A vehicle running road area recognition method and system based on GPS feature points

Abstract

The present application relates to the technical field of vehicle big data analysis, in particular to a vehicle running road area recognition method and system based on GPS feature points, which eliminates GPS signal mutation and noise interference by preprocessing driving information, improves the fault tolerance rate of abnormal points, sets the number of mutation feature points, combines the mandatory points to compress the vehicle driving information, reduces the computing power requirement, improves the computing efficiency while ensuring the accuracy.

Description

Technical Field

[0001] This invention relates to the field of vehicle big data analysis technology, specifically to a method and system for identifying vehicle operating road areas based on GPS feature points. Background Technology

[0002] The road area on which vehicles operate is one of the important indicators for constructing the cycle conditions of segmented markets. The types of roads on which vehicles travel during their operating cycles are generally based on the vehicle's original data, which includes information such as vehicle speed, GPS latitude and longitude, and operating time.

[0003] The distribution patterns of various loads borne by a vehicle during its operation under different times and road conditions are important basic data for studying vehicle durability, fatigue failure, and safety. The construction of the vehicle's full life cycle load spectrum needs to consider not only the differences in user driving habits but also the influence of road type.

[0004] According to national standards, urban road areas are divided into four types: urban areas, suburbs, mountainous areas, and expressways. Currently, the marking of driving road types is mainly done manually. The manual marking method is based on GPS map boundaries. People manually map the GPS information of each short driving segment onto the GPS map boundary to mark the driving road type of each short driving segment. This method is inefficient and prone to human error, resulting in low accuracy and therefore cannot meet the actual work requirements. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for identifying vehicle operating road areas based on GPS feature points, thereby solving the technical problems of low efficiency and low accuracy in current vehicle operating road area identification.

[0006] The solution of the present invention to the above-mentioned technical problems is as follows:

[0007] A method for identifying road areas where vehicles operate based on GPS feature points includes the following steps:

[0008] S1. Obtain the vehicle's driving information and preprocess the driving information;

[0009] S2. Divide the preprocessed driving information into multiple short-stroke segments;

[0010] S3. Filter the mutation feature points in the short-stroke segment and determine whether the number of mutation feature points meets the set amount. If yes, proceed to S4; otherwise, proceed to S5.

[0011] S4. Set mandatory points in the short-stroke segment matrix, and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points;

[0012] S5. Determine whether the interval supplementation threshold is met based on the total duration of the short-stroke segment. If yes, supplement the abrupt feature points by dividing the segment using an adaptive time interval, and combine the required points and abrupt feature points to obtain the feature point set. If no, proceed to step S6.

[0013] S6. Supplement abrupt feature points by segmenting and adding points using the velocity threshold, and obtain a feature point set by combining the required points and abrupt feature points.

[0014] S7. Construct a city boundary database, combine it with a set of feature points to determine the region corresponding to each short-distance segment, and obtain the driving area sequence;

[0015] S8. Correct outliers in the driving area sequence.

[0016] Further specifying, step S1 specifically includes:

[0017] S1.1 Obtain vehicle driving information and construct an observation matrix. :

[0018]

[0019] The driving information includes vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude. Represents the column vector of vehicle speeds. This represents the column vector of yaw rate. Represents the longitudinal acceleration column vector. Represents a latitude and longitude matrix. , Represents a longitude column vector. This represents a latitude column vector; the data size for vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is N;

[0020] S1.2. The Kalman filter algorithm is used to denoise the driving information to obtain the observation denoising matrix. .

[0021] Further specifying, step S2 specifically includes:

[0022] S2.1, Denoising the observation matrix Divide the path into multiple short-stroke segments to obtain the corresponding short-stroke segment matrix. :

[0023]

[0024] In the short-stroke segment, the data volume for vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is n; the short-stroke segment includes idling segment and driving segment. This is the observation matrix for the idling segment, containing denoised driving information during the idling period, with dimensions (n - m )×5; This is the driving segment observation matrix, containing denoised driving information during the driving process, with dimensions of [dimension number missing]. m ×5.

[0025] Further specifying, step S3 specifically includes:

[0026] S3.1, using the short-stroke segment matrix Denoising latitude and longitude matrix Determine the heading angle during the car's movement. :

[0027]

[0028] in, Let k be the longitude. Let k+1 be the longitude. Let k-1 be the longitude. Let k be the latitude at time k. Let k+1 be the latitude. Let be the latitude at time k-1;

[0029] S3.2, through The change in heading angle at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of the abrupt change in heading angle. The criterion value for the change in heading angle;

[0030] S3.3, Based on the short-stroke segment matrix Denoising latitude and longitude matrix By obtaining the latitude and longitude corresponding to three consecutive moments, the curvature of the vehicle's trajectory can be obtained. :

[0031]

[0032] in, , ;

[0033] S3.4, through The change in curvature at time k will satisfy the following conditions. The corresponding moment is taken as the feature point of curvature change. This is the criterion value for curvature change;

[0034] S3.5, Based on the short-stroke segment matrix Denoising the longitudinal acceleration column vector ,pass The change in acceleration at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of acceleration change. The criterion value for longitudinal acceleration change;

[0035] S3.6 Determine the short-stroke segment matrix Is the total number of abrupt change feature points in heading angle, curvature, and acceleration less than [a certain value]? If yes, proceed to step S5; otherwise, proceed to step S4. The set value for mutation feature points.

