years, calculates the for each record, and then calculates the average value to obtain the average value in the same historical period. For example, for road section A, its average value of the weather impact factors in the same historical period is ; Calculation process: Taking section A as an example, substitute the specific values obtained in the previous parameter acquisition steps for calculation: Given that , , , , , ; Calculate the event impact part: Event risk item ; Calculate the weather impact part: Weather risk item ; Calculate the total navigation risk score: ; This result indicates that the current section navigation risk score for section A is 2.35.

[0029] The steps to obtain the section speed deviation amount are as follows: Retrieve the current vehicle trajectory data, obtain the original trajectory data containing the timestamp field and the longitude and latitude coordinate fields from the on-vehicle recorder, parse the timestamp field and the longitude and latitude coordinate fields, and generate a vehicle trajectory coordinate sequence with timestamps; Based on the vehicle trajectory coordinate sequence with timestamps, traverse the trajectory coordinate points corresponding to each section unique identifier, calculate the displacement distance and time difference between adjacent coordinate points, divide the total displacement distance by the total time difference to obtain the average passing speed of the section, and generate a set of actual passing speeds for each section; Based on the set of actual passing speeds for each section, retrieve the legal speed limit or historical average speed value corresponding to the section unique identifier stored in the map data, calculate the absolute difference between the actual passing speed and the map reference speed section by section, and generate the section speed deviation amount.

[0030] Specifically, to retrieve the current vehicle trajectory data, the system first sends a data request to the vehicle's on-board data recording unit, such as the Telematics Control Unit (TCU) or a driving recorder with GPS recording function, or reads the stored logs to obtain the original trajectory data stream or file. These original data usually contain a series of records arranged in chronological order, with each record embedded with an accurate timestamp field, such as time information in Coordinated Universal Time (UTC) format, and corresponding geographical location information, i.e., longitude and latitude coordinate fields. After receiving these original data, the system executes a parsing program, reads each record one by one, extracts the timestamp information and longitude and latitude coordinate values. During this process, data validity checks are performed, such as checking whether the timestamps are in logical order and whether the coordinate values are within the expected geographical area range, discarding or marking invalid data points, and reorganizing the valid data points that pass the verification in chronological order to generate a timestamped vehicle trajectory coordinate sequence.

[0031] Based on the timestamped vehicle trajectory coordinate sequence generated in the previous step, which contains continuous position and time information of the vehicle during driving, the system then calculates the average driving speed of the vehicle when passing through each specific road section. First, it is necessary to match the points in the trajectory coordinate sequence with the road sections in the road network map (this matching process has been completed in the early steps, i.e., the unique identification code of the road section to which each coordinate point belongs is known). Then, the system traverses each unique identification code of the road sections in the road network. For a specific road section, it filters out all the vehicle trajectory coordinate points that fall within the geographical range of this road section and sorts them by time, denoted as the point sequence , where each point contains time and coordinates . If the number of trajectory points corresponding to this road section is less than two (i.e., ), then the effective speed cannot be calculated, and the actual passing speed of this road section is marked as invalid or null. If the number of points is sufficient ( ), then the system traverses all adjacent coordinate point pairs in the sequence . For each pair of points, using a geographical distance calculation method, such as the Haversine formula, based on their longitude and latitude coordinates and , it calculates the actual ground distance between the two points . At the same time, it calculates the time difference between the two points . It accumulates all the calculated distances between adjacent points within this road section to obtain the total displacement distance , and also accumulates all the time differences to obtain the total time difference . Finally, it divides the total displacement distance by the total time difference to calculate the average passing speed of this road section Finally, a set of actual passing speeds for each road segment is generated.

[0032] Based on the set of actual passing speeds for each road segment obtained in the previous step, which includes the average driving speed of current vehicles on each road segment , the system needs to further compare it with the reference speed to quantify the degree of speed deviation. The system accesses the map database that stores road network information. This database not only contains the geometric information of road segments but also stores the attribute data of each road segment (identified by its unique identification code ), including the speed value as a reference. The selection rule for this reference speed ( ) is as follows: preferentially adopt the legal speed limit value of the road segment recorded in the database. If the legal speed limit information is missing or not applicable (such as road sections without clearly marked speed limits like highway ramps), then adopt the historical average passing speed stored in the database. This historical average speed is obtained through statistical analysis based on long-term (such as the past year) accumulated vehicle passing data and may be dynamic, distinguishing the average speeds in different time periods (such as peak, off-peak, night) or different date types (weekday, weekend). The system will match the most appropriate historical average speed value according to the current time to obtain the map reference speed of the corresponding road segment After that, the system calculates its actual passing speed for each road segment and the map reference speed The absolute difference between them, and the calculation formula is . This difference is the speed deviation amount of the road segment. This calculation is performed for all road segments with actual passing speeds, and finally, a set of road segment speed deviation amounts containing the numerical values of speed deviation amounts for each road segment is generated.

