Vehicle navigation method, device, equipment and storage medium

Abstract

The present application discloses a vehicle navigation method, device, equipment and storage medium, which relates to the field of vehicle driving technology. The method includes: matching and correcting the map lane line and the visually perceived lane line corresponding to the vehicle to obtain the lane line matching information between the map lane line and the visually perceived lane line; correcting the vehicle positioning information based on the lane line matching information to obtain the corrected positioning information; obtaining the perceived lane line effective information and the map lane line effective information based on the vehicle's driving scene and lane line matching information; fusing the map center line and the perceived center line based on the perceived lane line effective information and the map lane line effective information to obtain the road center line, and navigating the vehicle through the road center line. The above technical solution solves the technical problem in the prior art that the vehicle navigation accuracy is not high due to certain errors in the vehicle positioning information when navigating the autonomous driving vehicle using a high-precision map.

Description

Technical Field

[0001] The present application relates to the field of vehicle driving technology, and in particular to a vehicle navigation method, device, equipment and storage medium. Background Art

[0002] As the autonomous driving capabilities of autonomous vehicles improve, they may become the mainstream mode of transportation for the public in the future. In order to ensure the normal autonomous driving of the vehicle, it is usually necessary to provide an accurate reference line for the vehicle's trajectory tracker. At present, the center line of the current lane can be calculated through high-precision map information and accurate positioning as the reference line for vehicle driving.

[0003] When using a high-definition map (HD Map) to navigate an autonomous vehicle, the latitude and longitude coordinates of the HD map data need to be converted into coordinates in the vehicle coordinate system. At this time, dynamic vehicle positioning information will be used. However, the positioning information of the GPS (Global Positioning System) is not accurate and usually has certain errors, resulting in errors in calculating the current lane centerline of the vehicle, thereby causing low vehicle navigation accuracy. Summary of the invention

[0004] The main purpose of this application is to provide a vehicle navigation method, device, equipment and storage medium, aiming to solve the technical problem in the prior art that when using high-precision maps to navigate autonomous vehicles, the vehicle navigation accuracy is not high due to certain errors in vehicle positioning information.

[0005] To achieve the above objectives, the present application proposes a vehicle navigation method, the method comprising:

[0006] Matching and correcting the lane line on the map and the visually perceived lane line corresponding to the vehicle to obtain lane line matching information between the lane line on the map and the visually perceived lane line;

[0007] Correcting the vehicle positioning information based on the lane line matching information to obtain corrected positioning information;

[0008] Acquiring sensed lane line valid information and map lane line valid information based on the driving scenario of the vehicle and the lane line matching information;

[0009] Determining a map centerline based on the corrected positioning information;

[0010] Determine a sensed center line based on vehicle travel information and sensed lane information;

[0011] The map center line and the perceived center line are fused based on the perceived lane line valid information and the map lane line valid information to obtain a road center line, and vehicle navigation is performed through the road center line.

[0012] In one embodiment, the step of matching and correcting the lane line on the map and the visually perceived lane line corresponding to the vehicle includes:

[0013] Obtain a first reference point cloud on a visual perception lane line corresponding to the vehicle and a second reference point cloud on a map lane line;

[0014] Constructing a target distance and a minimum function based on the first reference point cloud and the second reference point cloud;

[0015] Determine a target rotation matrix and a target translation matrix based on the target distance and the minimum function;

[0016] The map lane lines and the visually perceived lane lines are matched and corrected according to the target rotation matrix and the target translation matrix.

[0017] In one embodiment, the step of acquiring the sensed lane line valid information and the map lane line valid information based on the driving scene of the vehicle and the lane line matching information includes:

[0018] Segmenting the visually perceived lane line to obtain a plurality of road segments;

[0019] Generating a road topology structure diagram corresponding to the visually perceived lane line based on the connection relationship between the road segments;

[0020] Determining a driving scene of the vehicle based on the road topology diagram to obtain effective information of perceived lane lines;

[0021] The valid information of the lane lines on the map is obtained based on the lane line matching information.

[0022] In one embodiment, the step of obtaining valid map lane line information based on the lane line matching information includes:

[0023] Determining matching results between a plurality of first lane line point sets in the visually perceived lane lines and a plurality of second lane line point sets in the map lane lines based on the lane line matching information;

[0024] Determine whether the lane line on the map is a valid lane line according to the matching result;

[0025] Obtain valid map lane line information based on the judgment results.

[0026] In one embodiment, the step of determining the center line of the map based on the corrected positioning information includes:

[0027] Determining corrected coordinate information according to the corrected positioning information;

[0028] Performing error state Kalman filtering on the corrected coordinate information based on the vehicle body inertial measurement data to obtain filtered coordinate information;

[0029] A map centerline is determined based on the filtered coordinate information.

[0030] In one embodiment, the step of determining the sensed center line based on the vehicle driving information and the sensed lane information includes:

[0031] Determine a time score, a weather score, a scene score, and a camera score of the vehicle during a current driving process based on the vehicle driving information and the perceived lane information;

[0032] Determine confidence scores corresponding to left and right lane lines in the visually perceived lane line based on the time score, the weather score, the scene score, and the camera score;

[0033] When the confidence score exceeds a preset confidence score threshold, determining whether the lane line widths of the left and right lane lines exceed a preset width threshold;

[0034] If not, determining whether the left and right lane lines are in a parallel state, and obtaining a validity determination result of the left and right lane lines according to the determination result;

[0035] A sensed center line is determined based on the validity judgment result and the coordinate information of the left and right lane lines.

[0036] In one embodiment, the step of fusing the map centerline and the sensed centerline based on the sensed lane line valid information and the map lane line valid information to obtain the road centerline includes:

[0037] Determining a confidence level corresponding to the sensed center line according to the validity judgment result;

[0038] Performing low-confidence scene judgment on the visually perceived lane line according to the effective information of the perceived lane line to obtain a road scene recognition result;

[0039] Judging the validity of the lane line on the map according to the valid information of the lane line on the map, and obtaining a map validity recognition result;

[0040] The map centerline and the perceived centerline are fused based on the confidence level, the road scene recognition result, and the map validity recognition result to obtain a road centerline.

[0041] In addition, to achieve the above-mentioned purpose, the present application also proposes a vehicle navigation device, the device comprising:

[0042] A lane line matching module is used to match and correct the map lane line and the visually perceived lane line corresponding to the vehicle, and obtain lane line matching information between the map lane line and the visually perceived lane line;

[0043] A positioning information correction module, used to correct the vehicle positioning information based on the lane line matching information to obtain corrected positioning information;

[0044] An effective information acquisition module, used to acquire the effective information of the perceived lane line and the effective information of the map lane line based on the driving scene of the vehicle and the lane line matching information;

[0045] A centerline determination module, configured to determine a map centerline based on the corrected positioning information;

[0046] The centerline determination module is further used to determine the perceived centerline based on the vehicle driving information and the perceived lane information;

[0047] The vehicle navigation module is used to fuse the map center line and the perceived center line based on the perceived lane line valid information and the map lane line valid information to obtain a road center line, and perform vehicle navigation through the road center line.