[0036] Further specifying, step S4 specifically includes:

[0037] S4.1, in the short-stroke segment matrix In this process, time points k=1 and k=n are designated as mandatory points.

[0038] S4.2, convert the short-stroke segment matrix In the process, all abrupt change feature points from time k=2 to time k=n-2 are sieved to obtain the heading angle abrupt change feature matrix. Curvature mutation feature matrix and acceleration mutation feature matrix ;

[0039] S4.3, Merging the heading angle abrupt change feature matrix Curvature mutation feature matrix and acceleration mutation feature matrix The merged matrix is ​​obtained. ;

[0040] S4.4, In the merged matrix Selecting from top to bottom Each row vector will be selected. Each row vector is used as a candidate point at the corresponding time.

[0041] S4.5 Obtain the feature point set through the required points and candidate points. .

[0042] Further specifying, step S5 specifically includes:

[0043] S5.1, according to With short-stroke segment matrix The number of abrupt change feature points to be added is determined by the difference between the total number of feature points with abrupt changes in heading angle, curvature, and acceleration. L ;

[0044] S5.2 Calculate the total duration of the short-distance segment. :

[0045]

[0046] in, Short-stroke segment matrix Frequency of vehicle driving information collection;

[0047] S5.3 Determine the total duration of short-stroke segments Does it meet the requirements? If yes, proceed to step S5.4; otherwise, proceed to step S6. The threshold for inter-point supplementation;

[0048] S5.4, through Determine the short-distance inter-point interval ;

[0049] S5.5, In the short-stroke segment matrix In the middle, according to the short-distance supplementary point interval insert L A number of supplementary mutation feature points are identified, and driving information corresponding to the supplementary mutation feature points is generated through linear interpolation.

[0050] S5.6, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

[0051] Further specifying, step S6 specifically includes:

[0052] S6.1, Based on the short-stroke segment matrix medium speed , short-stroke segment matrix It is divided into high-speed and low-speed sections;

[0053] S6.2 Setting the low-speed section supplementary interval and high-speed section supplementary interval According to the low-speed section, the interval for supplementing points is as follows. Insert supplementary abrupt change feature points in all low-speed segments, following the same interval as the high-speed segments. Insert supplementary mutation feature points in all high-speed segments;

[0054] S6.3, Determine whether the supplementary abrupt change characteristic points in the high-speed segment meet the set error by integrating the acceleration. If not, then let... If so, repeat step S6.2; if yes, select supplementary mutation feature points for the low-speed segment, and if the number of supplementary mutation feature points for the low-speed segment is not less than the number of mutation feature points to be supplemented. L Then select according to time order. LIf no, select one supplementary mutation feature point; otherwise, select the remaining number of supplementary mutation feature points in the high-speed segment according to the time sequence.

[0055] S6.4, Generate respectively through linear interpolation L Driving information corresponding to each supplementary mutation feature point;

[0056] S6.5, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

[0057] Further specifying, step S7 specifically includes:

[0058] S7.1 Obtain a typical urban road GPS boundary database and use the urban area matrix Suburban area matrix High-speed region matrix and mountainous area matrix To represent;

[0059] S7.2, Based on the set of feature points Matrix each short-stroke segment Compression into short-stroke segment compression matrix :

[0060] ;

[0061] S7.3 Determine the compression matrix of short-stroke segments Check whether each latitude and longitude is within an urban area, suburban area, highway area, or mountainous area. If so, treat the corresponding latitude and longitude as an interior point and assign a value. =1, i For the first i If no latitude and longitude are specified, the assigned value will be discarded. =1 latitude and longitude, and proceed to step S7.4;

[0062] S7.4 Determine the distance between the remaining latitude and longitude and the boundary of the urban area, suburban area, highway area, or mountainous area. If the value does not exceed the buffer threshold, then the corresponding latitude and longitude will be used as the buffer point and assigned a value. If not, then the corresponding latitude and longitude will be used as the outer point of the buffer zone and assigned a value. =0;

[0063] S7.5 Determine whether the number of interior points in urban areas, suburban areas, highway areas, or mountainous areas is equal to or greater than the specified value. M +2, if yes, then the short-stroke segment belongs to the corresponding region; if no, then execute S7.6;

[0064] S7.6 Determine whether the number of consecutive interior points in the urban area, suburban area, highway area, or mountainous area exceeds the quantity threshold. If not, proceed to S7.7; if so, determine the distance between the consecutive interior points that meet the quantity threshold and the boundary of the urban area, suburban area, highway area, or mountainous area. Does it not exceed the buffer threshold? If yes, then execute S7.7; otherwise, the short-run segment belongs to the corresponding region.

