WO2024060420A1

WO2024060420A1 - Parking lot data matching method and apparatus, device, medium and vehicle

Info

Publication number: WO2024060420A1
Application number: PCT/CN2022/137917
Authority: WO
Inventors: 李江龙; 罗金辉; 单乐
Original assignee: 北京初速度科技有限公司
Priority date: 2022-09-19
Filing date: 2022-12-09
Publication date: 2024-03-28
Also published as: CN117789444A

Abstract

A parking lot data matching method and apparatus, a device, a medium, and a vehicle. The method comprises: randomly selecting, from parking lot traveling trajectories collected on multiple occasions, vehicle traveling trajectories collected on different occasions, taking one of the traveling trajectories as a current reference trajectory, and taking the other traveling trajectory as a current trajectory to be matched (S110); on the basis of alignment frames matched with vehicle poses in the current reference trajectory and the current trajectory to be matched, relative poses of trajectory frames to be matched, and an absolute position of a vehicle outside a parking lot, sequentially performing pose adjustment on the trajectory frames to be matched, and performing pose alignment on the pose-adjusted trajectory frames and a corresponding reference trajectory frame to obtain a current alignment result (S120); and returning to execute a matching operation between the current reference trajectory and the current trajectory to be matched until a predetermined number of parking lot traveling trajectory matching operations are completed (S130). Hence, the parking lot traveling trajectories collected on multiple occasions are quickly and accurately associated.

Description

A parking lot data matching method, device, equipment, medium and vehicle

Technical field

Embodiments of the present invention relate to the field of automatic driving technology, and specifically to a parking lot data matching method, device, equipment, medium and vehicle.

Background technique

At present, the functions of autonomous parking and automatic navigation of driverless vehicles in parking lots are getting more and more attention. However, realizing autonomous parking and automatic navigation in parking lots requires relying on high-precision maps. Therefore, it is necessary to accurately map the parking lot. appears particularly important.

A large multi-story parking lot often requires multiple data collections to fully cover it. Parking lots are often built underground and do not have GPS (Global Positioning System) signals. Most parking lots are multi-story, and the scenes on each floor are very similar. This brings great difficulties to the parking lot mapping. difficulty.

In related technologies, for parking lot data collected multiple times, the trajectory data collected each time are usually correlated manually. Due to the large amount of data of vehicle trajectory frames in multiple collected data, the efficiency and accuracy of manual data association are relatively low.

Contents of the invention

Embodiments of the present invention provide a parking lot data matching method, device, equipment, medium and vehicle to realize rapid and accurate association of multiple collected parking lot trajectories.

The specific technical solutions are as follows:

In a first aspect, embodiments of the present invention provide a parking lot data matching method, including:

Randomly select the vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, and use one of the driving trajectories as the current reference trajectory and the other driving trajectory as the current trajectory to be matched, where each driving trajectory includes The trajectory before entering the parking lot, the trajectory in the parking lot and the trajectory after exiting the parking lot;

Match the current reference trajectory with the current trajectory to be matched. During the matching process, based on the alignment frame matching the current reference trajectory and the vehicle pose in the current trajectory to be matched, and the relative relationship between each trajectory frame to be matched in the current trajectory to be matched. pose, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, adjust the pose of each trajectory frame to be matched in turn, and compare the adjusted trajectory frame with the corresponding reference trajectory in the current reference trajectory The frame is aligned to pose and pose, and the current alignment result corresponding to the current matching process is obtained;

Use the driving trajectory corresponding to the current alignment result as the new current reference trajectory, and select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and return to execute the new current reference trajectory. The matching operation is performed with the new current trajectory to be matched until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.

Through the above scheme, it can be known that in the process of matching the reference trajectory collected at different times with the trajectory to be matched, based on the alignment frame that matches the reference trajectory with the vehicle posture in the trajectory to be matched, the relative posture between each trajectory frame to be matched in the current trajectory to be matched, and the absolute position information of the vehicle before entering the parking lot and after leaving the parking lot, the posture of each trajectory frame to be matched can be adjusted in turn, and the trajectory frame after posture adjustment is aligned with the reference trajectory frame with the corresponding position, so that the driving trajectories collected multiple times can be accurately and quickly associated together. In the association process, by using a progressive posture adjustment method for each trajectory frame to be matched, the vehicle posture alignment can be made more accurate.

Optionally, based on the alignment frame that matches the current reference trajectory and the vehicle pose in the current trajectory to be matched, the relative pose between each trajectory frame to be matched in the current trajectory to be matched, and the vehicle before entering and exiting the parking lot. Based on the absolute position information after the field, the pose is adjusted for each trajectory frame to be matched in turn, and the pose is aligned with the corresponding reference trajectory frame in the current reference trajectory, and the current matching process corresponding to the current reference trajectory frame is obtained. Alignment results, including:

Traverse each to-be-matched trajectory frame in the current to-be-matched trajectory in sequence, and for each current to-be-matched trajectory frame, determine the current reference trajectory frame corresponding to the current to-be-matched trajectory frame position in the current reference trajectory;

Match the current reference trajectory frame with the current trajectory frame to be matched;

If the matching is successful, determine the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame;

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, the successful matching All matched trajectory frames to be matched are optimized in pose, and the reference trajectory frame corresponding to the optimized trajectory frame to be matched and the position where the matching is successful is used as an alignment frame;

Based on the current alignment frame, adjust the pose of the trajectory frame to be matched adjacent to the current alignment frame, and use the adjusted adjacent trajectory frame as the new current trajectory frame to be matched, and return to execute the new current trajectory frame to be matched. The matching operation of the trajectory frame and the current reference trajectory frame corresponding to the position is performed until all the trajectory frames to be matched in the current trajectory to be matched are traversed, and the current alignment result of the vehicle pose in the current matching process is obtained.

Optionally, after determining the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, the method provided by the embodiment of the present invention further includes:

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, the current reference All reference trajectory frames in the trajectory are optimized for pose, and the optimized trajectory frame to be matched and the reference trajectory frame after successful matching are used as alignment frames.

It can be seen from the above technical solution that when optimizing the topology map corresponding to the current trajectory frame to be matched, the pose of all reference trajectory frames in the current reference trajectory can be optimized at the same time, thereby achieving a more accurate relationship between the reference trajectory and the trajectory to be matched. Position alignment.

Optionally, match the current reference trajectory frame with the current trajectory frame to be matched, including:

Obtain local feature information of the parking lot, which includes: traffic signs, wall markings, road markings and lane markings;

If the local feature information of the parking lot corresponding to the current reference trajectory frame is consistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched, it is confirmed that the current reference trajectory frame and the current trajectory frame to be matched are successfully matched; or,

If the local feature information of the parking lot corresponding to the current reference trajectory frame is inconsistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched, it is determined that the matching between the current reference trajectory frame and the current trajectory frame to be matched fails.

Optionally, the method provided by the embodiment of the present invention also includes:

If the current reference trajectory frame fails to match the current trajectory frame to be matched, the current trajectory frame to be matched that fails to match is temporarily stored;

After obtaining the current alignment result, the temporarily stored target trajectory frames that failed to match are traversed in sequence. For any target trajectory frame that fails to match, based on the optimized pose of the target trajectory frame, the target trajectory with the optimized pose is repeatedly executed. The matching operation of the reference trajectory frame corresponding to the frame and position is performed until all temporarily stored target trajectory frames that have failed to match are traversed and a new current alignment result is obtained;

Accordingly, the driving trajectory corresponding to the current alignment result is used as the new current reference trajectory, and another unmatched driving trajectory is selected from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and the matching operation between the new current reference trajectory and the new current trajectory to be matched is returned, including:

The driving trajectory corresponding to the new current alignment result is used as the new current reference trajectory, and another unmatched driving trajectory is selected from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and the matching operation between the new current reference trajectory and the new current trajectory to be matched is returned.

It can be seen from the above technical solution that by querying and matching the reference trajectory frame of the corresponding position again with the failed target trajectory frame, the alignment result of the vehicle pose can be made more accurate.

