CN112200052B

CN112200052B - Track deviation recognition and vehicle running analysis method, device, equipment and medium

Info

Publication number: CN112200052B
Application number: CN202011062503.3A
Authority: CN
Inventors: 凌佳佳; 李世胤; 李龙
Original assignee: Uisee Shanghai Automotive Technologies Ltd
Current assignee: Uisee Shanghai Automotive Technologies Ltd
Priority date: 2020-09-30
Filing date: 2020-09-30
Publication date: 2023-06-09
Anticipated expiration: 2040-09-30
Also published as: CN112200052A

Abstract

The embodiment of the invention provides a track deviation recognition method, a vehicle running analysis method, a device, equipment and a medium, wherein the track deviation recognition method comprises the following steps: collecting sample point data of the vehicle running for a plurality of times on the same road section, wherein each sample point data comprises speed data, position data and course angle; determining, for each sample point, an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle; and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track. The technical scheme provided by the embodiment of the invention has obvious data clustering effect on the sample points, and can effectively and accurately identify the track deviation.

Description

Track deviation recognition and vehicle running analysis method, device, equipment and medium

Technical Field

The embodiment of the invention relates to the field of data analysis, in particular to a track deviation identification and vehicle running analysis method, device, equipment and medium.

Background

During testing of vehicle driving (e.g., unmanned), it is often necessary to provide a security officer with a supervision and assessment of the driving behavior of the vehicle. With the large expansion of the test scale and the popularization of unmanned test and operation, a small number of safety officers cannot consider all driving paths of all vehicles in a field, and then potential hidden hazards caused by abnormal track deviation of the vehicles are ignored. Thus, identifying abnormal trajectory shifts in an automated manner becomes a trend and necessity.

In the identification method of track offset in the related art, the effect of separating the offset track is not obvious, and the offset track is difficult to effectively identify.

Disclosure of Invention

The embodiment of the invention provides a track deviation identification method, a track deviation identification device, a track deviation analysis device and a track deviation medium, which can effectively and accurately identify track deviation.

In a first aspect, an embodiment of the present invention provides a track offset identification method, including:

collecting sample point data of the vehicle running for a plurality of times on the same road section, wherein each sample point data comprises speed data, position data and course angle;

determining, for each sample point, an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle;

and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track.

In a second aspect, an embodiment of the present invention further provides a vehicle running analysis method, including:

the method provided by the embodiment of the invention is adopted to determine the offset track;

determining a deviation reason of the vehicle in the driving process based on the associated data corresponding to the deviation track, or,

evaluating the service quality of the vehicle based on the associated data corresponding to the offset track; the associated data are sample point data in the offset trajectory and other data corresponding to the sample point data.

In a third aspect, an embodiment of the present invention further provides a track offset identifying device, including:

the system comprises an acquisition module, a control module and a control module, wherein the acquisition module is used for acquiring sample point data of a vehicle running for a plurality of times on the same road section, and each sample point data comprises speed data, position data and course angle;

an ellipse neighborhood determining module for determining, for each sample point, an ellipse neighborhood of the sample point based on the speed data, the position data, and the heading angle;

and the offset track determining module is used for performing DBSCAN clustering on the sample point data based on the elliptical neighborhood of each sample point to determine an offset track.

In a fourth aspect, an embodiment of the present invention provides an electronic device, including:

one or more processors;

storage means for storing one or more programs,

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the methods provided by the embodiments of the present invention.

In a fifth aspect, embodiments of the present invention provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method provided by embodiments of the present invention.

According to the technical scheme provided by the embodiment of the invention, for each sample point, the elliptical neighborhood of the sample point is determined based on the speed data, the position data and the course angle of the sample point, DBSCAN clustering is performed on the sample point data based on the elliptical neighborhood of each sample point, the offset track is determined, and the influence of the speed of the vehicle on the spatial density degree of the sample point data is considered, so that the neighborhood of the sample point changes along with the speed of the vehicle, the number of core sample points is more reasonable, the data clustering effect on the sample points is obvious, and the offset track can be accurately identified.

Drawings

FIG. 1a is a schematic diagram of a track section;

FIG. 1b is a schematic diagram of a neighborhood of a related art;

FIG. 1c is a schematic diagram of a neighborhood of the related art;

FIG. 1d is a flowchart of a track offset recognition method according to an embodiment of the present invention;

FIG. 1e is a schematic view of an elliptical neighborhood provided by an embodiment of the present invention;

FIG. 1f is a schematic diagram of cluster aggregation;

FIG. 1g is a schematic diagram of cluster aggregation;

FIG. 1h is a graph of trace effects obtained by the method provided by the embodiment of the invention;

FIG. 1i is a graph of track effects for different road segments of a road segment obtained by the method provided by the embodiment of the invention;

FIG. 2 is a flow chart of track offset identification provided by an embodiment of the present invention;

FIG. 3a is a flowchart of track offset identification according to an embodiment of the present invention;

FIG. 3b is a schematic diagram of density reachability relationship provided by an embodiment of the present invention;

FIG. 4 is a flow chart of a method for analyzing vehicle travel provided by an embodiment of the present invention;

FIG. 5a is a block diagram of a track offset recognition device according to an embodiment of the present invention;

FIG. 5b is a block diagram of a vehicle travel analysis apparatus according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.

