WO2020042642A1

WO2020042642A1 - Vehicle positioning method and device, and autonomous vehicle

Info

Publication number: WO2020042642A1
Application number: PCT/CN2019/084539
Authority: WO
Inventors: 刘威; 卫璐
Original assignee: 东软睿驰汽车技术（沈阳）有限公司
Priority date: 2018-08-29
Filing date: 2019-04-26
Publication date: 2020-03-05
Also published as: CN109115231A; WO2020042642A8; CN109115231B

Abstract

Provided are a vehicle positioning method and device, and an autonomous vehicle. The vehicle positioning method comprises: on the basis of the identification precision of a vehicle sensor for road characteristics, road characteristics in a first characteristic set are subjected to preliminary screening to obtain a second characteristic set composed of first type road characteristics; if the matching degree between the second characteristic set and a third characteristic set exceeds a first preset matching degree threshold, i.e. indicating that the road characteristics of a characteristic point detected by the vehicle sensor are similar to the road characteristics of the characteristic point on a vehicle driving route on a high-precision map, and that the matching of the road characteristics of the characteristic point is successful, the absolute position of the characteristic point and the relative positions of the vehicle and the characteristic point are obtained on the basis of the successfully matched characteristic point, so that a more accurate absolute position of the vehicle can be obtained. The method obtains the absolute position of the characteristic point on the basis of the high-precision map and further obtains the absolute position of the vehicle, and compared with the prior art, the accuracy of vehicle positioning is effectively improved.

Description

Vehicle positioning method and equipment, and autonomous driving vehicle

This application claims priority from a Chinese patent application filed with the Chinese Patent Office on August 29, 2018, with application number 201810994919.5, "A Vehicle Positioning Method, Device, and Autonomous Vehicle", the entire contents of which are incorporated herein by reference .

Technical field

The invention relates to the technical field of automobiles, and in particular, to a vehicle positioning method and device, and an autonomous driving vehicle.

Background technique

Nowadays, GPS and Beidou satellite navigation systems are often used to provide positioning and navigation services for vehicle users. However, the accuracy of these positioning and navigation systems in civil applications is currently limited. The positioning accuracy of the Global Positioning System and Beidou satellite navigation system is about ten meters, and the positioning accuracy of vehicles cannot be improved to the lane level. For example, according to the position information detected by the positioning system, the lane information of the lane where the vehicle is located obtained from the map database is inconsistent with the lane information of the lane where the vehicle is actually located.

It can be seen that the accuracy of the vehicle positioning method provided in the prior art is too low, and the positioning error caused affects the path planning of the autonomous vehicle.

Summary of the Invention

Based on the above technical problems, the present application provides a vehicle positioning method and device, and an autonomous vehicle to solve the problem of too low vehicle positioning accuracy.

This application proposes the following technical solutions:

According to a first aspect of the present application, a vehicle positioning method is provided. The method is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor; the vehicle-mounted sensor is used to identify road features;

The method includes:

Obtain road features of feature points on the vehicle's driving path from a high-precision map to form a first feature set;

Filtering the first type of road features from the first feature set to form a second feature set, and the first type of road features are road features that can be recognized by the vehicle-mounted sensor;

Using the vehicle-mounted sensor to identify road features of the feature points on the vehicle's driving path to obtain a third feature set;

Compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, then use the high-precision map Obtaining an absolute position of the feature point, and obtaining a relative position of the vehicle and the feature point from an on-board positioning system of the vehicle;

An absolute position of the vehicle is obtained according to an absolute position of the feature point and a relative position of the vehicle and the feature point.

As a possible implementation manner, before the comparing the second feature set and the third feature set, the method further includes:

Feature points are filtered based on the road features of the feature points, and invalid feature points are eliminated.

As a possible implementation manner, the filtering feature points according to the road features of the feature points and excluding invalid feature points specifically includes:

Determine whether the matching degree between the second feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a second preset matching degree threshold, and if so, the first feature passing by the vehicle in the neighboring feature points The feature points are used as effective feature points, and the remaining feature points except the effective feature points among the adjacent feature points are eliminated.

Obtaining the confidence level of each road feature in the third feature set of the feature points; the confidence level is a variable that changes with time;

Obtaining the confidence level of the feature point according to the confidence level of each road feature in the third feature set of the feature point;

If the matching degree between the third feature sets corresponding to adjacent feature points on the vehicle travel path exceeds a third preset matching degree threshold, and the matching degree between the confidence degrees corresponding to the adjacent feature points exceeds Preset a reliability matching threshold, and use the adjacent feature points as similar feature points;

The similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points are filtered, and the remaining feature points in the adjacent feature points are eliminated.

As a possible implementation manner, the obtaining the confidence level of the feature point according to the confidence level of each road feature in the third feature set of the feature point specifically includes:

Obtaining an average confidence level of each road feature in the third feature set of the feature points within a preset time;

Multiplying the average confidence level of each road feature within a preset time by the confidence level of each road feature obtained from the high-precision map to obtain a first confidence level of each road feature of the feature point ;

The first confidence levels of the road features of the feature points are added to obtain the confidence levels of the feature points.

According to a second aspect of the present application, a vehicle positioning device is provided. The device is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor; the vehicle-mounted sensor is used to identify road features;

The vehicle positioning device includes a first feature set acquisition module, a second feature set acquisition module, a third feature set acquisition module, a feature comparison and location acquisition module, and a positioning module;

The first feature set acquisition module is configured to obtain road features of feature points on a vehicle driving path from a high-precision map to form a first feature set;

The second feature set obtaining module is configured to filter the first type of road features from the first feature set to form a second feature set, and the first type of road features are road features that can be recognized by the vehicle-mounted sensor;

The third feature set acquisition module is configured to use the on-board sensor to identify road features of the feature points on the vehicle travel path to obtain a third feature set;

The feature comparison and position acquisition module is configured to compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset The matching degree threshold, using the high-precision map to obtain the absolute position of the feature point, and obtaining the relative position of the vehicle and the feature point from the on-board positioning system of the vehicle;

The positioning module is configured to obtain an absolute position of the vehicle according to an absolute position of the feature point and a relative position of the vehicle and the feature point.