[0033] The steps to obtain the comprehensive passing cost of a road segment are as follows: Based on the road segment speed deviation amount, retrieve the road segment length field and historical average passing time field corresponding to the unique identification code of the road segment in the map data, and combine with the real-time traffic flow density to generate a set of basic parameters; Based on the set of basic parameters, calculate the adjustment coefficient for each road segment. The calculation formula is: ; Among them, is the adjustment coefficient, is the relative speed deviation amount of the th road segment, is the road segment length, is the reference speed of the th road segment in the map data, is the historical average passing time, and are the real-time and historical traffic flow densities respectively; Based on the adjustment coefficient, superimpose the preset baseline passing cost parameters in the map data to generate the comprehensive passing cost of the road segment.

[0034] Specifically, based on the set of road segment speed deviation amounts calculated in the previous step, which contains the difference information between the actual passing speed and the map reference speed of each road segment, the system further integrates and calculates the basic parameters required for the comprehensive passing cost of the road segment. First, for each unique road segment identification code , the system queries the map database, retrieves and reads the accurate length field of this road segment ( ), and at the same time retrieves the average passing time field ( ) recorded for this road segment in a similar historical period (for example, the same weekday peak period), such as a numerical value in seconds. These two pieces of data ( and ) reflect the physical attributes and normal passing efficiency of the road segment. Then, the system also needs to obtain the real-time traffic flow density ( ) reflecting the current traffic congestion situation, which can be obtained by accessing a third-party real-time traffic information service platform (such as the traffic information interfaces of Amap and Baidu Map) to obtain the real-time density data of the corresponding road segment (for example, vehicles / km / lane), or by analyzing the data uploaded from roadside detectors (such as induction loops and microwave detectors), or by estimating using the real-time distribution and speed information of a large number of probe vehicles in the road segment in the vehicle networking platform. When obtaining the real-time traffic flow density, the system also needs to query or calculate from the historical traffic database the historical average traffic flow density ( ) corresponding to the current period as a comparison benchmark. Finally, the road segment length collected for the road segment , historical average passing time , real-time traffic flow density , historical average traffic flow density , together with the relative speed deviation amount obtained or calculated in the previous steps (defined as the ratio of the absolute speed deviation amount to the map reference speed, that is ) and the map reference speed , are jointly organized into a structured record to form the basic parameter set of this road segment.