A Density-based clustering method (Density-Based Spatial Clustering of Applications with Noise, DBSCAN) with noise is taken as a classical Density clustering algorithm, which can solve the problem of irregular shape clustering which cannot be solved by a k-means clustering method, and the method classifies the data of sample points with connected densities into the same class according to the compactness of sample distribution. If the DBSCAN clustering algorithm is applied to a scene of track deviation of a vehicle, the clustering method has the following defects:

The shape and the size of the E neighborhood for calculating the density reachable relation of the sample points are usually preset, and the size of the neighborhood is fixed in the whole clustering calculation process, so that the influence of the speed of the vehicle on the spatial density degree of the sample point data in the vehicle running track is not considered. In fact, under the condition of fixed collection frequency, the spatial density of sample points in the vehicle track is generally affected by the vehicle speed, when the speed is higher, the data of the collected sample points are less concentrated, and when the speed of the vehicle is lower, the data of the collected sample points are more concentrated. Treating these sample point data without difference results in a smaller number of core sample points in the straight-channel region with a higher average speed and a larger number of core sample points in the curve region with a lower average speed, which in turn results in a less obvious subsequent clustering effect and an inability to effectively identify the generation of the offset trajectory. For example, as in the track section shown in fig. 1a, an improper selection of a fixed neighborhood (e.g., a fixed circular neighborhood) will result in sample point data being divided into multiple meaningless categories (as shown in fig. 1 b), or in all sample point data being divided into the same category (as shown in fig. 1 c).

Fig. 1d is a flowchart of a track deviation identifying method according to an embodiment of the present invention, where the method may be performed by a track deviation device, where the device may be implemented by software and/or hardware, where the device may be configured in an electronic device such as a computer, a server, or where the vehicle may be configured in a cloud server, or where the device may be configured in a vehicle. The method can be applied to a scene of analyzing the process of the vehicle running on the same road section for a plurality of times. Alternatively, the vehicle may be an unmanned vehicle.

As shown in fig. 1d, the technical solution provided by the embodiment of the present invention includes:

s110: sample point data of the vehicle running on the same road section for a plurality of times are collected, wherein each sample point data comprises speed data, position data and course angle.

In the embodiment of the invention, a plurality of times means more than one time. The sample point may be understood as a point of position at which the vehicle is traveling. The location data may be latitude and longitude data, or may be other forms of location data. The heading angle may refer to an angle between a traveling direction of the vehicle and a lateral axis (x-axis) in a ground coordinate.

In the embodiment of the present invention, the collected sample point data of the road section may be segmented according to a fixed Trip section road section, and S120 and S130 are executed for each sample point in each Trip section road section to obtain an offset track. Therefore, by dividing the sample point data, more data can be prevented from being processed at one time, and system resources are saved. Or the sample point data corresponding to the road segment may not be divided, and S120 and S130 may be executed for each sample point in the road segment to obtain the offset track.

S120: for each sample point, determining an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle.

In one implementation of the present embodiments, the major and minor half axes may be determined based on the speed data, the center position may be determined based on the position data, and the elliptical direction may be determined based on the heading angle; an ellipse neighborhood is determined based on the major half axis, minor half axis, center position, and ellipse direction.

Wherein, the elliptical direction may refer to the angle between the major axis and the transverse axis (x-axis) of the ellipse. Alternatively, an average speed of the section of road centered on the sample point may be determined based on the speed data of the sample point, and the major or minor half axes may be determined based on the average speed. Or the velocity data of the sample points may be taken as the instantaneous velocity, based on which the major and minor axes are directly determined. The manner in which a particular ellipse neighborhood is determined may be described in detail in the following embodiments.

In one implementation of the embodiment of the invention, the influence of the acquisition frequency of the sample points on the ellipse neighborhood can be considered, the long half axis can be determined based on the acquisition frequency and the speed data, and the short half axis can be determined based on the speed data; the center position may be determined based on the position data, and the elliptical direction may be determined based on the heading angle; thus determining an ellipse neighborhood based on the major half axis, minor half axis, center position and ellipse direction. Alternatively, the average speed of the road section centered on the sample point may be determined based on the speed data of the sample point, the long half axis may be determined based on the average speed and the acquisition frequency, or the speed data of the sample point may be used as the instantaneous speed, and the long axis may be directly determined based on the instantaneous speed and the acquisition frequency. The method for determining the specific ellipse neighborhood can be described in detail in the following embodiments.

In one implementation of the present embodiment, the major half axis may be determined based on the speed data and the minor half axis set to the set point; and the center position may be determined based on the position data, and the elliptical direction may be determined based on the heading angle, such that the elliptical neighborhood is determined based on the major half axis, the minor half axis, the center position, and the elliptical direction. Wherein, the setting value can be set according to actual demands. If the vehicle runs for multiple times on a road section, the road section is a straight road section, the short half shaft can be set as a set value, and then an ellipse neighborhood can be determined. Thus, by setting the shorter half axis of the ellipse to a fixed set value in the straight line section, the algorithm can be simplified and the calculation speed can be improved.