As a possible implementation manner, the vehicle positioning device further includes:

The feature point screening module is configured to filter feature points according to the road features of the feature points and eliminate invalid feature points.

As a possible implementation manner, the feature point screening module specifically includes:

The feature point first screening unit is configured to determine whether a matching degree between second feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a second preset matching degree threshold, and if yes, compare the phase The first feature point passed by the vehicle in the adjacent feature points is used as the effective feature point, and the remaining feature points except the effective feature points in the adjacent feature points are eliminated.

The road feature confidence degree obtaining unit is configured to obtain the confidence degree of each road feature in the third feature set of the feature points; the confidence degree is a variable that changes with time;

A feature point confidence obtaining unit, configured to obtain the confidence of the feature point according to the confidence of each road feature in the third feature set of the feature point;

A similar feature point determining unit is configured to: if a matching degree between a third feature set corresponding to an adjacent feature point on the vehicle travel path exceeds a third preset matching degree threshold, and the adjacent feature points respectively correspond to The degree of matching between confidences exceeds a preset reliability matching threshold, and the adjacent feature points are used as similar feature points;

The second feature point screening unit is used for filtering similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points, and excluding the remaining feature points in the adjacent feature points.

According to a third aspect of the present application, there is provided an autonomous driving vehicle including the vehicle positioning device provided in the second aspect of the present application, and further comprising: an automatic driving system;

The vehicle positioning device, configured to send the absolute position of the vehicle to the automatic driving system;

The automatic driving system is configured to control the vehicle for automatic driving according to the absolute position of the vehicle.

Compared with the prior art, the present invention has at least the following advantages:

The vehicle positioning method provided in this application is applied to a vehicle equipped with a vehicle-mounted sensor capable of recognizing road characteristics. First, road characteristics of feature points on a driving path of a vehicle are obtained from a high-precision map to form a first feature set. The first type of road features are filtered from a feature set to form the second feature set. The first type of road features are road features that can be recognized by on-board sensors; then, on-vehicle sensors are used to identify road features at feature points on the vehicle's driving path to obtain a third feature Set; then, comparing the second feature set and the third feature set, if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, using an accurate map to obtain the absolute position of the feature points, The relative position of the vehicle and the feature point is obtained from the vehicle's on-board positioning system; finally, the absolute position of the vehicle is obtained according to the absolute position of the feature point and the relative position of the vehicle and the feature point.

This method is based on the recognition accuracy of road features by on-board sensors, preliminary screening of road features in the first feature set, and obtaining a second feature set composed of the first type of road features. After screening, it avoids road features that cannot be recognized by the on-board sensors. Affects the exact matching of road features at feature points. If the matching degree between the second feature set and the third feature set exceeds the first preset matching degree threshold value, it indicates that the road features of the feature points detected by the on-board sensors and the roads of the feature points on the vehicle driving path on the high-precision map The features are similar, and the road features of the feature points are successfully matched. Based on the successfully matched feature points, the absolute position and the relative position of the vehicle and the feature points can be obtained to obtain a more accurate absolute position of the vehicle. This method obtains the absolute position of feature points based on a high-precision map, and further obtains the absolute position of the vehicle. Compared with the prior art, the accuracy of vehicle positioning is effectively improved.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to explain the technical solutions in the embodiments of the present application or the prior art more clearly, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely These are some of the embodiments described in this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without paying creative labor.

FIG. 1 is a flowchart of a vehicle positioning method according to a first embodiment of the present application; FIG.

2 is a flowchart of a vehicle positioning method according to a second embodiment of the present application;

3 is a flowchart of a vehicle positioning method according to a third embodiment of the present application;

4 is a schematic structural diagram of a vehicle positioning device according to a fourth embodiment of the present application;

FIG. 5 is a schematic structural diagram of an autonomous driving vehicle according to a fifth embodiment of the present application.

detailed description

In order to solve the problem that the foregoing vehicle positioning accuracy is too low and it is difficult to meet the needs of map-based automatic driving, the present application provides a vehicle positioning method, device, and automatic driving vehicle.

The technical solution of the present application will be clearly and completely described in the following with reference to the embodiments and the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

First embodiment

Refer to FIG. 1, which is a flowchart of a vehicle positioning method according to an embodiment of the present application. First of all, it should be noted that the vehicle positioning method provided in this embodiment is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor capable of identifying road characteristics.

As shown in FIG. 1, the vehicle positioning method provided in this embodiment includes:

S101: Obtain road features of feature points on a vehicle's driving path from a high-precision map to form a first feature set.

The vehicle travel path may be a travel path determined by a travel start point and a travel destination point of the vehicle. The vehicle travel path includes at least two characteristic points. In this embodiment, the characteristic points may be various types of intersections and landmarks, such as intersections, T-junctions, restaurants, shopping malls, schools, stations, bridges, and the like.

Road features at feature points include traffic signs, pavement features, pavement conditions, and traffic conditions near feature points. As an example, the road characteristics of the feature point may be: intersection stop line, lane line shape, lane curvature, intersection characteristics, no-stop sign, speed limit sign, speed limit release sign, sign, signal light, left turn, right turn, Go straight, change lanes, drive to the left, drive to the right, construction ahead, merge, traffic accidents ahead, etc.

The first feature set of feature points contains all road features of the feature points in the high-precision map.

It should be noted that, for each feature point on the vehicle's travel path, the technical solution provided by this embodiment is adopted. Each feature point corresponds to a first feature set. In this embodiment, description is made by taking only a case of any one of the feature points as an example.

S102: Filter road features of the first type from the first feature set to form a second feature set.