Optionally, determining a current reference trajectory frame corresponding to the position of the current trajectory frame to be matched in the current reference trajectory includes:

When the current trajectory frame to be matched belongs to the trajectory before entering the parking lot or the trajectory after exiting the parking lot:

Obtain the absolute position corresponding to the current trajectory frame to be matched, and search within the first set distance range from the absolute position in the current reference trajectory to obtain the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched;

Or, when the current trajectory frame to be matched belongs to the trajectory in the parking lot:

According to the current position information corresponding to the current trajectory frame to be matched, in the current reference trajectory, search within the first set distance range from the current position information to obtain the candidate reference trajectory frame;

Determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot;

If they belong to the same layer, the candidate reference trajectory frame is used as the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched.

Optionally, determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot, including:

Determine the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame;

Detect whether there is a slope segment in the shortest topological path;

If there is no slope segment, determine that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot; or,

If there is a slope segment, and if the difference between the height information of the uphill segment and the height information of the downhill segment in the slope segment is within the second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the parking lot. of the same layer; or,

If the difference between the height information of the uphill section and the height information of the downhill section in the slope section exceeds the second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched do not belong to the same layer of the parking lot.

It can be seen from the above technical solution that by determining the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame, complications between various paths from the current trajectory frame to be matched to the candidate reference trajectory frame are avoided. By detecting slope segments in the shortest topological path and based on the height information of the uphill and downhill segments in the slope segments, it can be effectively determined whether the candidate reference trajectory frame in the current reference trajectory and the current trajectory frame to be matched belong to the same parking level. field to avoid mismatching of trajectory frames that do not belong to the same parking lot, and solve the problem of incorrect correlation of cross-layer data.

Optionally, detect whether there are slope segments in the shortest topological path, including:

Determine whether the pitch angle of the vehicle relative to the horizontal plane is greater than the preset angle;

The road segment corresponding to the trajectory frame with a pitch angle greater than or equal to the preset angle is regarded as a slope segment, and the road segment corresponding to the trajectory frame with the pitch angle smaller than the preset angle is regarded as a non-slope segment.

Optionally, based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched. , perform pose optimization on all successfully matched trajectory frames to be matched, including:

Constructing a first topological graph corresponding to the current trajectory to be matched and a second topological graph corresponding to the current reference trajectory, wherein the nodes of the first topological graph are the trajectory frames to be matched, and adjacent trajectory frames to be matched are connected by edges, and the nodes of the second topological graph are the reference trajectory frames, and adjacent reference trajectory frames are connected by edges;

Based on the first relative posture, the absolute position information of the vehicle in each driving trajectory before entering and exiting the parking lot, and the second relative posture between each trajectory frame to be matched in the current driving trajectory to be matched, the first topological graph is optimized by a posture graph optimization method, and the trajectory frame to be matched after posture optimization and the successfully matched reference trajectory frame are used as alignment frames, wherein the alignment frame nodes in the first topological graph and the second topological graph are connected by edges.

In the above technical solution, by using the posegraph optimization method to optimize the topology map corresponding to the current trajectory frame to be matched and the topology map corresponding to the reference trajectory, the vehicle pose can be accurately aligned, thereby The reference trajectory and the trajectory to be matched are accurately associated together, which solves the problem of insufficient data correlation.

Optionally, before matching the current reference trajectory with the current trajectory to be matched, the method provided by the embodiment of the present invention also includes:

Based on the absolute position information before entering the parking lot and after exiting the parking lot, the first topology map and the second topology map are optimized respectively to reduce the deviation in absolute position between the current trajectory to be matched and the current reference trajectory.

Optionally, the vehicle pose corresponding to each reference trajectory frame or the vehicle pose corresponding to each trajectory frame to be matched is obtained through any of the following sensor data, or by fusing the following multiple sensor data:

Inertial measurement unit IMU data, image data, radar point cloud data and odometer data.

In a second aspect, embodiments of the present invention also provide a parking lot data matching device, including:

A driving trajectory selection module is configured to arbitrarily select vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, and use one of the driving trajectories as the current reference trajectory and another driving trajectory as the current trajectory to be matched, wherein each driving trajectory includes a trajectory before entering the parking lot, a trajectory in the parking lot, and a trajectory after leaving the parking lot;

The matching module is configured to match the current reference trajectory with the current trajectory to be matched. During the matching process, based on the alignment frame matching the current reference trajectory with the vehicle pose in the current trajectory to be matched, each of the current trajectories to be matched. The relative pose between the trajectory frames to be matched, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, adjust the pose of each trajectory frame to be matched in turn, and compare the adjusted trajectory frame with The corresponding reference trajectory frames in the current reference trajectory are pose aligned to obtain the current alignment result corresponding to the current matching process;

The data association module is configured to use the driving trajectory corresponding to the current alignment result as the new current reference trajectory, and select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and Return to perform the matching operation of the new current reference trajectory and the new current trajectory to be matched until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.

Optional matching modules include:

The current reference trajectory frame determination unit is configured to sequentially traverse each trajectory frame to be matched in the current trajectory to be matched, and for each current trajectory frame to be matched, determine the position corresponding to the current trajectory frame to be matched in the current reference trajectory. The current reference trajectory frame;

A matching unit configured to match the current reference trajectory frame with the current trajectory frame to be matched;

A relative pose determination unit configured to determine a first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame if the current reference trajectory frame matches the current trajectory frame to be matched successfully;

The first posture optimization unit is configured to be based on the first relative posture, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the space between each to-be-matched trajectory frame in the current driving trajectory to be matched. For the second relative pose, perform pose optimization on all successfully matched trajectory frames to be matched, and use the reference trajectory frame corresponding to the optimized trajectory frame to be matched and the position where the matching is successful as an alignment frame;

The pose alignment unit is configured to adjust the pose of the trajectory frame to be matched adjacent to the current alignment frame based on the current alignment frame, and use the adjacent trajectory frame after the pose adjustment as the new current trajectory frame to be matched, and Return to perform the matching operation of matching the new current trajectory frame to be matched with the current reference trajectory frame corresponding to the position, until all trajectory frames to be matched in the current trajectory to be matched are traversed, and the current alignment result of the vehicle pose in the current matching process is obtained. .

Optionally, the device provided by the embodiment of the present invention also includes:

The second pose optimization unit is configured to, after determining the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, based on the first relative pose, the vehicle entering the parking lot in each driving trajectory The absolute position information before and after leaving the parking lot, the second relative pose between the trajectory frames to be matched in the current driving trajectory to be matched, the pose optimization of all reference trajectory frames in the current reference trajectory, and the pose The optimized trajectory frame to be matched and the optimized reference trajectory frame of the successfully matched pose are used as alignment frames.

Optional, matching units include:

The local feature information acquisition subunit is configured to acquire local feature information of the parking lot, where the local feature information includes: traffic signs, wall markings, road markings and lane markings;

The matching subunit is configured to determine that the current reference trajectory frame and the current trajectory frame to be matched are successfully matched if the local feature information of the parking lot corresponding to the current reference trajectory frame is consistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched. ;or,

The matching failed frame temporary storage module is configured to temporarily store the failed matching current trajectory frame to be matched if the current reference trajectory frame fails to match the current trajectory frame to be matched;

The re-matching module is configured to sequentially traverse the temporarily stored target trajectory frames that failed to match. For any target trajectory frame that failed to match, based on the optimized pose of the target trajectory frame, repeatedly execute the optimized target trajectory. The matching operation of the reference trajectory frame corresponding to the frame and position is performed until all temporarily stored target trajectory frames that have failed to match are traversed and a new current alignment result is obtained;

Correspondingly, the data association module is specifically configured as:

Use the driving trajectory corresponding to the new current alignment result as the new current reference trajectory, and select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and return to execute the new current trajectory. The reference trajectory is matched with the new current trajectory to be matched until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.

Optional, the current reference trajectory frame determination unit includes:

The first search subunit is configured to obtain the absolute position corresponding to the current trajectory frame to be matched when the current trajectory frame to be matched belongs to the trajectory before entering the parking lot or the trajectory after exiting the parking lot, and in the current reference trajectory, Search within the first set distance range from the absolute position to obtain the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched;

or,

The second search subunit is configured to, in the case where the current trajectory frame to be matched belongs to the trajectory in the parking lot, based on the current location information corresponding to the current trajectory frame to be matched, in the current reference trajectory, the first distance from the current location information is Search within a certain distance range to obtain candidate reference trajectory frames;

The same layer determination subunit is configured to determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot;

The current reference trajectory frame determination subunit is configured to use the candidate reference trajectory frame as the position corresponding to the current trajectory frame to be matched if the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot. The current reference trajectory frame.