S130: and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track.

In one implementation of the embodiment of the present invention, optionally, performing DBSCAN clustering on the sample point data based on an elliptical neighborhood of each sample point includes: taking each sample point as a current sample point, adding the sample points in an elliptical neighborhood of the current sample point to an elliptical neighborhood sub-sample set aiming at each current sample point, and adding the current sample points to a core sample point set if the number of the sample points in the elliptical neighborhood sub-sample set is larger than the minimum number of direct sample points; randomly selecting one core sample point from the core sample point set to serve as a target core sample point; determining all sample points which are communicated with the target core sample point density to form a cluster set; removing the target core sample point from the set of core sample points and other core sample points contained in the cluster; judging whether the core sample point set is an empty set or not; if not, returning to the operation of arbitrarily selecting one core sample point in the core sample point set until the core sample point set is an empty set, and taking the sample point in each cluster set as a classification.

The method for determining the core sample point set may specifically be: taking each sample point as a current sample point, and executing the following operations aiming at each current sample point: converting the position data of other sample points into the position data in the coordinate system of the elliptical neighborhood of the current sample point; based on the position data obtained by conversion, adding the sample points in the elliptical neighborhood of the current sample point into an elliptical neighborhood sub-sample set; if the number of the sample points in the oval neighborhood sub-sample set is greater than the minimum direct sample point number, adding the current sample point into the sample pointA set of core sample points. The coordinate system of the position data of the sample point is a ground coordinate system, as shown in FIG. 1e, the ground coordinate system is a coordinate system formed by an x-axis and a coordinate axis perpendicular to the x-axis, and the ellipse neighborhood coordinate system is formed by x ^′ And y ^′ And (5) forming a coordinate system. The position data of other sample points are required to be converted into the position data in the coordinate system where the elliptical neighborhood of the current sample point is located, and the sample points in the elliptical neighborhood of the current sample point are added into the elliptical neighborhood sub-sample set.

In the embodiment of the invention, the minimum up to sample point number and the sample point density are in positive correlation. When the collected sample point data is data of the vehicle during the running process, the minimum number of direct sample points may be set based on the number of running turns, for example, the minimum number of direct sample points mincls=1.5m, where M is the number of running turns.

In the related art, when the density reachable relation of the sample points is calculated, the minimum sample number of the core sample points is judged to be a fixed preset parameter, and the parameter is unchanged in the whole calculation process, and the influence of the density degree of the sample points of a fixed road section or the running circle number on the distribution of the sample point data is not considered, so that the formed clustering is unreasonable, and the phenomenon of unobvious clustering is caused. According to the embodiment of the invention, the minimum number of direct sample points is set through the density degree of the sample points, or the minimum number of direct sample points is set according to the number of running circles, so that the influence of the density degree of the sample points or the number of the running circles on the sample point data distribution is considered, and the cluster set can be reasonably determined, so that a better clustering effect is generated.

In an embodiment of the present invention, density connectivity may be understood as: if sample points can be related by an elliptical neighborhood of one or more core sample points (forming a cluster set), then the sample points can be considered to be in dense communication.

Wherein, in one implementation manner of the embodiment of the present invention, optionally, determining that all sample points that are in communication with the target core sample point density form a cluster set may include: if no other core sample points exist in the oval neighborhood of the target core sample point, forming a cluster set by sample points which are not divided into other cluster sets in the oval neighborhood of the target core sample point; if there are other core sample points in the elliptical neighborhood of the target core sample point, determining a cluster set for each of the other core sample points based on sample points that are within the elliptical neighborhood of the other core sample point and that are not partitioned into other cluster sets, and sample points that are within the elliptical neighborhood of the target core sample point and that are not partitioned into other cluster sets.

In one implementation of the embodiment of the present invention, the determining a cluster based on the sample points that are within the oval neighborhood of the other core sample points and that are not divided into other cluster sets, and the sample points that are within the oval neighborhood of the target core sample point and that are not divided into other cluster sets includes: if a first-stage other core sample point exists in the oval neighborhood of the target core sample point, determining the sample points in the oval neighborhood of the first-stage other core sample point; judging whether the next stage of other core sample points exist in the elliptical neighborhood of the first stage of other core sample points; if yes, the next-stage other core sample point is used as a first-stage other core sample point, and the operation of judging whether the next-stage other core sample point exists in the oval neighborhood of the first-stage other core sample point is returned until no new core sample point is contained in the oval neighborhood of the last-stage other core sample point; sample points which are not divided into other cluster sets and sample points which are not divided into other cluster sets in the oval neighborhood of the target core sample point and in oval neighborhood of other core sample points of all levels form one cluster set.

Specifically, to illustrate the determination of clustering, as shown in fig. 1f, if point a is a selected target core sample point, there are points B and C in the oval neighborhood of point a, and neither point B nor point C is a core sample point, and are not divided into other cluster sets, so that point a, point B, and point C form one cluster set.