[0065] S7.7 Calculate the percentage of inland points in the urban area, suburban area, highway area, and mountainous area respectively. Number of points outside the buffer zone and Gaussian decay weighted score Determine whether each region simultaneously satisfies:

[0066]

[0067] If yes, then determine that the short-stroke segment belongs to the corresponding region; if not, then mark the corresponding short-stroke segment as an unknown region and remove it; where , The Gaussian decay coefficient;

[0068] S7.8. Collect short-distance segments of the defined region to obtain the travel area sequence:

[0069]

[0070] in, To determine the number of short-stroke segments in the region, For the first A region consisting of short-distance segments.

[0071] Further defined, step S8 specifically comprises:

[0072] S8.1 Determining the short-stroke segment region If a mutation has occurred, the process ends; otherwise, the mutation region is calculated. Compared with the previous area Minimum spherical distance between trajectory points corresponding to short-distance segments :

[0073]

[0074] Determine the minimum spherical distance If the distance is less than the spacing threshold, then the short-stroke segment region is determined. If the region is a geographically discontinuous abrupt change, proceed with step S8.2; otherwise, terminate.

[0075] S8.2 Determining the short-stroke segment region If the proportion of the first y short-journey segment regions and the next y short-journey segment regions in the current 2y+1 regions meets the correction threshold, then the short-journey segments are... If the corresponding region is corrected to meet the correction threshold, the driving region sequence is updated and step S8.3 is executed; otherwise, the process ends.

[0076] S8.3, Short-stroke segment area The first y short-stroke segment regions and the next y short-stroke segment regions are used to construct 2y+1 segment windows. Weighted confidence scores are obtained based on the frequency of occurrence of each region type within the window. The region type with the highest weighted confidence score is selected as the short-stroke segment region. The type.

[0077] A GPS-based feature point-based vehicle operation road area identification system, used to implement the above-mentioned GPS-based feature point-based vehicle operation road area identification method, includes:

[0078] The driving information preprocessing module is used to acquire the vehicle's driving information and preprocess it.

[0079] The short-stroke segmentation module is used to divide the preprocessed driving information into multiple short-stroke segments;

[0080] The mutation feature point filtering module is used to filter mutation feature points in short-stroke segments and determine whether the number of mutation feature points meets the set quantity. If yes, the first feature point set module is activated; otherwise, the second feature point set module is activated.

[0081] The first feature point set module is used to set mandatory points in the short-stroke segment matrix and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points.

[0082] The second feature point set module is used to determine whether the interval supplementation threshold is met based on the total duration of the short-journey segment. If so, the abrupt feature points are supplemented by adaptive time interval division, and the feature point set is obtained by combining the required points and the abrupt feature points. If not, the third feature point set module is activated.

[0083] The third feature point set module is used to supplement abrupt feature points by segmenting points according to the velocity threshold, and to obtain a feature point set by combining the required points and abrupt feature points.

[0084] The driving area sequence acquisition module is used to build a city boundary database, combine it with a set of feature points to determine the region corresponding to each short travel segment, and obtain the driving area sequence.

[0085] The driving area sequence correction module is used to correct outliers in the driving area sequence.

[0086] The beneficial effects of this invention are as follows:

[0087] This invention preprocesses driving information to eliminate GPS signal mutations and noise interference, thereby improving the fault tolerance rate of outliers. It sets the number of mutation feature points and compresses vehicle driving information in combination with mandatory points, reducing computing power requirements and improving computational efficiency while ensuring accuracy. Detailed Implementation

[0088] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0089] Example 1

[0090] This embodiment provides a method for identifying vehicle operating road areas based on GPS feature points, including the following steps:

[0091] S1. Obtain the vehicle's driving information and preprocess the driving information;

[0092] S2. Divide the preprocessed driving information into multiple short-stroke segments;

[0093] S3. Filter the mutation feature points in the short-stroke segment and determine whether the number of mutation feature points meets the set amount. If yes, proceed to S4; otherwise, proceed to S5.

[0094] S4. Set mandatory points in the short-stroke segment matrix, and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points;

[0095] S5. Determine whether the interval supplementation threshold is met based on the total duration of the short-stroke segment. If yes, supplement the abrupt feature points by dividing the segment using an adaptive time interval, and combine the required points and abrupt feature points to obtain the feature point set. If no, proceed to step S6.

[0096] S6. Supplement abrupt feature points by segmenting and adding points using the velocity threshold, and obtain a feature point set by combining the required points and abrupt feature points.

[0097] S7. Construct a city boundary database, combine it with a set of feature points to determine the region corresponding to each short-distance segment, and obtain the driving area sequence;

[0098] S8. Correct outliers in the driving area sequence.

[0099] Among them, short stroke refers to the movement segment of a vehicle from one idle speed to another, which generally consists of an idle segment and a running segment.