The in-vehicle sensor has limited road feature recognition accuracy for feature points on the vehicle's driving path, and may not be able to recognize all road features of the feature points included in the first feature set, but only part of the road features. In order to achieve high-precision positioning, the high-precision map feature points and the feature points detected by the on-board sensor will be matched based on the detection and recognition results of the road features by the on-board sensor. The exact matching caused by this method affects. In this step, the road features that can be recognized by the vehicle-mounted sensors, that is, the road features of the first type, are selected from the first feature set, and the road features of the first type constitute the second feature set. Furthermore, the road features included in the second feature set are all road features that can be recognized based on the feature point on-board sensors.

As an example, each road feature in the first feature set has a confidence level, and the confidence level of each road feature indicates the degree to which the road feature can be reliably recognized by the vehicle-mounted sensor. For example, the feature recognition confidence threshold of the vehicle-mounted sensor is T _r . In the first feature set, road features with a confidence level greater than the feature recognition confidence threshold T _r are used as the first type of road features, and the first type of road features constitutes the second feature. set.

It should be noted that, in the embodiment of the present application, the vehicle-mounted sensors include, but are not limited to, any one or more of the following combinations: Global Positioning System (Global Navigation Navigation Satellite System, GNSS), odometer, lane recognition system, traffic sign recognition system And other road feature recognition systems.

Among them, the odometer includes but is not limited to any one or more of the following combinations: a transmission system output shaft speed sensor, a wheel speed sensor, and a vehicle speed sensor;

The identification content of the lane recognition system includes, but is not limited to, any one or more of the following combinations: lane line type, lane line curvature, lane-based offset, and lane-based heading deviation;

The identification content of the traffic sign recognition system includes, but is not limited to, any one or more of the following: a speed limit sign and a traffic sign;

The identification content of other road feature recognition systems includes, but is not limited to: water horse and gantry.

S103: Use the in-vehicle sensor to identify road features of the feature points on the vehicle travel path to obtain a third feature set.

Each road feature in the third feature set is a road feature whose confidence degree actually recognized by the vehicle-mounted sensor exceeds the feature recognition confidence threshold T _r .

S104: Compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, use the high precision The map obtains the absolute position of the feature point, and obtains the relative position of the vehicle and the feature point from a vehicle's on-board positioning system.

To determine whether the second feature set and the third feature set of the feature points are successfully matched, in the embodiment of the present application, the matching degree between the second feature set and the third feature set corresponding to the feature points is compared with a first preset matching degree threshold. If the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, it indicates that the second feature set of the feature points and the third feature set are successfully matched, and the feature detected by the vehicle sensor The road features of the points are similar to the road features of the feature points on the vehicle's driving path on the high-precision map, that is, the road features are successfully matched.

It should be noted that when determining whether the second feature set of the feature points and the third feature set are successfully matched, the consistency of the order of road features in the second feature set and the third feature set is not required.

As an example, the second feature set of feature points is {L1, L2, L3, L4}, the third feature set is {L3, L1, L4}, and L1, L2, L3, and L4 respectively represent different kinds of feature points. Road characteristics. In this example, the first preset matching degree threshold is 72%. Since the road features of the same kind included in the second feature set and the third feature set occupy 75% of all road feature types in the second feature set, the features are The matching degree between the second feature set and the third feature set of the point exceeds 75% of the first preset matching degree threshold. In addition, since the confidence of each road feature in the third feature set exceeds the feature recognition confidence threshold of the vehicle-mounted sensor, it can be determined that the second feature set is similar to the third feature set.

The second feature set of the feature points is similar to the third feature set, and the second feature set representing the feature points and the road features in the third feature set are successfully matched with each other. This feature point can be used to achieve high-precision positioning of the vehicle. As an example, the absolute position P _{J of the} feature point J in the high-precision map is obtained, and the position of the feature point J obtained by the vehicle's in-vehicle positioning system such as the Global Positioning System (GPS) is not P _J. Because the positioning accuracy of the vehicle positioning system is insufficient, there must be an error in the vehicle W position obtained only by the vehicle positioning system. In order to overcome the positioning error of the vehicle W, the relative position ΔP _{WJ of the} vehicle W and the feature point J is also obtained from the vehicle-mounted positioning system, and ΔP _WJ can be used to obtain the absolute position of the vehicle W.

S105: Obtain the absolute position of the vehicle according to the absolute position of the feature point and the relative position of the vehicle and the feature point.

As an optional implementation manner, S105 may specifically include:

S1051: Convert the relative position in a coordinate system of an absolute position of the feature point to obtain a first relative position of the vehicle and the feature point.

S1052: Move the absolute position of the feature point to the first relative position to obtain the absolute position of the vehicle. The absolute position of the vehicle includes a longitudinal coordinate, a lateral coordinate, and a heading of the vehicle.

Following the example in S104, the relative position ΔP _{WJ of} the vehicle W and the characteristic point J is transformed in the coordinate system of the absolute position P _J of the characteristic point J, and the first relative position ΔP ′ _{WJ of} the vehicle W and the characteristic point J is obtained; The absolute position P _{J of the} feature point J is moved by the first relative position ΔP ′ _WJ to obtain the absolute position P _{W of the} vehicle W. Relative to the position of the vehicle W obtained directly by the vehicle-mounted positioning system, the accuracy of the absolute position P _W of the vehicle W obtained by using the vehicle positioning method provided in the embodiment of the present application is improved.

The above is the vehicle positioning method provided by the embodiment of the present application. This method is based on the recognition accuracy of road features by on-board sensors, preliminary screening of road features in the first feature set, and obtaining a second feature set composed of the first type of road features. After screening, it avoids road features that cannot be recognized by the on-board sensors. Affects the exact matching of road features at feature points. If the matching degree between the second feature set and the third feature set exceeds the first preset matching degree threshold value, it indicates that the road features of the feature points detected by the on-board sensors and the roads of the feature points on the vehicle driving path on the high-precision map The features are similar, and the road features of the feature points are successfully matched. Based on the successfully matched feature points, the absolute position and the relative position of the vehicle and the feature points can be obtained to obtain a more accurate absolute position of the vehicle. This method obtains the absolute position of feature points based on a high-precision map, and further obtains the absolute position of the vehicle. Compared with the prior art, the accuracy of vehicle positioning is effectively improved.