As shown in fig. 1g, if point a is the selected target core sample point, there are point B and point C within the elliptical neighborhood of point a, point B is the core sample point, and point C is not the core sample point. Judging whether other core sample points exist in the ellipse neighborhood of the point B, if the other core sample points exist in the ellipse neighborhood of the point B, namely the point D, and the point E which exists in the ellipse neighborhood of the point D is not the core sample point, judging whether other core sample points exist in the ellipse neighborhood of the point D, and if the other core sample points do not exist in the ellipse neighborhood of the point D, the point F is not the core sample point. Sample points in the oval neighborhood of point a and not divided into other clusters, sample points in the oval neighborhood of point B and not divided into other clusters, and sample points in the oval neighborhood of point D and not divided into other clusters form one cluster set, i.e., points a-F form one cluster set.

In one implementation of the embodiment of the present invention, optionally, each sample point data further includes time data; the determining an offset trajectory includes: if the number of the sample points in the target cluster set is less than the set number, determining that the sample points in the target cluster set are sample points with track offset, and taking the sample points which are not divided into the cluster set as the sample points with track offset; and connecting the sample points with the track offset based on the corresponding time data to form an offset track. The set number is one tenth of the number of the sample points in the cluster set with the largest number of the sample points, or the set number can be set according to the situation that the sample points are clustered actually. Specifically, sample points in a cluster set with fewer sample points and sample points which are not divided into clusters can be used as sample points with track offset, and the sample points are connected according to time data to form offset tracks, namely the embodiment of the invention adopts the clustering method to the characteristics of data which are not classified in the clustering process and the characteristics of data in the cluster set with fewer sample points in the clustering process, and is used for finding track offset in the vehicle running track. For example, fig. 1h is a track effect diagram, fig. 1i is a track effect diagram of different road segment sections of one road segment, and as shown in fig. 1h and fig. 1i, points deviating from the curve are sample points of track offset.

According to the technical scheme provided by the embodiment of the invention, for each sample point, the elliptical neighborhood of the sample point is determined based on the speed data, the position data and the course angle of the sample point, DBSCAN clustering is performed on the sample point data based on the elliptical neighborhood of each sample point, the offset track is determined, and the influence of the speed of the vehicle on the spatial density degree of the sample point data is considered, so that the elliptical neighborhood changes along with the speed of the vehicle, the number of core sample points is more reasonable, the data clustering effect on the sample points is obvious, and the offset track can be accurately identified.

Fig. 2 is a flowchart of track offset recognition according to an embodiment of the present invention, where in this embodiment, the collection frequency of the sample point data is a fixed value.

Optionally, the determining an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle includes:

determining a major half axis and a minor half axis based on the speed data; if the acquisition frequency of the sample point data is a fixed value, the long half shaft and the speed show positive correlation, and the short half shaft and the speed show inverse correlation; the speed is an average speed of a section of road centered on a sample point determined based on the speed data, or an instantaneous speed of a sample point, the instantaneous speed being the speed data.

Determining the position data as a central position and the course angle as an elliptical direction;

an ellipse neighborhood is determined based on the major half axis, the minor half axis, the center position, and the ellipse direction.

As shown in fig. 2, the technical solution provided by the embodiment of the present invention includes:

s210: sample point data of the vehicle running on the same road section for a plurality of times are collected, wherein each sample point data comprises speed data, position data and course angle.

S220: determining, for each sample point, a major and minor half axis based on the speed data; if the acquisition frequency of the sample point data is a fixed value, the long half shaft and the speed show positive correlation, and the short half shaft and the speed show inverse correlation; the speed is an average speed of a section of road centered on a sample point determined based on the speed data, or an instantaneous speed of a sample point, the instantaneous speed being the speed data.

In one implementation of the embodiment of the present invention, optionally, determining the major and minor half axes includes: determining the major and minor half axes based on the following formula:

wherein a is the long half shaft; b is the short half shaft; a, a ₀ And b ₀ Is a fixed coefficient;

and f is the acquisition frequency, and is the average speed of a road section taking the sample point as the center or the instantaneous speed of the sample point.

In an embodiment of the invention, a ₀ Can be set according to actual needs, b ₀ Can be set according to practical factors such as road width, for example, the wider the road width is, the more b is ₀ The larger may be. Under the condition that the vehicle runs for many times on the same road section, the speed is different in the process of passing through the road section taking the sample point as the center, so that the collected sample points are not the same in density, and in order to ensure that a reasonable number of sample points exist in an elliptical neighborhood of the sample points, so that better clustering is facilitated, the long axis of the elliptical neighborhood can be calculated by the average speed of the road section taking the sample point as the center.

In the embodiment of the invention, if the acquisition frequency of the sample points is a fixed value, the influence of the speed on the elliptical neighborhood is considered as can be known by the calculation formulas of the long half axis and the short half axis. The long half axis and the speed show positive correlation, when the speed is higher, the long axis is higher, more sample points can be ensured to be in the oval neighborhood, when the speed is lower, the long axis is smaller, more reasonable sample points can be ensured to be in the oval neighborhood, more reasonable core sample points can be determined, and better clustering can be carried out on the data of the sample points.

S230: the position data is determined as a center position and the heading angle is determined as an elliptical direction.

S240: an ellipse neighborhood is determined based on the major half axis, the minor half axis, the center position, and the ellipse direction.

