Method and system for automatic parking in connection with digital roads

Abstract

The application discloses an automatic parking method and system combined with a digital road and relates to the technical field of automatic parking. The method comprises the following steps: acquiring a layout data source of the digital road, establishing a feature association relationship between a digital twin space and the layout data source and a vehicle communication data source, positioning an interactive parking vehicle in a region based on the digital twin space, establishing an intention sharing perception subspace, performing path time sequence tracking on the interactive parking vehicle, performing self-verification and interactive verification on perception data, and identifying an abnormal space; according to the abnormal space, performing space risk evaluation by using data results of the self-verification and the interactive verification of the perception data; and when a space risk evaluation result meets a parking space requirement, performing collaborative parking control according to a space verification result and a risk evaluation result. The technical problem that the interaction between vehicles is insufficient during automatic parking in the prior art, resulting in low parking safety and reliability is solved, and the technical effect of improving parking safety and reliability is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic parking, in particular to an automatic parking method and system combined with a digital road. BACKGROUND

[0002] Automatic parking, as one of the important functions of assisted driving, greatly improves the driving experience. However, the existing automatic parking system still has problems such as insufficient information interaction, low environmental perception accuracy, and limited risk prediction ability in complex road environments or multi-vehicle parking scenarios. Due to insufficient understanding of the surrounding environment, obstacles, and the behavior of other vehicles, the vehicle may not avoid obstacles in time, and the parking action may be unstable during the parking path planning and execution process, thereby affecting the safety and reliability of parking. SUMMARY

[0003] The present application provides an automatic parking method and system combined with a digital road, which solves the technical problem of insufficient interaction between vehicles during automatic parking in the prior art, resulting in low safety and reliability of parking.

[0004] In a first aspect, the present application provides an automatic parking method combined with a digital road, the method comprising:

[0005] Obtaining a layout data source of a digital road, establishing a feature association relationship between a digital twin space and the layout data source and a vehicle communication data source; positioning an interactive parking vehicle in a region based on the digital twin space, establishing an intention sharing perception subspace, tracking the path timing of the interactive parking vehicle in the intention sharing perception subspace, and performing perception data self-verification and interaction verification to identify an abnormal space; based on the abnormal space, performing spatial risk evaluation using data results of the perception data self-verification and interaction verification; when the spatial risk evaluation result meets the parking space requirement, performing collaborative parking control based on the spatial verification result and the risk evaluation result.

[0006] In a second aspect, the present application provides an automatic parking system combined with a digital road, the system comprising:

[0007] The data source association module acquires the layout data source of the digital road, establishes a feature association relationship between the digital twin space and the layout data source and the vehicle communication data source; the tracking module: based on the digital twin space, locates the interactive parking vehicle in the region, establishes an intention sharing perception subspace, performs path time sequence tracking on the interactive parking vehicle in the intention sharing perception subspace, and performs perception data self-verification and interaction verification to identify an abnormal space; the risk evaluation module: according to the abnormal space, using the data results of the perception data self-verification and interaction verification to perform space risk evaluation; the parking control module: when the space risk evaluation result meets the parking space requirement, according to the space verification result and the risk evaluation result, performs collaborative parking control.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] Firstly, the layout data source of the digital road is acquired, and a feature association relationship between the digital twin space and the layout data source and the vehicle communication data source is established. Then, based on the digital twin space, the interactive parking vehicle in the region is located, an intention sharing perception subspace is established, path time sequence tracking is performed on the interactive parking vehicle in the intention sharing perception subspace, and perception data self-verification and interaction verification are performed to identify an abnormal space. Then, according to the abnormal space, using the data results of the perception data self-verification and interaction verification to perform space risk evaluation. Finally, when the space risk evaluation result meets the parking space requirement, according to the space verification result and the risk evaluation result, performs collaborative parking control. The technical problem of low parking safety and reliability caused by insufficient interaction between vehicles during automatic parking in the prior art is solved, and the technical effect of significantly improving parking safety and reliability by constructing a digital twin space and jointly verifying multi-source perception data to realize efficient interaction between vehicles and controllable space risk is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0011] Figure 1 The automatic parking method flowchart combined with the digital road provided by the embodiment of the present application is shown in the figure.

[0012] Figure 2 The automatic parking system structure combined with the digital road provided by the embodiment of the present application is shown in the figure.

[0013] Explanation of reference signs: data source association module 11, tracking module 12, risk evaluation module 13, parking control module 14. DETAILED DESCRIPTION

[0014] The present application solves the technical problem of low parking safety and reliability caused by insufficient interaction between vehicles during automatic parking in the prior art by providing an automatic parking method and system combined with a digital road.