[0048] In addition, to achieve the above objectives, the present application also proposes a vehicle navigation device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vehicle navigation method described above.

[0049] In addition, to achieve the above objectives, the present application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the vehicle navigation method described above are implemented.

[0050] The present application provides a vehicle navigation method, which discloses matching and correcting a map lane line with a visually perceived lane line corresponding to a vehicle, obtaining lane line matching information between the map lane line and the visually perceived lane line; correcting the vehicle positioning information based on the lane line matching information, obtaining the corrected positioning information; obtaining the perceived lane line effective information and the map lane line effective information based on the vehicle's driving scene and the lane line matching information; determining the map center line based on the corrected positioning information; determining the perceived center line based on the vehicle's driving information and the perceived lane information; fusing the map center line and the perceived center line based on the perceived lane line effective information and the map lane line effective information to obtain the road center line, and performing vehicle navigation through the road center line; compared with the prior art During the process, when using a high-precision map to navigate an autonomous vehicle, due to the low accuracy of GPS positioning information, there will be errors in calculating the vehicle's current lane center line, which in turn leads to low vehicle navigation accuracy. Since the present invention can correct the vehicle positioning information according to the lane line matching information between the map lane line and the visually perceived lane line, and determine the map center line based on the corrected positioning information, and then fuse the map center line and the perceived center line based on the effective information of the perceived lane line and the effective information of the map lane line, the road center line is obtained for vehicle navigation, thereby solving the technical problem in the prior art of using a high-precision map to navigate an autonomous vehicle, due to certain errors in the vehicle positioning information, resulting in low vehicle navigation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0053] Figure 1 A flowchart of the first embodiment of the vehicle navigation method of the present application is provided;

[0054] Figure 2 A schematic diagram of the original map line and the visual lane line in the vehicle navigation method of this application;

[0055] Figure 3 A schematic diagram of matching point sampling in the vehicle navigation method of the present application;

[0056] Figure 4 A schematic diagram showing the comparison of lane lines before and after correction in the vehicle navigation method of this application;

[0057] Figure 5 A flowchart diagram of the second embodiment of the vehicle navigation method of the present application is provided;

[0058] Figure 6 A flowchart of the third embodiment of the vehicle navigation method of the present application is provided;

[0059] Figure 7 A road topology diagram in the vehicle navigation method of this application;

[0060] Figure 8 A schematic diagram of road separation and merging in the vehicle navigation method of the present application;

[0061] Fig. 9 A schematic diagram of the process of generating a lane centerline in the vehicle navigation method of the present application;

[0062] Fig.10 A diagram showing centerline fusion in the vehicle navigation method of this application;

[0063] Fig.11 This is a schematic diagram of the module structure of a vehicle navigation device according to an embodiment of the present application;

[0064] Fig.12 It is a schematic diagram of the device structure of the hardware operating environment involved in the vehicle navigation method in the embodiment of the present application.

[0065] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0066] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0067] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0068] In the prior art, when using high-precision maps to navigate autonomous vehicles, dynamic vehicle positioning information is required to convert the latitude and longitude coordinates of the high-precision map data into coordinates in the vehicle coordinate system. However, since the accuracy of GPS positioning information is not high, there will be errors in calculating the current lane centerline of the vehicle, resulting in low vehicle navigation accuracy.

[0069] The present application provides a solution, which can correct the vehicle positioning information according to the lane line matching information between the map lane line and the visually perceived lane line, determine the map center line based on the corrected positioning information, and then fuse the map center line and the perceived center line based on the effective information of the perceived lane line and the effective information of the map lane line to obtain the road center line for vehicle navigation, thereby solving the technical problem in the prior art that when using high-precision maps to navigate autonomous driving vehicles, the vehicle positioning information has certain errors, resulting in low vehicle navigation accuracy.

[0070] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an ADAS device capable of realizing the above functions, a vehicle navigation device, etc. The following takes a vehicle navigation device as an example (hereinafter referred to as the device) to illustrate this embodiment and the following embodiments.

[0071] Based on this, the embodiment of the present application provides a vehicle navigation method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the vehicle navigation method of the present application.

[0072] In this embodiment, the vehicle navigation method includes steps S10 to S60:

[0073] Step S10: Match and correct the map lane line and the visually perceived lane line corresponding to the vehicle to obtain lane line matching information between the map lane line and the visually perceived lane line.

[0074] It should be noted that the above-mentioned map lane lines can be lane lines near the vehicle on the high-precision map. In this embodiment, the preprocessing module in the device can calculate the map boundary lines in the vehicle coordinate system before correction based on the vehicle's positioning information and the high-precision map information, that is, the above-mentioned map lane lines. Specifically, the device can convert the longitude and latitude points of the lane lines near the vehicle to the vehicle coordinate system based on the vehicle's positioning information (latitude and longitude, heading angle, etc.), and the following formula, so that the lane line points on the map and the perceived lane lines can be expressed in the same coordinate system:

[0075]

[0076] in, is the latitude of the point, lon is the longitude of the point, is the length of one degree of latitude, The length of one degree of longitude. Among them, the length of one degree of latitude is the shortest distance between two points on the earth's surface with a latitude difference of 1 degree; the length of one degree of longitude is the shortest distance between two points on the earth's surface with a longitude difference of 1 degree.

[0077]

[0078] in, and is the point trace of the lane line in the vehicle coordinate system, , and They are respectively the longitude, latitude and direction of the vehicle detected by GPS before correction.

[0079] It should be understood that the above-mentioned visually perceived lane line can be the lane line of the vehicle perceived by the front camera of the vehicle.

[0080] It should be noted that the above lane line matching information may be information used to characterize the matching situation between the map lane line and the visually perceived lane line.

[0081] It should be noted that, refer to Figure 2 , Figure 2 This is a schematic diagram of the original map line and visual lane line in the vehicle navigation method of this application. In actual applications, the perceived lane line is also composed of point traces. In this embodiment, the latitude and longitude points of the lane line are converted to the vehicle coordinate system to obtain the following Figure 2 The points of the left, left-left, right, and right-right lane lines of the perceived lane lines and the map lane lines are shown by Figure 2 It can be seen that the map boundary line ( Figure 2 The solid dots in the image) and the visual lane lines ( Figure 2 There is an obvious deviation (the dotted line in the figure), and this deviation is caused by the error of the positioning data. At this time, in order to improve the accuracy of subsequent vehicle navigation, the positioning data can be corrected to perform vehicle navigation based on the corrected positioning information.