Optionally, the same-layer judgment subunit includes:

a shortest topological path determination component configured to determine the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame;

a slope segment detection component configured to detect whether there is a slope segment in the shortest topological path;

The same layer determination component is configured to determine that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot if there is no slope segment in the shortest topological path;

Alternatively, if there is a ramp section in the shortest topological path, and if the difference between the height information of the uphill section and the height information of the downhill section in the ramp section is within a second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot;

Alternatively, if the difference between the height information of the uphill section and the height information of the downhill section in the slope section exceeds the second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched do not belong to the same layer of the parking lot.

Optional, the same layer determines the component, which is specifically configured as:

The road segment corresponding to the trajectory frame with a pitch angle greater than or equal to the preset angle is regarded as a slope segment, and the road segment corresponding to the trajectory frame with a pitch angle smaller than the preset angle is regarded as a non-slope segment.

Optional, the first attitude optimization unit is specifically configured as:

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, through the pose The graph optimization method optimizes the first topology graph, and uses the optimized trajectory frame to be matched and the successfully matched reference trajectory frame as the alignment frame, where the alignment frame node in the first topology graph and the second topology graph is Connected by edges.

Optionally, the device provided in this embodiment also includes:

The initial optimization module is configured to optimize the first topology map and the second topology map respectively based on the absolute position information before entering the parking lot and after exiting the parking lot before matching the current reference trajectory with the current trajectory to be matched. , to reduce the absolute position deviation between the current trajectory to be matched and the current reference trajectory.

In a third aspect, an embodiment of the present invention provides an electronic device, which includes:

one or more processors;

a storage device for storing one or more programs,

When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the matching method of parking lot data as provided in any embodiment of the present invention.

In a fourth aspect, embodiments of the present invention provide a storage medium on which a computer program is stored, characterized in that when the program is executed by a processor, the parking lot data matching method as provided in any embodiment of the present invention is implemented. .

In a fifth aspect, embodiments of the present invention provide a vehicle, which includes the parking lot data matching device provided by any embodiment of the present invention, or the electronic device provided by any embodiment of the present invention.

In a sixth aspect, an embodiment of the present invention provides a computer program. The computer program includes program instructions. When the program instructions are executed by a computer, the parking lot data matching method provided by any embodiment of the present invention is implemented.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts.

Figure 1a is a flow chart of a parking lot data matching method provided in Embodiment 1 of the present invention;

FIG1b is a topological diagram of vehicle driving trajectories including a first-floor underground parking lot provided in Embodiment 1 of the present invention;

Figure 1c is a topological diagram of vehicle driving trajectories including a two-story underground parking lot provided by Embodiment 1 of the present invention;

FIG. 1d is a topological diagram after initial posture optimization of the reference driving trajectory and the trajectory to be matched provided by the first embodiment of the present invention;

Figure 1e is a topological diagram for matching a trajectory to be matched with a reference trajectory provided in Embodiment 1 of the present invention;

Figure 1f is another topological diagram for matching a trajectory to be matched with a reference trajectory provided by Embodiment 1 of the present invention;

FIG2 is a flow chart of a parking lot data matching method provided by Embodiment 2 of the present invention;

Figure 3 is a structural block diagram of a parking lot data matching device provided in Embodiment 3 of the present invention;

Figure 4 is a structural block diagram of an electronic device provided in Embodiment 4 of the present invention;

Figure 5 is a schematic diagram of a vehicle provided in Embodiment 5 of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without exerting creative efforts fall within the scope of protection of the present invention.

It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes Other steps or units inherent to such processes, methods, products or devices.

The embodiment of the invention discloses a parking lot data matching method, device, equipment, medium and vehicle. Each is explained in detail below.

Embodiment 1

Figure 1a is a flow chart of a parking lot data matching method provided in Embodiment 1 of the present invention. This method can be applied to vehicle-mounted computers, vehicle-mounted industrial personal computers (IPC) and other vehicle-mounted terminals, and can also be applied to servers. , the embodiment of the present invention does not limit this. The method provided in this embodiment can be applied to the application scenario of correlating the driving trajectories of multi-story parking lots collected multiple times. The method provided in this embodiment can be executed by a parking lot data matching device, which can be implemented by software and/or hardware. As shown in Figure 1a, the method provided by this embodiment specifically includes:

S110: Randomly select vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, use one of the driving trajectories as the current reference trajectory, and use the other driving trajectory as the current trajectory to be matched.

Among the vehicle driving trajectories collected at different times, the vehicle pose corresponding to each frame trajectory can be obtained by collecting vehicle-mounted sensors, or can also be obtained by fusing multiple sensor data. Among them, vehicle-mounted sensors can include inertial measurement units (Inertial Measurement Unit, IMU), image sensors, radars and odometers, etc.

In this embodiment, the vehicle driving trajectories collected multiple times include the driving trajectory before entering the parking lot, the driving trajectory in the parking lot, and the driving trajectory after leaving the parking lot. Among them, there will be GPS (Global Positioning System) signals before the vehicle enters the parking lot and after it comes out of the parking lot. The three trajectories before entering the parking lot, the driving trajectory inside the parking lot, and the trajectory after leaving the parking lot are continuous trajectories. By applying GPS constraints to the two trajectories of entering and leaving the parking lot, and using the continuity of the three trajectories, the entire trajectory can obtain an absolute position positioning accuracy of meters.

In this embodiment, matching the current reference trajectory with the current trajectory to be matched actually means matching the current reference trajectory with the vehicle pose corresponding to each trajectory frame in the current trajectory frame to be matched, and optimizing the vehicle pose to optimize the vehicle pose. The frame to be matched trajectory is aligned with the reference trajectory at the corresponding position. In order to reduce the amount of calculation, multiple key frames can be selected from the current reference trajectory as reference trajectory frames, and multiple key frames can be selected from the current trajectory to be matched as trajectory frames to be matched. Among them, the selection of each key frame may be: use the trajectory frame whose driving distance is the preset distance as the key frame. Among them, the preset distance can be any distance within the range of 1 to 3 meters.

In this embodiment, in order to achieve accurate alignment of poses between the current reference trajectory and the current trajectory to be matched, the poses of each reference trajectory frame and each trajectory frame to be matched may be initially optimized respectively. Specifically, a first topology graph corresponding to the current trajectory to be matched and a second topology graph corresponding to the current reference trajectory can be constructed, where the nodes of the first topology graph are each trajectory frame to be matched, and the nodes between adjacent trajectory frames to be matched are They are connected by edges. The nodes of the second topology graph are each reference trajectory frame, and adjacent reference trajectory frames are connected by edges. Before matching the current reference trajectory with the current trajectory to be matched, the first topology map and the second topology map are optimized respectively based on the absolute position information before entering the parking lot and after exiting the parking lot, which can reduce the current trajectory to be matched. Deviation in absolute position from the current reference trajectory.

Specifically, Figure 1b is a topological diagram of a vehicle driving trajectory including a one-story underground parking lot (B1) provided in Embodiment 1 of the present invention, and Figure 1c is a topological diagram of a vehicle driving trajectory including a second-level underground parking lot (B2) provided in Embodiment 1 of the present invention. ) is a topological map of the vehicle's trajectory. As shown in Figures 1b and 1c, the nodes in each driving trajectory are vehicle driving trajectory frames, that is, optimization variables. Adjacent nodes are connected by edges, and adjacent nodes are constrained by relative poses. The GPS signal generated by GPS measurement is constrained before entering the parking lot and after exiting the parking lot. One of the driving trajectories in Figure 1b and Figure 1c is used as the trajectory to be matched, and the other driving trajectory is used as the reference trajectory. Figure 1d shows the reference driving trajectory and the trajectory to be matched after optimizing the initial pose provided by Embodiment 1 of the present invention. Topology. By optimizing the pose graph of Figure 1b and Figure 1c, the initial optimization of the vehicle pose can be achieved to reduce the deviation in absolute position between the trajectory to be matched and the reference trajectory. As shown in Figure 1d, after the initial optimization is completed, the nodes of each trajectory frame in the reference driving trajectory and the trajectory to be matched are not completely aligned, and there are still meter-level errors in their positions. Therefore, subsequent steps need to be carried out to The trajectory frame is adjusted in pose to achieve pose alignment between the trajectory frame to be matched and the reference trajectory frame.