[0035] Formula: , The advantage of this formula is that it is designed to quantify the comprehensive deviation degree of the current road condition relative to the normal or historical average condition, generating an adjustment coefficient for correcting the baseline passing cost, combining information from two dimensions: speed deviation and density deviation. The first term compares the relative speed deviation Acting on the relationship between the expected travel time (roughly determined by and ) and the historical average travel time , it reflects the degree of influence of speed changes on travel time. The second term directly quantifies the deviation ratio of the real-time traffic flow density relative to the historical average density, focusing on the abnormality of the congestion situation. Adding these two parts together can more comprehensively capture the change in travel cost caused by speed reduction and / or density increase (or decrease); The steps to obtain the parameter are as follows: The relative speed deviation of the section ( ) is calculated based on the speed deviation of the section calculated in the previous step and the map reference speed ( ). The specific calculation formula is . It represents the relative proportion of the current actual speed deviating from the reference speed and is a dimensionless value. For example, for a section B, the reference speed km / h, and the current actual speed km / h, then its absolute speed deviation is 20 km / h, and the relative speed deviation ; The steps to obtain the parameter are as follows: The section length ( ) of the th section is static attribute data directly retrieved from the map database, representing the physical length of the section. This data is provided and maintained by the map provider. For example, by querying the map database, the length of section B is obtained as meters, that is, 0.5 kilometers; The steps to obtain the parameter are as follows: The reference speed ( ) of the th section in the map data is the speed value already obtained from the map database when calculating the speed deviation of the section. Its selection logic is to give priority to the legal speed limit and then the historical average speed. For example, the reference speed (historical average speed) of section B is km / h; The steps to obtain the parameter are as follows: The historical average travel time ( ) of the th section is retrieved from the map database, which reflects the time usually required to pass through this section during a time period similar to the current one. This data is obtained based on long-term historical traffic data statistics. For example, the historical average travel time of section B is seconds; The steps to obtain the parameter are as follows: The real-time traffic flow density ( ) of the is a dynamic parameter obtained by accessing the real-time traffic information service interface or analyzing roadside sensor and vehicle-to-everything (V2X) detection vehicle data, etc., which reflects the density of vehicles on this road section at the current moment. For example, the real-time traffic flow density of road section B obtained is vehicles per kilometer; The steps for obtaining the parameter are as follows: The historical average traffic flow density of the road section ( is queried or calculated from the traffic database, which represents the normal traffic flow density level of this road section during a time period similar to the current one and serves as a comparison benchmark for the real-time density, with the same unit as the real-time density. For example, the average traffic flow density of road section B during the same historical period queried is vehicles per kilometer, and this value must be ensured to be non-zero; Calculation process: Taking road section B as an example, substitute the specific values obtained in the above parameter acquisition steps for calculation: Given , km, km / h, s = h = h, veh / km, veh / km; Calculate the first item (speed deviation impact item): ; Calculate the second item (density deviation impact item): ; Calculate the adjustment coefficient: ; The result shows that the adjustment coefficient of road section B at present is approximately 1.333, and the value 1.333 comprehensively reflects the negative impacts on the traffic conditions from two aspects: the low speed (contributing 1 / 3) and the traffic density significantly higher than the historical average level (contributing 1).

[0036] Based on the adjustment coefficients of each road section calculated in the previous step , this coefficient reflects the deviation degree of the current road condition from the benchmark. Next, the system uses this coefficient to correct the preset benchmark traffic cost, so as to generate a road section comprehensive traffic cost that better conforms to the real-time situation. First, the system needs to retrieve the preset benchmark traffic cost parameters for each road section from the map database ( ), this benchmark cost represents the cost of passing through this section under ideal or average conditions. Its setting can be based on various factors. A commonly used method is to directly adopt the historical average travel time of this section ( ), that is . It can also be a comprehensive cost value calculated based on factors such as distance, road grade, combined with average fuel consumption or wear, or considering the fees of toll sections, etc. This benchmark cost parameter ( ) is preset and stored in the map data attributes. After obtaining the benchmark cost, the system performs an "overlay" operation, applying the adjustment coefficient to the benchmark cost. The specific overlay method usually adopts a multiplicative adjustment model, and the calculation formula is: the comprehensive travel cost of the section . This method enables the adjustment coefficient to scale the benchmark cost up or down proportionally (theoretically, if could be negative). When (indicating that the road condition is worse than the benchmark), the comprehensive cost will be higher than the benchmark cost. When (indicating that the road condition meets the benchmark), the comprehensive cost is equal to the benchmark cost. This calculation formula ensures the positive correlation between cost adjustment and the degree of deviation . For example, for section B with the previously calculated adjustment coefficient , if its benchmark travel cost is set as its historical average travel time seconds, then its comprehensive travel cost of the section is calculated as seconds. The system performs this calculation for all sections and finally generates a set of comprehensive travel cost values for each section.

[0037] The steps to obtain the candidate path cost list are as follows: Based on the comprehensive travel cost of the section, obtain the real-time longitude and latitude coordinate data of the current vehicle, receive the input of the destination name or coordinates, parse and convert them into the standard geographical coordinate format, and generate the navigation starting point coordinates and the navigation destination coordinates; Based on the navigation starting point coordinates and the navigation destination coordinates, call the topological structure data of the road network map, traverse all feasible paths connecting the starting point and the destination, disassemble each path into an ordered permutation and combination of the unique section identification codes, and generate a set of path section sequences; Based on the set of path section sequences, traverse the sequence of unique section identification codes of each path, extract the comprehensive travel cost of the section corresponding to the unique section identification code, and accumulate the comprehensive travel costs of all sections for each path one by one to generate a candidate path cost list.