[0100] To further explain, step S1 specifically involves:

[0101] S1.1 Obtain vehicle driving information and construct an observation matrix. :

[0102]

[0103] The driving information includes vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude. Represents the column vector of vehicle speeds. This represents the column vector of yaw rate. Represents the longitudinal acceleration column vector. Represents a latitude and longitude matrix. , Represents a longitude column vector. The data volume for the latitude column vectors representing vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is N.

[0104] Specifically, the vehicle's driving status information can be obtained from the CAN bus.

[0105] S1.2. The Kalman filter algorithm is used to denoise the driving information to obtain the observation denoising matrix. .

[0106] Specifically, the Kalman filter algorithm is used to denoise the collected data and ensure its smoothness.

[0107] Step S2 is as follows:

[0108] S2.1, Denoising the observation matrix Divide the path into multiple short-stroke segments to obtain the corresponding short-stroke segment matrix. :

[0109]

[0110] In the short-stroke segment, the data volume for vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is n; the short-stroke segment includes idling segment and driving segment. This is the observation matrix for the idling segment, containing denoised driving information during the idling period, with dimensions ( n - m )×5; This is the driving segment observation matrix, containing denoised driving information during the driving process, with dimensions of [dimension number missing]. m ×5.

[0111] To further explain, step S3 specifically involves:

[0112] S3.1, using the short-stroke segment matrix Denoising latitude and longitude matrix Determine the heading angle during the car's movement. :

[0113]

[0114] in, Let k be the longitude. Let k+1 be the longitude. Let k-1 be the longitude. Let k be the latitude at time k. Let k+1 be the latitude. Let be the latitude at time k-1;

[0115] S3.2, through The change in heading angle at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of the abrupt change in heading angle. The criterion value for the change in heading angle; where, .

[0116] S3.3, Based on the short-stroke segment matrix Denoising latitude and longitude matrix By obtaining the latitude and longitude corresponding to three consecutive moments, the curvature of the vehicle's trajectory can be obtained. :

[0117]

[0118] in, , ;

[0119] S3.4, through The change in curvature at time k will satisfy the following conditions. The corresponding moment is taken as the feature point of curvature change. This is the criterion value for curvature change;

[0120] S3.5, Based on the short-stroke segment matrix Denoising the longitudinal acceleration column vector ,pass The change in acceleration at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of acceleration change. The criterion value for longitudinal acceleration change; .

[0121] S3.6 Determine the short-stroke segment matrix Is the total number of abrupt change feature points in heading angle, curvature, and acceleration less than [a certain value]? If yes, proceed to step S5; otherwise, proceed to step S4. The set amount of mutation feature points, in order to M Let's take 8 as an example to illustrate.

[0122] To further explain, step S4 specifically involves:

[0123] S4.1, in the short-stroke segment matrix In this process, time points k=1 and k=n are designated as mandatory points.

[0124] S4.2, convert the short-stroke segment matrix In the process, all abrupt change feature points from time k=2 to time k=n-2 are sieved to obtain the heading angle abrupt change feature matrix. Curvature mutation feature matrix and acceleration mutation feature matrix ;

[0125] S4.3, Merging the heading angle abrupt change feature matrix Curvature mutation feature matrix and acceleration mutation feature matrix The factors influencing the accuracy of subsequent road type determination are related as follows: abrupt changes in heading angle. Acceleration mutation Due to the abrupt change in curvature, the construction dimension is... The sorting matrix is ​​used to obtain the merge matrix. ;

[0126] S4.4, In the merged matrix Selecting from top to bottom Each row vector will be selected. Each row vector is used as a candidate point at the corresponding time.

[0127] S4.5 Obtain the feature point set through the required points and candidate points. .

[0128] To further explain, step S5 specifically involves:

[0129] S5.1, according to With short-stroke segment matrix The number of abrupt change feature points to be added is determined by the difference between the total number of feature points with abrupt changes in heading angle, curvature, and acceleration. L ;

[0130] S5.2 Calculate the total duration of the short-distance segment. :

[0131]

[0132] in, Short-stroke segment matrix Frequency of vehicle driving information collection;

[0133] S5.3 Determine the total duration of short-stroke segments Does it meet the requirements? If yes, proceed to step S5.4; otherwise, proceed to step S6. The threshold for inter-point supplementation; .

[0134] S5.4, through Determine the short-distance inter-point interval ;

[0135] S5.5, In the short-stroke segment matrix In the middle, according to the short-distance supplementary point interval insert L A number of supplementary mutation feature points are identified, and driving information corresponding to the supplementary mutation feature points is generated through linear interpolation.

[0136] Specifically, taking latitude and longitude as an example, it is in The points supplemented after linear interpolation at time step are:

[0137]

[0138]

[0139] Other data such as vehicle speed yaw rate acceleration This method was used for all of them.