S120. Match the current reference trajectory with the current trajectory to be matched. During the matching process, based on the alignment frame that matches the current reference trajectory and the vehicle pose in the current trajectory to be matched, the alignment frame between each trajectory frame to be matched in the current trajectory to be matched. The relative pose of the vehicle, as well as the absolute position information of the vehicle before entering the parking lot and after leaving the parking lot, adjust the pose of each trajectory frame to be matched in turn, and compare the adjusted trajectory frame with the corresponding one in the current reference trajectory. Perform pose alignment with reference to the trajectory frame, and obtain the current alignment result corresponding to the current matching process.

In this embodiment, for each trajectory frame to be matched in the trajectory to be matched, the pose of each trajectory frame to be matched can be optimized through progressive optimization. Specifically, this can be achieved through the following steps (1) to (3).

(1) Sequentially traverse each trajectory frame to be matched in the current trajectory to be matched, and for each current trajectory frame to be matched, determine the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched in the current reference trajectory.

Wherein, the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched means that it belongs to the same parking lot as the current trajectory frame to be matched and is close to the location of the current trajectory frame to be matched, that is, the distance is within the first set distance range. Reference trajectory frame. The number of reference trajectory frames that meet the above conditions can be one or multiple.

In this embodiment, the current trajectory frame to be matched may belong to the trajectory before entering the parking lot or the trajectory after exiting the parking lot, or may also belong to the trajectory within the parking lot.

Among them, when the current trajectory frame to be matched belongs to the trajectory before entering the parking lot or the trajectory after exiting the parking lot, since the vehicle can determine the absolute position information of the vehicle based on the received GPS signal, the position can be directly compared based on the absolute position information. The search for the corresponding reference trajectory frame can be implemented in the following ways:

Obtain the absolute position corresponding to the current trajectory frame to be matched, and search within the first set distance range from the absolute position in the current reference trajectory to obtain the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched. Among them, the first set distance range may be 8 to 12 meters.

Or, when the current trajectory frame to be matched belongs to the trajectory in the parking lot, the vehicle cannot receive the GPS signal, or the received GPS signal is weak. At this time, the following steps A to C can be used to determine the position of the current trajectory frame. The corresponding current reference trajectory frame.

A. According to the current position information corresponding to the current trajectory frame to be matched, in the current reference trajectory, search within the first set distance range from the current position information to obtain the candidate reference trajectory frame.

B. Determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot. If they belong to the same layer, use the candidate reference trajectory frame as the current reference trajectory frame corresponding to the current trajectory frame position.

Among them, there are many ways to determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot. This embodiment uses the method of detecting the slope segment from the shortest topological path and determining the uphill road in the slope segment. The height difference between the segment and the downhill segment is used to determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot. This can be achieved in the following ways:

Determine the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame; detect whether there is a slope segment in the shortest topology path; if there is no slope segment, determine that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the parking lot on the same layer; or, if there is a slope segment, and if the difference between the height information of the uphill segment and the height information of the downhill segment in the slope segment is within the second set distance range, determine the candidate reference trajectory frame and the current waiting The matching trajectory frame belongs to the same layer of the parking lot; or, if the difference between the height information of the uphill section and the height information of the downhill section in the slope section exceeds the second set distance range, determine the candidate reference trajectory frame and the current trajectory to be matched The frame does not belong to the same level of the parking lot. Among them, the second set distance range may be 3 to 5 meters.

In this embodiment, whether it is a slope segment can be determined by the pitch angle of the vehicle relative to the horizontal plane. Specifically, the segment corresponding to the trajectory frame whose pitch angle is greater than or equal to the preset angle can be regarded as the slope segment, and the pitch angle is less than The road segment corresponding to the trajectory frame at the preset angle is regarded as a non-slope segment. The range of the preset angle may be 5° to 10°. The pitch angle of the vehicle relative to the horizontal plane can be determined by the IMU, or the attitude of the vehicle can also be determined based on the fused data of multiple sensors, thereby determining the pitch angle of the vehicle relative to the horizontal plane. Among them, multi-sensors can include IMU, odometer, image sensor, etc.

In this embodiment, the height information of the uphill section or the height information of the downhill section can be calculated based on the height difference between the vehicle posture at the top of the ramp section and the vehicle posture at the bottom of the ramp section.

In this embodiment, a candidate reference trajectory frame that belongs to the same parking lot as the current trajectory to be matched is used as the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched. For candidate reference trajectory frames that do not belong to the same parking lot level as the current trajectory to be matched, they will not be matched with the current trajectory to be matched. In this case, continue to traverse backward to the next new trajectory frame to be matched, and return to perform the operation of determining the current reference trajectory frame corresponding to the position of the new current trajectory frame to be matched, that is, return to step (1) operate.

For example, Figure 1e is a topological diagram for matching a trajectory to be matched with a reference trajectory provided in Embodiment 1 of the present invention. As shown in Figure 1e, K2, K4, K8, K7 and K6 are the reference trajectory frames in the current reference trajectory. K1, K3, K9, K10, K5 and K11 are the trajectory frames to be matched in the current trajectory to be matched, where K1 and K2, K3 and K4 are alignment frames whose poses have been aligned. During the progressive positioning matching process, since the GPS signal cannot be received after entering the parking lot, the vehicle trajectory will obviously drift in the absolute position, and may even cause misalignment. As shown in Figure 1e, K5 and K6, because the trajectory of K5 gradually drifts, its position is very close to K6, which will lead to an erroneous correlation between the two. This situation will cause the entire mapping to fail. In order to avoid this problem, each time two frame matching is initiated, the shortest topological path between the two needs to be queried from the topology map. For example, the shortest topological path path_0 between K5 and K6 is: K5-K10-K9-K3- K4-K8-K7-K6.

In the shortest topological path path_0, the uphill and downhill conditions on the shortest topological path are queried based on the posture of the trajectory frame to be matched. As shown in Figure 1e, K3-K9 descended directly from outside the parking lot to the B2 floor, for example, they dropped 6m, and K4-K8 descended from outside the parking lot to the B1 floor, for example, they dropped 3m. Then the current trajectory frame to be matched can be determined. K5 and the candidate reference trajectory frame K6 are not in the same layer, so there is no need to match K5 and K6, thereby eliminating erroneous matching situations. At this time, the candidate reference trajectory frame K11, which is the next trajectory frame to be matched, needs to be searched again. If the candidate reference trajectory frame K6 is found, continue to determine whether K11 and K6 belong to the same layer, that is, calculate K11 Uphill and downhill conditions on the shortest topological path to K6. For example, K4-K8 dropped by 3m, K3-K9 dropped by 6m, and K5-K11 rose by 3m. Therefore, it can be judged that K11 and K6 are on the same floor. At this time, the reference trajectory frame K6 can be used as the current key frame K11 to be matched. The current reference trajectory frame corresponding to the position, and you can continue to perform the following step (2), that is, use the local feature information of the parking lot to match K6 and K11, and after K6 and K11 become alignment frames that match the pose, Connect K6 and K11 in the topology diagram.

In this embodiment, by determining the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame, complications between various paths from the current trajectory frame to be matched to the candidate reference trajectory frame are avoided. By detecting slope segments in the shortest topological path and based on the height information of the uphill and downhill segments in the slope segments, it can be effectively determined whether the candidate reference trajectory frame in the current reference trajectory and the current trajectory frame to be matched belong to the same parking level. field to avoid mismatching of trajectory frames that do not belong to the same parking lot, and solve the problem of incorrect correlation of cross-layer data.

(2) Match the current reference trajectory frame with the current trajectory frame to be matched. If the matching is successful, determine the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame.

Those skilled in the art can understand that vehicle trajectories obtained by driving different times in the same parking lot will have deviations in posture, but the local characteristics of the parking lot corresponding to the corresponding trajectory frames in different driving trajectories Information is fixed. Among them, the local characteristic information of the parking lot may include traffic signs, wall markings, road markings, lane markings, etc. The local characteristic information of the parking lot can be collected through vehicle-mounted sensors (such as image sensors, radar, etc.).

Based on the above reasons, in this embodiment, the current reference trajectory frame can be matched with the current trajectory frame to be matched based on the acquired local feature information of the parking lot. If the local feature information of the parking lot corresponding to the current reference trajectory frame is consistent with the current If the local feature information of the parking lot corresponding to the trajectory frame to be matched is consistent, it is determined that the current reference trajectory frame and the current trajectory frame to be matched are successfully matched; if the local feature information of the parking lot corresponding to the current reference trajectory frame is consistent with the current trajectory frame to be matched. If the local feature information of the parking lot is inconsistent, it is determined that the current reference trajectory frame and the current trajectory frame to be matched have failed to match.