[0038] Specifically, based on the comprehensive travel cost data of each section calculated in the previous steps, when the user initiates a navigation request, the system first obtains the real-time longitude and latitude coordinate data of the vehicle's current location through the on-vehicle positioning hardware , meanwhile, receive the destination information input by the user through the human-computer interaction interface. This input information can be the name of a point of interest, a detailed address, or a direct longitude and latitude coordinate. After receiving the destination information, if the input is a name or address, the system calls the geocoding service for parsing and converts it into a standard geographical coordinate. , if the parsing fails or is inaccurate, prompt the user to re-enter or select a map point. After obtaining the original starting and ending coordinates, perform the map matching process. Apply matching algorithms such as those based on the Hidden Markov Model or simple nearest neighbor search algorithms to precisely attach these two coordinate points in free space to the points on the closest passable nodes or road segments in the road network graph structure, thereby determining the corresponding navigation starting point node in the road network topology. and the navigation destination node , generating the navigation starting point coordinates and the navigation destination coordinates.

[0039] Based on the navigation starting point coordinates determined in the previous step and the navigation destination coordinates , the system then calls the topological structure data of the road network graph stored locally or in the cloud for path search. This topological structure data is represented in the form of a graph (Graph), where vertices represent intersections, road segment endpoints, or other key locations, and edges represent drivable road segments connecting the vertices. Each edge is associated with a unique road segment identification code and has a key attribute: namely, the previously calculated comprehensive road segment passing cost. , this cost value will be used as the weight or cost of the edge in the path search algorithm. Adopt Yen's K shortest path algorithm and set a target number of paths (for example or ), the goal of the algorithm is to search in the road network graph for the to paths with the lowest cumulative comprehensive road segment passing cost. During the path search process, accumulate and calculate the total path cost based on the value of each edge, and consider road traffic rules (such as one-way streets, prohibited turns, etc.), traverse and explore different routes connecting the starting point and the destination, and finally find the paths with the lowest cost and significantly different path compositions. For each feasible path found, the system represents it as an ordered sequence of unique road segment identification codes, records the identifiers of all road segments passed from the starting node to the destination node in sequence, and finally generates a path segment sequence set containing the path segment sequences of these paths.

[0040] Based on the set of path segment sequences generated in the previous step, this set contains multiple (e.g., ) candidate paths connecting the navigation starting point to the navigation destination. Each path is composed of an ordered list of unique segment identification codes. Next, the system performs a cost aggregation calculation for each path in this set. The specific process is as follows: Traverse each path sequence in the set of path segment sequences , where is the total number of segments contained in the th path. Initialize a total cost accumulator for this path. Then, sequentially traverse each unique segment identification code in the path (where ranges from 1 to ). According to this , query and extract the corresponding comprehensive passage cost value of this segment from the previously calculated and stored segment comprehensive passage cost data . Add the extracted segment cost to the total cost accumulator of the current path, that is, execute . When all segments in a path sequence have been traversed, the final value in the accumulator is the total comprehensive passage cost of this candidate path. The system records each candidate path and its calculated total cost as an entry, and finally generates a candidate path cost list containing all candidate paths and their corresponding total costs.

[0041] The steps to obtain the optimal adaptive navigation path are as follows: Based on the candidate path cost list and the segment navigation risk score, traverse each candidate path in the candidate path cost list, extract the total path cost and the set of segment navigation risk scores corresponding to all unique segment identification codes included in the path, and generate a path total cost - segment risk score association set; Based on the path total cost - segment risk score association set, calculate the comprehensive screening index for each candidate path. The calculation formula is: ; where, is the comprehensive screening index, is the total cost of the th path, is the segment navigation risk score of the th segment in the th path, is the th path, and Based on the comprehensive screening metrics, sort the comprehensive screening metrics of all candidate paths in ascending order, and select the path with the smallest comprehensive screening metric as the optimal adaptive navigation path with both low total cost and low risk.