[0082] In practical applications, refer to Figure 3 , Figure 3 Schematic diagram of matching point sampling in the vehicle navigation method of this application. Figure 3As shown, in this embodiment, the perceived lane lines (left, left-left, right, right-right) can be sampled at intervals of 5-10 meters to obtain four sets of reference point sets, and then the point sets are used to find projection points on the map lane lines (left, left-left, right, right-right) as four sets of source point sets, and then the point sets of the four perceived lane lines are matched with the four lane line point sets of the map in turn. If the distance between the two point sets is less than the threshold, it is considered that the perceived lane line can be matched with the map lane line. After matching the four lane lines in turn, the matching relationship between the perceived lane lines and the map lane lines can be obtained, and the matching points are used as the final matching point set. Then, the transformation relationship between the two sets of point sets - rotation matrix R and translation matrix T - is obtained through the two-dimensional ICP (Iterative Closest Point, point cloud registration) matching method based on SVD (Singular Value Decomposition), so that the two sets of data meet the optimal matching relationship, thereby realizing the correction of the vehicle positioning information and obtaining the corrected map boundary line, where, Figure 3 The angle that needs to be corrected is 0.01328 radians, the lateral error that needs to be corrected is 0.011 meters, and the confidence level is 0.96846. The device can sample reference points on the visual lane lines (here only two lane lines are considered), and then find the projection point of each reference point on the nearest map boundary line as the source point. At this time, two sets of one-to-one matching points can be obtained (such as Figure 3 The hollow circle and asterisk on the right side) are finally obtained through the ICP algorithm to obtain the relative pose relationship R and T between the two sets of matching points. Since this method samples in the vertical direction, the obtained translation matrix mainly uses the lateral error to correct the positioning. Figure 4 , Figure 4 Schematic diagram of the comparison before and after lane line correction in the vehicle navigation method of this application. Figure 4 The map reference lines shown have been corrected to an angle so that they almost completely coincide with the visual lane lines.

[0083] In this embodiment, the step of matching and correcting the lane line on the map and the visually perceived lane line corresponding to the vehicle includes:

[0084] Step S101: Obtain a first reference point cloud on a visually perceived lane line corresponding to the vehicle and a second reference point cloud on a map lane line.

[0085] It should be understood that the first reference point cloud may be a point set consisting of reference points on visually perceived lane lines; and the second reference point cloud may be a point set consisting of reference points on map lane lines.

[0086] Step S102: constructing a target distance and a minimum function based on the first reference point cloud and the second reference point cloud.

[0087] It should be noted that the above target distance and minimum function may be a function for ensuring that the distance between the first reference point cloud and the second reference point cloud is minimized. In this embodiment, the sum of squared Euclidean distances of all the closest points in the first reference point cloud and the second reference point cloud may be defined as the target function, i.e., the target distance and minimum function.

[0088] Step S103: determining a target rotation matrix and a target translation matrix based on the target distance and the minimum function.

[0089] It should be understood that the above-mentioned target rotation matrix can be a matrix used to ensure that the sum of the point distances between the second reference point cloud and the first reference point cloud after rotation is minimized; the above-mentioned target translation matrix can be a matrix used to ensure that the sum of the point distances between the second reference point cloud and the first reference point cloud after translation is minimized.

[0090] In practical applications, such as Figure 2 As shown, there is an obvious deviation between the map boundary line and the visual lane line, and this deviation is caused by the error of the positioning data. Therefore, in order to improve the accuracy of subsequent vehicle navigation, this embodiment can correct the positioning data. In this embodiment, the device can first sample the reference point set A (i.e., the first reference point cloud mentioned above) on all the lane lines of perception fusion, and then find the projection point of each reference point on the closest map lane line as the source point, which is the reference point set B (i.e., the second reference point cloud mentioned above), and finally match the point sets of A and B through the ICP algorithm to obtain the conversion relationship between the two sets of point sets, specifically:

[0091] The ICP algorithm is essentially an optimal registration method based on the least squares method, that is, repeatedly selecting corresponding point pairs and calculating the optimal rigid body transformation [R, T] until the convergence accuracy requirements of correct registration are met. Its purpose is to find the rotation matrix R and translation matrix T between the point cloud data to be registered and the reference cloud data, so that the two point data meet the optimal match under a certain metric.

[0092] For two point clouds A and B, when the vehicle is positioned, A is the point cloud of perception fusion, and B is the point cloud on the map. Then the sum of the squares of the Euclidean distances of all the nearest points can be defined as the objective function. The main purpose is to ensure that the objective function is minimized, that is, the distance between the points of the two point sets is minimized. is the rotation matrix, is the translation matrix, that is, the distance between the points of point set B and the points of point set A is minimized after the points are rotated and translated. In this embodiment, the target distance and minimum function It can be defined as:

[0093]

[0094] Then we can calculate the target distance and the minimum function To simplify, specifically, first, each point in the A point cloud can be calculated by normalization To their center Deviation , and each point in point cloud B To their center Deviation , we can ignore the translation matrix T at this time:

[0095]

[0096]

[0097] Then the target distance and minimum function are simplified to:

[0098]

[0099] Decomposing the target distance and the minimum function yields:

[0100]

[0101] Even if the function G is the largest:

[0102]

[0103] Then solve the transformation matrix: take the rotation matrix R and the translation matrix T, is the angle to be rotated, and The distance to be translated:

[0104]

[0105] Then the function G:

[0106]

[0107]

[0108] For the function G To find the extreme value, just set its derivative to 0, then:

[0109]

[0110]

[0111]

[0112] Just ask for:

[0113]

[0114] The rotation angle can be calculated , find The rotation matrix R can be obtained, and then the translation matrix T can be obtained based on the rotation matrix R:

[0115]

[0116] Then continue to iterate until the target distance is considered small enough, and finally determine the target rotation matrix and target translation matrix.

[0117] Step S104: matching and correcting the map lane lines and the visually perceived lane lines according to the target rotation matrix and the target translation matrix.

[0118] It can be understood that after determining the target rotation matrix and the target translation matrix, the points in the second reference point cloud can be rotated and translated based on the target rotation matrix and the target translation matrix, thereby achieving matching correction of the map lane lines and the visually perceived lane lines.

[0119] Step S20: Correcting the vehicle positioning information based on the lane line matching information to obtain corrected positioning information.

[0120] It is understandable that the vehicle positioning information may be information such as the vehicle's own vehicle orientation and longitude and latitude. Correspondingly, the corrected positioning information may be information such as the vehicle's own vehicle orientation and longitude and latitude after correction, which is not limited in this embodiment.

[0121] In this embodiment, the map lane line and the visual perception lane line can be corrected by the following formula to obtain the corrected vehicle direction: , and the corrected longitude and latitude and , and recalculate the map reference line in the vehicle coordinate system:

[0122]

[0123]

[0124]

[0125]

[0126]

[0127] Where R is the target rotation matrix, T is the target translation matrix, is the corrected vehicle orientation, is the vehicle orientation before correction, and are the corrected longitude and latitude, and are the latitude and longitude before correction, is the angle to be rotated, and The distance to be translated.