In this embodiment, for the successfully matched current to-be-matched trajectory frame and the corresponding current reference trajectory frame, the relative pose of the two is calculated to associate the two. For the current track frame to be matched that fails to match, it will be temporarily stored, and after the traversal of all track frames to be matched is completed, the track frame to be matched that fails to match will be re-matched.

(3) Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, The pose is optimized for all the successfully matched trajectory frames to be matched, and the reference trajectory frame corresponding to the position of the optimized trajectory frame to be matched and the position where the matching is successful is used as an alignment frame.

In this embodiment, the posture of the vehicle is optimized by constructing a posture topology map and optimizing the posture topology map, so as to achieve posture alignment between the trajectory frame to be matched and the reference trajectory frame corresponding to the position where the match is successful. This can be achieved through the following steps 1 to 2:

1. Construct a first topological graph corresponding to the current trajectory to be matched and a second topological graph corresponding to the current reference trajectory.

For example, the topology map shown in Figure 1b is used as the topology map corresponding to the current reference trajectory, and the topology map shown in Figure 1c is used as the topology map corresponding to the current trajectory to be matched. As shown in Figure 1b, the nodes of the second topology graph corresponding to the current reference trajectory are each reference trajectory frame, and adjacent reference trajectory frames are connected by edges. As shown in Figure 1c, the nodes of the first topology graph corresponding to the current trajectory to be matched are each trajectory frame to be matched, and adjacent trajectory frames to be matched are connected by edges.

2. Based on the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the current driving trajectory to be matched. For the second relative pose between the trajectory frames to be matched, the first topology graph is optimized through the pose graph optimization method, and the optimized trajectory frame to be matched and the successfully matched reference trajectory frame are used as alignment frames, Wherein, the aligned frame nodes in the first topology graph and the second topology graph are connected by edges.

In this embodiment, by using a posegraph optimization method to optimize the topology graph corresponding to the current trajectory frame to be matched, the variable to be optimized represented by the node in Figure 1c, that is, the vehicle pose, can be optimized. Specifically, during the optimization process, if the current traversal is the Nth trajectory frame to be matched, the trajectory frame for pose optimization is obtained from the 1st trajectory frame to be matched to the current Nth trajectory frame to be matched. is the optimized position corresponding to the first to N nodes, and the trajectory frames to be matched after the optimization of these N poses and the successfully matched reference trajectory frame can be used as alignment frames. In this embodiment, by adopting the pose graph optimization method, the vehicle poses in the trajectory to be matched and the reference trajectory can be accurately aligned, so that the reference trajectory and the trajectory to be matched are accurately associated together. , which solves the problem of insufficient data correlation.

Furthermore, in order to make the alignment of the current trajectory frame to be matched and the current reference trajectory frame more accurate, when optimizing the topology map corresponding to the current trajectory frame to be matched, the current trajectory frame to be matched and the corresponding current reference trajectory can be optimized. The first relative pose between frames is the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot in each driving trajectory, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched. All reference trajectory frames in the current reference trajectory undergo pose optimization to achieve more accurate pose alignment between the reference trajectory and the trajectory to be matched.

For example, Figure 1f is another topological diagram for matching a trajectory to be matched with a reference trajectory provided in Embodiment 1 of the present invention. Based on Figure 1d, Figure 1f shows the further optimization process of the vehicle pose. For the first trajectory frame K1 to be matched in the current trajectory data2 to be matched, the trajectory frame K1 is the trajectory frame before entering the parking lot. According to the absolute position information of the trajectory frame K1 before entering the parking lot, the reference trajectory frame K2 corresponding to the posture of the trajectory frame K1 to be matched within the first set distance range can be queried in the reference driving trajectory data1. Use the local feature information of the parking lot corresponding to K1 and K2 to match K1 and K2. If the matching is successful, calculate the relative pose of the vehicle corresponding to the K1 trajectory frame and the K2 trajectory frame, and after completing the calculation of the relative pose, pass Optimizing the first topology map and the second topology map can achieve a more accurate posture configuration between K1 and K2. The current trajectory frame K1 to be matched and the corresponding reference trajectory frame K2 after pose optimization are the alignment frames. In the topology diagram, connect the nodes corresponding to K1 and K2 so that the current trajectory data2 to be matched and the current reference trajectory data1 obtain initial registration.

(4) Based on the current alignment frame, adjust the pose of the trajectory frame to be matched adjacent to the current alignment frame, and use the adjusted adjacent trajectory frame as the new current trajectory frame to be matched, and return to execute the new The matching operation of the current trajectory frame to be matched and the current reference trajectory frame corresponding to the position is performed until all the trajectory frames to be matched in the current trajectory to be matched are traversed, and the current alignment result of the vehicle pose in the current matching process is obtained.

Those skilled in the art can understand that since the relative pose between two adjacent trajectory frames in the same driving trajectory frame is fixed, when the Nth currently traversed to-be-matched object is obtained through topology graph optimization, After the optimized position of the trajectory frame, the next trajectory frame to be matched adjacent to the currently traversed Nth trajectory frame to be matched, that is, the pose of the N+1th frame also needs to be adjusted accordingly, and the pose is The adjusted N+1th trajectory frame to be matched is used as the current new trajectory frame to be matched, and the operation of step (1) above is returned to obtain the current reference trajectory frame corresponding to the position of the new current trajectory frame to be matched. , and continue to perform the matching operation of the new current trajectory frame to be matched with the current reference trajectory frame corresponding to the position in step (2) above, and then continue to perform steps (3) and (4) until the current trajectory frame to be matched is traversed. All the trajectory frames to be matched in the trajectory are used to obtain the current alignment result of the vehicle pose in the current matching process.

Specifically, as shown in Figure 1f, the trajectory frame K1 to be matched and the trajectory frame K2 to be matched have been aligned. At this time, the pose of the trajectory frame K1 to be matched has been optimized. Since the relative pose between the trajectory frame K1 to be matched and the adjacent trajectory frame K3 is fixed, after the pose of the trajectory frame K1 to be matched is adjusted, the adjacent trajectory frame K3 (solid black circle) According to the relative pose between it and K1, the pose can be updated to the black dotted circle, that is, the position of the adjacent trajectory frame K3 is adjusted. Use the posture-adjusted trajectory frame K3 as the new current trajectory frame to be matched, and search based on the position again to obtain the corresponding reference trajectory frame K4 in the reference trajectory, and use the local features of the parking lot to perform trajectory frame K4 and trajectory Match between frames K3, and calculate the relative pose after successful matching, so as to associate the trajectory frame K4 with the trajectory frame K3. The above process is repeated for all subsequent trajectory frames to be matched until the processing of all frames to be matched is completed.

S130, taking the driving trajectory corresponding to the current alignment result as the new current reference trajectory, and selecting another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and returning to perform the matching operation between the new current reference trajectory and the new current trajectory to be matched, until a predetermined number of parking lot driving trajectory matching operations are completed, and obtaining a target matching result of the vehicle posture.

The predetermined number can be set according to the number of actually collected parking lot driving trajectories, for example, it can be set to the total number of collected parking lot driving trajectories, which is not specifically limited in this embodiment.

The technical solution provided by this embodiment is, in the process of matching the reference trajectories collected at different times with the trajectories to be matched, based on the alignment frames that match the reference trajectories and the vehicle poses in the trajectories to be matched, and each of the current trajectories to be matched. By matching the relative poses between trajectory frames, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, the poses of each trajectory frame to be matched can be adjusted in turn, and the adjusted trajectory frames are combined with The position is aligned with the reference trajectory frame so that the driving trajectories collected multiple times can be accurately associated together. During the association process, by adopting a progressive pose adjustment method for each trajectory frame to be matched, the vehicle pose alignment can be made more accurate. Compared with the method of manually correlating parking lot data collected multiple times in the related art, the method provided by this embodiment improves the efficiency and accuracy of parking lot data correlation.

Embodiment 2

Figure 2 is a flow chart of a parking lot data matching method provided in Embodiment 2 of the present invention. Based on the above embodiment, this embodiment performs a test on the situation where the current reference trajectory frame and the current trajectory frame to be matched fail to match. To refine, as shown in Figure 2, the method provided by this embodiment includes:

S200: Randomly select the vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, use one of the driving trajectories as the current reference trajectory, and use the other driving trajectory as the current trajectory to be matched.