[0042] Specifically, based on the candidate path cost list generated in the previous steps and the calculated navigation risk score data for each road segment, the system begins to integrate these two parts of information to prepare for the final path screening. Specifically, the system traverses each candidate path entry in the candidate path cost list. For the th candidate path, the system extracts the previously calculated total cost of this path (denoted as ), and the sequence of unique identifiers of the road segments that make up this path, arranged in order , where is the total number of road segments included in the th path. Then, the system uses this road segment sequence to query the stored navigation risk score data for each road segment, and for each unique identifier of the road segment (where ranges from 1 to ), searches for and extracts its corresponding road segment navigation risk score value (denoted as ). Collect all the road segment navigation risk score values extracted for the th path to form a risk score set (such as a list or an array) . Finally, the system associates the extracted total cost of the path with the corresponding road segment navigation risk score set, and can create a data structure containing path identification, total cost, and risk score set. Repeat this process for all candidate paths to generate a path total cost-road segment risk score association set.

[0043] Formula: , The advantage of the formula is that it aims to balance the total cost of the path (such as time) and the overall risk of the path through a comprehensive screening metric , to achieve a dual optimization selection of cost and risk. It does not simply add the cost and risk scores, but uses a multiplicative structure, multiplying the total cost of the path by a factor representing the average risk level of the path. Using the geometric mean instead of the arithmetic mean can better reflect the impact of high-risk road segments in the path and avoid being overly diluted by a large number of low-risk road segments, giving a greater penalty to paths containing extremely high-risk road segments. At the same time, adding 1 to the risk score and then taking the logarithm ensures that the parameter of the logarithm operation is always positive (because ), and can handle the situation of . The final metric is in terms of cost is measured in units (such as seconds), and its numerical value reflects both cost and risk. The smaller the value, the better the path performs in terms of the comprehensive consideration of cost and risk; The steps to obtain the parameter are as follows: For the total cost of the th path ( ), it is the value obtained by accumulating the comprehensive passing costs of all road segments included in this path in the previous step of calculating the candidate path cost list. This value represents the main cost expected to be paid for passing this path. For example, the total cost of the 1st path obtained from the candidate path cost list is seconds; The steps to obtain the parameter are as follows: The navigation risk score of the th road segment in the th path ( ) is the risk quantification value calculated for each road segment based on factors such as real-time events, weather, and historical data in the previous step of calculating the navigation risk score of the road segment. For example, the risk scores of the 3 road segments included in the 1st path are obtained as , , ; The steps to obtain the parameter are as follows: The total number of road segments included in the th path ( ) is the length of the road segment sequence that constitutes this candidate path, that is, how many road segments this path consists of. For example, if the 1st path consists of 3 road segments, then ; Calculation process: Taking the 1st candidate path as an example, substitute the specific values obtained in the previous parameter acquisition steps for calculation: Given seconds, , and the set of road segment risk scores on the path is ; Calculate the logarithmic part : ; ; ; Calculate the logarithm sum : ; Calculate the average value of the logarithm sum : ; Calculate the risk factor (geometric mean part) : ; Calculate the comprehensive screening index : ; The result shows that the comprehensive screening index of the first candidate path is approximately 2263.75. This value combines the original passing cost of the path and its overall risk level, providing a comprehensive evaluation. The lower the value, the better. This index will be used to compare with the indexes of other candidate paths to select the path with the best combination of cost and risk.

[0044] Based on the set of comprehensive screening indexes calculated for all candidate paths in the previous step the system then performs the final path selection step. First, organize all candidate paths and their corresponding comprehensive screening indexes into a list or a similar data structure. Then, call a standard sorting algorithm to process this list. The sorting is based on the numerical values of the comprehensive screening indexes in ascending order (from smallest to largest). After sorting, the first element in the list corresponds to the candidate path with the lowest comprehensive screening index value. The system selects the path with the smallest comprehensive screening index in this sorting result. This path performs best after balancing the passing cost (such as time) and the navigation risk (considering factors such as events and weather), and is recognized as the optimal choice with both low total cost and risk under the current conditions. Finally, determine this selected path as the preferred adaptive navigation path for this navigation task and pass it to the navigation engine for subsequent path guidance.