[0128] Step S30: Acquire the perceived lane line valid information and the map lane line valid information based on the driving scenario of the vehicle and the lane line matching information.

[0129] It should be noted that the above-mentioned effective information of the sensed lane line is used to characterize the information of the effectiveness of the sensed lane line. In practical applications, various scenarios may be encountered during vehicle driving, and the effectiveness of the high-precision information and the lane line information obtained by perception may vary in different scenarios. Therefore, this embodiment may obtain the effectiveness of the sensed lane line according to the vehicle driving scenario, and then obtain the effective information of the sensed lane line.

[0130] It should be noted that the above-mentioned map lane line validity information may be information used to characterize the validity of the map lane line. In this embodiment, the device may determine the matching of the four point sets of the four lane lines (left, right, left-left, right-right) perceived and the four lane line point sets of the map in turn according to the lane line matching information, and determine the validity of the map lane line according to the matching situation, thereby obtaining the map lane line validity information. Specifically, if the distance between the point sets of two lane lines is within the threshold range, it is considered that the point sets can be matched, and if there is no lane line at this position in the perception or map, it cannot be matched.

[0131] Step S40: Determine the map centerline based on the corrected positioning information.

[0132] It should be understood that the map center line may be a map reference line in the vehicle coordinate system after the positioning information is corrected. In this embodiment, the map reference line in the vehicle coordinate system after the positioning information is corrected may be calculated based on the corrected positioning information and the high-precision map information.

[0133] Step S50: Determine the perceived center line based on the vehicle driving information and the perceived lane information.

[0134] It should be noted that the above-mentioned perceived center line may be a center line obtained by merging the left and right lane lines of the visually perceived lane lines.

[0135] It is understandable that the above vehicle driving information refers to the scene, time, weather and other information of the vehicle's current driving, such as: highway scene, city scene, rural scene, night, sunny day, etc., and this embodiment does not limit this. The above sensed lane information can be the information of the lane the vehicle is currently driving, such as: whether the lane line is valid, the shape of the lane line, etc., and this embodiment does not limit this.

[0136] In this embodiment, a perception fusion module may be provided in the device. When calculating the perceived center line, the perception fusion module may send information such as the polynomial and effective length of the perceived lane line. At this time, the validity of the left and right lane lines in the visually perceived lane lines may be judged based on this information, and then Kalman filtering may be performed. Finally, the left and right lane lines may be merged to obtain the perceived center line.

[0137] Step S60: Based on the perceived lane line valid information and the map lane line valid information, the map center line and the perceived center line are merged to obtain a road center line, and vehicle navigation is performed through the road center line.

[0138] It should be understood that the road centerline can be a reference centerline for a vehicle to travel. In this embodiment, the device can fuse the map centerline and the sensed centerline based on the valid information of the sensed lane line and the map lane line, calculate the road centerline, and then the vehicle can travel according to the road centerline.

[0139] The present embodiment provides a vehicle navigation method, which discloses matching and correcting a map lane line with a visually perceived lane line corresponding to a vehicle, obtaining lane line matching information between the map lane line and the visually perceived lane line; correcting the vehicle positioning information based on the lane line matching information, obtaining the corrected positioning information; obtaining the perceived lane line valid information and the map lane line valid information based on the vehicle's driving scene and the lane line matching information; determining the map center line based on the corrected positioning information; determining the perceived center line based on the vehicle's driving information and the perceived lane information; fusing the map center line and the perceived center line based on the perceived lane line valid information and the map lane line valid information to obtain the road center line, and performing vehicle navigation through the road center line; compared with the prior art During the operation, when a high-precision map is used to navigate an autonomous vehicle, due to the low accuracy of GPS positioning information, errors will occur when calculating the vehicle's current lane centerline, which in turn leads to low vehicle navigation accuracy. Since the present embodiment can correct the vehicle positioning information based on the lane line matching information between the map lane line and the visually perceived lane line, and determine the map centerline based on the corrected positioning information, and then fuse the map centerline and the perceived centerline based on the effective information of the perceived lane line and the effective information of the map lane line, the road centerline is obtained for vehicle navigation, thereby solving the technical problem in the prior art of using high-precision maps to navigate an autonomous vehicle, due to certain errors in the vehicle positioning information, resulting in low vehicle navigation accuracy.

[0140] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 5 , Figure 5 A flow chart of the second embodiment of the vehicle navigation method of the present application is provided.

[0141] In this embodiment, step S30 further includes steps S301 to S304:

[0142] Step S301: segmenting the visually perceived lane line to obtain a plurality of road segments.

[0143] Step S302: Generate a road topology structure diagram corresponding to the visually perceived lane line based on the connection relationship between the road segments.

[0144] It should be noted that the above-mentioned road topology structure diagram may be a structure diagram for representing the connection relationship between lanes on the road on which the vehicle is traveling.

[0145] In actual applications, the map will divide the road into segments, and then determine the topological connection relationship of the road based on the connection relationship between the lanes in each segment, and construct a road topology diagram. Figure 7 , Figure 7 A road topology diagram in the vehicle navigation method of this application, Figure 7 The map will be divided into segments according to the number of lanes and line type changes. When a lane change or line type change occurs, a segmentation will be performed, and each lane in each segment will be regarded as a node, such as Figure 7 Shown in green. Figure 8 , Figure 8 This is a schematic diagram of road separation and merging in the vehicle navigation method of this application. Figure 8 This is a scene of a road-separated off-ramp. The orange line is divided into four segments according to the road. Each green dot represents a node. When a node is connected to two nodes in the next segment, it means that the road is separated. When two nodes in a segment are connected to the same node in the next segment, it means that the road is merged.

[0146] Step S303: Determine the driving scenario of the vehicle based on the road topology diagram to obtain effective information of perceived lane lines.

[0147] It should be noted that the above driving scenarios may be scenarios encountered by the vehicle during driving, such as ordinary highway / elevated scenarios, scenarios with lane separation / lane merging, and scenarios where the road is renovated, etc., and this embodiment does not limit this. In this embodiment, the device can determine the current driving scenario of the vehicle based on the topological structure diagram of the perceived lane line, and determine the confidence level of the perceived lane line in the current driving scenario, as well as information such as validity, based on the current driving scenario of the vehicle, to obtain valid information of the perceived lane line.

[0148] In actual applications, the confidence of high-precision information and lane line information acquired by perception will be different in different scenarios. In this embodiment, it can be specifically divided into the following scenarios: In ordinary highway / elevated scenarios, the roads are parallel and the effective perception length is far, greater than 80 meters. The lane lines perceived in this scenario will be more accurate; in scenarios with lane separation and lane merging, such as on and off ramps, high-speed switching, etc. According to data analysis, the lane lines perceived in these scenarios are likely to have problems such as missed detection, short effective length, and inability to calculate the drivable center line. Therefore, this embodiment can determine this scenario as a low perception confidence scenario; in the scenario of road renovation, since the collection of high-precision maps is effective, they are generally collected and updated regularly. If the road is repaired during this period, there will be inconsistencies between the perceived lane lines and the map lane lines. Therefore, this embodiment can determine this scenario as an invalid map scenario. This solution can identify the latter two scenarios, among which the lane separation / merging scenario can be judged by the topological information of the high-precision map.