S210. Match the current reference trajectory with the current trajectory to be matched. During the matching process, traverse each trajectory frame to be matched in the current trajectory to be matched in sequence. For each current trajectory frame to be matched, determine the current reference trajectory and the current trajectory frame. The current reference trajectory frame corresponding to the position of the trajectory frame to be matched.

S220. Match the current reference trajectory frame with the current trajectory frame to be matched, and determine whether the two match successfully. If so, execute step S230; otherwise, execute step S240.

S230: Determine the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, and continue to step S250.

S240: Temporarily store the current trajectory frame to be matched that fails to match, and continue to execute step S270.

S250. Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, All trajectory frames to be matched that have been successfully matched, as well as all reference trajectory frames in the current reference trajectory, are optimized for pose, and the optimized trajectory frames to be matched are aligned with the successfully matched reference trajectory frames after pose optimization. frame, continue to step S260.

In this embodiment, when using the pose graph optimization method for pose optimization, by optimizing the pose of all successfully matched trajectory frames to be matched and all reference trajectory frames in the current reference trajectory, the current to-be-matched The trajectory frame is more accurately aligned with the current reference trajectory frame pose.

S260. Based on the current alignment frame, adjust the pose of the trajectory frame to be matched adjacent to the current alignment frame, and use the adjacent trajectory frame after the pose adjustment as the new current trajectory frame to be matched, and return to step S210 until After traversing all the trajectory frames to be matched in the current trajectory to be matched, the current alignment result corresponding to the current matching process is obtained, and step S270 is continued.

S270. Traverse the temporarily stored target trajectory frames that fail to match in sequence. For any target trajectory frame that fails to match, obtain the optimized pose of the target trajectory frame, and return to step S210 until all temporarily stored failed matching targets are traversed. The target trajectory frame is obtained, the new current alignment result is obtained, and step S280 is continued.

S280. Use the driving trajectory corresponding to the new current alignment result as the new current reference trajectory, select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and return to execute the new The matching operation between the current reference trajectory and the new current trajectory to be matched is to return to step S210 until all parking lot driving trajectories are traversed and the target matching result of the vehicle pose is obtained.

In this embodiment, after completing the matching between the current reference trajectory and the current trajectory to be matched, many frames to be matched have been correctly matched. The relative pose between the trajectory to be matched and the reference trajectory is more accurate than before matching. Based on this , by querying and matching the reference trajectory frame of the corresponding position again for the target trajectory frame that failed to match, the alignment result of the vehicle pose can be made more accurate. In addition, by using the matching results obtained each time as a new reference trajectory and repeatedly matching the new reference trajectory with the new trajectory to be matched, the driving trajectories collected multiple times can be accurately associated. Compared with the method of manually correlating parking lot data collected multiple times in the related art, the method provided in this embodiment further improves the efficiency and accuracy of parking lot data correlation.

Embodiment 3

Figure 3 is a structural block diagram of a parking lot data matching device provided in Embodiment 3 of the present invention. As shown in Figure 3, the device includes: a driving trajectory selection module 310, a matching module 320 and a data association module 330, where,

The driving trajectory selection module 310 is configured to arbitrarily select the vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, and use one of the driving trajectories as the current reference trajectory and the other driving trajectory as the current trajectory to be matched. , where each driving trajectory includes the trajectory before entering the parking lot, the trajectory in the parking lot, and the trajectory after exiting the parking lot;

The matching module 320 is configured to match the current reference trajectory with the current trajectory to be matched. During the matching process, based on the alignment frame matching the current reference trajectory with the vehicle pose in the current trajectory to be matched, the current trajectory to be matched. The relative pose between each trajectory frame to be matched, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, adjust the pose of each trajectory frame to be matched in turn, and the adjusted trajectory frame Perform pose alignment with the corresponding reference trajectory frame in the current reference trajectory to obtain the current alignment result corresponding to the current matching process;

The data association module 330 is configured to use the driving trajectory corresponding to the current alignment result as a new current reference trajectory, select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as a new current trajectory to be matched, and return to perform a matching operation between the new current reference trajectory and the new current trajectory to be matched until a predetermined number of parking lot driving trajectory matching operations are completed to obtain a target matching result of the vehicle posture.

Optional, matching module 320 includes:

Optional, matching units include:

The matching subunit is configured to determine that the current reference trajectory frame matches the current trajectory frame to be matched successfully if the parking lot local feature information corresponding to the current reference trajectory frame is consistent with the parking lot local feature information corresponding to the current trajectory frame to be matched; or

If the parking lot local feature information corresponding to the current reference trajectory frame is inconsistent with the parking lot local feature information corresponding to the current trajectory frame to be matched, it is determined that the current reference trajectory frame fails to match the current trajectory frame to be matched.

Correspondingly, the data association module is specifically configured as:

Optional, the current reference trajectory frame determination unit includes:

or,

Optional, same-layer judgment subunits include:

The same layer determination component is configured to determine that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot if there is no slope segment in the shortest topological path; or,

If there is a slope segment in the shortest topological path, and if the difference between the height information of the uphill segment and the height information of the downhill segment in the slope segment is within the second set distance range, determine the candidate reference trajectory frame and the current trajectory to be matched. The frame belongs to the same level of the parking lot; or,

Determine whether the pitch angle of the vehicle relative to the horizontal plane is greater than a preset angle;

Optional, the first attitude optimization unit is specifically configured as:

Construct a first topology graph corresponding to the current trajectory to be matched, and a second topology graph corresponding to the current reference trajectory, where the nodes of the first topology graph are each trajectory frame to be matched, and adjacent trajectory frames to be matched are connected by edges, The nodes of the second topological graph are each reference trajectory frame, and adjacent reference trajectory frames are connected by edges;

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched, through the pose The graph optimization method optimizes the first topology graph, and uses the optimized trajectory frame to be matched and the successfully matched reference trajectory frame as the alignment frame, where the alignment frame node in the first topology graph and the second topology graph Connected by edges.

Optionally, the device provided in this embodiment also includes:

The initial optimization module is configured to optimize the first topological map and the second topological map respectively based on the absolute position information before entering the parking lot and after exiting the parking lot before matching the current reference trajectory with the current trajectory to be matched, so as to reduce the deviation in absolute position between the current trajectory to be matched and the current reference trajectory.

The parking lot data matching device provided by the embodiment of the present invention can execute the parking lot data matching method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method. For technical details that are not described in detail in the above embodiments, please refer to the parking lot data matching method provided by any embodiment of the present invention.

Embodiment 4

Figure 4 is a structural block diagram of an electronic device provided in Embodiment 4 of the present invention. As shown in Figure 4, the electronic device includes:

Memory 510 storing executable program code;

processor 520 coupled to memory 510;

The processor 520 calls the executable program code stored in the memory 510 to execute the parking lot data matching method provided by any embodiment of the present invention.

Embodiment 5

Based on the above embodiments, another embodiment of the present invention provides a vehicle, which includes the device as described in any of the above embodiments, or includes the electronic device as described above.

Figure 5 is a schematic diagram of a vehicle provided in Embodiment 5 of the present invention. As shown in Figure 5, the vehicle includes a speed sensor 61, an ECU (Electronic Control Unit) 62, a GPS (Global Positioning System) positioning device 63, and a T-Box (Telematics Box). 64. Among them, the speed sensor 61 is used to measure the vehicle speed and uses the vehicle speed as the empirical speed for model training; the GPS positioning device 63 is used to obtain the current geographical location of the vehicle; the T-Box64 can be used as a gateway to communicate with the server; the ECU62 can perform the above-mentioned parking Field data matching method.

In addition, the vehicle may also include: a V2X (Vehicle-to-Everything) module 65, a radar 66 and a camera 67. The V2X module 65 is used to communicate with other vehicles, roadside equipment, etc.; the radar 66 or the camera 67 is used to sense the road environment information in the front and/or other directions to obtain original point cloud data; the radar 66 and/or the camera 67 can be configured at the front and/or rear of the vehicle body.

Based on the above method embodiments, another embodiment of the present invention provides a storage medium on which executable instructions are stored. When executed by the processor, the instructions enable the processor to implement the parking lot data as described in any of the above embodiments. matching method.

