Map path matching method and device, electronic device, and storage medium

Abstract

The present invention relates to a map path matching method and apparatus, an electronic device, and a storage medium. The map path matching method includes: determining a navigation path based on a navigation map, where the navigation path is an ordered sequence of line points including a plurality of line points; determining a set of alternative points for each of the line points by using a high-precision map; performing a decentralization operation on the navigation path and the set of alternative points to obtain a decentralized set of alternative points; and performing an optimal matching of the decentralized set of alternative points with the corresponding line points to obtain a high-precision map anchor point sequence.

Description

Technical Field

[0001] The present invention relates to navigation technology, and in particular to navigation using high-precision maps. Background Art

[0002] Autonomous driving technology usually regards high-precision maps (hereinafter referred to as HD maps) as prior knowledge of the world, thereby assisting in making reasonable long-distance planning and decisions. People usually use navigation maps (hereinafter referred to as SD maps) to predict the direction of the road and make reasonable advance actions.

[0003] In recent years, the combination of SD maps and HD maps to provide navigation services to users has become the mainstream business of autonomous driving. However, when the sources of SD maps and HD maps are different, that is, when the roads on the SD map and the roads on the HD map do not have a one-to-one matching relationship, the navigation path on the SD map cannot be directly converted into the HD map path that the autonomous driving system can use. This problem needs to be solved urgently. Summary of the invention

[0004] The present invention is made in view of the above situation of the prior art, to overcome or alleviate one or more problems of the prior art, and at least provide a beneficial choice.

[0005] According to one aspect of the present invention, a map path matching method is provided, comprising: determining a navigation path based on a navigation map, the navigation path being an ordered line point sequence including a plurality of line points; determining a set of candidate points for each of the line points using a high-precision map; performing a decentralized operation on the navigation path and the set of candidate points to obtain a decentralized set of candidate points; and performing optimal matching of the decentralized set of candidate points with corresponding line points to obtain a high-precision map anchor point sequence.

[0006] According to another aspect of the present invention, a map path matching device is provided, comprising: a navigation path determination unit, which determines a navigation path based on a navigation map, wherein the navigation path is an ordered line point sequence including a plurality of line points; an alternative point set determination unit, which determines an alternative point set for each of the line points using a high-precision map; a decentralization unit, which performs a decentralization operation on the navigation path and the alternative point set to obtain a decentralized alternative point set; and a matching unit, which performs optimal matching on the decentralized alternative point set with corresponding line points to obtain a high-precision map anchor point sequence.

[0007] According to an embodiment, each of the line points has one or more attributes selected from the following combination: heading angle, road name and / or lane number, and whether the uplink and downlink are separated.

[0008] According to one embodiment, the alternative point set determination unit sets weights for the alternative points of each row point according to the attribute, and the matching unit performs the best match between the decentralized alternative point set and the corresponding row point according to the weights.

[0009] According to one embodiment, the decentralization unit performs decentralization as follows: Assume that the alternative point set of a certain row point is P = {p1, p2, p3, …, p n}, and its center point is

[0010]

[0011] Then the decentralized alternative point set is

[0012]

[0013] According to one embodiment, the alternative point set determination unit uses the points on the road center line or lane center line within a certain range around each row point as alternative points.

[0014] According to one embodiment, the matching unit performs the best match between the decentralized alternative point set and the corresponding row point as follows: For each row point in the decentralized navigation path, determine the nearest point in its corresponding decentralized alternative point set; obtain the translation parameter and rotation parameter according to the rigid body transformation with the minimum average distance calculated by the point pair; use the translation parameter and rotation parameter to perform rotation and translation on the decentralized navigation path to obtain a new point set after rotation and translation; for each row point on the new point set, determine the nearest point in its corresponding decentralized alternative point set, and obtain the high-precision map anchor point sequence with the set of these nearest points.

[0015] According to one embodiment, when the matching unit performs the best match between the decentralized alternative point set and the corresponding row point, it uses the iterative closest point (ICP) algorithm with a scale scaling factor.