In the embodiment of the invention, since the elliptical direction may refer to an included angle between a major axis and a transverse axis (x axis) of the ellipse, if the heading angle is taken as the elliptical direction, the heading direction of the vehicle is the major axis direction of the elliptical neighborhood, i.e., the major axis direction of the elliptical neighborhood is taken as the heading direction of the vehicle.

In the related art, in the process of adopting a DBSCAN clustering algorithm, the shape and the size of an ellipse neighborhood are usually preset, the influence of the speed is not considered in the clustering calculation method in the related art, the long half shaft and the short half shaft are determined through speed data, the long half shaft and the speed show positive correlation, a course angle is taken as an ellipse direction, the ellipse neighborhood is determined based on the long half shaft, the short half shaft, the central position and the ellipse direction, and in the related art, the clusters are gathered in the course direction of a vehicle by means of the course angle of the vehicle, so that the transverse deviation in the running process of the vehicle can be highlighted; different component weights are adopted in the direction of the heading of the vehicle and the direction perpendicular to the direction of the heading, the distance component in the direction of the heading is reduced according to the speed of the vehicle, and the distance component perpendicular to the direction of the heading is amplified. Thus, a sample point along the long axis may itself be farther from the current sample point, but it may still be in the elliptical neighborhood, while a sample point along the short axis may itself be closer to the current sample point, but it is likely that it is not in the elliptical neighborhood, counteracting the effects of non-uniformities in the sample points in the trajectory resulting from the varying speed of the vehicle. According to the method provided by the embodiment of the invention, the influence caused by the transverse offset is considered, and when judging whether other sample points are in the elliptical neighborhood, the short axis of the elliptical neighborhood is determined based on the speed, so that the distance component perpendicular to the heading direction can be adaptively adjusted according to the speed, and the offset sample points are better distinguished from the non-offset sample points.

S250: and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track.

Reference is made to the above embodiments for an introduction of further steps.

FIG. 3a is a flow chart of track offset identification provided by an embodiment of the present invention, in which, optionally, frequency variations are collected, and long and short half axes are determined based on the speed data, comprising:

the major half axis is determined based on the speed data and the acquisition frequency, and the minor half axis is determined based on the speed data.

As shown in fig. 3a, the technical solution provided by the embodiment of the present invention includes:

s310: sample point data of the vehicle running on the same road section for a plurality of times are collected, wherein each sample point data comprises speed data, position data and course angle.

S320: determining, for each sample point, the major half axis based on the speed data and acquisition frequency, and determining the minor half axis based on the speed data; if the acquisition frequency changes, the long half shaft and the speed show positive correlation, and show inverse correlation with the acquisition frequency; the short half shaft and the speed show an inverse relation; the speed is an average speed of a section of road centered on a sample point determined based on the speed data, or an instantaneous speed of a sample point, the instantaneous speed being the speed data.

/>

In an embodiment of the invention, a ₀ Can be set according to actual needs, b ₀ Can be set according to practical factors such as road width, for example, the wider the road width is, the more b is ₀ The larger may be.

In the embodiment of the invention, if the acquisition frequency changes, the calculation modes of the long axis and the short axis show positive correlation between the long axis and the speed, and show inverse correlation between the long axis and the acquisition frequency; the minor half axis exhibits an inverse relationship with speed.

In the related art, in the process of adopting the DBSCAN clustering algorithm, the shape and the size of the ellipse neighborhood are usually preset and have fixed shapes and sizes. According to the method provided by the embodiment of the invention, the long half axis of the elliptical neighborhood is determined through the speed data and the acquisition frequency, the non-uniformity of the degree of the density of the sample point data in the space position caused by the non-uniformity of the speed of the vehicle is considered, the long axis of the elliptical neighborhood can be adaptively adjusted based on the speed and the acquisition frequency of the vehicle under the condition that other sample points are judged to be in the elliptical neighborhood, and the distance component in the heading direction is adjusted, so that the reasonable number of sample points in the elliptical neighborhood is ensured, and better clustering is facilitated. According to the method provided by the embodiment of the invention, the influence caused by the transverse offset is considered, and when judging whether other sample points are in the elliptical neighborhood, the short axis of the elliptical neighborhood is determined based on the speed, so that the distance component perpendicular to the course direction can be adaptively adjusted according to the speed, and the offset sample points are better distinguished from the non-offset sample points, so that reasonable clustering is realized.

S330: the position data is determined as a center position and the heading angle is determined as an elliptical direction.

S340: an ellipse neighborhood is determined based on the major half axis, the minor half axis, the center position, and the ellipse direction.

S350: and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track.

Other relevant steps may be referred to the above embodiments.

In order to explain the technical solution of the embodiment of the present invention in more detail, the method provided by the embodiment of the present invention may include the following steps:

sample point data of the unmanned vehicle in the driving process are collected, the sample point data are time sequence data sets, the time sequence data comprise position data, course angle and speed data in the driving process of the vehicle, and the collection frequency f of data collection is recorded. And then preprocessing the time sequence data set, dividing the time sequence data set into a plurality of sub-time sequence data sets according to the fixed Trip interval section, and recording the running circle number M of each divided Trip interval section. Clustering is performed for each sub-time series data set, and one sub-time series data set is taken as an example in the following operation.