Those of ordinary skill in the art can understand that the accompanying drawing is only a schematic diagram of an embodiment, and the modules or processes in the accompanying drawing are not necessarily necessary for implementing the present invention.

Those of ordinary skill in the art can understand that the modules in the device in the embodiment may be distributed in the device in the embodiment according to the description of the embodiment, or may be correspondingly changed and located in one or more devices different from this embodiment. The modules of the above embodiments can be combined into one module, or further divided into multiple sub-modules.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be used Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent substitutions are made to some of the technical features; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

A parking lot data matching method, which is characterized by including:

Randomly select the vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, and use one of the driving trajectories as the current reference trajectory and the other driving trajectory as the current trajectory to be matched, where each driving trajectory includes The trajectory before entering the parking lot, the trajectory in the parking lot and the trajectory after exiting the parking lot;

The current reference trajectory is matched with the current trajectory to be matched. During the matching process, based on the alignment frame matching the current reference trajectory with the vehicle pose in the current trajectory to be matched, the current trajectory to be matched. The relative pose between each trajectory frame to be matched, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, adjust the pose of each trajectory frame to be matched in turn, and add the adjusted trajectory The frame is aligned with the corresponding reference trajectory frame in the current reference trajectory to obtain the current alignment result corresponding to the current matching process;

Use the driving trajectory corresponding to the current alignment result as the new current reference trajectory, select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and return to execute the new current reference trajectory. The reference trajectory is matched with the new current trajectory to be matched until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.
The method according to claim 1 is characterized in that, based on the alignment frame matching the current reference trajectory with the vehicle posture in the current trajectory to be matched, the relative posture between each trajectory frame to be matched in the current trajectory to be matched, and the absolute position information of the vehicle before entering and leaving the parking lot, each trajectory frame to be matched is adjusted in turn, and the trajectory frame after the posture adjustment is aligned with the corresponding reference trajectory frame in the current reference trajectory to obtain the current alignment result corresponding to the current matching process, including:

Sequentially traverse each to-be-matched trajectory frame in the current to-be-matched trajectory, and for each current to-be-matched trajectory frame, determine the current reference trajectory frame corresponding to the current to-be-matched trajectory frame position in the current reference trajectory;

Match the current reference trajectory frame with the current trajectory frame to be matched;

If the matching is successful, determine the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame;

Based on the first relative posture, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after leaving the parking lot, and the second relative posture between each trajectory frame to be matched in the current driving trajectory to be matched, the posture of all trajectory frames to be matched that have been successfully matched are optimized, and the trajectory frame to be matched after the posture optimization and the reference trajectory frame corresponding to the position of the successful match are used as alignment frames;

Based on the current alignment frame, adjust the posture of the trajectory frame to be matched adjacent to the current alignment frame, and use the posture-adjusted adjacent trajectory frame as the new current trajectory frame to be matched, and return to execute the new current trajectory frame. The matching operation of the trajectory frame to be matched and the current reference trajectory frame corresponding to the position is performed until all the trajectory frames to be matched in the current trajectory to be matched are traversed, and the current alignment result of the vehicle pose in the current matching process is obtained.
The method according to claim 2, characterized in that, after determining the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, the method further includes:

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched. , perform pose optimization on all reference trajectory frames in the current reference trajectory, and use the pose-optimized trajectory frame to be matched and the successfully matched reference trajectory frame after pose optimization as an alignment frame.
The method according to claim 2, wherein matching the current reference trajectory frame with the current trajectory frame to be matched includes:

Obtain local feature information of the parking lot, which includes: traffic signs, wall markings, road markings and lane markings;

If the local feature information of the parking lot corresponding to the current reference trajectory frame is consistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched, it is determined that the current reference trajectory frame matches the current trajectory frame to be matched. Success; or,

If the local feature information of the parking lot corresponding to the current reference trajectory frame is inconsistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched, it is determined that the current reference trajectory frame matches the current trajectory frame to be matched. fail.
The method according to any one of claims 2-4, characterized in that the method further includes:

If the current reference trajectory frame fails to match the current trajectory frame to be matched, the current trajectory frame to be matched that fails to match is temporarily stored;

After obtaining the current alignment result, the temporarily stored target trajectory frames that failed to match are traversed in sequence. For any target trajectory frame that fails to match, based on the optimized pose of the target trajectory frame, the optimized pose is repeatedly executed. The matching operation of the target trajectory frame and the reference trajectory frame corresponding to the position is performed until all temporarily stored target trajectory frames that have failed to match are traversed, and a new current alignment result is obtained;

Correspondingly, the driving trajectory corresponding to the current alignment result is used as a new current reference trajectory, and another unmatched driving trajectory is selected from the parking lot driving trajectories collected multiple times as a new current trajectory to be matched, and the matching operation between the new current reference trajectory and the new current trajectory to be matched is returned, including:

The driving trajectory corresponding to the new current alignment result is used as a new current reference trajectory, and another unmatched driving trajectory is selected from the parking lot driving trajectories collected multiple times as a new current trajectory to be matched, and the matching operation between the new current reference trajectory and the new current trajectory to be matched is returned.
The method according to claim 2, wherein determining the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched in the current reference trajectory includes:

When the current trajectory frame to be matched belongs to the trajectory before entering the parking lot or the trajectory after exiting the parking lot:

Obtain the absolute position corresponding to the current trajectory frame to be matched, and search within the first set distance range from the absolute position in the current reference trajectory to obtain the current reference corresponding to the position of the current trajectory frame to be matched. track frame;

Or, when the current trajectory frame to be matched belongs to the trajectory in the parking lot:

According to the current position information corresponding to the current trajectory frame to be matched, in the current reference trajectory, search within the first set distance range from the current position information to obtain a candidate reference trajectory frame;

Determining whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same floor of the parking lot;

If they belong to the same layer, the candidate reference trajectory frame is used as the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched.
The method according to claim 6, wherein determining whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot includes:

Determine the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame;

Detect whether there is a slope segment in the shortest topological path;

If there is no slope segment, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot; or,

If there is a slope segment, and if the difference between the height information of the uphill segment and the height information of the downhill segment in the slope segment is within the second set distance range, it is determined that the candidate reference trajectory frame is the same as the current to-be-subjected trajectory frame. Matching trajectory frames belong to the same level of the parking lot;

If the difference between the height information of the uphill section and the height information of the downhill section in the slope section exceeds the second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched do not belong to the parking lot. same floor.
The method according to claim 7, wherein detecting whether there is a slope segment in the shortest topological path includes:

Determine whether the pitch angle of the vehicle relative to the horizontal plane is greater than the preset angle;

The road segment corresponding to the trajectory frame with the pitch angle greater than or equal to the preset angle is regarded as a slope segment, and the road segment corresponding to the trajectory frame with the pitch angle smaller than the preset angle is regarded as a non-slope segment.
The method according to claim 2, characterized in that, based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, the current driving to be matched The second relative pose between the trajectory frames to be matched in the trajectory, and pose optimization for all trajectory frames to be matched that have been successfully matched, including:

Constructing a first topological graph corresponding to the current trajectory to be matched and a second topological graph corresponding to the current reference trajectory, wherein the nodes of the first topological graph are the trajectory frames to be matched, and adjacent trajectory frames to be matched are connected by edges, and the nodes of the second topological graph are the reference trajectory frames, and adjacent reference trajectory frames are connected by edges;

Based on the first relative pose, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the second relative pose between each to-be-matched trajectory frame in the current driving trajectory to be matched. , the first topology map is optimized through the pose graph optimization method, and the pose-optimized trajectory frame to be matched and the successfully matched reference trajectory frame are used as alignment frames, where the first topology map and the second topology map The aligned frame nodes in are connected by edges.
The method according to claim 9, characterized in that, before matching the current reference trajectory with the current trajectory to be matched, the method further includes:

Based on the absolute position information before entering the parking lot and after exiting the parking lot, the first topology map and the second topology map are optimized respectively to narrow the absolute position between the current trajectory to be matched and the current reference trajectory. deviation.
The method according to claim 1, characterized in that the vehicle posture corresponding to each reference trajectory frame or the vehicle posture corresponding to each trajectory frame to be matched is obtained through any one of the following sensor data, or by combining the following multiple types of sensor data: The sensor data is fused to obtain:

Inertial measurement unit IMU data, image data, radar point cloud data and odometer data.
A parking lot data matching device, which is characterized by including:

The driving trajectory selection module is configured to arbitrarily select the vehicle driving trajectories collected at different times from the parking lot driving trajectories collected multiple times, and use one of the driving trajectories as the current reference trajectory and the other driving trajectory as the current trajectory to be matched. Among them, each driving trajectory includes the trajectory before entering the parking lot, the trajectory in the parking lot, and the trajectory after exiting the parking lot;

A matching module configured to match the current reference trajectory with the current trajectory to be matched. During the matching process, based on the alignment frame that matches the vehicle pose in the current reference trajectory and the current trajectory to be matched, The relative pose between each track frame to be matched in the current track to be matched, as well as the absolute position information of the vehicle before entering the parking lot and after exiting the parking lot, adjust the pose of each track frame to be matched in turn, and The posture-adjusted trajectory frame is aligned with the corresponding reference trajectory frame in the current reference trajectory to obtain the current alignment result corresponding to the current matching process;

The data association module is configured to use the driving trajectory corresponding to the current alignment result as the new current reference trajectory, and select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched. , and return to perform the matching operation of the new current reference trajectory and the new current trajectory to be matched, until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.
The device according to claim 12, characterized in that the matching module includes:

The current reference trajectory frame determination unit is configured to match the current reference trajectory with the current trajectory to be matched, and during the matching process, sequentially traverse each trajectory frame to be matched in the current trajectory to be matched, for each The current trajectory frame to be matched, the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched is determined in the current reference trajectory;

A matching unit configured to match the current reference trajectory frame with the current trajectory frame to be matched;

A relative pose determination unit configured to determine a first relative position between the current trajectory frame to be matched and the corresponding current reference trajectory frame if the current reference trajectory frame and the current trajectory frame to be matched are successfully matched. posture;

The first posture optimization unit is configured to be based on the first relative posture, the absolute position information of the vehicle in each driving trajectory before entering the parking lot and after exiting the parking lot, and the current driving trajectory to be matched. For the second relative pose between trajectory frames, pose optimization is performed on all trajectory frames to be matched that have been successfully matched, and the optimized trajectory frame to be matched is used as the alignment with the reference trajectory frame corresponding to the position where the matching is successful. frame;

The pose alignment unit is configured to adjust the pose of the trajectory frame to be matched adjacent to the current alignment frame based on the current alignment frame, and use the adjacent trajectory frame after the pose adjustment as the new current trajectory frame to be matched. , and return to perform the matching operation of matching the new current trajectory frame to be matched with the current reference trajectory frame corresponding to the position, until all the trajectory frames to be matched in the current trajectory to be matched are traversed, and the vehicle pose in the current matching process is obtained. the current alignment result.
The device according to claim 13, characterized in that the device further includes:

The second pose optimization unit is configured to, after determining the first relative pose between the current trajectory frame to be matched and the corresponding current reference trajectory frame, based on the first relative pose, the vehicle in each driving trajectory The absolute position information before entering the parking lot and after exiting the parking lot, the second relative pose between the trajectory frames to be matched in the current driving trajectory to be matched, and all reference trajectory frames in the current reference trajectory are processed. The pose is optimized, and the trajectory frame to be matched after the pose is optimized and the successfully matched reference trajectory frame after the pose is optimized are used as alignment frames.
The device according to claim 13, characterized in that the matching unit includes:

The local feature information acquisition subunit is configured to acquire local feature information of the parking lot, where the local feature information includes: traffic signs, wall markings, road markings and lane markings;

The matching subunit is configured to determine that the current reference trajectory frame is consistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched. The current trajectory frame to be matched is successfully matched; or,

If the local feature information of the parking lot corresponding to the current reference trajectory frame is inconsistent with the local feature information of the parking lot corresponding to the current trajectory frame to be matched, it is determined that the current reference trajectory frame matches the current trajectory frame to be matched. fail.
The device according to any one of claims 13 to 15, characterized in that the device further comprises:

The matching failed frame temporary storage module is configured to temporarily store the failed matching current trajectory frame to be matched if the current reference trajectory frame fails to match the current trajectory frame to be matched;

The re-matching module is configured to sequentially traverse the temporarily stored target trajectory frames that failed to match. For any target trajectory frame that failed to match, based on the optimized pose of the target trajectory frame, repeatedly execute the optimized target trajectory. The matching operation of the reference trajectory frame corresponding to the frame and position is performed until all temporarily stored target trajectory frames that have failed to match are traversed and a new current alignment result is obtained;

Correspondingly, the data association module is specifically configured as:

Use the driving trajectory corresponding to the new current alignment result as the new current reference trajectory, select another unmatched driving trajectory from the parking lot driving trajectories collected multiple times as the new current trajectory to be matched, and return to execute the new The current reference trajectory and the new current trajectory to be matched are matched until a predetermined number of parking lot driving trajectory matching operations are completed, and the target matching result of the vehicle pose is obtained.
The device according to claim 13, characterized in that the current reference trajectory frame determination unit includes:

The first search subunit is configured to obtain the absolute position corresponding to the current trajectory frame to be matched when the current trajectory frame to be matched belongs to the trajectory before entering the parking lot or the trajectory after exiting the parking lot, and obtain the absolute position corresponding to the current trajectory frame to be matched in the current reference In the trajectory, search within the first set distance range from the absolute position to obtain the current reference trajectory frame corresponding to the position of the current trajectory frame to be matched;

or,

The second search subunit is configured to, when the current trajectory frame to be matched belongs to the trajectory in the parking lot, based on the current location information corresponding to the current trajectory frame to be matched, in the current reference trajectory, within the distance from the Search within the first set distance range of the current position information to obtain candidate reference trajectory frames;

The same layer determination subunit is configured to determine whether the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot;

The current reference trajectory frame determination subunit is configured to use the candidate reference trajectory frame as the current trajectory frame to be matched if the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot. The current reference trajectory frame corresponding to the position.
The device according to claim 17, characterized in that the same layer determination subunit includes:

a shortest topological path determination component configured to determine the shortest topological path from the current trajectory frame to be matched to the candidate reference trajectory frame;

a slope segment detection component configured to detect whether there is a slope segment in the shortest topological path;

A same-layer determination component configured to determine that the candidate reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot if there is no slope segment in the shortest topological path;

Alternatively, if there is a slope segment in the shortest topological path, and if the difference between the height information of the uphill segment and the height information of the downhill segment in the slope segment is within a second set distance range, determine the candidate The reference trajectory frame and the current trajectory frame to be matched belong to the same layer of the parking lot;

Alternatively, if the difference between the height information of the uphill section and the height information of the downhill section in the slope section exceeds the second set distance range, it is determined that the candidate reference trajectory frame and the current trajectory frame to be matched do not belong to parking. same layer of the field.
The device according to claim 18, wherein the peer determination component is specifically configured as:

Determine whether the pitch angle of the vehicle relative to the horizontal plane is greater than the preset angle;

The road segment corresponding to the trajectory frame with the pitch angle greater than or equal to the preset angle is regarded as a slope segment, and the road segment corresponding to the trajectory frame with the pitch angle smaller than the preset angle is regarded as a non-slope segment.
The device according to claim 13, characterized in that the first attitude optimization unit is specifically configured to:

Construct a first topology graph corresponding to the current trajectory to be matched, and a second topology graph corresponding to the current reference trajectory, wherein the nodes of the first topology graph are each trajectory frame to be matched, and between adjacent trajectory frames to be matched Connected by edges, the nodes of the second topology graph are each reference trajectory frame, and adjacent reference trajectory frames are connected by edges;

Based on the first relative posture, the absolute position information of the vehicle in each driving trajectory before entering and exiting the parking lot, and the second relative posture between each trajectory frame to be matched in the current driving trajectory to be matched, the first topological graph is optimized by a posture graph optimization method, and the trajectory frame to be matched after posture optimization and the successfully matched reference trajectory frame are used as alignment frames, wherein the alignment frame nodes in the first topological graph and the second topological graph are connected by edges.
An electronic device, characterized by including:

one or more processors;

a storage device for storing one or more programs,

When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method as described in any one of claims 1-11.
A storage medium on which a computer program is stored, characterized in that when the program is executed by a processor, the method according to any one of claims 1-11 is implemented.
A vehicle, characterized in that the vehicle includes the system according to any one of claims 1 to 11, or the electronic device according to claim 21.