[0140] S5.6, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

[0141] To further explain, step S6 specifically involves:

[0142] S6.1, Based on the short-stroke segment matrix medium speed , short-stroke segment matrix It is divided into high-speed and low-speed sections;

[0143] S6.2 Setting the low-speed section supplementary interval and high-speed section supplementary interval According to the low-speed section, the interval for supplementing points is as follows. Insert supplementary abrupt change feature points in all low-speed segments, following the same interval as the high-speed segments. Insert supplementary mutation feature points in all high-speed segments;

[0144] S6.3, Determine whether the supplementary abrupt change characteristic points in the high-speed segment meet the set error by integrating the acceleration. If not, then let... If so, repeat step S6.2; if yes, select supplementary mutation feature points for the low-speed segment, and if the number of supplementary mutation feature points for the low-speed segment is not less than the number of mutation feature points to be supplemented. L Then select according to time order. L If no, select one supplementary mutation feature point; otherwise, select the remaining number of supplementary mutation feature points in the high-speed segment according to the time sequence.

[0145] S6.4, Generate respectively through linear interpolation L Driving information corresponding to each supplementary mutation feature point;

[0146] S6.5, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

[0147] If a short-stroke segment contains multiple low-speed segments, consider the following special cases:

[0148] (1) If the low-speed segment time is greater than If the high-speed segment lasts longer than a certain amount, it is recorded as an effective low-speed segment; if the high-speed segment lasts longer than a certain amount, it is recorded as an effective low-speed segment. If the speed is high, it is recorded as an effective high-speed segment. The total number of effective low-speed segments and high-speed segments in this short-distance segment is then counted. .

[0149] The number of supplementary mutation feature points to be inserted is known. If the total number of valid fragments The remaining supplementary mutation feature points are determined according to step S5.4, which specifies the short-range supplementary point interval. And obtain the set of feature points .

[0150] (2) The low-speed segment time is greater than If the time is greater than a certain value, it is recorded as an effective low-speed segment; if the high-speed segment time is greater than a certain value, it is recorded as an effective low-speed segment. If the total number of valid fragments The number of points obtained by segmenting and supplementing points according to the speed threshold is counted. If there is:

[0151]

[0152] in It is a positive integer. Indicates rounding down; if Then the interval time and Enlarged to Double the time, and then retake points at this interval. The number of points taken this time should be less than [a certain value]. The remaining mutation feature points are then completed according to step S5.4.

[0153] To further explain, step S7 specifically involves:

[0154] S7.1. Use ArcGIS software to output a typical urban road label GPS delimitation library, obtain a typical urban road GPS delimitation library, and then use the urban area matrix. Suburban area matrix High-speed region matrix and mountainous area matrix To represent;

[0155] S7.2, Based on the set of feature points Matrix each short-stroke segment Compression into short-stroke segment compression matrix :

[0156] ;

[0157] Dimensions .

[0158] S7.3 Determine the compression matrix of short-stroke segments Check whether each latitude and longitude is within an urban area, suburban area, highway area, or mountainous area. If so, treat the corresponding latitude and longitude as an interior point and assign a value. =1, i For the first i If no latitude and longitude are specified, the assigned value will be discarded. =1 latitude and longitude, and proceed to step S7.4;

[0159] Specifically, for example, short-stroke segment compression matrix If every latitude and longitude point is located within a high-speed zone, then... Each latitude and longitude is assigned a value.

[0160] S7.4 Determine the distance between the remaining latitude and longitude and the boundary of the urban area, suburban area, highway area, or mountainous area. Does it not exceed the buffer threshold? The buffer threshold is 50m, i.e. , If the value is greater than 0, a buffer zone with a width of 50m is constructed outside each region; otherwise, the corresponding latitude and longitude are used as buffer points and assigned values. If not, then the corresponding latitude and longitude will be used as the outer point of the buffer zone and assigned a value. =0;

[0161] Specifically, if the number of latitude and longitude coordinates located within the high-speed area is... If one is found, then determine whether the remaining 3 latitude and longitude coordinates are located in the buffer zone to avoid having two different values ​​for the same latitude and longitude coordinate.

[0162] S7.5 Determine whether the number of interior points in urban areas, suburban areas, highway areas, or mountainous areas is equal to or greater than the specified value. If yes, then the short-stroke segment belongs to the corresponding region; otherwise, execute S7.6.

[0163] S7.6 Determine whether the number of consecutive interior points in urban areas, suburban areas, highway areas, or mountainous areas exceeds a threshold. The threshold is... If not, proceed to step S7.7; if yes, determine the distance between the continuous interior points that meet the quantity threshold and the boundary of the urban area, suburban area, highway area, or mountainous area. Does it not exceed the buffer threshold? If yes, then execute S7.7; otherwise, the short-run segment belongs to the corresponding region.

[0164] That is, when the number of consecutive latitude and longitude coordinates within any region, such as an urban area, is greater than or equal to... If the continuous latitude and longitude is located in the buffer zone of other areas, then determine whether it is. If there are any continuous latitude and longitude in the city that are located in the buffer zone of other areas, then execute step S7.7. If there are no continuous latitude and longitude in the city that are located in the buffer zone of other areas, then the current short trip segment is considered to be in the city.