[0015] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0016] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] Embodiment one, as shown in the present application provides an automatic parking method combined with a digital road, wherein the method comprises: Figure 1 Obtaining the layout data source of the digital road, establishing a feature association relationship between the digital twin space and the layout data source and the vehicle communication data source.

[0018]

[0019] ​In the embodiments of the present application, road structure information, road marking information, parking space position and size information, traffic sign information, road slope and road surface material information, real-time traffic state information and other original data related to the target parking area are obtained from various data providing ends such as road infrastructure management platform, road side unit (RSU), high-precision map server, etc., and the data is uniformly processed in format and corrected in time stamp to ensure the spatio-temporal alignment of data from different sources; after the data preprocessing is completed, a digital twin space is constructed, the physical space of the target parking area is virtually modeled in the digital environment according to a unified coordinate system, and the road structure elements and environmental elements corresponding to the above layout data sources are mapped in the model; further, vehicle communication data sources are called, including real-time position information, speed information, acceleration information, steering angle information, vehicle body attitude information, sensor detection information and other dynamic running data from the vehicle terminal, through feature extraction and multi-source association algorithm, the feature parameters in the vehicle communication data source are matched and associated with the corresponding road, parking space and environmental features in the digital twin space, forming the feature association relationship between the digital twin space and the layout data source and the vehicle communication data source, thereby providing basic data support for subsequent parking path analysis, perception data verification and cooperative control.

[0020] Based on the digital twin space, the interactive parking vehicles are positioned, an intention sharing perception subspace is established, the interactive parking vehicles are tracked in path time sequence in the intention sharing perception subspace, and the perception data is self-verified and interactively verified to identify abnormal space.

[0021] Based on the digital twin space, the interactive parking vehicles in the target parking area are positioned, specifically including using the real-time position, speed, attitude and perception information provided by the vehicle communication data source, through coordinate conversion and time synchronization, mapping the vehicle running state to the unified coordinate system of the digital twin space, determining the position and motion trajectory of each interactive parking vehicle in the virtual space; after positioning is completed, according to the spatial proximity, motion trend and potential parking target area between vehicles, an intention sharing perception subspace is constructed, which is used to carry the dynamic behavior information and environmental perception data of multiple interactive parking vehicles in a specific area; in the intention sharing perception subspace, path time sequence tracking is performed on the interactive parking vehicles, that is, within a set time window, the historical and predicted path point sequences of each vehicle are continuously recorded and analyzed, and path time sequence data containing time stamp, spatial coordinate and motion state are generated; at the same time, the continuity of the vehicle's own perception data in the time dimension is verified, the consistency anomaly of the perception information in the longitudinal sequence is identified, and the perception results of different vehicles in the spatial perception overlapping area are interactively verified to judge the conflict or loss of the perception data in the horizontal multi-source comparison; when the self-verification and interactive verification both determine that there is a perception anomaly in a certain spatial area, the area is marked as an abnormal space, which provides a basis for subsequent risk assessment and cooperative parking control.

[0022] Further, in the intention sharing perception subspace, the interactive parking vehicles are tracked in path time sequence, and the perception data is self-verified and interactively verified to identify abnormal spaces, including:

[0023] Synchronize the layout data source and the vehicle communication data source of each interactive parking vehicle; perform time sequence longitudinal comparison according to the layout data source of each parking vehicle to obtain self-verification data of the perception data, generate a self-verification abnormal label; according to the overlapping perception verification relationship of the layout data source and the vehicle communication data source of the interactive parking vehicle, obtain the interactive verification data of the perception data, generate an interactive verification abnormal label; if the self-verification abnormal label and the interactive verification abnormal label are activated at the same time, mark the overlapping perception area as a confirmed abnormal space.

[0024] The layout data source and the vehicle communication data source of each interactive parking vehicle in the target parking area are time-synchronized and space-aligned, wherein the layout data source includes environmental perception data provided by road infrastructure, and the vehicle communication data source includes multi-modal perception data collected by vehicle sensors and vehicle operating state parameters, and through unified timestamp and coordinate system, the spatio-temporal consistency between different data sources is ensured; for each parking vehicle, based on the historical perception frame sequence of the layout data source in the continuous time window, longitudinal time sequence comparison is performed, and consistency indexes in spatial coordinates, perception target types, target boundaries, etc. between adjacent frames are calculated, when the consistency index exceeds the dynamic threshold or shows a mutation trend, it is determined that the perception is abnormal, and a self-verification abnormal label is generated in the perception data record of the vehicle; according to the layout data source and the vehicle communication data source of multiple interactive parking vehicles, a spatial overlap verification relationship of perception results is established, that is, a spatial intersection region of multiple vehicle perception results is identified in an intent sharing perception subspace, and consistency detection is performed on the perception results in the region, when there is data conflict, missing or significant difference, an interactive verification abnormal label is generated; if a certain spatial region triggers the self-verification abnormal label and the interactive verification abnormal label at the same time, the spatial region is marked as a confirmed abnormal space.