[0149] Step S304: Obtaining valid map lane line information based on the lane line matching information.

[0150] Specifically, the step S304 includes:

[0151] Step S304a: Determine the matching results between the first lane line point sets in the visually perceived lane lines and the second lane line point sets in the map lane lines based on the lane line matching information.

[0152] It can be understood that the first lane line point set and the second lane line point set are point sets composed of points in visually perceived lane lines and map lane lines, respectively.

[0153] Step S304b: Determine whether the map lane line is a valid lane line based on the matching result.

[0154] In actual applications, if the visually perceived lane line and the left and right lane lines in the map lane line can match, for example, the perceived left lane line matches the left lane line of the map, and the perceived right lane line matches the right lane line of the map; or the perceived left-left lane line (right-right lane line) matches the left lane line (right lane line) of the map, and the perceived left lane line (right lane line) matches the right lane line (left lane line) of the map, and the distance between the left lane line (right lane line) of the perceived lane line is very close, the above two situations are both valid map situations, as long as the two adjacent lane lines in the map can match the two adjacent lane lines in the perception. In addition, if the visually perceived lane line and the lane line on one side of the map can match, and the other side cannot match, and the effective length of the perceived lane line is better, then the result of the perceived lane line is more trusted at this time, so the map lane line is considered invalid, and the confidence of the matching result is also reduced. If the visually perceived lane line and the left and right sides of the map lane line cannot match and it is not a scenario where the confidence of the above perception information is reduced, then the map lane line is considered invalid and no matching is performed. If the visually perceived lane lines and the four lane lines in the map lane lines, namely the left lane line, the left-left lane line, the right lane line, and the right-right lane line, can all match, then the map lane lines are valid, and the confidence in the matching result is increased to the maximum.

[0155] Step S304c: Obtain valid map lane line information based on the judgment result.

[0156] In this embodiment, whether the map lane line is valid can be determined based on whether the sensed lane line and the map lane line can be successfully matched, thereby obtaining valid information of the map lane line.

[0157] In this embodiment, it is disclosed that the visually perceived lane lines are segmented to obtain a plurality of road segments; a road topology structure diagram corresponding to the visually perceived lane lines is generated based on the connection relationship between the road segments; the driving scene of the vehicle is determined based on the road topology structure diagram to obtain valid information of the perceived lane lines; and valid information of the map lane lines is obtained based on the lane line matching information, so that the validity of the map lane lines can be accurately determined based on the valid information of the map lane lines, thereby improving the calculation accuracy of the subsequent road center lines, so that the vehicle navigation accuracy can be further improved.

[0158] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the above embodiments can be referred to the above introduction, and will not be described in detail later. Figure 6 , Figure 6 A flowchart diagram of the third embodiment of the vehicle navigation method of the present application is provided.

[0159] In this embodiment, the step S40 includes steps S401 to S403:

[0160] Step S401: Determine corrected coordinate information according to the corrected positioning information.

[0161] It is understandable that the above-mentioned corrected coordinate information may be the latitude and longitude coordinate information in the corrected positioning information. In this embodiment, after obtaining the corrected positioning information, the corrected longitude and latitude information may be determined according to the corrected positioning information to obtain the corrected coordinate information.

[0162] Step S402: performing error state Kalman filtering on the corrected coordinate information based on the vehicle body inertial measurement data to obtain filtered coordinate information.

[0163] It should be noted that the above-mentioned vehicle body inertial measurement data may be data provided by the IMU (Inertial Measurement Unit) in the vehicle. In practical applications, the IMU is mainly used on the vehicle to measure the attitude, dynamic motion and navigation positioning of the vehicle. It can provide information about the vehicle attitude (roll angle, pitch angle, yaw angle) and dynamic motion (acceleration, deceleration, turning, etc.) by measuring the three-axis angular velocity and three-axis acceleration of the object. In this embodiment, since the coordinates after direct correction will be jagged and not smooth, the corrected coordinates can be subjected to an Error-State Kalman Filter (ESKF) to eliminate errors, smooth the results, and finally obtain the filtered coordinate information.

[0164] In practical applications, the model system of ESKF describes the system error rather than directly describing the system state. The system state needs to be obtained through conversion, which belongs to indirect filtering. The Error-State value of ESKF (the value that represents the difference between the actual state and the expected state) is generally close to 0, which can avoid some singularities that may appear in the first-order part. It can be applied to the universal lock problem of inertial navigation systems and provide linear validity guarantees at all times. Moreover, the Error-State is very small, which can ensure that the quadratic part in the Taylor expansion is negligible, making the calculation of the Jacobian simple and fast. Since the dynamic system of the Error-State is usually slow, all larger components are concentrated to calculate the nominal state, making the update process of the Kalman filter less frequent than the prediction process.

[0165] In the specific implementation, the implementation process of ESKF filtering can mainly include the steps of constructing a true state model, constructing a nominal state model, estimating the error state, constructing a measurement model, filtering update and correcting the nominal state. In this embodiment, the mathematical form of ESKF can be elaborated in detail by taking the fusion of matching positioning as an example, and the specific implementation process is as follows:

[0166] 1) True value state model

[0167] For the true state model, this step is mainly to determine the true value relationship of the information of each sensor of the vehicle:

[0168]

[0169]

[0170]

[0171]

[0172]

[0173] For the true value of IMU and wheel speed meter , , which is consistent with the measured value The relationship between them is:

[0174]

[0175]

[0176] in, is the true value of the east position, is the eastward position change, i.e., the eastward velocity , is the true value of the north position, is the northward position change, i.e., northward velocity , is the speed measurement value, is the true value of the heading angle, is the angular velocity , is the angular velocity disturbance of IMUz axis, is the wheel speed meter disturbance, is the bias of the angular velocity meter, is the offset of the wheel speed meter, is the noise measured by the wheel speed meter and IMU, is the disturbance measured by the wheel speed meter and IMU, is the angular velocity measurement, is the angular velocity noise, is the speed value measured by the car's wheel speedometer, is the disturbance of the wheel speed meter.

[0177] 2) Nominal state model:

[0178]

[0179] Compared with the true value, the nominal state has no noise and disturbance. The calculation formula of the nominal state is as follows:

[0180]

[0181]

[0182]

[0183]

[0184]

[0185] In the formula, is the eastward position change, is the change in north position, is the angle change, is the angular velocity disturbance of IMUz axis, is the wheel speed meter disturbance, is the speed measurement value, where , is the speed value measured by the car's wheel speedometer, is the calculated velocity bias, is the Y-axis angular velocity measured by the IMU, is the calculated angular velocity bias. This formula can be used to infer the new position from the previous frame calculation result and the sensor measurement value, that is, ,in, is the result of the previous frame calculation, For position changes, The new position is calculated by the calculation results of the previous frame and the sensor measurement values.