Wherein, each sample point data is set as a group of four-dimensional data, wherein, the group of four-dimensional data x of the ith sample point _i ＝(lat _i ,lon _i ,θ _i ,v _i ) Wherein lat _i Is latitude, lon _i Longitude, theta _i V is course angle _i The total sample set is d= (x) ₁ ,x ₂ ,…,x _n )。

The clustering method comprises the following detailed steps:

(1) Initializing a set of core sample points

Initializing the cluster number to k=0, initializing the unvisited sample set l=d, clustering the partition +.>

Initializing the major axis 2a of an ellipse neighborhood ₀ Short axis 2b of ellipse ₀ 。

(2) For j=1, 2,.. all core sample points are found as follows.

a) Firstly, an ellipse neighborhood of a sample point is adaptively determined according to the current sample point data, and parameters are included: the center position 0 of the ellipse, the major axis 2a of the ellipse, the minor axis 2b of the ellipse, and the rotation angle α (elliptical direction) of the ellipse.

As shown in fig. 1e, for the current sample point x _j ＝(lat _j ，lon _j ，θ _j ，v _j ) The above parameters can be calculated by the following formula:

wherein, the liquid crystal display device comprises a liquid crystal display device,

for the average speed of the road section centered on the current sample point, the coordinate system established with x 'y' is the coordinate system established by the ellipse neighborhood as shown in fig. 1 e.

b) And mapping the position coordinates of other sample points into an elliptical neighborhood coordinate system established by the sample points, and calculating the relation with the current sample point. If the sample point x _i At sample point x _j Is within the ellipse neighborhood of (x) _i Added to x _i E-neighborhood subsampled set N _∈ (x _j ) The density reachable relationship is shown in fig. 3 b.

C) If the number of sub-sample set sample points satisfies |N _∈ (x _j ) I is not less than MinCls, and the sample point x is obtained _j Added to the core sample point set: Ω=Ω { x } U } _j }. Wherein MinCls is the core sample point should be fullThe minimum number of direct samples can be set according to the density of the sample points, for example, mincls=1.5m is set according to the number of running turns M.

3) If the core sample point set

Step 7) is performed, otherwise step 4) is performed.

4) Selecting a core sample point p from the set of core sample points Ω and adding the sample point to the new cluster: c (C) _k = { p }. Updating the set of unvisited samples l=l- { p }, while defining clusters to be searched: omega shape _p ＝{p}。

5) Updating the core sample point set: Ω=Ω -C _k . If omega _p Is empty set, then current cluster C is described _k After the generation is finished, cluster C _k And adding the C set, executing the step 3, and otherwise, executing the next step.

6) At Ω _p At selecting a core sample point p', finding out an elliptical neighborhood sample set N of the sample point _∈ (p') updating the current cluster set: c (C) _k ＝C _k ∪(N _∈ (p') ∈L), update the unvisited set: l=l- (N) _∈ (p'). U.L), update Ω _p ＝Ω _p ∪(N _∈ (p '). U.L. Omega.) to p', go to step 5.

Wherein, the steps (3) - (6) can be specifically:

when core sample point set

(if the set is not empty, the set enters the circulation), and a core sample point p epsilon omega is selected;

adding the sample points to the new cluster: c (C) _k = { p } (only itself initially).

Updating the set of non-access sample points l=l- { p };

defining clusters to be searched: omega shape _p = { p } (initially only itself).

And then starting to expand cluster scale:

when clustering to be searched

Selecting a core object p' e omega _p ；

Find out the ellipse neighborhood sample set N of the sample point _∈ (p ') (find all sample points within the elliptical neighborhood of the core sample point p');

updating the current cluster set: c (C) _k ＝C _k ∪(N _∈ (p') ∈l) (among the found sample points, if the sample point is not accessed, i.e., is not divided into other cluster sets, it is attributed to the current cluster set);

updating the unvisited collection: l=l- (N) _∈ (p') ∈l) (sample points belonging to the current cluster set need to be removed from the unvisited set to prevent access by other clusters);

updating omega _p ＝Ω _p ∪(N _∈ (p '). U.l. U.omega.) p ' (other core sample points within the elliptical neighborhood of core sample point p ' are added to. Omega. First _p The current core sample point p' is removed again because the current core object has been processed up to this point).

7) Outputting a final classification result: c= { C ₁ ，...，C _k The number of sample points of each category is n= { N } ₁ ，...，N _k -a }; wherein N is the number of sample points, N _k The number of sample points for the kth category.

8) After the clustering result is obtained, the sample points in the cluster set with the smaller sample points and the sample points in the unaccessed sample set L are divided into sample points with track offset, and the discrimination modes are as follows:

9) Fig. 1h shows a graph of the track effect obtained by the method, and as shown in fig. 1h, the point deviating from the curve is a sample point of track deviation.