[0025] Further, according to the layout data source of each parking vehicle, time sequence longitudinal comparison is performed to obtain perception data self-verification data, generate a self-verification abnormal label, including:

[0026] Extracting a perception data frame sequence in a continuous time window of each parking vehicle; based on the perception data frame sequence, calculating the spatial consistency index of adjacent frames, if the index exceeds the dynamic threshold, determining that the vehicle perception fails; using longitudinal historical time sequence data frames for perception data failure tracing, positioning abnormal data spatial coordinates, and generating the self-verification abnormal label. The perception data frame sequence in the continuous time window is extracted from the layout data source of each parking vehicle, wherein the perception data frame sequence includes road environment information, target object position, shape and attribute parameters collected by the vehicle at adjacent time points.

[0027] Based on the sequence of perception data frames, the same type of target in adjacent frames is matched, and the spatial consistency index of its position coordinates, attitude angle, size boundary and detection confidence is calculated. The consistency index can be obtained by weighted calculation of multi-dimensional features such as Euclidean distance, attitude difference and boundary overlap rate. When the spatial consistency index exceeds the dynamic threshold adaptively adjusted according to the scene complexity and vehicle dynamic characteristics, it is determined that the vehicle has perception failure at the corresponding time point, that is, the perception result of the current frame and the continuous historical frame have abnormal deviation in spatial performance. On this basis, the longitudinal historical time sequence data frame is called to trace the source of perception data failure, the starting time and duration of abnormal deviation are analyzed by backtracking multiple frames of data before the abnormal frame, and the corresponding abnormal data space coordinates are located by using the coordinate projection method mapped with the digital twin space. Finally, the abnormal space coordinates and their related perception failure information are recorded as self-verification abnormal labels.

[0028] Further, according to the overlapping perception verification relationship of the layout data source and the vehicle communication data source of the interactive parking vehicle, the perception data interaction verification data is obtained, and the interaction verification abnormal label is generated, including:

[0029] Based on the layout data source and the vehicle communication data source of the interactive parking vehicle, the signal distribution map of each interactive vehicle is established; the signal distribution map of the interactive vehicle is overlapped according to the intention sharing perception subspace relationship by using path time sequence tracking, and the conflict perception data in the overlapping space range is identified; based on the perception credibility of the conflict perception data, the collaborative verification is carried out, the abnormal perception space coordinates of the abnormal parking vehicle are located, and the interaction verification abnormal label is generated.

[0030] Based on the layout data source of each interactive parking vehicle and the vehicle communication data source, a signal distribution map of the vehicle is established in a unified digital twin space coordinate system, and the signal distribution map is used to describe the target distribution, signal strength, detection confidence and timestamp information of the vehicle within the current perception range; using path timing tracking results, the signal distribution maps of multiple interactive vehicles are superimposed and reconstructed according to the spatial boundary and time correlation of the intention sharing perception subspace, to obtain the alignment result of the perception data of each vehicle in the same space region and time period; in the superimposed result, the conflict perception data in the overlapping space range is detected, and the conflict perception data includes significant differences in target category, spatial coordinates, size or attitude of the perception results of the same space position by different vehicles, or some vehicles have perception missing in the space region; the perception data is verified based on the perception confidence of the conflict perception data, wherein the perception confidence can be calculated by weighting factors such as sensor type, historical detection accuracy, consistency degree with other vehicle data, etc.; by comparing the perception confidence of different vehicles, the conflict data is screened for credibility, the perception result with a credibility lower than a threshold is determined, and the abnormal parking vehicle and its abnormal perception space coordinates are located according to the abnormal perception space coordinates and the corresponding vehicle identifier; the abnormal perception space coordinates and the corresponding vehicle identifier are generated to generate an interactive verification abnormal label.

[0031] According to the abnormal space, the spatial risk evaluation is performed on the data results of the perception data self-verification and interactive verification.

