Method, device, storage medium and electronic device for correcting driving trajectory

By receiving and analyzing trajectory verification information, the machine learning model and HMM model are used to correct deviations in the vehicle's driving trajectory, solving the problem of driving trajectory deviation caused by inaccurate positioning, and achieving accurate billing and improving user experience.

CN111831768BActive Publication Date: 2024-11-01DITU (BEIJING) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010436474.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-05-21
Publication Date
2024-11-01
Estimated Expiration
2040-11-01

AI Technical Summary

Technical Problem

In the case of inaccurate positioning, there is a deviation between the vehicle's driving trajectory and the real driving trajectory, resulting in inaccurate billing and poor user experience.

Method used

By receiving the trajectory verification information and the trajectory points uploaded by the service mobile terminal, a predetermined machine learning model is used to determine whether the trajectory point is offset, and in the event of an offset, the trajectory point of the deviation is corrected through the trajectory verification information. , the observation probability matrix and state transition probability matrix of the HMM model are determined using auxiliary position information and pose information, and Viterbi decoding is performed to determine the true trajectory that is most matched with the vehicle's travel process and correct the driving trajectory.

Benefits of technology

Ensure that the obtained driving trajectory is the corrected driving trajectory, and then accurately bills based on the corrected driving trajectory to improve user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111831768B_ABST
    Figure CN111831768B_ABST
Patent Text Reader

Abstract

The disclosed embodiment provides a method, device, storage medium and electronic device for correcting a driving trajectory, the method comprising: receiving trajectory verification information and trajectory points uploaded by a mobile terminal of a service party; inputting the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points is offset; in the event of an offset, correcting the deviated trajectory points through the trajectory verification information to correct the driving trajectory; wherein the trajectory verification information at least comprises: the posture information output by the mobile terminal of the service party, the real road network information and the auxiliary location information, the auxiliary location information at least comprises: the location information uploaded by the mobile terminal of the request party and / or the network location information uploaded by an external positioning device. The disclosed embodiment can correct the offset driving trajectory in real time, and then can accurately charge according to the corrected driving trajectory, and the user experience is good.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of data processing, and in particular to a method, device, storage medium and electronic device for correcting a driving trajectory. Background Art

[0002] With the rapid development of mobile Internet and the increase in smartphone users, people are becoming more and more dependent on location-based services. For example, when a user uses an online car-hailing platform to take a taxi, the billing method after the use of the taxi is based on the driving trajectory of the service provider's (driver) mobile client. At present, Beidou positioning, GPS (Global Positioning System) positioning and other methods can be used to achieve all-round positioning, and can realize information interconnection functions with mobile terminals and cloud backends.

[0003] However, when the service provider's mobile terminal is not accurately positioned due to quality problems, or when the positioning signal is poor at a certain place, the final positioning will be inaccurate, so the uploaded trajectory points will be offset (commonly known as drift).

[0004] In the case of inaccurate positioning, the vehicle's driving trajectory deviates from the actual driving trajectory, resulting in inaccurate billing and poor user experience. Summary of the invention

[0005] In view of this, the embodiments of the present disclosure propose a method, device, storage medium and electronic device for correcting a driving trajectory, so as to solve the following problems in the prior art: in the case of inaccurate positioning, there is a deviation between the vehicle's driving trajectory and the actual driving trajectory, resulting in inaccurate billing and poor user experience.

[0006] On the one hand, an embodiment of the present disclosure proposes a method for correcting a driving trajectory, comprising: receiving trajectory verification information and trajectory points uploaded by a service party mobile terminal; inputting the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points is offset; in the event of an offset, correcting the deviated trajectory points through the trajectory verification information to correct the driving trajectory; wherein the trajectory verification information includes at least: posture information output by the service party mobile terminal, real road network information and auxiliary position information, and the auxiliary position information includes at least: position information uploaded by the requesting party mobile terminal and / or network position information uploaded by an external positioning device.

[0007] In some embodiments, the trajectory verification information and the trajectory points are input into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points has deviated, including: determining the driving trajectory based on the trajectory points; verifying whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; and determining that a deviation has occurred if the predetermined matching degree is not reached.

[0008] In some embodiments, the correcting of the deviated trajectory point by using the trajectory verification information includes: determining the deviated trajectory point in the offset driving trajectory; replacing the position information of the deviated trajectory point with the requesting party position information and / or the network position information.