[0165] S7.7 Calculate the percentage of inland points in the urban area, suburban area, highway area, and mountainous area respectively. Number of points outside the buffer zone and Gaussian decay weighted score Determine whether each region simultaneously satisfies:

[0166]

[0167] If yes, then determine that the short-stroke segment belongs to the corresponding region; if not, then mark the corresponding short-stroke segment as an unknown region and discard it, as this segment has no practical significance for constructing the segmented market cycle, and therefore is discarded; , The Gaussian attenuation coefficient is... This indicates that the weight decays to 50 meters. .

[0168] S7.8. Collect short-distance segments of the defined region to obtain the travel area sequence:

[0169] ,

[0170] in, To determine the number of short-stroke segments in the region, For the first A region consisting of short-distance segments.

[0171] To further explain, step S8 specifically involves:

[0172] S8.1 Determining the short-stroke segment region If a mutation has occurred, the process ends; otherwise, the mutation region is calculated. Compared with the previous area Minimum spherical distance between trajectory points corresponding to short-distance segments :

[0173]

[0174] Determine the minimum spherical distance If the distance is less than the spacing threshold (5km), it is determined to be a short-distance segment area. This threshold covers the minimum ramp spacing of urban expressways / highways to avoid frequent changes in area type within short distances. If the region is a geographically discontinuous abrupt change, proceed with step S8.2; otherwise, terminate.

[0175] S8.2 Determining the short-stroke segment region If the proportion of the first y short-range segment regions and the next y short-range segment regions in the current 2y+1 regions is not less than a correction threshold of 80%, then the short-range segments are... If the corresponding region is corrected to meet the correction threshold, the driving region sequence is updated and step S8.3 is executed; otherwise, the process ends.

[0176] S8.3, Short-stroke segment area The first y short-stroke segment regions and the next y short-stroke segment regions are used to construct 2y+1 segment windows. Weighted confidence scores are obtained based on the frequency of occurrence of each region type within the window. The region type with the highest weighted confidence score is selected as the short-stroke segment region. The type.

[0177] Example 2

[0178] Based on the GPS feature point-based vehicle road area identification method described in Embodiment 1, this embodiment provides a GPS feature point-based vehicle road area identification system, including:

[0179] The driving information preprocessing module is used to acquire the vehicle's driving information and preprocess it.

[0180] The short-stroke segmentation module is used to divide the preprocessed driving information into multiple short-stroke segments;

[0181] The mutation feature point filtering module is used to filter mutation feature points in short-stroke segments and determine whether the number of mutation feature points meets the set quantity. If yes, the first feature point set module is activated; otherwise, the second feature point set module is activated.

[0182] The first feature point set module is used to set mandatory points in the short-stroke segment matrix and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points.

[0183] The second feature point set module is used to determine whether the interval supplementation threshold is met based on the total duration of the short-journey segment. If so, the abrupt feature points are supplemented by adaptive time interval division, and the feature point set is obtained by combining the required points and the abrupt feature points. If not, the third feature point set module is activated.

[0184] The third feature point set module is used to supplement abrupt feature points by segmenting points according to the velocity threshold, and to obtain a feature point set by combining the required points and abrupt feature points.

[0185] The driving area sequence acquisition module is used to build a city boundary database, combine it with a set of feature points to determine the region corresponding to each short travel segment, and obtain the driving area sequence.

[0186] The driving area sequence correction module is used to correct outliers in the driving area sequence.

[0187] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present invention.