[0032] Further, according to the abnormal space, the spatial risk evaluation is performed on the data results of the perception data self-verification and interactive verification, including:

[0033] According to the abnormal perception data source and the abnormal deviation amount of the abnormal perception space coordinates, the path time and space abnormal positioning is performed in combination with the parking path interaction space of the interactive parking vehicle; taking the parking vehicle as a target, the path time and space abnormal positioning is fitted in the digital twin space, to obtain the risk probability and risk path node corresponding to the abnormal perception data of the vehicle in the parking path; according to the risk path node, risk space coordinates and risk probability of the parking vehicle, the spatial risk evaluation result is obtained.

[0034] Firstly, the abnormal perception spatial coordinates determined by self-verification and interactive verification are obtained, and the abnormal perception data source information corresponding to the spatial coordinates is called, the data source information including sensor type, collection time, detection object category, spatial position and deviation parameter when the abnormality occurs; then, the abnormal deviation amount of the abnormal perception data in spatial coordinates, target size, attitude and the like relative to the normal perception result is extracted, and the abnormal positioning in path time and space dimensions is performed in combination with the parking path interaction space data of the interactive parking vehicle, wherein the path time abnormal positioning is used to determine the specific time and duration when the abnormality occurs, and the spatial abnormal positioning is used to determine the accurate position range of the abnormality on the parking path; then, taking the parking vehicle as an analysis target, the path time and spatial abnormal positioning results are mapped into the digital twin space, the spatial state of the vehicle when the abnormality occurs is determined through path fitting and trajectory playback, and the risk probability of the abnormal position is calculated based on the abnormal deviation amount and the historical perception accuracy, and the corresponding risk path node is marked; finally, the risk path node, risk spatial coordinates and risk probability of the parking vehicle are comprehensively considered to generate the spatial risk evaluation result, and the spatial risk evaluation result includes the risk level distribution, cumulative risk probability and weight value of the risk space of each risk path node.

[0035] In the spatial risk evaluation, the risk probability of each risk path node can be calculated based on the abnormal deviation amount weight, historical perception accuracy and path importance coefficient, and the formula is as follows: wherein, is an abnormal deviation amount vector, including spatial position deviation , attitude deviation and size deviation ; is a deviation amount normalization function, used for uniformly mapping the deviations of different dimensions to the interval [0, 1], and can be expressed as: wherein, , are the allowed thresholds of each deviation dimension, , , are the weight coefficients of each dimension and add up to 1; is the historical perception accuracy, with a value range of [0, 1], which can be obtained by comparing the historical perception result with the labeled data; is the path importance coefficient, with a value range of [0, 1], which can be calculated comprehensively according to the position criticality, traffic substitution and traffic flow influence degree of the path node in the parking process; , , are the deviation amount weight, accuracy weight and path importance weight respectively, and satisfy the sum of 1.

[0036] Further, the spatial risk evaluation is performed using the data results of the self-verification and interaction verification of the perception data, and then further includes:

[0037] When the spatial risk evaluation result does not meet the parking space requirement, the risk space orientation and risk probability are identified, the parking path of the interactive parking vehicle and the path interaction space are connected, the path risk timing analysis is performed on the parking path and the path interaction space according to the risk space orientation and the risk probability, and the risk timing chain is obtained, wherein the risk timing chain has a risk space coordinate identifier and a space risk probability; and the corresponding parking path node is adaptively adjusted according to the corresponding risk space coordinate and the space risk probability of the risk timing chain, and the parking control mode is switched.

[0038] When the spatial risk evaluation result is lower than the preset parking space availability threshold, it is determined that the current parking space does not meet the parking space requirement, and the spatial orientation information of the risk space and the corresponding risk probability value are extracted from the risk evaluation result, the spatial orientation information includes the two-dimensional or three-dimensional position coordinate range of the risk area in the digital twin space unified coordinate system and the azimuth angle relative to the vehicle parking path; then, the parking path data of the interactive parking vehicle and the path interaction space data are called, the risk space is mapped into the vehicle path node and the path interaction area; according to the orientation and risk probability of the risk space, the path risk timing analysis is performed, that is, the relationship between each node in the parking path and the risk space is dynamically calculated in the time dimension, and the risk timing chain is generated, the risk timing chain includes a plurality of risk space coordinate identifiers and corresponding space risk probability values arranged in time sequence; after obtaining the risk timing chain, the adaptive path adjustment algorithm is called for each parking path node corresponding to the risk space coordinate and the risk probability, the corresponding control constraint strategy is matched according to the risk probability level, the path node is adjusted, such as offset adjustment, path replacement or speed parameter adjustment, and the original parking control mode is switched to the parking control mode suitable for the risk level, such as low-speed fine parking mode, obstacle avoidance priority mode or safe parking mode, thereby reducing the influence of the risk space on the parking process and improving the safety and stability of parking.