[0016] According to one embodiment, it further includes a path matching unit, which obtains the path on the high-precision map according to the high-precision map anchor point sequence;

[0017] According to one embodiment, it further includes a storage unit, which saves the correspondence between the navigation path and the high-precision map path, and the alternative point set determination unit uses the correspondence to determine the alternative point set of each row point.

[0018] According to another aspect of the present invention, there is also provided an electronic device, including: a processor; a memory for storing the executable instructions of the processor; wherein, the processor is configured to execute the instructions to implement the method of the present invention.

[0019] According to another aspect of the present invention, there is also provided a computer-readable storage medium, characterized in that a device control program is stored on the readable storage medium, and when the device control program is executed by a processor, the method of the present invention is implemented.

[0020] According to the technical solution of the present invention, the navigation map and the high-precision map can be effectively utilized to provide more accurate navigation information. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The embodiments of the present invention can be better described in conjunction with the accompanying drawings. The accompanying drawings are only schematic, not drawn to scale, and are not a limitation on the protection scope of the present invention. In the drawings,

[0022] Figure 1 is a schematic flow chart showing a map path matching method according to an embodiment of the present invention.

[0023] Figure 2 is a schematic diagram showing a road center line and a lane center line.

[0024] Figure 3 is a schematic block diagram showing a map path matching device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] The following describes in detail the specific embodiments of the present invention in conjunction with the accompanying drawings. These embodiments are all exemplary and not a limitation on the protection scope of the present invention.

[0026] Figure 1 is a schematic flow chart showing a map path matching method according to an embodiment of the present invention.

[0027] As Figure 1 shown, first in step S100, according to the starting point and the ending point given by the user, based on the navigation map, the navigation system performs path calculation to obtain a navigation map (SD) path S. This process can be performed locally or in the cloud. The navigation map path S is actually an ordered sequence of line points {s1, s2, s3, …, s n}. Optionally, the navigation map path S can also be divided by road ID or other means. The attributes carried by each line point include but are not limited to GPS position, heading angle, road name, length, number of lanes, whether the up and down directions are separated, etc.

[0028] Subsequently, in step S200, using a high-precision map, a set of candidate points for each row point in the navigation map path S is determined. According to one embodiment, since the matching process may use the entire high-precision map and the data volume is large, the navigation system will initiate a matching request to upload the navigation map path S to the matching server, and the cloud matching server will perform the matching, including the determination of the set of candidate points here and the decentralized and ICP matching described below. Of course, when the processing power is high, the high-precision map can also be used locally, and all operations are performed locally. The high-precision map (HD) can be regarded as a binary set {L, K}, where the road segment set L = {l1, l2, l3, …, l n} contains several road segments, and each road segment l i = {p1, p2, p3, …, p n} contains several points, and K is the set of attributes and relationships within the road segment L, such as the specified number of lanes, whether the up and down directions are separated, the heading angle, the speed limit, etc.

[0029] According to one embodiment, according to the GPS positions of each row point Q, the HD map engine is called to query the points within a certain range r (determined according to the absolute accuracy difference between the SD map and the HD map) around each GPS position and add them to the set of candidate points T. When the matching relationship between the road names on the SD path and the road IDs on the HD map is unknown, the GPS position can be used, but when the matching relationship between the high-definition map road numbers and the navigation map road names is obtained in subsequent steps, the matching relationship can be directly used for searching. For example, if it is known that HD road 10092 corresponds to Yunjin Road, then the candidate points including Yunjin Road are selected from road 10092.

[0030] According to one embodiment, these candidate points should be points on the center line of the lane or the center line of the road. Figure 2 The schematic diagrams of the center line of the road and the center line of the lane are shown.

[0031] According to one embodiment, a screening step is further included. Using the attributes of the row points (i.e., the GPS position, heading angle, road name, length, number of lanes, whether the up and down directions are separated, etc. mentioned above), the obtained candidate points are screened. For example, the road name information of the row points can be used to retain only the candidate points on the road corresponding to the road name in the high-precision map. There may be no road name information on the high-precision map, but only road number information. In this case, the association relationship between the road number information on the high-precision map and the road name information on the navigation map can be established, so that this screening can be conveniently completed. For example, in the above example, if it is known that HD road 10092 corresponds to Yunjin Road, then the candidate points including Yunjin Road are selected from road 10092.