Claims

1. A method for identifying vehicle operating road areas based on GPS feature points, characterized in that, Includes the following steps: S1. Obtain the vehicle's driving information and preprocess the driving information; S2. Divide the preprocessed driving information into multiple short-stroke segments; S3. Filter the mutation feature points in the short-stroke segment and determine whether the number of mutation feature points meets the set amount. If yes, proceed to S4; otherwise, proceed to S5. S4. Set mandatory points in the short-stroke segment matrix, and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points; S5. Determine whether the interval supplementation threshold is met based on the total duration of the short-stroke segment. If yes, supplement the abrupt feature points by dividing the segment using an adaptive time interval, and combine the required points and abrupt feature points to obtain the feature point set. If no, proceed to step S6. S6. Supplement abrupt feature points by segmenting and adding points using the velocity threshold, and obtain a feature point set by combining the required points and abrupt feature points. S7. Construct a city boundary database, combine it with a set of feature points to determine the region corresponding to each short-distance segment, and obtain the driving area sequence; S8. Correct outliers in the driving area sequence; Step S3 specifically involves: S3.1, using the short-stroke segment matrix Denoising latitude and longitude matrix Determine the heading angle during the car's movement. : in, Let k be the longitude. Let k+1 be the longitude. Let k-1 be the longitude. Let k be the latitude at time k. Let k+1 be the latitude. Let be the latitude at time k-1; S3.2, through The change in heading angle at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of the abrupt change in heading angle. The criterion value for the change in heading angle; S3.3, Based on the short-stroke segment matrix Denoising latitude and longitude matrix By obtaining the latitude and longitude corresponding to three consecutive moments, the curvature of the vehicle's trajectory can be obtained. : in, , ; S3.4, through The change in curvature at time k will satisfy the following conditions. The corresponding moment is taken as the feature point of curvature change. This is the criterion value for curvature change; S3.5, Based on the short-stroke segment matrix Denoising the longitudinal acceleration column vector ,pass The change in acceleration at time k will satisfy the following conditions. The corresponding moment is taken as the characteristic point of acceleration change. The criterion value for longitudinal acceleration change; S3.6 Determine the short-stroke segment matrix Is the total number of abrupt change feature points in heading angle, curvature, and acceleration less than [a certain value]? If yes, proceed to step S5; otherwise, proceed to step S4. The set value for mutation feature points; Step S4 specifically involves: S4.1, in the short-stroke segment matrix In this process, time points k=1 and k=n are designated as mandatory points. S4.2, convert the short-stroke segment matrix In the process, all abrupt change feature points from time k=2 to time k=n-2 are sieved to obtain the heading angle abrupt change feature matrix. Curvature mutation feature matrix and acceleration mutation feature matrix ; S4.3, Merging the heading angle abrupt change feature matrix Curvature mutation feature matrix and acceleration mutation feature matrix The merged matrix is ​​obtained. ; S4.4, In the merged matrix Selecting from top to bottom Each row vector will be selected. Each row vector is used as a candidate point at the corresponding time. S4.5 Obtain the feature point set through the required points and candidate points. ; Step S5 specifically involves: S5.1, according to With short-stroke segment matrix The number of abrupt change feature points to be added is determined by the difference between the total number of feature points with abrupt changes in heading angle, curvature, and acceleration. L ; S5.2 Calculate the total duration of the short-distance segment. : in, Short-stroke segment matrix Frequency of vehicle driving information collection; S5.3 Determine the total duration of short-stroke segments Does it meet the requirements? If yes, proceed to step S5.4; otherwise, proceed to step S6. The threshold for inter-point supplementation; S5.4, through Determine the short-distance inter-point interval ; S5.5, In the short-stroke segment matrix In the middle, according to the short-distance supplementary point interval insert L A number of supplementary mutation feature points are identified, and driving information corresponding to the supplementary mutation feature points is generated through linear interpolation. S5.6, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

2. The method for identifying vehicle operating road areas based on GPS feature points according to claim 1, characterized in that, Step S1 specifically involves: S1.1 Obtain vehicle driving information and construct an observation matrix. : The driving information includes vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude. Represents the column vector of vehicle speeds. This represents the column vector of yaw rate. Represents the longitudinal acceleration column vector. Represents a latitude and longitude matrix. , Represents a longitude column vector. This represents a latitude column vector; the data size for vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is N; S1.

2. The Kalman filter algorithm is used to denoise the driving information to obtain the observation denoising matrix. .

3. The method for identifying vehicle operating road areas based on GPS feature points according to claim 2, characterized in that, Step S2 specifically involves: S2.1, Denoising the observation matrix Divide the path into multiple short-stroke segments to obtain the corresponding short-stroke segment matrix. : In the short-stroke segment, the data volume for vehicle speed, yaw rate, longitudinal acceleration, and latitude and longitude is n; the short-stroke segment includes idling segment and driving segment. This is the observation matrix for the idling segment, containing denoised driving information during the idling period, with dimensions ( n m )×5; This is the driving segment observation matrix, containing denoised driving information during the driving process, with dimensions of [dimension number missing]. m ×5.

4. The method for identifying vehicle operating road areas based on GPS feature points according to claim 1, characterized in that, Step S6 specifically involves: S6.1, Based on the short-stroke segment matrix medium speed , short-stroke segment matrix It is divided into high-speed and low-speed sections; S6.2 Setting the low-speed section supplementary interval and high-speed section supplementary interval According to the low-speed section, the interval for supplementing points is as follows. Insert supplementary abrupt change feature points in all low-speed segments, following the same interval as the high-speed segments. Insert supplementary mutation feature points in all high-speed segments; S6.3, Determine whether the supplementary abrupt change characteristic points in the high-speed segment meet the set error by integrating the acceleration. If not, then let... If so, repeat step S6.2; if yes, select supplementary mutation feature points for the low-speed segment, and if the number of supplementary mutation feature points for the low-speed segment is not less than the number of mutation feature points to be supplemented. L Then select according to time order. L If no, select one supplementary mutation feature point; otherwise, select the remaining number of supplementary mutation feature points in the high-speed segment according to the time sequence. S6.4, Generate respectively through linear interpolation L Driving information corresponding to each supplementary mutation feature point; S6.5, using the short-stroke segment matrix The two required points, all mutation feature points and L The feature point set is obtained by supplementing mutation feature points. .