[0039] Further, the adaptive adjustment of the corresponding parking path node and the switching of the parking control mode include:

[0040] An adaptive mode adjustment library is established, which includes the risk evaluation level, the corresponding risk probability, and the control constraint parameters of the parking control mode; the space risk probability of the parking path node is matched with the adaptive mode adjustment library to obtain the node parking control mode; and the node parking control mode is smoothed according to the parking path node to determine the switching of the parking control mode.

[0041] The adaptive mode adjustment library is hierarchically configured according to different risk evaluation levels, each risk evaluation level is associated with a corresponding spatial risk probability range and a parking control mode and control constraint parameters matched therewith, wherein the control constraint parameters include a maximum parking speed of the vehicle, a steering angle change rate, a brake response threshold, an acceleration limit value, an obstacle avoidance trigger distance, etc.

[0042] For each parking path node, the spatial risk probability obtained in the spatial risk evaluation stage is extracted, and the risk probability is matched with the preset risk probability range in the adaptive mode adjustment library to obtain the node parking control mode corresponding to the path node; based on the continuity principle of the parking path, the parking control modes corresponding to the path nodes are smoothed, that is, the mode transition optimization is performed between the path nodes to prevent the vehicle from being dynamically unstable due to frequent switching of the parking control mode in a short time; the actual switched parking control mode is determined according to the smoothed result, and the mode parameters are sent to the vehicle control execution unit, so that the vehicle can smoothly and safely complete the switching of the control mode when passing through the path nodes of different risk levels.

[0043] Further, the parking control mode corresponding to the parking path node is smoothed, and the switched parking control mode is determined, including:

[0044] The parking control mode time sequence consistency of the parking path node is identified; when the consistency is not satisfied, the parameter smoothing fitting is performed according to the control constraint parameters of the corresponding parking control mode based on the path switching distance relationship of the parking path node, the smoothing control parameter with the best safety is determined, and the switched parking control mode is obtained.

[0045] Based on the node sequence of the parking path, the corresponding parking control mode of each node is read in sequence, and the mode switching between adjacent nodes on the time axis is analyzed, and the time sequence consistency index of the parking control mode is calculated, which can comprehensively consider the number of mode type changes, change frequency and switching interval time and other parameters; When the time sequence consistency index is lower than the preset consistency threshold, it is determined that the current parking path does not meet the smoothness requirement in terms of control mode switching; Subsequently, the path switching distance relationship between adjacent nodes is extracted, which represents the actual vehicle driving distance required to switch from the current node to the next node control mode; On this basis, combined with the control constraint parameters (including speed limit, steering angle change rate, acceleration limit, brake response delay, etc.) of the corresponding parking control mode, the transition optimization of adjacent mode parameters is carried out by using the parameter smoothing fitting algorithm, and the smoothing fitting can adopt piecewise linear interpolation, weighted polynomial fitting or nonlinear fitting method based on vehicle dynamics model; The continuous change curve of the control variable within the path switching distance range is obtained by fitting calculation, and the smooth control parameter under the condition of comprehensive optimization of safety and comfort is determined; Finally, the smooth control parameter is applied to the node parking control mode switching execution process, realizing the gradual transition of the parking control mode, thereby avoiding the dynamic instability of the vehicle caused by the sudden change of the control mode, and improving the safety and controllability of the parking process.

[0046] Further, based on the path switching distance relationship of the parking path node, the parameter smoothing fitting is carried out according to the control constraint parameters of the corresponding parking control mode, and the smooth control parameter with the best safety is determined, including:

[0047] Based on the path switching distance relationship of the parking path node, the conflict items of the control constraint parameters of adjacent nodes are compared; According to the safety influence of the conflict items, the priority weight is set, the conflict items are calculated by weighted compromise according to the priority weight, and the transition parameter is generated; According to the close distance of the transition parameter to the control constraint parameters of the parking control mode, the parking control mode of the node is determined.

[0048] Firstly, the path switching distance of any two adjacent path nodes in the target parking path is obtained, and the path switching distance represents the spatial distance that the vehicle needs to travel in the process of switching from the previous node to the next node control mode; on this basis, the control constraint parameter set corresponding to the parking control mode of the two nodes is extracted, the control constraint parameters include the maximum allowed speed, the steering angle change rate, the maximum acceleration, the brake delay time, the obstacle avoidance trigger distance, etc., and the difference values of the adjacent node control constraint parameters are compared, and the conflict items with significant differences and possible influence on the dynamic stability or safety of the vehicle are identified; then, the priority weight is set according to the influence degree of each conflict item on the safety of the vehicle, and the influence degree can be quantitatively evaluated according to the sensitivity of the control parameter to the vehicle attitude change, the collision risk, the path deviation probability and other factors, and the weight value can be allocated in the range of [0, 1]; subsequently, the parameter difference of each conflict item is weighted and compromised according to the priority weight, and the transition parameter is generated, which represents the transition target value of the control variable within the path switching distance; finally, according to the close distance relationship (i.e. the parameter value proximity) between the transition parameter and the control constraint parameter in the target parking control mode, the control parameter that meets the safety requirement is selected as the parking control mode parameter of the current node, and the mode parameter is applied to the vehicle control execution unit, so as to realize the smooth switching and dynamic optimization of the parking control mode.