Because there may be several feature points on the vehicle's driving path that are consecutively adjacent and have similar road features. These adjacent feature points with highly similar road features may cause interference to the road feature matching of the second feature set and the third feature set for the same feature point. For example, on the vehicle's driving path, there are adjacent feature points K and feature points F, the second feature sets corresponding to feature points K and feature points F are highly similar, and the third features corresponding to feature points K and feature points F, respectively. The collections are highly similar. If the second feature set of the feature point K is highly similar to the third feature set of the feature point F, it is likely that the feature point F identified by the in-vehicle sensor is regarded as the feature point K on the driving path of the vehicle on the high-precision map, and then affect Vehicle positioning accuracy.

To avoid the above problems, the present application further provides another vehicle positioning method. The specific implementation of the vehicle positioning method is described in detail below with reference to the embodiments and the accompanying drawings.

Second embodiment

Referring to FIG. 2, this figure is a flowchart of a vehicle positioning method according to this embodiment.

As shown in FIG. 2, the vehicle positioning method provided in this embodiment includes:

S201: Obtain road features of feature points on a vehicle's driving path from a high-precision map to form a first feature set.

S202: Filter road features of the first type from the first feature set to form a second feature set.

S203: Use the in-vehicle sensor to identify road features of the feature points on the vehicle travel path to obtain a third feature set.

In this embodiment, the implementation manners of S201 to S203 are the same as the implementation manners of S101 to S103 in the foregoing embodiment, respectively. For a detailed description of S201 to S203, refer to the foregoing embodiment, which will not be described in detail in this embodiment.

S204: Filter feature points according to the road features of the feature points, and remove invalid feature points.

In this step, the neighboring feature points are filtered and eliminated according to the matching degree of the road features of the neighboring feature points. It should be noted that, in this embodiment, adjacent feature points may be two adjacent feature points, or two or more feature points having an adjacent relationship.

As a first implementation manner, adjacent feature points can be filtered and eliminated from the perspective of road feature matching degrees including feature points in a high-precision map. details as follows:

In this embodiment, if the matching degree between the second feature sets corresponding to the adjacent feature points exceeds a second preset matching degree threshold value, it indicates that the adjacent feature points are adjacent feature points with similar road features. As an example, the second preset matching degree threshold may be 90%.

As an example, the matching degree between the second feature set corresponding to the adjacent feature points K and the feature points F on the vehicle travel path is greater than the second preset matching degree threshold, that is, the first corresponding feature points K and the feature points F respectively correspond to In the two feature sets, the road features are highly similar. The first feature point K passed by the vehicle is taken as the valid feature point, and the following S205 to S206 are continued; the feature point F will be removed as an invalid feature point and will not be executed. S205 to S206 are described.

As a second implementation manner, adjacent feature points can be filtered and eliminated from the perspective of road feature matching of the feature points actually recognized by the vehicle-mounted sensor. details as follows:

Determine whether the matching degree between the third feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a second preset matching degree threshold, and if so, the first feature passing by the vehicle in the adjacent feature points The feature points are used as effective feature points, and the remaining feature points except the effective feature points among the adjacent feature points are eliminated.

If the matching degree between the third feature sets corresponding to the adjacent feature points exceeds the second preset matching degree threshold value, it indicates that the adjacent feature points are adjacent feature points with similar road features. As an example, the second preset matching degree threshold may be 90%.

As an example, the matching degree between the third feature set corresponding to the adjacent feature point M and the feature point N on the vehicle travel path is greater than the second preset matching degree threshold, that is, the first corresponding feature point M and the feature point N respectively correspond to Among the three feature sets, road features are highly similar. The first feature point M passed by the vehicle is taken as the effective feature point, and the following S205 to S206 are continued; the feature point N is eliminated as the invalid feature point, and the following S205 to S206 are not executed.

It can be understood that S204 can also combine the above two implementation methods to filter and remove adjacent feature points. In other words, from the perspective of the road feature matching degree of the high-precision map containing the feature points, and the angle of the road feature matching degree of the feature points actually recognized by the vehicle sensor, the adjacent feature points can be filtered and eliminated. Here, the implementation is not described in detail, and reference may be made to the descriptions in the two implementations of the foregoing S204.

S205: Compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, use the high precision The map obtains the absolute position of the feature point, and obtains the relative position of the vehicle and the feature point from a vehicle's on-board positioning system.

S206: Obtain the absolute position of the vehicle according to the absolute position of the feature point and the relative position of the vehicle and the feature point.

In this embodiment, the implementation manners of S205 and S206 are the same as the implementation manners of S104 to S105 in the foregoing embodiment, respectively. For a detailed description of S205 to S206, refer to the foregoing embodiment, which will not be repeated in this embodiment.

The above is a vehicle positioning method provided by this embodiment. In the process of achieving high-precision positioning, the method screens feature points according to road characteristics of the feature points, and eliminates invalid feature points. Specifically, adjacent feature points whose matching degree between the second feature sets exceeds a second preset matching degree threshold, and / or neighboring feature points whose matching degree between the third feature sets exceeds a second preset matching degree threshold The filtering is performed to keep the first feature point passed by the vehicle in the adjacent feature points, and to remove the remaining feature points except the effective feature points in the adjacent feature points. Therefore, the feature points that are highly similar to the road features in the second feature set of the first feature point after the first feature point in the adjacent feature points, and / or after the first feature point in the adjacent feature points are filtered out. Feature points that are highly similar to the road features of the third feature set of the first feature point, to avoid mismatching of road features of the feature points from the second feature set to the third feature set. Therefore, this method improves the accuracy of road feature matching between the second feature set and the third feature set of feature points, thereby avoiding the mismatch of feature points and further ensuring the accuracy of vehicle positioning.