Fig. 4 is a flowchart of a vehicle running analysis method provided by an embodiment of the present invention, where the method may be performed by a vehicle running analysis apparatus, where the apparatus may be implemented by software and/or hardware, where the apparatus may be configured in an electronic device such as a computer, a server, or where the vehicle may be configured in a cloud server, or where the apparatus may be configured in a vehicle. The method can be applied to a scene of analyzing the process of the vehicle running on the same road section for a plurality of times. Alternatively, the vehicle may be an unmanned vehicle.

As shown in fig. 4, the technical solution provided by the embodiment of the present invention may include:

S410: collecting sample point data of the vehicle running for a plurality of times on the same road section, wherein each sample point data comprises speed data, position data and course angle;

s420: determining, for each sample point, an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle;

s430: and performing DBSCAN clustering on the sample point data based on the ellipse neighborhood of each sample point, and determining an offset track.

S440: determining a deviation reason of the vehicle in the running process based on the associated data corresponding to the deviation track, or evaluating the service quality of the vehicle based on the associated data corresponding to the deviation track; the associated data are sample point data in the offset trajectory and other data corresponding to the sample point data.

Wherein S410-S430 may refer to the description of the above embodiments. Other data corresponding to the sample point data may include body data, etc.

In the embodiment of the invention, the associated data corresponding to the offset track can be used as the characteristic data of the driving behavior of the vehicle, and the offset reason in the driving process of the vehicle is determined according to the data, so that a worker maintains the vehicle, for example, the reason for causing the offset of the vehicle can be the offset caused by the performance of each part such as a controller or the like, or can be other reasons. Or the service quality of the vehicle can be evaluated through the associated data corresponding to the offset track, so that the staff can improve the vehicle according to the evaluation result.

determining a major half axis and a minor half axis based on the speed data;

determining an elliptical neighborhood based on the major half axis, the minor half axis, the center position, and the elliptical direction;

if the acquisition frequency of the sample point data is a fixed value, the long half shaft and the speed show positive correlation, and the short half shaft and the speed show inverse correlation; the speed is an average speed of a road section taking a sample point as a center, or an instantaneous speed of the sample point, which is determined based on the speed data.

Optionally, determining the long half shaft and the short half shaft based on the speed data includes:

determining the long half shaft based on the speed data and the acquisition frequency, and determining the short half shaft based on the speed data; if the acquisition frequency changes, the long half shaft and the speed show positive correlation, and show inverse correlation with the acquisition frequency; the minor half axis exhibits an inverse relationship to the speed.

Optionally, determining the major and minor half axes includes:

determining the major and minor half axes based on the following formula:

wherein the method comprises the steps ofA is the long half shaft; b is the short half shaft; a, a ₀ And b ₀ Is a fixed coefficient;

and f is the sampling frequency, and is the average speed of a road section taking the sample point as the center or the instantaneous speed of the sample point.

Optionally, the performing DBSCAN clustering on the sample point data based on the elliptical neighborhood of each sample point includes:

taking each sample point as a current sample point, adding the sample points in an elliptical neighborhood of the current sample point to an elliptical neighborhood sub-sample set aiming at each current sample point, and adding the current sample points to a core sample point set if the number of the sample points in the elliptical neighborhood sub-sample set is larger than the minimum number of direct sample points; randomly selecting one core sample point from the core sample point set to serve as a target core sample point;

determining all sample points which are communicated with the target core sample point density to form a cluster set;

removing the target core sample point from the set of core sample points and other core sample points contained in the cluster;

Judging whether the core sample point set is an empty set or not;

if not, returning to the operation of arbitrarily selecting one core sample point in the core sample point set until the core sample point set is an empty set, and taking the sample point in each cluster set as a classification.

Optionally, each sample point data further comprises time data;

the determining an offset trajectory includes:

if the number of the sample points in the target cluster set is less than the set number, determining that the sample points in the target cluster set are sample points with track offset, and taking the sample points which are not divided into the cluster set as the sample points with track offset;

and connecting the sample points with the track offset based on the corresponding time data to form an offset track.

Optionally, the set number is one tenth of the number of sample points in the cluster set with the largest number of sample points.

Optionally, the number of the minimum up sample points and the density of the sample points are in positive correlation.

According to the technical scheme provided by the embodiment of the invention, the reason of the offset of the vehicle in the running process is determined based on the associated data corresponding to the offset track, or the service quality of the vehicle is evaluated based on the associated data corresponding to the offset track, so that a worker can know the condition of the vehicle conveniently, and the vehicle is improved or maintained conveniently.

Fig. 5a is a block diagram of a track offset recognition device according to an embodiment of the present invention, where the device includes an acquisition module 510, an ellipse neighborhood determination module 520, and an offset track determination module 530.

The collecting module 510 is configured to collect sample point data of multiple traveling of the vehicle on the same road section, where each sample point data includes speed data, position data and heading angle;

an ellipse neighborhood determination module 520 for determining, for each sample point, an ellipse neighborhood of the sample point based on the speed data, the position data, and the heading angle;

an offset trajectory determination module 530, configured to perform DBSCAN clustering on the sample point data based on an ellipse neighborhood of each sample point, and determine an offset trajectory.

determining a major half axis and a minor half axis based on the speed data;

Optionally, determining the major and minor half axes includes:

determining the major and minor half axes based on the following formula:

judging whether the core sample point set is an empty set or not;

Optionally, each sample point data further comprises time data;

The determining an offset trajectory includes:

The device can execute the method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the method.