[0186] 3) Error state estimation:

[0187] The dynamics of a discrete system usually satisfies the above formula, which can be written as the state equation:

[0188] , , ,

[0189] In the formula, represents the nominal state variable, represents the error state variable, Represents the input parameters, represents the noise term of the disturbance, where For the east position, For the north position, is the angle, is the angular velocity bias, is the speed bias, is the angular velocity measurement, is the speed measurement value, is the disturbance of the wheel speed meter, is the angular velocity noise, is the noise measured by the wheel speed meter and IMU, are the disturbances measured by the wheel speed meter and IMU.

[0190] Among them, the motion equation of the error state can be:

[0191]

[0192] At this time, the forward error covariance P is:

[0193]

[0194] In the formula, , represent the Jacobian and disturbance of the state variable, respectively. Represents a noise matrix:

[0195]

[0196]

[0197]

[0198] in:

[0199]

[0200]

[0201]

[0202] The error covariance can be calculated by the above equation and error status Update to get .

[0203] 4) Measurement model

[0204] The observed value used in this embodiment can be the difference between the position and yaw angle given by the map and after correction and the predicted value, and the formula for measurement update can be:

[0205]

[0206] In the formula, is the observed data, is the observation noise, is the covariance matrix of the noise.

[0207] Among them, the calculation formula of the Jacobian matrix H of the observation equation relative to the error state can be:

[0208]

[0209] 5) Filter update

[0210] Kalman Gain The calculation formula is:

[0211]

[0212] Where P is the error covariance and H is the Jacobian matrix of the observation equation relative to the error state.

[0213] At this time, the observation update state error state can be:

[0214]

[0215] in, is the error state, is the updated error state, is the position and heading angle obtained by GPS, It is the result of forward calculation through the corrected position and heading angle of the previous frame.

[0216] in:

[0217]

[0218] In the formula, Indicates GPS .

[0219] 6) Correction of nominal status

[0220] The formula for correcting the nominal state in this embodiment can be:

[0221]

[0222] In the formula, is the latitude and longitude of the final filter, is the longitude and latitude before filtering, is the error value of the calculated longitude and latitude.

[0223] Step S403: Determine the center line of the map based on the filtered coordinate information.

[0224] In practical applications, the device can obtain the corrected data after correcting the positioning information before correction through the rotation matrix R and the translation matrix T, but the corrected data has relatively large burrs, so the corrected positioning needs to be filtered. In this embodiment, the new positioning position can be calculated as the predicted value by the vehicle speed measured by the wheel speed meter and the yaw rate measured by the IMU, and the corrected GPS measurement positioning is used as the observed value. Among them, the positioning error value is a state quantity, which can update the error value through the state transfer matrix, and calculate the covariance of the error by measuring the error and noise. Then the Kalman gain is calculated by the covariance, the measurement matrix and the measurement noise, where the validity of the measurement value and the observation noise will be affected by the map validity and matching results. Finally, the error state is updated by the measurement value and the Kalman gain, and the error is finally added to the predicted value calculated by the sensor to obtain the final positioning data, that is, the above-mentioned filtered coordinate information, and then the map lane line can be calculated according to the filtered coordinate information.

[0225] Furthermore, the step S50 includes steps S501 to S505:

[0226] Step S501: Determine the time score, weather score, scene score and camera score of the vehicle during the current driving process based on the vehicle driving information and the perceived lane information.

[0227] It should be understood that the above-mentioned time score can be a score used to characterize the state of time during the current driving process of the vehicle, for example, the night score is 0, the strong light score is 40, the other scores are 100, etc.; the above-mentioned weather score can be a score used to characterize the weather state during the driving process of the vehicle, for example, the sunny score is 100, the cloudy score is 70, the heavy rain score is 20, the rainstorm score is 10, etc.; the above-mentioned scene score can be a score used to characterize the state of the scene during the driving process of the vehicle, for example, the high-speed scene score is 100, the urban scene score is 50, and the rural scene score is 20; the above-mentioned camera score can be a score used to characterize the state of the camera during the driving process of the vehicle, for example, the full score is 100 points, and there are occlusion, glare and other situations and the level of the situation, such as minor, serious, etc. If a serious situation occurs, 20 points will be deducted, and if a minor situation occurs, 10 points will be deducted, which can be calculated after receiving the camera status.

[0228] Step S502: Determine the confidence scores corresponding to the left and right lane lines in the visually perceived lane lines based on the time score, the weather score, the scene score, and the camera score.

[0229] It should be noted that the above confidence score can be a score used to characterize the effectiveness of the left and right lane lines in the visual perception lane lines. In this embodiment, the confidence of the left and right lane lines can be determined by the scene, time, weather, and camera status. A confidence score greater than 50 is effective. The confidence score calculation formula is:

[0230]

[0231] in, is the confidence score, Score the scene, with 100 for high-speed scenes, 50 for urban scenes, and 20 for rural scenes; For time score, 0 for darkness, 40 for bright light, and 100 for other times; Score the weather, with sunny days scoring 100, cloudy days scoring 70, heavy rain scoring 20, and rainstorms scoring 10; The camera status score has a full score of 100 points, which includes occlusion, glare and other situations, as well as the level of the situation, such as minor, severe, etc. If a severe situation occurs, 20 points will be deducted, and if a minor situation occurs, 10 points will be deducted.

[0232] Step S503: When the confidence score exceeds a preset confidence score threshold, determine whether the lane line widths of the left and right lane lines exceed a preset width threshold.

[0233] It is understandable that the above-mentioned preset confidence score threshold can be the minimum score used to characterize the validity of the left and right lane lines. In this embodiment, the preset confidence score threshold can be set to 50. If the confidence score corresponding to the left and right lane lines in the visually perceived lane lines is greater than 50, it means that the left and right lane lines are valid. At this time, it can continue to determine whether the lane line width of the lane line exceeds the preset width threshold. In practical applications, in this embodiment, the preset width threshold can be set to 5.5 meters. If the width of the left and right lane lines is greater than 5.5 meters, the lane line farther away from the vehicle is invalid at this time, otherwise the lane line is valid.

[0234] Step S504: If not, determine whether the left and right lane lines are in a parallel state, and obtain the validity determination result of the left and right lane lines according to the determination result.

[0235] In this embodiment, if the lane width of the left and right lane lines does not exceed the preset width threshold, it is possible to continue to determine whether the left and right lane lines are parallel. Specifically, the device can determine the distance between the left and right lane lines at 0 meters and 50 meters. If the width difference is greater than 0.7 meters, it is considered that the two lane lines are not parallel, and the width of 50 meters is greater than that of 0 meters. The shape of the left and right lane lines is an open trumpet shape (often appears when 3 lanes become 4 lanes), and then the curvature variance of the left and right lane lines is calculated. A larger variance indicates that the line is more curved. In order to allow the car to continue to drive along a certain lane line, the more curved line is considered invalid, and the straighter line is valid. According to the above judgment method, the validity judgment result of the left and right lane lines can be obtained.