Because there may be several feature points on the vehicle's driving path that are adjacent and have similar road characteristics. Because these feature points are closely located and the road features are similar, they may interfere with the road feature matching of the second feature set and the third feature set for the same feature point.

In order to avoid the above problems and ensure the accuracy of vehicle positioning, the present application further provides a vehicle positioning method. The specific implementation of the vehicle positioning method is described in detail below with reference to the embodiments and the accompanying drawings.

Third embodiment

Refer to FIG. 3, which is a flowchart of a vehicle positioning method according to this embodiment.

As shown in FIG. 3, the vehicle positioning method provided in this embodiment includes:

S301: Obtain road features of feature points on a vehicle's driving path from a high-precision map to form a first feature set.

S302: Filter road features of the first type from the first feature set to form a second feature set.

S303: Use the vehicle-mounted sensor to identify road features of the feature points on the vehicle's travel path to obtain a third feature set.

In this embodiment, the implementation manners of S301 to S303 are the same as the implementation manners of S101 to S103 in the foregoing embodiment, respectively. For a detailed description of S301 to S303, refer to the foregoing embodiment, which will not be repeated in this embodiment.

S304: Obtain the confidence level of each road feature in the third feature set of the feature points.

As a specific implementation manner, when the road feature of the feature point is identified by the vehicle-mounted sensor, the confidence level of each road feature can be obtained. It can be understood that, as the vehicle passes through the feature point, the distance between each road feature of the feature point and the vehicle-mounted sensor continuously changes, so the confidence of each identified road feature is a variable that changes with time.

The confidence of each road feature in the third feature set represents the credibility of each road feature in the third feature set. Road features with higher confidence have higher credibility.

Of course, the confidence level of each road feature in the third feature set can also be processed and calculated by certain parameters of the road feature. Here, the specific obtaining method of the confidence level of each road feature in the third feature is not limited.

S305: Obtain the confidence level of the feature point according to the confidence level of each road feature in the third feature set of the feature point.

In order to filter the adjacent feature points on the vehicle's driving path and the third feature set is similar, in this embodiment, the confidence of each road feature in the third feature set is used to calculate the confidence of the feature points corresponding to the third feature set degree. In this step, the confidence of the feature points can be obtained as follows:

S3051: Obtain the average confidence of each road feature in the third feature set of the feature points within a preset time.

As an example, the in-vehicle sensor recognizes the road features of the feature points, and the confidence levels of m road features in the third feature set of the feature points are f _c1 (t), f _c2 (t), f _c3 (t), ... f _cm (t), within a predetermined time T m th average confidence road features are _{_{_{f ca1, f ca2, f ca3}}} ... f cam.

S3052: Multiply the average confidence level of each road feature within a preset time by the confidence level of each road feature obtained from the high-precision map to obtain a first level of each road feature of the feature point. Confidence.

As an example, the confidence levels of the above m road features at the feature points obtained from the high-precision map are f ₁ , f ₂ , f ₃ … f _{m, respectively} .

The first confidence level of m road features of the feature point can be obtained by using formula (1).

f _j1 = f _j * f _caj formula (1)

In formula (1), j represents the j-th road feature among m road features at the feature point, m is an integer greater than 1, j is any integer in [1, m], and f _j is the j-th road feature On the high-precision map, f _caj is the average confidence of the j-th road feature within a preset time T, and f _j1 represents the first confidence of the j-th road feature.

S3053: Add the first confidence levels of the road features of the feature points to obtain the confidence levels of the feature points.

Following the example in S3052, the confidence level of the feature points can be expressed by formula (2).

In formula (2), j represents the j-th road feature among the m road features of the feature point, f _j1 represents the first confidence level of the j-th road feature, and F represents the confidence level of the feature point.

S306: If the matching degree between the third feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a third preset matching degree threshold, and the matching between the confidence degrees corresponding to the adjacent feature points respectively If the degree exceeds a preset reliability matching threshold, the adjacent feature points are regarded as similar feature points.

In this embodiment, if the matching degree between the third feature sets corresponding to the adjacent feature points exceeds a third preset matching degree threshold, it indicates that the road features of the third feature set corresponding to the adjacent feature points are similar. As an example, the third preset matching degree threshold may be 70%.

As an example, in this embodiment, it can be determined whether the third feature set of the adjacent feature points are similar according to the number and type of road features in the third feature set of the adjacent feature points. For example, a third preset matching degree threshold Ts is set in advance, there are adjacent feature points A and B, and the corresponding third feature sets are L _A3 = {L1, L2, L3, L4}, and L _B3 = {L1 , L2, L5, L3}. According to L _A3 and L _B3 , the number of road features in the third feature set of the adjacent feature points A and B is the same, and the number of the same road feature in each third feature set occupies 75% of the total number of road features. The determined matching degree between the third feature sets corresponding to the adjacent feature points A and B is 75%. The similarity between adjacent feature points A and B is 75% and the third preset matching degree threshold Ts is compared. If Ts is exceeded, it can be determined that the third feature sets of adjacent feature points A and B are similar.

In this embodiment, the preset reliability matching threshold is a maximum matching degree for determining that the confidence degrees of adjacent feature points do not match. If the matching degree between the respective confidence points corresponding to the adjacent feature points exceeds the preset value confidence matching threshold value, it means that the confidence degrees corresponding to the feature points respectively match each other, and the corresponding confidence degrees corresponding to the neighboring feature points are similar.