[0009] In some embodiments, the correcting of deviated trajectory points through the trajectory verification information includes: determining the observation probability matrix of the HMM model based on the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information and the position information of the trajectory point; determining the state transition probability matrix of the HMM model based on the relative change of the posture information, the real road network information and the position information of the trajectory point; performing Viterbi decoding on the observation probability matrix and the state transition probability matrix, determining the real trajectory that best matches the vehicle's travel process based on the result of the Viterbi decoding, and correcting the deviated trajectory points in the driving trajectory through the real trajectory.

[0010] On the other hand, an embodiment of the present disclosure proposes a device for correcting a driving trajectory, comprising: a receiving module for receiving trajectory verification information and trajectory points uploaded by a service party mobile terminal; a judgment module for inputting the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points is offset; a correction module for correcting the deviated trajectory points through the trajectory verification information in case of offset, so as to correct the driving trajectory; wherein the trajectory verification information includes at least: angle information output by a gyroscope of the service party mobile terminal, real road network information and auxiliary position information, and the auxiliary position information includes at least: position information uploaded by the requesting party mobile terminal and / or network position information uploaded by an external positioning device.

[0011] In some embodiments, the judgment module is specifically used to: determine the driving trajectory based on the trajectory points; verify whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; and determine that a deviation has occurred if the predetermined matching degree is not reached.

[0012] In some embodiments, the correction module is specifically used to: determine the deviated trajectory point in the deviated driving trajectory; and replace the position information of the deviated trajectory point with the requesting party position information and / or the network position information.

[0013] In some embodiments, the correction module is specifically used to: determine the observation probability matrix of the HMM model based on the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information and the position information of the trajectory points; determine the state transition probability matrix of the HMM model based on the relative change of the posture information, the real road network information and the position information of the trajectory points; perform Viterbi decoding on the observation probability matrix and the state transition probability matrix, determine the real trajectory that best matches the vehicle's travel process based on the result of the Viterbi decoding, and correct the deviated trajectory points in the driving trajectory through the real trajectory.

[0014] On the other hand, an embodiment of the present disclosure provides a storage medium storing a computer program, which implements the method provided by any embodiment of the present disclosure when the computer program is executed by a processor.

[0015] On the other hand, an embodiment of the present disclosure provides an electronic device, which includes at least a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the method provided in any embodiment of the present disclosure when executing the computer program in the memory.

[0016] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] Figure 1 A flow chart of a method for correcting a driving trajectory provided in a first embodiment of the present disclosure;

[0019] Figure 2 A flow chart of a method for correcting a driving trajectory provided in a second embodiment of the present disclosure;

[0020] Figure 3 A schematic diagram of a deviation trajectory provided for the second embodiment of the present disclosure;

[0021] Figure 4 A schematic diagram of WIFI positioning provided in the second embodiment of the present disclosure;

[0022] Figure 5 A schematic diagram of a deviation trajectory after correction provided by the second embodiment of the present disclosure;

[0023] Figure 6 A schematic diagram of the structure of a device for correcting a driving trajectory provided in a third embodiment of the present disclosure;

[0024] Figure 7 A schematic diagram of the structure of an electronic device provided in the fifth embodiment of the present disclosure. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0026] Unless otherwise defined, the technical terms or scientific terms used in the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0027] In order to keep the following description of the embodiments of the present disclosure clear and concise, the present disclosure omits detailed descriptions of well-known functions and well-known components.

[0028] The first embodiment of the present disclosure provides a method for correcting a driving trajectory. The process of the method is as follows: Figure 1 As shown, steps S101 to S103 are included:

[0029] S101, receiving trajectory verification information and trajectory points uploaded by a mobile terminal of a service provider.

[0030] The above-mentioned track points are usually GPS points obtained by the service provider's mobile terminal. GPS can determine the location of the vehicle in real time. Of course, the uploaded track points may be inaccurate, so verification is required.

[0031] The above-mentioned trajectory verification information may be information related to the vehicle where the service provider is located. In order to ensure the accuracy of the verification, it should at least include: the posture information output by the service provider's mobile terminal, the real road network information and the auxiliary position information, wherein the posture information output by the mobile terminal may be the angle information output by the gyroscope in the mobile terminal, etc.; the auxiliary position information at least includes: the requester's (e.g., passenger) mobile terminal uploaded requester's position information and / or the network position information uploaded by the external standard positioning device (e.g., WIFI positioning information uploaded by the WIFI device). Of course, in order to further increase the accuracy of the verification, other information that can be uploaded can also be uploaded to verify whether the uploaded trajectory points are correct.

[0032] When using the above track verification information, more information can be inferred based on the track verification information that has been uploaded, thereby increasing the accuracy of the verification.