[0032] According to an embodiment, weights are set for the alternative points of each of the row points according to attributes. Due to different precisions, for some attribute information, such as the number of lanes, whether the up and down lanes are separated, etc., the two maps may not be exactly the same. If alternative points are removed only based on these attributes, the matching precision may be reduced. Therefore, different weight values can be given to the alternative points according to these attributes. A larger weight value for an alternative point means that the point is more similar to the row point.

[0033] After that, in step S300, a centering operation is performed on the navigation map path S and each set of alternative points to obtain a centered set of alternative points.

[0034] Assume that the navigation map path S has an ordered point set P = {p1, p2, p3, …, p n}, and its center point is:

[0035]

[0036] Perform a centering operation on the ordered point set P, which is

[0037]

[0038] The centering operation on each set of alternative points can refer to the centering operation on the navigation map path S.

[0039] The centering operation can eliminate too large coordinate system amplitudes for the subsequent ICP matching, and can improve the matching precision.

[0040] Then, in step S400, the best match between the centered set of alternative points (centered set of alternative points) and the corresponding row points is performed to obtain a high-precision map anchor point sequence.

[0041] According to an embodiment, this best match is completed using ICP matching (point cloud registration method). According to an embodiment, it is performed as follows:

[0042] a) For each row point in the centered navigation path S', determine the nearest alternative point in its corresponding centered set of alternative points.

[0043] According to an embodiment, optionally, when the sampling process includes weight values, weight considerations can be added to improve accuracy. The formula is as follows:

[0044] D = wd

[0045] d is the original Euclidean distance between the row point and the alternative point, w is the weight value, ranging from 0 to 1, and D is the weighted distance value.

[0046] b) Calculate the rigid body transformation with the minimum average distance based on point pairs to obtain the translation parameters and rotation parameters. A rigid body transformation means that the distance between two points remains unchanged before and after the transformation.

[0047] The point pairs can be determined according to the principle of proximity. For example, it can be carried out as described in step a). The following distance difference function can be used and solved by the SVD decomposition method to obtain the translation parameter t and the rotation parameter R.

[0048]

[0049] q in the formula i and p i are the corresponding point pairs.

[0050] c) Use the translation parameters and rotation parameters to perform rotation and translation on the decentralized navigation path S', and obtain a new point set after rotation and translation. The rotation and translation can be carried out by various methods known now or in the future. For example, the points in the decentralized navigation path S' are rotated and translated in the following way

[0051] p' i = Rp i + t, p ∈ S'

[0052] d) For each row of points in the new point set, determine the nearest points in their corresponding decentralized alternative point sets, and form a set of these nearest points into the high-precision map anchor point sequence A. That is to say, the path on the high-definition map must pass through these anchor points. Note that since the points in S' are ordered, the points in the anchor point sequence A still remain ordered.

[0053] According to one embodiment, considering that there may be a slight difference in scale between the SD map and the HD map, that is, although S' and T' are similar in shape, there is a scaling relationship in the road length, so the above ICP algorithm can also be replaced by an ICP algorithm with a scale scaling factor.

[0054] According to one embodiment, before step d), there is also a distance judgment step e). If the average distance between the corresponding points of the new point set S' and T' is less than a certain threshold, stop; if it is still greater than the threshold, return to step a). Preferably, the number of iterations is also determined: if the current number of iterations is greater than the maximum number of iterations, stop the loop of steps a), b), c), and e).

[0055] Finally, in step S500, path matching is performed. In step S500, for the ordered anchor point sequence A, continue to perform a constrained (i.e., must pass through the anchor point sequence A) path finding algorithm on the HD map. Common A* algorithms etc. can be used to obtain the path S on the high-precision map hd, thus completing the matching with path S.

[0056] According to one embodiment, it may include the step of saving the correspondence R between the navigation path S and the high-precision map path. Storing this correspondence R in a database can greatly accelerate the subsequent path matching or the selection of alternative points in step S200, and these relationships can be gradually established during actual operation, accurately and avoiding a large amount of manual work.