In this embodiment, the confidence of adjacent feature points may be used as a difference and an absolute value may be taken, and the obtained value may be used as the matching degree of the confidence of neighboring feature points. For example, the confidence levels of adjacent feature points are differenced and taken as absolute values, and the absolute value results are compared with a preset confidence level matching threshold T _F to determine whether the confidence levels of adjacent feature points are similar. For example, the confidence levels of adjacent feature points A and B are 78% and 91%, respectively, and the absolute value of the difference between the confidence levels of adjacent feature points A and B is 13%. The absolute value result of the degree difference value is 13% compared with the confidence matching threshold value T _F. If the absolute value result exceeds T _F , it can be determined that the confidence levels of the adjacent feature points A and B are similar.

It can be understood that the method for determining the matching degree of the third feature set respectively corresponding to the adjacent feature points and the method of determining the confidence degree matching degree of the adjacent feature points provided above are merely examples. In specific implementation, other methods may be used to determine the matching degree of the third feature set corresponding to the adjacent feature points, and the confidence degree matching degree of the adjacent feature points, respectively. Here, the specific implementation manner of determining the matching degree of the third feature set respectively corresponding to the adjacent feature points and determining the matching degree of the confidence degree corresponding to the adjacent feature points are not limited.

S307: Filter the similar feature points with the highest confidence among the similar feature points and exceeding the confidence threshold, and remove the remaining feature points from the adjacent feature points.

As an example, in this embodiment, a confidence threshold F _thre = 83% is set in advance. If there are four adjacent feature points A, B, C, and D, and the feature points A, B, C, and D are similar feature points to each other. The confidence level F _{A of the} feature point A is 78%, the confidence level F _{B of the} feature point B is 91%, the confidence level F _{C of the} feature point C is 79%, and the confidence level F _{D of the} feature point D is 85%. Among them, the similar feature point with the highest confidence level and exceeding the confidence threshold value F _thre is the feature point B, so only the feature point B is retained, and the remaining feature points A, C, and D in the neighboring feature points are eliminated.

For the remaining feature points B, the following S308 to S309 are continued; while the feature points A, C, and D will be removed as invalid feature points, and the following S308 to S309 will not be executed.

S308: Compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, use the high precision The map obtains the absolute position of the feature point, and obtains the relative position of the vehicle and the feature point from a vehicle's on-board positioning system.

S309: Obtain the absolute position of the vehicle according to the absolute position of the feature point and the relative position of the vehicle and the feature point.

In this embodiment, the implementation manners of S308 and S309 are the same as the implementation manners of S104 to S105 in the foregoing embodiment, respectively. For detailed descriptions of S308 to S309, refer to the foregoing embodiment, which will not be repeated in this embodiment.

The above is a vehicle positioning method provided by this embodiment. In the process of achieving high-precision positioning, the method determines whether the adjacent feature points are similar according to the matching degree of the third feature set corresponding to the adjacent feature points and the matching degree of the confidence level corresponding to the adjacent feature points, respectively. Feature points; thereafter, the respective confidence levels corresponding to the similar feature points are compared with the confidence threshold values, and only the similar feature points with the highest confidence among the similar feature points and exceeding the confidence threshold value are retained. As a result, screening of similar feature points on the vehicle's travel path is realized. Avoid mismatching of road features between the second feature set and the third feature set of feature points. Therefore, this method also improves the accuracy of road feature matching between the second feature set and the third feature set of feature points, thereby avoiding the mismatch of feature points and further ensuring the accuracy of vehicle positioning.

It can be understood that, since the confidence level of the feature points identified by the vehicle-mounted sensor is a variable that changes with time, in this embodiment, when selecting similar feature points in adjacent feature points, the feature point confidence level update interval t _S can be set, Every t _s time interval, the confidence level of the feature point is updated and calculated. Furthermore, the vehicle positioning method provided by the vehicle positioning method provided in this embodiment has higher real-time accuracy.

In order to solve the foregoing technical problem, the present application is based on the foregoing method embodiment, and also provides a vehicle positioning device. The vehicle positioning device is described in detail below with reference to the drawings and embodiments.

Fourth embodiment

Refer to FIG. 4, which is a schematic structural diagram of a vehicle positioning device according to an embodiment of the present application. It should be noted that the vehicle positioning device provided in this embodiment is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor capable of identifying road characteristics.

As shown in FIG. 4, the vehicle positioning device 40 provided in this embodiment includes:

The first feature set acquisition module 401, the second feature set acquisition module 402, the third feature set acquisition module 403, the feature comparison and position acquisition module 404, and the positioning module 405.

The first feature set obtaining module 401 is configured to obtain road features of feature points on a driving path of a vehicle from a high-precision map to form a first feature set;

The second feature set acquisition module 402 is configured to filter the first type of road features from the first feature set to form a second feature set, where the first type of road features are road features that can be recognized by the vehicle-mounted sensor;

The third feature set acquisition module 403 is configured to use the on-board sensor to identify road features of the feature points on the vehicle's driving path to obtain a third feature set;

The feature comparison and location acquisition module 404 is configured to compare the second feature set and the third feature set, and if the degree of matching between the second feature set and the third feature set exceeds a first prediction Set a matching threshold, use the high-precision map to obtain the absolute position of the feature point, and obtain the relative position of the vehicle and the feature point from the vehicle's on-board positioning system;

The positioning module 405 is configured to obtain an absolute position of the vehicle according to an absolute position of the feature point and a relative position of the vehicle and the feature point.

The above is the vehicle positioning device provided by the embodiment of the present application. Based on the recognition accuracy of road features by on-board sensors, the device performs a preliminary screening of road features in the first feature set to obtain a second feature set composed of first-class road features. After screening, it avoids road features that cannot be recognized by on-board sensors. Affects the exact matching of road features at feature points. If the matching degree between the second feature set and the third feature set exceeds the first preset matching degree threshold value, it indicates that the road features of the feature points detected by the on-board sensors and the roads of the feature points on the vehicle driving path on the high-precision map The features are similar, and the road features of the feature points are successfully matched. Based on the successfully matched feature points, the absolute position and the relative position of the vehicle and the feature points can be obtained to obtain a more accurate absolute position of the vehicle. The vehicle positioning device obtains the absolute position of the feature points based on the high-precision map, and further obtains the absolute position of the vehicle. Compared with the prior art, the accuracy of vehicle positioning is effectively improved.