[0236] Step S505: Determine the perceived center line based on the validity judgment result and the coordinate information of the left and right lane lines.

[0237] In this embodiment, after determining the validity of the lane line, the lane line of this frame and the lane line of the previous frame may be filtered to filter the point traces of the lane line:

[0238]

[0239]

[0240]

[0241]

[0242] in, To pass The predicted value of coordinate transformation is calculated as follows:

[0243]

[0244] in, , , They are the position and angle change values ​​of this frame and the previous frame sent by the perception fusion module; is the Kalman gain, is the updated process covariance, is the covariance of the previous frame process, is the process noise, calculated based on the calibration quantity, generally The bigger the noise, the louder the noise; for of Frame covariance, is the measured noise, For Frame the coordinate points on the new lane line , for The final coordinates in the state.

[0245] Through the above steps, Kalman filter calculation is performed on each point of the lane line to obtain the calculation of the perceived center line. If the lane line of this frame is invalid, it will not be updated and only the coordinate conversion will be performed; if the lane line has been invalid for ten consecutive frames, the lane line will not be calculated.

[0246] After calculating the left and right lane lines, you can use The coordinate point of the center line is calculated by . If only one lane line is valid, the lane center line is calculated by translating the lane line by half the lane width, and finally the perceived center line is determined. The confidence level of the center line can be divided into three levels: high (both left and right lane lines are valid), medium (one-side lane line is valid), and low (both left and right lane lines are invalid).

[0247] Furthermore, the step S60 includes steps S601 to S604:

[0248] Step S601: Determine the confidence level corresponding to the perceived center line according to the validity judgment result.

[0249] It can be understood that the above confidence levels can be values ​​used to characterize the effectiveness of the perceived center line. In this embodiment, the perceived center line can include three levels: high (both left and right lane lines are valid), medium (single-side lane line is valid), and low (neither left nor right lane lines are valid).

[0250] Step S602: Perform low-confidence scene judgment on the visually perceived lane line according to the perceived lane line valid information to obtain a road scene recognition result.

[0251] In this embodiment, the device can determine whether the vehicle is currently in a low-confidence scenario based on the perceived lane line information, thereby obtaining a road scene recognition result. Specifically, if the vehicle is currently in a scenario where lane separation / lane merging exists based on the effective information of the perceived lane line, since in the scenario where lane separation / lane merging exists, the perceived lane line is likely to be missed, the effective length is short, and the drivable center line cannot be calculated, the confidence is low, so it can be determined that the current vehicle is in a low-confidence scenario, and the road scene recognition result is obtained at this time.

[0252] Step S603: judging the validity of the map lane line according to the map lane line validity information to obtain a map validity recognition result.

[0253] Step S604: Based on the confidence level, the road scene recognition result and the map validity recognition result, the map centerline and the perception centerline are fused to obtain a road centerline.

[0254] It should be noted that in this embodiment, the lane centerline can be calculated by sensing the lane line, and the lane centerline can also be calculated through filtered positioning and high-precision map information. Therefore, in order to improve the accuracy of the lane centerline, this embodiment can calculate the factors of the two separately through scene information and matching results, and then calculate the final lane centerline through fusion filtering.

[0255] In this embodiment, the calculation formula of the road centerline can be:

[0256]

[0257]

[0258] Where y is the y coordinate value of the point trace on the center line of the road; The confidence level of the perceived center line ranges from 0 to 1. The larger the value, the more confident the center line of the map is. is 0, 0.5 when high, and 1 when low; Indicates the scenario where the confidence level of the perceived lane line is low. If this is the scenario, this value is 1, otherwise it is 0; Whether the map is valid, 1 if valid, 0 if invalid; is the effective length of the perceived lane line; For dotted Coordinate value; The center line of the high-precision map Axis coordinate values; The y-axis coordinate value of the perception center line.

[0259] In the specific implementation, refer to Fig. 9 , Fig. 9 This is a schematic diagram of the process of generating the lane centerline in the vehicle navigation method of this application. Fig. 9As shown, first, the pre-processing module in the device can calculate the map boundary line in the self-vehicle coordinate system before correction based on the vehicle's positioning information and high-precision map information. Then, the ICP correction module can sample the map boundary line and the visual lane line, and obtain the transformation matrix of the two sets of sampling points by calculation, correct the positioning data, and then judge the current scene based on the map information, such as lane separation and lane merging, and combine the current matching results to judge the validity of the map lane line and the perceived lane line. Finally, the vehicle body IMU data is used as a priori to perform error state Kalman filtering on the positioning data to obtain the final corrected positioning information. Then, the map reference line in the self-vehicle coordinate system after correction can be recalculated based on the corrected positioning information and high-precision map information to obtain the map centerline. Finally, the map centerline and the perceived centerline can be fused and filtered to obtain the road centerline, so that the vehicle can travel according to the road centerline. In this embodiment, the example diagram of road centerline fusion can refer to Fig.10 , Fig.10 The red dashed line in the figure represents the perceived centerline, the blue dashed line represents the map centerline, and the circle represents the road centerline obtained after fusion.

[0260] In this embodiment, it is disclosed that corrected coordinate information is determined based on the corrected positioning information; error state Kalman filtering is performed on the corrected coordinate information based on the vehicle's body inertial measurement data to obtain filtered coordinate information; and the map center line is determined based on the filtered coordinate information. Since this embodiment can perform an error state Kalman filtering on the corrected coordinate information to eliminate errors, the filtered coordinate information finally obtained is more accurate, so that the map center line can be accurately determined, thereby improving the navigation accuracy of subsequent vehicles.

[0261] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the vehicle navigation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0262] This application also provides a vehicle navigation device, please refer to Fig.11 , the vehicle navigation device comprises:

[0263] A lane line matching module 10 is used to match and correct the map lane line and the visually perceived lane line corresponding to the vehicle, and obtain lane line matching information between the map lane line and the visually perceived lane line;

[0264] A positioning information correction module 20, used to correct the vehicle positioning information based on the lane line matching information to obtain corrected positioning information;

[0265] An effective information acquisition module 30, for acquiring effective information of perceived lane lines and effective information of map lane lines based on the driving scene of the vehicle and the lane line matching information;

[0266] A centerline determination module 40, configured to determine a map centerline based on the corrected positioning information;

[0267] The centerline determination module 40 is further used to determine the perceived centerline based on the vehicle driving information and the perceived lane information;

[0268] The vehicle navigation module 50 is used to fuse the map center line and the perceived center line based on the perceived lane line valid information and the map lane line valid information to obtain a road center line, and perform vehicle navigation through the road center line.