5. The method for identifying vehicle operating road areas based on GPS feature points according to claim 4, characterized in that, Step S7 specifically involves: S7.1 Obtain a typical urban road GPS boundary database and use the urban area matrix Suburban area matrix High-speed region matrix and mountainous area matrix To represent; S7.2, Based on the set of feature points Matrix each short-stroke segment Compression into short-stroke segment compression matrix : ； S7.3 Determine the compression matrix of short-stroke segments Check whether each latitude and longitude is within an urban area, suburban area, highway area, or mountainous area. If so, treat the corresponding latitude and longitude as an interior point and assign a value. =1, i For the first i If no latitude and longitude are specified, the assigned value will be discarded. =1 latitude and longitude, and proceed to step S7.4; S7.4 Determine the distance between the remaining latitude and longitude and the boundary of the urban area, suburban area, highway area, or mountainous area. If the value does not exceed the buffer threshold, then the corresponding latitude and longitude will be used as the buffer point and assigned a value. If not, then the corresponding latitude and longitude will be used as the outer point of the buffer zone and assigned a value. =0; S7.5 Determine whether the number of interior points in urban areas, suburban areas, highway areas, or mountainous areas is equal to or greater than the specified value. M +2, if yes, then the short-stroke segment belongs to the corresponding region; if no, then execute S7.6; S7.6 Determine whether the number of consecutive interior points in the urban area, suburban area, highway area, or mountainous area exceeds the quantity threshold. If not, proceed to S7.7; if so, determine the distance between the consecutive interior points that meet the quantity threshold and the boundary of the urban area, suburban area, highway area, or mountainous area. Does it not exceed the buffer threshold? If yes, then execute S7.7; otherwise, the short-run segment belongs to the corresponding region. S7.7 Calculate the percentage of inland points in the urban area, suburban area, highway area, and mountainous area respectively. Number of points outside the buffer zone and Gaussian decay weighted score Determine whether each region simultaneously satisfies: If yes, then determine that the short-stroke segment belongs to the corresponding region; if not, then mark the corresponding short-stroke segment as an unknown region and remove it; where , The Gaussian decay coefficient; S7.

8. Collect short-distance segments of the defined region to obtain the travel area sequence: in, To determine the number of short-stroke segments in the region, For the first A region consisting of short-distance segments.

6. The method for identifying vehicle operating road areas based on GPS feature points according to claim 5, characterized in that, Step S8 specifically involves: S8.1 Determining the short-stroke segment region If a mutation has occurred, the process ends; otherwise, the mutation region is calculated. Compared with the previous area Minimum spherical distance between trajectory points corresponding to short-distance segments : Determine the minimum spherical distance If the distance is less than the spacing threshold, then the short-stroke segment region is determined. If the region is a geographically discontinuous abrupt change, proceed with step S8.2; otherwise, terminate. S8.2 Determining the short-stroke segment region If the proportion of the first y short-journey segment regions and the next y short-journey segment regions in the current 2y+1 regions meets the correction threshold, then the short-journey segments are... If the corresponding region is corrected to meet the correction threshold, the driving region sequence is updated and step S8.3 is executed; otherwise, the process ends. S8.3, Short-stroke segment area The first y short-stroke segment regions and the next y short-stroke segment regions are used to construct 2y+1 segment windows. Weighted confidence scores are obtained based on the frequency of occurrence of each region type within the window. The region type with the highest weighted confidence score is selected as the short-stroke segment region. The type.

7. A vehicle operation road area identification system based on GPS feature points, characterized in that, The method for identifying vehicle operating road areas based on GPS feature points as described in any one of claims 1 to 6 includes: The driving information preprocessing module is used to acquire the vehicle's driving information and preprocess it. The short-stroke segmentation module is used to divide the preprocessed driving information into multiple short-stroke segments; The mutation feature point filtering module is used to filter mutation feature points in short-stroke segments and determine whether the number of mutation feature points meets the set quantity. If yes, the first feature point set module is activated; otherwise, the second feature point set module is activated. The first feature point set module is used to set mandatory points in the short-stroke segment matrix and obtain the feature point set based on the mandatory points and the corresponding number of mutation feature points. The second feature point set module is used to determine whether the interval supplementation threshold is met based on the total duration of the short-journey segment. If so, the abrupt feature points are supplemented by adaptive time interval division, and the feature point set is obtained by combining the required points and the abrupt feature points. If not, the third feature point set module is activated. The third feature point set module is used to supplement abrupt feature points by segmenting points according to the velocity threshold, and to obtain a feature point set by combining the required points and abrupt feature points. The driving area sequence acquisition module is used to build a city boundary database, combine it with a set of feature points to determine the region corresponding to each short travel segment, and obtain the driving area sequence. The driving area sequence correction module is used to correct outliers in the driving area sequence.

Images