Because there may be several adjacent points on the vehicle's driving path, and the road features are highly similar or similar. These adjacent feature points with high or similar road features may interfere with the road feature matching of the second feature set and the third feature set for the same feature point, thereby affecting the positioning accuracy of the vehicle.

To solve the above problems, the vehicle positioning device provided in this embodiment may further include:

A feature point screening module 406 is configured to filter feature points according to road characteristics of the feature points and eliminate invalid feature points.

As an implementation manner, the feature point screening module 406 may specifically include:

The feature point first filtering unit 4061 is configured to determine whether the matching degree between the second feature sets corresponding to adjacent feature points on the driving path of the vehicle exceeds a second preset matching degree threshold. The first feature point passed by the vehicle in the adjacent feature points is taken as the effective feature point, and the remaining feature points except the effective feature points in the adjacent feature points are eliminated.

In this implementation, the adjacent feature points whose matching degree between the second feature sets exceeds the second preset matching degree threshold are filtered, the first feature point passing by the vehicle among the neighboring feature points is retained, and the neighboring feature points are eliminated. The remaining feature points except the effective feature points are thus filtered out. The feature points that are highly similar to the road features of the second feature set of the first feature point after the first feature point of the adjacent feature points are filtered to avoid the feature points The road feature of the second feature set is mismatched with the third feature set. Therefore, the device improves the accuracy of road feature matching between the second feature set and the third feature set of feature points, thereby avoiding mismatching of feature points and further ensuring the accuracy of vehicle positioning.

As another implementation manner, the feature point filtering module 406 may specifically include: a road feature confidence obtaining unit 4062, a feature point confidence obtaining unit 4063, a similar feature point determining unit 4064, and a feature point second filtering unit 4065.

The road feature confidence obtaining unit 4062 is configured to obtain the confidence of each road feature in the third feature set of the feature points; the confidence is a variable that changes with time;

A feature point confidence obtaining unit 4063, configured to obtain the confidence degree of the feature point according to the confidence degree of each road feature in the third feature set of the feature point;

A similar feature point determining unit 4064 is configured to: if a matching degree between a third feature set corresponding to an adjacent feature point on the vehicle travel path exceeds a third preset matching degree threshold, and the adjacent feature points respectively correspond to The degree of matching between the confidence levels exceeds a preset reliability matching threshold, and the adjacent feature points are used as similar feature points;

The second feature point screening unit 4065 is configured to filter the similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points, and remove the remaining feature points from the adjacent feature points.

In this implementation manner, according to the matching degree of the third feature set corresponding to the adjacent feature points and the matching degree of the confidence degree corresponding to the adjacent feature points, respectively, it is determined whether the adjacent feature points are similar feature points, and similar features are determined. The confidence level corresponding to each point is compared with the confidence threshold value, and only the similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points are retained. As a result, screening of similar feature points on the vehicle's travel path is realized. Avoid interference with road feature matching between the second feature set and the third feature set of feature points. Therefore, the device can also improve the accuracy of road feature matching between the second feature set and the third feature set of feature points, thereby avoiding mismatching of feature points and further ensuring the accuracy of vehicle positioning.

Based on the vehicle positioning device provided in the foregoing embodiment, the present application further provides an autonomous driving vehicle. The specific implementation of the vehicle will be described below with reference to the drawings and embodiments.

Fifth Embodiment

Refer to FIG. 5, which is a schematic structural diagram of an autonomous driving vehicle according to an embodiment of the present application.

As shown in FIG. 5, the autonomous driving vehicle 50 provided in this embodiment includes:

The vehicle positioning device 40 and the automatic driving system 501 provided in the fourth embodiment.

Wherein, the vehicle positioning device 40 is configured to send the absolute position of the vehicle to the automatic driving system;

The automatic driving system 501 is configured to control a vehicle for automatic driving according to an absolute position of the vehicle.

The above is the autonomous driving vehicle 50 provided in the embodiment of the present application. Among them, the vehicle positioning device 40 based on the recognition accuracy of road features by the on-board sensors, performs preliminary screening on the road features in the first feature set, and obtains the second feature set composed of the first type of road features. The identified road features affect the exact matching of feature points and road features. If the matching degree between the second feature set and the third feature set exceeds the first preset matching degree threshold value, it indicates that the road features of the feature points detected by the on-board sensors and the roads of the feature points on the vehicle driving path on the high-precision map The features are similar, and the road features of the feature points are successfully matched. Based on the successfully matched feature points, the absolute position and the relative position of the vehicle and the feature points can be obtained to obtain a more accurate absolute position of the vehicle. The vehicle positioning device 40 obtains the absolute position of the feature points based on the high-precision map, and further obtains the absolute position of the vehicle. Compared with the prior art, the accuracy of vehicle positioning is effectively improved.

In the self-driving vehicle 50, the auto-driving system 501 controls the vehicle to perform automatic driving based on the absolute position of the self-driving vehicle 50 obtained by the vehicle positioning device 40. Since the absolute position accuracy and accuracy of the vehicle is improved, the automatic driving system 501 can obtain the accurate position of the vehicle in the lane, which is conducive to the path planning and operation decision of the automatic driving of the vehicle.

It should be understood that in the present application, "at least one (item)" means one or more, and "multiple" means two or more. "And / or" is used to describe the association relationship between related objects, which means that there can be three kinds of relationships, for example, "A and / or B" can mean: only A, only B, and both A and B Where A and B can be singular or plural. The character "/" generally indicates that the related objects are an "or" relationship. "At least one or more of the following" or similar expressions means any combination of these items, including any combination of single or plural items. For example, at least one (a), a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c" ", Where a, b, and c can be single or multiple.