[0057] Figure 3 is a schematic block diagram showing a map path matching device according to an embodiment of the present invention. As Figure 3 shown, the map path matching device according to an embodiment of the present invention includes: a navigation path determination unit 100 for determining a navigation path based on a navigation map, the navigation path being an ordered sequence of line points including a plurality of line points; an alternative point set determination unit 200 for determining an alternative point set for each of the line points by using a high-precision map; a decentralization unit 300 for performing a decentralization operation on the navigation path and the alternative point set to obtain a decentralized alternative point set; an anchor point sequence acquisition unit 400 for performing an optimal match of the decentralized alternative point set with the corresponding line points to obtain a high-precision map anchor point sequence; and a path matching unit 500 for obtaining a path on the high-precision map according to the high-precision map anchor point sequence. Preferably, the matching unit can also save the correspondence between the navigation path and the high-precision map path.

[0058] According to one embodiment, the alternative point set determination unit 200 sets weights for the alternative points of each line point according to the attributes of each line point, and the anchor point sequence acquisition unit 400 performs an optimal match of the decentralized alternative point set with the corresponding line points according to the weights. The anchor point sequence acquisition unit 400 can perform the optimal match by using the steps of step S400 and its preferred steps.

[0059] According to one embodiment, each of the line points has one or more attributes selected from the following combinations: heading angle, road name, and / or whether the number of lanes is separated for up and down traffic.

[0060] According to one embodiment, the alternative point set determination unit 200 sets weights for the alternative points of each line point according to the attributes, and the matching unit performs an optimal match of the decentralized alternative point set with the corresponding line points according to the weights.

[0061] According to one embodiment, the decentralization unit 300 performs decentralization as follows: Assume that the alternative point set of a certain line point is P = {p1, p2, p3, …, p n}, and its center point is

[0062]

[0063] The alternative point set after decentralization is then

[0064]

[0065] According to one embodiment, the alternative point set determining unit 200 uses the points on the road center line or lane center line within a certain range around each row of points as alternative points.

[0066] According to one embodiment, the matching unit 400 performs the best matching of the decentralized alternative point set with the corresponding row of points as follows:

[0067] a. For each row of points in the decentralized navigation path, determine the nearest point in its corresponding decentralized alternative point set;

[0068] b. Obtain the translation parameters and rotation parameters according to the rigid body transformation with the smallest average distance calculated from the point pairs;

[0069] c. Rotate and translate the decentralized navigation path using the translation parameters and rotation parameters to obtain a new point set after rotation and translation;

[0070] d. For each row of points in the new point set, determine the nearest point in its corresponding decentralized alternative point set, and obtain a high-precision map anchor point sequence from the set of these nearest points.

[0071] According to one embodiment, when the matching unit 400 performs the best matching of the decentralized alternative point set with the corresponding row of points, it uses the ICP algorithm with a scale scaling factor.

[0072] According to one embodiment, the device further includes a storage unit (not shown), and the storage unit saves the correspondence between the navigation path and the high-precision map path. The alternative point set determining unit 200 uses the correspondence to determine the alternative point sets of each row of points.

[0073] Unless otherwise specifically indicated by the context, the description order of the method steps does not represent the actual execution order.

[0074] Those skilled in the art should understand that the above devices can be implemented by dedicated hardware, such as field programmable gate arrays, single-chip microcomputers, or microchips, etc., or can also be implemented by a combination of software and hardware.

[0075] The present invention also provides an electronic device, including: a processor; a memory for storing executable instructions of the processor; wherein, the processor is configured to execute the instructions to implement the method of the present invention.

[0076] The present invention also relates to a computer software, which can implement the method of the present invention when executed by a computing device (such as a single-chip microcomputer, a computer, a CPU, etc.).

[0077] The present invention also relates to a computer software storage device, such as a hard disk, a floppy disk, a flash memory, etc., which stores the above-mentioned computer software.

[0078] The description of the method or steps of the present invention can be used to understand the description of the device, equipment and unit, and the description of the device and equipment can also be used to understand the method of the present invention.

[0079] The above description is only illustrative and not a limitation on the protection scope of the present invention. Any change or replacement within the scope of the claims of the present invention is within the protection scope of the present invention.