[0269] The vehicle navigation device provided by the present application adopts the vehicle navigation method in the above embodiment, which can solve the technical problem in the prior art that the vehicle navigation accuracy is not high due to certain errors in the vehicle positioning information when navigating the autonomous driving vehicle using a high-precision map. Compared with the prior art, the beneficial effects of the vehicle navigation device provided by the present application are the same as the beneficial effects of the vehicle navigation method provided by the above embodiment, and the other technical features in the vehicle navigation device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0270] The present application provides a vehicle navigation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the vehicle navigation method in the above-mentioned embodiment 1.

[0271] Reference below Fig.12 , which shows a schematic diagram of the structure of a vehicle navigation device suitable for implementing an embodiment of the present application. The vehicle navigation device in the embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.12 The vehicle navigation device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0272] like Fig.12As shown, the vehicle navigation device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the vehicle navigation device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the vehicle navigation device to communicate wirelessly or wired with other devices to exchange data. Although the vehicle navigation device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0273] The vehicle navigation device provided by the present application adopts the vehicle navigation method in the above embodiment to solve the technical problem of vehicle navigation. Compared with the prior art, the beneficial effects of the vehicle navigation device provided by the present application are the same as the beneficial effects of the vehicle navigation method provided by the above embodiment, and other technical features in the vehicle navigation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0274] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0275] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the vehicle navigation method in the above-mentioned embodiment.

[0276] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned vehicle navigation method, and can solve the technical problem in the prior art that when using high-precision maps to navigate autonomous driving vehicles, the vehicle positioning information has certain errors, resulting in low vehicle navigation accuracy. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the vehicle navigation method provided in the above-mentioned embodiment, and will not be repeated here.

[0277] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A vehicle navigation method, characterized in that: The method includes: Matching and correcting the lane line on the map and the visually perceived lane line corresponding to the vehicle to obtain lane line matching information between the lane line on the map and the visually perceived lane line; Correcting the vehicle positioning information based on the lane line matching information to obtain corrected positioning information; Acquiring sensed lane line valid information and map lane line valid information based on the driving scenario of the vehicle and the lane line matching information; Determining a map centerline based on the corrected positioning information; Determine a sensed center line based on vehicle travel information and sensed lane information; Based on the sensed lane line valid information and the map lane line valid information, the map center line and the sensed center line are merged to obtain a road center line, and vehicle navigation is performed through the road center line; The step of determining the sensed center line based on the vehicle driving information and the sensed lane information comprises: Determine a time score, a weather score, a scene score, and a camera score of the vehicle during a current driving process based on the vehicle driving information and the perceived lane information; Determine confidence scores corresponding to left and right lane lines in the visually perceived lane line based on the time score, the weather score, the scene score, and the camera score; When the confidence score exceeds a preset confidence score threshold, determining whether the lane line widths of the left and right lane lines exceed a preset width threshold; If not, determining whether the left and right lane lines are in a parallel state, and obtaining a validity determination result of the left and right lane lines according to the determination result; A sensed center line is determined based on the validity judgment result and the coordinate information of the left and right lane lines.

2. The method according to claim 1, characterized in that The step of matching and correcting the lane line on the map and the visually perceived lane line corresponding to the vehicle includes: Obtain a first reference point cloud on a visual perception lane line corresponding to the vehicle and a second reference point cloud on a map lane line; Constructing a target distance and a minimum function based on the first reference point cloud and the second reference point cloud; Determine a target rotation matrix and a target translation matrix based on the target distance and the minimum function; The map lane lines and the visually perceived lane lines are matched and corrected according to the target rotation matrix and the target translation matrix.

3. The method according to claim 1, characterized in that The step of acquiring the sensed lane line valid information and the map lane line valid information based on the driving scene of the vehicle and the lane line matching information comprises: Segmenting the visually perceived lane line to obtain a plurality of road segments; Generating a road topology structure diagram corresponding to the visually perceived lane line based on the connection relationship between the road segments; Determining a driving scene of the vehicle based on the road topology diagram to obtain effective information of perceived lane lines; The valid information of the lane lines on the map is obtained based on the lane line matching information.

4. The method according to claim 3, characterized in that The step of obtaining valid information of map lane lines based on the lane line matching information includes: Determining matching results between a plurality of first lane line point sets in the visually perceived lane lines and a plurality of second lane line point sets in the map lane lines based on the lane line matching information; Determine whether the lane line on the map is a valid lane line according to the matching result; Obtain valid map lane line information based on the judgment results.

5. The method according to claim 1, characterized in that The step of determining the center line of the map based on the corrected positioning information comprises: Determining corrected coordinate information according to the corrected positioning information; Performing error state Kalman filtering on the corrected coordinate information based on the vehicle body inertial measurement data to obtain filtered coordinate information; A map centerline is determined based on the filtered coordinate information.

6. The method according to claim 1, characterized in that The step of fusing the map center line and the sensed center line based on the sensed lane line valid information and the map lane line valid information to obtain the road center line comprises: Determining a confidence level corresponding to the sensed center line according to the validity judgment result; Performing low-confidence scene judgment on the visually perceived lane line according to the effective information of the perceived lane line to obtain a road scene recognition result; Judging the validity of the lane line on the map according to the valid information of the lane line on the map, and obtaining a map validity recognition result; The map centerline and the perceived centerline are fused based on the confidence level, the road scene recognition result, and the map validity recognition result to obtain a road centerline.

7. A vehicle navigation device, characterized in that: The device comprises: A lane line matching module is used to match and correct the map lane line and the visually perceived lane line corresponding to the vehicle, and obtain lane line matching information between the map lane line and the visually perceived lane line; A positioning information correction module, used to correct the vehicle positioning information based on the lane line matching information to obtain corrected positioning information; An effective information acquisition module, used to acquire the effective information of the perceived lane line and the effective information of the map lane line based on the driving scene of the vehicle and the lane line matching information; A centerline determination module, configured to determine a map centerline based on the corrected positioning information; The centerline determination module is further used to determine the perceived centerline based on the vehicle driving information and the perceived lane information; A vehicle navigation module, configured to fuse the map centerline and the sensed centerline based on the sensed lane line valid information and the map lane line valid information, obtain a road centerline, and perform vehicle navigation through the road centerline; The center line determination module is also used to determine the time score, weather score, scene score and camera score of the vehicle in the current driving process based on the vehicle driving information and the perceived lane information; determine the confidence score corresponding to the left and right lane lines in the visually perceived lane lines based on the time score, the weather score, the scene score and the camera score; when the confidence score exceeds a preset confidence score threshold, determine whether the lane line width of the left and right lane lines exceeds a preset width threshold; if not, determine whether the left and right lane lines are in a parallel state, and obtain the validity judgment result of the left and right lane lines based on the judgment result; determine the perceived center line based on the validity judgment result and the coordinate information of the left and right lane lines.

8. A vehicle navigation device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vehicle navigation method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the vehicle navigation method according to any one of claims 1 to 6 are implemented.

Images