[0049] When the space risk evaluation result meets the parking space requirement, the collaborative parking control is performed according to the space verification result and the risk evaluation result.

[0050] When the risk probability of all risk path nodes in the space risk evaluation result is lower than the preset safety threshold, and the parking space size, orientation and passability meet the vehicle parking conditions, the space verification result and the risk evaluation result are associated and mapped to form the available parking path and the control safety parameter set, and the coordinated control is performed on the parking vehicle to ensure that the vehicle motion matches the environment state in the parking process, so that the parking operation is safely completed.

[0051] In summary, the embodiments of the present application have at least the following technical effects:

[0052] Firstly, a layout data source of a digital road is acquired, and a feature association relationship between a digital twin space and the layout data source and a vehicle communication data source is established. Then, an interactive parking vehicle is positioned in a region based on the digital twin space, an intention sharing perception subspace is established, path time sequence tracking of the interactive parking vehicle is performed in the intention sharing perception subspace, perception data self-verification and interaction verification are performed, and an abnormal space is identified. Then, according to the abnormal space, spatial risk evaluation is performed by using data results of the perception data self-verification and the interaction verification. Finally, when a spatial risk evaluation result meets a parking space requirement, collaborative parking control is performed according to a spatial verification result and a risk evaluation result. The technical problem that in the prior art, during automatic parking, interaction between vehicles is insufficient, resulting in low parking safety and reliability is solved, and the technical effect that through digital twin space construction and multi-source perception data joint verification, efficient interaction between vehicles and controllable spatial risk are realized, so that parking safety and reliability are significantly improved is achieved.

[0053] In the embodiment two, based on the same inventive concept as the automatic parking method combined with the digital road in the foregoing embodiments, as shown in the embodiment two, Figure 2 The application provides an automatic parking system combined with a digital road, wherein the system comprises:

[0054] The data source association module 11 acquires a layout data source of a digital road, and establishes a feature association relationship between a digital twin space and the layout data source and a vehicle communication data source. The tracking module 12 positions an interactive parking vehicle in a region based on the digital twin space, establishes an intention sharing perception subspace, performs path time sequence tracking of the interactive parking vehicle in the intention sharing perception subspace, and performs perception data self-verification and interaction verification, and identifies an abnormal space. The risk evaluation module 13 performs spatial risk evaluation by using data results of the perception data self-verification and the interaction verification according to the abnormal space. The parking control module 14 performs collaborative parking control according to a spatial verification result and a risk evaluation result when a spatial risk evaluation result meets a parking space requirement.

[0055] Further, the parking control module 14 is used to perform the following method:

[0056] When the spatial risk evaluation result does not meet the parking space requirement, a risk space orientation and a risk probability are identified, a parking path of the interactive parking vehicle and a path interaction space are connected, path risk time sequence analysis is performed on the parking path and the path interaction space according to the risk space orientation and the risk probability, a risk time sequence chain is obtained, the risk time sequence chain has a risk space coordinate identifier and a spatial risk probability thereof, a parking path node corresponding to the risk time sequence chain is adaptively adjusted according to a risk space coordinate corresponding to the risk time sequence chain and a spatial risk probability thereof, and a parking control mode is switched.

[0057] Further, the parking control module 14 is configured to execute the following method:

[0058] establishing an adaptive mode adjustment library, wherein the adaptive mode adjustment library comprises risk evaluation grades and corresponding risk probabilities, and control constraint parameters of parking control modes; matching the spatial risk probability of the parking path node with the adaptive mode adjustment library to obtain a node parking control mode; and performing smoothing processing on the node parking control mode according to the parking path node to determine a switching parking control mode.

[0059] Further, the parking control module 14 is configured to execute the following method:

[0060] identifying the parking control mode time sequence consistency of the parking path node; when the consistency is not met, performing parameter smoothing fitting on the control constraint parameters according to the path switching distance relationship of the parking path node, to determine the smoothing control parameters with the best safety, and to obtain the switching parking control mode.