The above descriptions are merely preferred embodiments of the present invention, and do not limit the present invention in any form. Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make many possible changes and modifications to the technical solution of the present invention or modify it to the equivalent equivalent without departing from the scope of the technical solution of the present invention. Examples. Therefore, without departing from the content of the technical solution of the present invention, any simple modifications, equivalent changes, and modifications made to the above embodiments according to the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

A vehicle positioning method, which is characterized in that the method is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor; the vehicle-mounted sensor is used to identify road features;

Obtain road features of feature points on the vehicle's driving path from a high-precision map to form a first feature set;

Filtering the first type of road features from the first feature set to form a second feature set, and the first type of road features are road features that can be recognized by the vehicle-mounted sensor;

Using the vehicle-mounted sensor to identify road features of the feature points on the vehicle's driving path to obtain a third feature set;

Compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset matching degree threshold, then use the high-precision map Obtaining an absolute position of the feature point, and obtaining a relative position of the vehicle and the feature point from an on-board positioning system of the vehicle;

An absolute position of the vehicle is obtained according to an absolute position of the feature point and a relative position of the vehicle and the feature point.
The vehicle positioning method according to claim 1, wherein before the comparing the second feature set and the third feature set, the method further comprises:

Feature points are filtered based on the road features of the feature points, and invalid feature points are eliminated.
The vehicle positioning method according to claim 2, wherein the filtering feature points based on road features of the feature points and excluding invalid feature points specifically includes:

Determine whether the matching degree between the second feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a second preset matching degree threshold, and if so, the first feature passing by the vehicle in the neighboring feature points The feature points are used as effective feature points, and the remaining feature points except the effective feature points among the adjacent feature points are eliminated.
The vehicle positioning method according to claim 2, wherein the filtering feature points based on road features of the feature points and excluding invalid feature points specifically includes:

Obtaining the confidence level of each road feature in the third feature set of the feature points; the confidence level is a variable that changes with time;

Obtaining the confidence level of the feature point according to the confidence level of each road feature in the third feature set of the feature point;

If the matching degree between the third feature sets corresponding to adjacent feature points on the vehicle travel path exceeds a third preset matching degree threshold, and the matching degree between the confidence degrees corresponding to the adjacent feature points exceeds Preset a reliability matching threshold, and use the adjacent feature points as similar feature points;

The similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points are filtered, and the remaining feature points in the adjacent feature points are eliminated.
The vehicle positioning method according to claim 4, wherein the obtaining the confidence level of the feature point according to the confidence level of each road feature in the third feature set of the feature point specifically includes:

Obtaining an average confidence level of each road feature in the third feature set of the feature points within a preset time;

Multiplying the average confidence level of each road feature within a preset time by the confidence level of each road feature obtained from the high-precision map to obtain a first confidence level of each road feature of the feature point ;

The first confidence levels of the road features of the feature points are added to obtain the confidence levels of the feature points.
A vehicle positioning device is characterized in that it is applied to a vehicle, and the vehicle includes a vehicle-mounted sensor; the vehicle-mounted sensor is used to identify road features; the vehicle positioning device includes: a first feature set acquisition module and a second feature set An acquisition module, a third feature set acquisition module, a feature comparison and location acquisition module, and a positioning module;

The first feature set acquisition module is configured to obtain road features of feature points on a vehicle driving path from a high-precision map to form a first feature set;

The second feature set obtaining module is configured to filter the first type of road features from the first feature set to form a second feature set, and the first type of road features are road features that can be recognized by the vehicle-mounted sensor;

The third feature set acquisition module is configured to use the on-board sensor to identify road features of the feature points on the vehicle travel path to obtain a third feature set;

The feature comparison and position acquisition module is configured to compare the second feature set and the third feature set, and if the matching degree between the second feature set and the third feature set exceeds a first preset The matching degree threshold, using the high-precision map to obtain the absolute position of the feature point, and obtaining the relative position of the vehicle and the feature point from the on-board positioning system of the vehicle;

The positioning module is configured to obtain an absolute position of the vehicle according to an absolute position of the feature point and a relative position of the vehicle and the feature point.
The vehicle positioning device according to claim 6, wherein the vehicle positioning device further comprises:

The feature point screening module is configured to filter feature points according to the road features of the feature points and eliminate invalid feature points.
The vehicle positioning device according to claim 7, wherein the feature point screening module specifically comprises:

The feature point first screening unit is configured to determine whether a matching degree between second feature sets corresponding to adjacent feature points on the vehicle's driving path exceeds a second preset matching degree threshold, and if yes, compare the phase The first feature point passed by the vehicle in the adjacent feature points is used as the effective feature point, and the remaining feature points except the effective feature points in the adjacent feature points are eliminated.
The vehicle positioning device according to claim 7, wherein the feature point screening module specifically comprises:

The road feature confidence degree obtaining unit is configured to obtain the confidence degree of each road feature in the third feature set of the feature points; the confidence degree is a variable that changes with time;

A feature point confidence obtaining unit, configured to obtain the confidence of the feature point according to the confidence of each road feature in the third feature set of the feature point;

A similar feature point determining unit is configured to: if a matching degree between a third feature set corresponding to an adjacent feature point on the vehicle travel path exceeds a third preset matching degree threshold, and the adjacent feature points respectively correspond to The degree of matching between confidences exceeds a preset reliability matching threshold, and the adjacent feature points are used as similar feature points;

The second feature point screening unit is used for filtering similar feature points with the highest confidence level and exceeding the confidence threshold value among the similar feature points, and excluding the remaining feature points in the adjacent feature points.
An automatic driving vehicle, comprising the vehicle positioning device according to any one of claims 6-9, further comprising: an automatic driving system;

The vehicle positioning device, configured to send the absolute position of the vehicle to the automatic driving system;

The automatic driving system is configured to control the vehicle for automatic driving according to the absolute position of the vehicle.