Fig. 5b is a block diagram of a vehicle running analysis device according to an embodiment of the present invention, where the device includes an acquisition module 510, an ellipse neighborhood determination module 520, an offset trajectory determination module 530, and an analysis module 540.

an offset trajectory determination module 530, configured to perform DBSCAN clustering on the sample point data based on an ellipse neighborhood of each sample point, and determine an offset trajectory;

an analysis module 540 for:

determining a major half axis and a minor half axis based on the speed data;

Optionally, determining the major and minor half axes includes:

determining the major and minor half axes based on the following formula:

judging whether the core sample point set is an empty set or not;

Optionally, each sample point data further comprises time data;

the determining an offset trajectory includes:

The device can execute the vehicle running analysis method provided by the embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

Fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 6, where the device includes:

one or more processors 610, one processor 610 being illustrated in fig. 6;

a memory 620;

the apparatus may further include: an input device 630 and an output device 640.

The processor 610, memory 620, input 630 and output 640 of the device may be connected by a bus or other means, for example in fig. 6.

The memory 620 is used as a non-transitory computer readable storage medium for storing software programs, computer executable programs, and modules, such as program instructions/modules corresponding to a trajectory offset recognition method in an embodiment of the present invention (e.g., the acquisition module 510, the elliptical neighborhood determination module 520, and the offset trajectory determination module 530 shown in fig. 5 a), or program instructions/modules corresponding to a vehicle driving analysis method in an embodiment of the present invention (e.g., the acquisition module 510, the elliptical neighborhood determination module 520, the offset trajectory determination module 530, and the analysis module 540 shown in fig. 5 b). The processor 610 executes various functional applications of the computer device and data processing by running software programs, instructions and modules stored in the memory 620, i.e. implements a track offset identification method of the above-described method embodiment, namely:

Or the method for analyzing the running of the vehicle provided by the embodiment of the invention is realized, namely:

performing DBSCAN clustering on the sample point data based on the elliptical neighborhood of each sample point, and determining an offset track;

Memory 620 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created according to the use of the computer device, etc. In addition, memory 620 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 620 optionally includes memory remotely located relative to processor 610, which may be connected to the terminal device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The input device 630 may be used to receive entered numeric or character information and to generate key signal inputs related to user settings and function control of the computer device. The output device 640 may include an output interface or the like.

The embodiment of the invention provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements a track offset identification method as provided in the embodiment of the invention:

Any combination of one or more computer readable media may be employed. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

Note that the above is only a preferred embodiment of the present invention and the technical principle applied. Those skilled in the art will appreciate that the invention is not limited to the specific embodiments described herein, and that various obvious changes, rearrangements and substitutions can be made by those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.

Claims

1. A track offset recognition method, comprising:

collecting sample point data of the vehicle running for a plurality of times on the same road section, wherein each sample point data comprises speed data, position data, course angle and time data;

performing DBSCAN clustering on the sample point data based on an elliptical neighborhood of each sample point, determining an offset trajectory, comprising:

taking each sample point as a current sample point, adding the sample points in an elliptical neighborhood of the current sample point to an elliptical neighborhood sub-sample set aiming at each current sample point, and adding the current sample points to a core sample point set if the number of the sample points in the elliptical neighborhood sub-sample set is larger than the minimum number of direct sample points;

randomly selecting one core sample point from the core sample point set to serve as a target core sample point;

Judging whether the core sample point set is an empty set or not;

if not, returning to the operation of arbitrarily selecting one core sample point in the core sample point set until the core sample point set is an empty set, and taking the sample point in each cluster set as a classification;

the determining an offset trajectory includes:

2. The method of claim 1, wherein the determining an elliptical neighborhood of the sample point based on the speed data, the position data, and the heading angle comprises:

determining a major half axis and a minor half axis based on the speed data;

3. The method of claim 2, wherein the determining major and minor half axes based on the speed data comprises:

4. A method according to claim 2 or 3, wherein determining the major and minor half axes comprises:

determining the major and minor half axes based on the following formula:

/>

5. The method of claim 1, wherein the set number is one tenth of the number of sample points in the cluster set having the largest number of sample points.

6. The method of any of claims 1-5, wherein the minimum number of up to sample points is positively correlated with the density of sample points.

7. A vehicle travel analysis method, comprising:

determining an offset trajectory using the method of any one of claims 1-6;

8. A track shift identifying device, comprising:

the system comprises an acquisition module, a control module and a control module, wherein the acquisition module is used for acquiring sample point data of a vehicle running for a plurality of times on the same road section, and each sample point data comprises speed data, position data, course angle and time data;

the offset track determining module is used for performing DBSCAN clustering on the sample point data based on the elliptical neighborhood of each sample point to determine an offset track;

the performing DBSCAN clustering on the sample point data based on the elliptical neighborhood of each sample point includes:

Judging whether the core sample point set is an empty set or not;

the determining an offset trajectory includes:

9. An electronic device, comprising:

one or more processors;

storage means for storing one or more programs,

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-7.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-7.