[0045] The above is only a preferred embodiment of the present invention, and it does not impose other forms of limitations on the present invention. Any person skilled in the relevant art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A vehicle adaptive navigation method with high positioning accuracy, characterized in that, The steps include the following: Obtain real-time vehicle positioning data, real-time event information, and weather data, associate and match them with road segments geographically, establish data records for each road segment, and establish a set of real-time road segment states; Based on the set of real-time road segment states, extract the real-time vehicle positioning data, real-time event information, and weather data of each road segment, compare them with preset benchmarks respectively to obtain road segment factor quantization values, set weight coefficients for event impacts and weather impacts based on the road segment factor quantization values, and establish a road segment navigation risk score; Retrieve the current vehicle trajectory data, calculate the passing speed of the vehicle trajectory data on each road segment, compare the passing speed with the reference speed of the corresponding road segment provided by the map data to generate a road segment speed deviation amount, determine an adjustment coefficient according to the road segment speed deviation amount, perform correction, and generate a comprehensive road segment passing cost; Based on the comprehensive road segment passing cost, use the current vehicle positioning data to set a navigation starting point, determine a navigation destination in combination with user input, search for all paths connecting the starting point and the destination in the road network map, decompose each path into a road segment sequence, summarize the comprehensive road segment passing costs of each path sequence package, obtain a candidate path cost list, and perform comparison and screening based on the candidate path cost list and the road segment navigation risk score to obtain an optimal adaptive navigation path.

2. The high-precision vehicle adaptive navigation method according to claim 1, wherein The steps for obtaining the set of real-time road segment states are as follows: Obtain real-time vehicle positioning data, real-time event information, and weather data, collect vehicle longitude and latitude coordinates, receive accident and construction event description texts, and obtain temperature and precipitation intensity values through a meteorological interface to obtain unmatched vehicle positioning data, event information data, and weather data; Based on the unmatched vehicle positioning data, event information data, and weather data, parse the set of vertex coordinates of the geographical road segment polygon, traverse the spatial relationship between the vehicle longitude and latitude coordinates and the road segment polygon vertices. If the vehicle coordinates are within a certain road segment polygon, bind the corresponding event description text and weather values to the unique identifier of the road segment to obtain associated road segment-vehicle positioning data, road segment-event information, and road segment-weather information; Based on the associated road segment-vehicle positioning data, road segment-event information, and road segment-weather information, append the UTC timestamp of the collection moment to each data record according to the timestamp field, integrate the timestamp field, vehicle positioning coordinate field, event description text field, and meteorological value field under the same road segment unique identifier to form a set of real-time road segment states.

3. The vehicle adaptive navigation method with high positioning accuracy according to claim 1, characterized in that The steps for obtaining the road segment factor quantization value are as follows: Based on the set of real-time road segment states, traverse the data records corresponding to each road segment unique identifier in the set, separately isolate the set of vehicle positioning coordinates associated with the timestamp field, the set of event description text fields, and the set of meteorological value fields to obtain the real-time vehicle positioning data set, real-time event information data set, and weather data set of each road segment; Based on the real-time vehicle positioning dataset, real-time event information dataset, and weather dataset, call the standard coordinate set, standard event level set, and standard meteorological threshold set in the preset benchmark, calculate the Euclidean distance difference between the vehicle positioning coordinates and the standard coordinates in the preset benchmark, match the grade difference value between the event description text and the corresponding event type in the standard event level set, and compare the deviation degree of the meteorological value from the temperature and precipitation intensity in the standard meteorological threshold set to obtain the road section factor quantization value.

4. The high-precision vehicle adaptive navigation method according to claim 1, characterized in that The steps for obtaining the road section navigation risk score are as follows: Based on the road section factor quantization value, parse the factor quantization value data fields of each road section, separate the sub-factor quantization value representing the event impact intensity and the sub-factor quantization value representing the weather impact degree, and obtain the road section event impact factor quantization value set and the road section weather impact factor quantization value set; Based on the road section event impact factor quantization value set and the road section weather impact factor quantization value set, according to the correlation frequency distribution of event types and meteorological conditions in the historical accident data, allocate the weight ratios of event impact and weather impact to generate the event impact dynamic weight parameter set and the weather impact dynamic weight parameter set; Based on the event impact dynamic weight parameter set and the weather impact dynamic weight parameter set, calculate the road section navigation risk score for each road section.