Claims

1. A map path matching method, characterized in that, Including: Based on the starting point and ending point of the current target itinerary given by the user, determine the navigation path of the current target itinerary and the ordered sequence of waypoints composed of multiple waypoints on this navigation path based on the navigation map; Using a high-precision map, determine the set of alternative points for each of the waypoints; Perform a decentralization operation on the navigation path and the set of alternative points to obtain a decentralized set of alternative points; and Perform an optimal matching of the decentralized set of alternative points with the corresponding waypoints to obtain a high-precision map anchor point sequence, and the high-precision map anchor point sequence is used to determine the high-precision map path corresponding to the navigation path of the current target itinerary in the high-precision map; Among them, the step of performing a decentralization operation on the navigation path and the set of alternative points to obtain a decentralized set of alternative points includes: Perform the decentralization operation as follows: Suppose the alternative point set of a certain line of points is P = {p1, p2, p3, …, p n}, and its center point is Then the decentralized set of alternative points is The step of performing an optimal matching of the decentralized set of alternative points with the corresponding waypoints includes: setting weights for the alternative points of each waypoint according to the attributes of each waypoint, and performing an optimal matching of the decentralized set of alternative points with the corresponding waypoints according to the weights.

2. The map path matching method according to claim 1, wherein After the step of using a high-precision map to determine the set of alternative points for each of the waypoints, it further includes: Using at least one attribute of the waypoint to screen the obtained alternative points, and the attributes include at least one of GPS position, heading angle, road name, length, number of lanes, and whether the up and down lanes are separated.

3. The map path matching method according to claim 1, characterized in that Take the points on the road center line or lane center line within a certain range around each waypoint as alternative points.

4. The map path matching method according to claim 1, wherein Perform an optimal matching of the decentralized set of alternative points with the corresponding waypoints as follows: a. For each waypoint in the decentralized navigation path, determine the nearest point in its corresponding decentralized set of alternative points; b. Obtain the translation parameter and rotation parameter according to the rigid body transformation with the smallest average distance calculated from the point pairs; c. Rotate and translate the decentralized navigation path using the translation parameter and rotation parameter to obtain a new set of points after rotation and translation; d. For each waypoint on the new set of points, determine the nearest point in its corresponding decentralized set of alternative points, and obtain the high-precision map anchor point sequence from the set of these nearest points.

5. The map path matching method according to claim 1, characterized in that Among them, When performing an optimal matching of the decentralized set of alternative points with the corresponding waypoints, use the iterative closest point (ICP) algorithm with a scale scaling factor; Among them, the method further includes the step of obtaining the path on the high-precision map according to the high-precision map anchor point sequence; Among them, the method further includes the step of saving the correspondence between the navigation path and the high-precision map path, and the map path matching method uses this correspondence to determine the set of alternative points for each of the waypoints.

6. A map path matching device, characterized in that, Including: A navigation path determination unit, which based on the starting point and ending point of the current target itinerary given by the user, determines the navigation path of the current target itinerary and the ordered sequence of waypoints composed of multiple waypoints on this navigation path based on the navigation map; An alternative point set determination unit, which uses a high-precision map to determine the set of alternative points for each of the waypoints; A decentralized unit that performs a decentralization operation on the navigation path and the set of alternative points to obtain a decentralized set of alternative points; A matching unit that performs an optimal match of the decentralized set of alternative points with the corresponding row points to obtain a high-precision map anchor point sequence, and the high-precision map anchor point sequence is used to determine the high-precision map path corresponding to the navigation path of the current target itinerary in the high-precision map; Wherein, the decentralized unit is used to perform the decentralization operation as follows: Suppose the alternative point set of a certain line of points is P = {p1, p2, p3, …, p n}, and its center point is Then the decentralized set of alternative points is In the process that the matching unit performs an optimal match of the decentralized set of alternative points with the corresponding row points: weights are set for the alternative points of each row point according to the attributes of each row point, and an optimal match of the decentralized set of alternative points with the corresponding row points is performed according to the weights.

7. An electronic device, characterized in that, Including: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to execute the instructions to implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A device control program is stored on the readable storage medium, and when the device control program is executed by the processor, the method according to any one of claims 1 to 5 is implemented.

Images