[0061] Further, the parking control module 14 is configured to execute the following method:

[0062] comparing the conflict items of the control constraint parameters of adjacent nodes based on the path switching distance relationship of the parking path node; setting a priority weight according to the safety influence of the conflict items, and performing weighted compromise calculation on the conflict items according to the priority weight to generate transition parameters; and determining the parking control mode of the node according to the proximity distance of the transition parameters to the control constraint parameters of the parking control mode.

[0063] Further, the tracking module 12 is configured to execute the following method:

[0064] synchronizing the layout data source and the vehicle communication data source of each interactive parking vehicle; performing time sequence longitudinal comparison on the layout data source of each parking vehicle to obtain perception data self-verification data and generate a self-verification abnormal label; obtaining perception data interactive verification data according to the overlapping perception verification relationship of the layout data source and the vehicle communication data source of the interactive parking vehicle, and generating an interactive verification abnormal label; and if the self-verification abnormal label and the interactive verification abnormal label are activated at the same time, marking the overlapping perception area as a confirmed abnormal space.

[0065] Further, the tracking module 12 is configured to execute the following method:

[0066] extracting the perception data frame sequence within the continuous time window of each parking vehicle; calculating the spatial consistency index of adjacent frames based on the perception data frame sequence, and determining that the vehicle perception is invalid if the index exceeds a dynamic threshold; and performing perception data failure tracing using longitudinal historical time sequence data frames to locate abnormal data space coordinates and generate the self-verification abnormal label.

[0067] Further, the tracking module 12 is configured to perform the following method:

[0068] Based on the layout data source of the interactive parking vehicle and the vehicle communication data source, a signal distribution map of each interactive vehicle is established; the signal distribution map of the interactive vehicle is overlapped according to the intention shared perception subspace relationship by using path timing tracking, and the conflict perception data in the overlapping space range is identified; the abnormal perception space coordinates of the abnormal parking vehicle are located based on the perception credibility of the conflict perception data, and the interactive verification abnormal label is generated.

[0069] Further, the risk evaluation module 13 is configured to perform the following method:

[0070] According to the abnormal perception data source of the abnormal perception space coordinates and the abnormal deviation amount, the path time and space abnormal positioning is performed in combination with the parking path interaction space of the interactive parking vehicle; the parking vehicle is taken as a target, and the risk probability and risk path node corresponding to the abnormal perception data of the vehicle in the parking path are obtained by fitting in the digital twin space according to the path time and space abnormal positioning; and the spatial risk evaluation result is obtained according to the risk path node, risk space coordinates and risk probability of the parking vehicle.

[0071] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0072] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0073] The present application is only an exemplary description of the present application, and should be considered as covering any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalents, the present application intends to include these modifications and changes.

Claims

1. An automated parking method incorporating digital roads, characterized in that: The method includes: Obtain the deployment data source of digital roads and establish the characteristic correlation between digital twin space and deployment data source and vehicle communication data source; Based on the digital twin spatial positioning area interactive parking vehicle, establish an intent sharing perception subspace, perform path timing tracking of interactive parking vehicle in the intent sharing perception subspace, and perform perception data self-verification and interactive verification to identify abnormal spaces. Based on the abnormal space, spatial risk assessment is performed using the data results from the self-verification and interactive verification of the perceived data. When the spatial risk assessment results meet the parking space requirements, collaborative parking control is carried out based on the spatial verification results and the risk assessment results. By performing a time-series longitudinal comparison of the data sources for each parked vehicle, self-verification data of the perception data is obtained, and self-verification anomaly labels are generated, including: Extract the sequence of perception data frames within a continuous time window for each parked vehicle; Based on the sensing data frame sequence, the spatial consistency index of adjacent frames is calculated. If the index exceeds the dynamic threshold, the vehicle's sensing is deemed to have failed. By using longitudinal historical time-series data frames to trace the source of perceived data failures, the spatial coordinates of abnormal data are located, and the self-verifying abnormal labels are generated. Based on the overlapping perception verification relationship between the deployment data source and the vehicle communication data source of the interactive parking vehicle, perception data interaction verification data is obtained, and interaction verification anomaly labels are generated, including: Based on the deployment data source and vehicle communication data source of interactive parking vehicles, establish a signal distribution map for each interactive vehicle. By using path timing tracking, the signal distribution maps of interactive vehicles are overlaid according to the intention-shared perception subspace relationship to identify conflicting perception data within the overlapping space. Based on the perception credibility of conflict perception data, collaborative verification is performed to locate the abnormal perception spatial coordinates of abnormally parked vehicles and generate the interactive verification abnormal label. Based on the abnormal space, a spatial risk assessment is performed using the data results from the self-verification and interactive verification of the perceived data, including: Based on the abnormal perception data source and abnormal deviation amount of the abnormal perception spatial coordinates, and combined with the parking path interaction space of the interactive parking vehicle, path time and spatial anomaly localization is performed. Taking parked vehicles as the target, the system fits the path time and spatial anomaly localization in the digital twin space to obtain the risk probability and risk path nodes corresponding to the anomaly perception data of the space where the vehicle is located in the parking path. Based on the risk path nodes, risk spatial coordinates, and risk probabilities of parked vehicles, spatial risk assessment results are obtained.