5. The high-precision vehicle adaptive navigation method according to claim 1, characterized in that The steps for obtaining the road section speed deviation amount are as follows: Retrieve the current vehicle trajectory data, obtain the original trajectory data containing the timestamp field and longitude and latitude coordinate fields from the on-vehicle recorder, parse the timestamp field and longitude and latitude coordinate fields, and generate a vehicle trajectory coordinate sequence with timestamps; Based on the vehicle trajectory coordinate sequence with timestamps, traverse the trajectory coordinate points corresponding to each road section unique identification code, calculate the displacement distance and time difference between adjacent coordinate points, divide the total displacement distance by the total time difference to obtain the average passing speed of the road section, and generate the actual passing speed set for each road section; Based on the actual passing speed set for each road section, retrieve the legal speed limit or historical average speed value corresponding to the road section unique identification code stored in the map data, and calculate the absolute difference between the actual passing speed and the map reference speed for each road section to generate the road section speed deviation amount.

6. The high-precision vehicle adaptive navigation method according to claim 1, wherein The steps for obtaining the road section comprehensive passing cost are as follows: Based on the road section speed deviation amount, retrieve the road section length field and historical average passing time field corresponding to the road section unique identification code in the map data, and combine with the real-time traffic flow density to generate the basic parameter set; Based on the basic parameter set, calculate the adjustment coefficient for each road section; Based on the adjustment coefficient, superimpose the preset benchmark passing cost parameter in the map data to generate the road section comprehensive passing cost.

7. The high-precision vehicle adaptive navigation method according to claim 1, characterized in that The steps for obtaining the candidate path cost list are as follows: Based on the road section comprehensive passing cost, obtain the current vehicle longitude and latitude coordinate data in real time, receive the destination name or coordinate input, parse and convert it into the standard geographic coordinate format, and generate the navigation starting point coordinates and the navigation destination coordinates; Based on the navigation starting point coordinates and the navigation destination coordinates, call the topological structure data of the road network graph, traverse all feasible paths connecting the starting point and the destination, decompose each path into an ordered permutation and combination of the unique road section identification codes, and generate a set of path road section sequences; Based on the set of path road section sequences, traverse the unique road section identification code sequences of each path, extract the comprehensive passing cost of the road section corresponding to the unique road section identification code, accumulate the comprehensive passing costs of all road sections path by path, and generate a candidate path cost list.

8. The high-precision vehicle adaptive navigation method according to claim 1, characterized in that The steps for obtaining the optimal adaptive navigation path are as follows: Based on the candidate path cost list and the road section navigation risk score, traverse each candidate path in the candidate path cost list, extract the total path cost and the set of road section navigation risk scores corresponding to all the unique road section identification codes included in the path, and generate a path total cost-road section risk score association set; Based on the path total cost-road section risk score association set, calculate the comprehensive screening index for each candidate path; Based on the comprehensive screening index, sort the comprehensive screening indexes of all candidate paths in ascending order, and select the path with the smallest comprehensive screening index as the optimal adaptive navigation path with both low total cost and low risk.

9. The vehicle navigator for the high positioning accuracy vehicle adaptive navigation method according to any one of claims 1-8, characterized in that, Including: A real-time data fusion module that obtains real-time vehicle positioning data, real-time event information, and weather data, associates and matches them with road sections on the geography, creates a data record for each road section, and establishes a set of real-time road section states; A risk assessment modeling module that extracts the real-time vehicle positioning data, real-time event information, and weather data of each road section, compares them with a preset benchmark respectively to obtain the road section factor quantization value, sets weight coefficients for the event impact and the weather impact based on the road section factor quantization value, and establishes a road section navigation risk score; A dynamic cost optimization module that retrieves the current vehicle trajectory data, calculates the passing speed of the vehicle trajectory data on each road section, compares the passing speed with the reference speed of the corresponding road section provided by the map data to generate a road section speed deviation amount, determines an adjustment coefficient according to the road section speed deviation amount, makes corrections, and generates a comprehensive passing cost for the road section; An intelligent path generation module that, based on the comprehensive passing cost of the road section, uses the current vehicle positioning data to set the navigation starting point, determines the navigation destination in combination with the user input, searches for all paths connecting the starting point and the destination in the road network graph, decomposes each path into a road section sequence, summarizes the comprehensive passing costs of the road section sequence packages of each path, obtains a candidate path cost list, and performs comparison and screening based on the candidate path cost list and the road section navigation risk score to obtain the optimal adaptive navigation path.

Images