2. The automatic parking method incorporating digital roads according to claim 1, characterized in that, Spatial risk assessment is performed using the self-verification and interactive verification results of the perceived data, and then includes: When the spatial risk assessment results do not meet the parking space requirements, identify the location and probability of the risk space. Connect the parking paths and path interaction spaces of the interactive parking vehicles; Based on the risk spatial orientation and risk probability, a path risk temporal analysis is performed on the parking path and path interaction space to obtain a risk temporal chain, wherein the risk temporal chain has risk spatial coordinate identifiers and spatial risk probabilities. Based on the risk spatial coordinates and spatial risk probability corresponding to the risk time sequence chain, the corresponding parking path nodes are adaptively adjusted to switch the parking control mode.

3. The automatic parking method incorporating digital roads according to claim 2, characterized in that, Adaptively adjust the corresponding parking path nodes and switch parking control modes, including: Establish an adaptive mode adjustment library, which includes risk assessment levels and their corresponding risk probabilities, as well as control constraint parameters for parking control modes; The spatial risk probability of the parking path node is matched with the adaptive mode adjustment library to obtain the node parking control mode; The parking control mode of the node is smoothed according to the parking path node, and the parking control mode is switched.

4. The automatic parking method incorporating digital roads according to claim 3, characterized in that, Smoothing the parking control mode of the nodes according to the parking path nodes, and determining the switching of the parking control mode, including: Identify the timing consistency of the parking control mode for the parking path nodes; When consistency is not met, based on the path switching distance relationship of the parking path nodes, the parameters are smoothly fitted according to the control constraint parameters of the corresponding parking control mode to determine the smooth control parameters with the best safety, and the switching parking control mode is obtained.

5. The automatic parking method incorporating digital roads according to claim 4, characterized in that, Based on the path switching distance relationship of the parking path nodes, parameter smoothing fitting is performed according to the control constraint parameters of the corresponding parking control mode to determine the smooth control parameters with optimal safety, including: Based on the path switching distance relationship of the parking path nodes, compare the conflict items of the control constraint parameters of adjacent nodes; Priority weights are set according to the security impact of conflicting items, and weighted compromise calculations are performed on conflicting items according to priority weights to generate transition parameters; The parking control mode of the node is determined based on the proximity distance of the transition parameters to the corresponding control constraint parameters in the parking control mode.

6. The automatic parking method incorporating digital roads according to claim 1, characterized in that, In the intent-sharing perception subspace, path-time tracking of interactively parked vehicles is performed, along with self-verification and interactive verification of perception data, to identify abnormal spaces, including: Synchronize the deployment data source and vehicle communication data source of each interactive parking vehicle; By performing a longitudinal comparison of the data sources deployed for each parked vehicle over time, we can obtain self-verifying perception data and generate self-verifying anomaly labels. Based on the overlapping perception and verification relationship between the deployment data source and the vehicle communication data source of the interactive parking vehicle, perception data interaction verification data is obtained, and interaction verification anomaly labels are generated. If the self-verification anomaly label and the interactive verification anomaly label are activated at the same time, the overlapping perception area will be marked as the confirmed anomaly space.

7. An automated parking system integrated with digital roads, characterized in that: For implementing the automated parking method incorporating digital roads as described in any one of claims 1-6, the system comprises: Data source association module: acquires the deployment data source of digital roads and establishes the characteristic association relationship between the digital twin space and the deployment data source and vehicle communication data source; Tracking module: Based on the digital twin spatial positioning area of ​​the interactive parking vehicle, establish an intent-sharing perception subspace, perform path-time tracking of the interactive parking vehicle in the intent-sharing perception subspace, as well as self-verification and interactive verification of perception data, and identify abnormal spaces; Risk assessment module: Based on the abnormal space, spatial risk assessment is performed using the data results of self-verification and interactive verification of the perceived data; Parking control module: When the space risk assessment results meet the parking space requirements, collaborative parking control is performed based on the space verification results and risk assessment results.

Images