[0033] S102, inputting the trajectory verification information and the trajectory point into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory point is offset. The predetermined machine learning model is used to determine the quality of the driving trajectory uploaded by the service provider's mobile terminal, that is, to determine whether there is an offset trajectory point.

[0034] During implementation, the predetermined machine learning model can first determine the driving trajectory based on the trajectory points, and then match the location information of the trajectory points uploaded by the service-side mobile terminal with the requester location information uploaded by the requester mobile terminal or the high-precision network location information calculated using an external standard positioning device. If the degree of match is high, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is good, and the driving trajectory formed by the trajectory points can be directly used for subsequent cost calculation and other processes; if the degree of match is poor, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is poor, and it needs to be corrected. During implementation, it is only necessary to input the trajectory verification information and the trajectory points into the predetermined machine learning model to verify whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches the predetermined matching degree, and if the predetermined matching degree is not reached, it is determined that an offset has occurred.

[0035] Of course, using the requester location information uploaded by the requester mobile terminal or the network location information uploaded by an external standard positioning device as the matching basis of the predetermined machine learning model is only an example. Those skilled in the art can set the predetermined machine learning model according to needs. For example, matching the driving trajectory corresponding to the trajectory point with the angle information output by the gyroscope. If the gyroscope recognizes a right turn at the intersection, but the trajectory is still straight, it can be determined that there is a problem with the driving trajectory, and the trajectory points on the driving trajectory need to be corrected.

[0036] The above-mentioned predetermined machine learning model can be created according to the training method of the existing machine learning model, that is, to create some labeled data sets, which include correct trajectory verification information, correct trajectory points uploaded by the service party's mobile terminal without offset, and trajectory points uploaded by the service party's mobile terminal with offset. By training a basic model with these data sets, the required predetermined machine learning model can be obtained. Of course, based on different requirements for the predetermined machine learning model, the data content labeled in the labeled data set can be transformed according to the needs. Those skilled in the art can refer to the existing methods for implementation, which will not be repeated here.

[0037] S103, in the event of deviation, correcting the deviated trajectory points through trajectory verification information to correct the driving trajectory.

[0038] In the process of correcting the trajectory points, since the trajectory verification information is a reference for verifying the driving trajectory, the trajectory verification information is also used for correction during the correction process to maximize the information utilization. For example, in the event of an offset, the deviated trajectory point is determined in the deviated driving trajectory, and then the location information of the deviated trajectory point can be directly replaced with the requesting party location information and / or network location information, thereby achieving rapid correction.

[0039] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience.

[0040] The second embodiment of the present disclosure provides a method for correcting a driving trajectory. The process of the method is as follows: Figure 1 As shown, steps S201 to S205 are included:

[0041] S201, receiving trajectory verification information and trajectory points uploaded by a service provider mobile terminal.

[0042] The above-mentioned track points are usually GPS points obtained by the service provider's mobile terminal. GPS can determine the location of the vehicle in real time. Of course, the uploaded track points may be inaccurate, so verification is required.

[0043] The above-mentioned trajectory verification information can be information related to the vehicle where the service provider is located. In order to ensure the accuracy of the verification, it should at least include: the posture information output by the service provider's mobile terminal, the real road network information and the auxiliary position information, wherein the posture information output by the mobile terminal can be the angle information output by the gyroscope in the mobile terminal, etc.; the auxiliary position information at least includes: the requester's position information uploaded by the requester's mobile terminal and / or the network position information uploaded by the external standard positioning device (for example, the WIFI positioning information uploaded by the WIFI device). When specifically selecting, the posture information output by the service provider's mobile terminal is preferably the angle information output by the mobile terminal gyroscope. Of course, in order to further increase the accuracy of the verification, other information that can be uploaded can also be uploaded to verify whether the uploaded trajectory points are correct.

[0044] When using the above track verification information, more information can be inferred based on the track verification information that has been uploaded, thereby increasing the accuracy of the verification.

[0045] S202, inputting the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points is offset.

[0046] The above-mentioned predetermined machine learning model is used to determine the quality of the driving trajectory uploaded by the service provider's mobile terminal, that is, to determine whether there are any offset trajectory points.

[0047] During implementation, the predetermined machine learning model can first determine the driving trajectory based on the trajectory points, and then match the location information of the trajectory points uploaded by the service-side mobile terminal with the requester location information uploaded by the requester mobile terminal or the high-precision network location information calculated using an external standard positioning device. If the degree of match is high, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is good, and the driving trajectory formed by the trajectory points can be directly used for subsequent cost calculation and other processes; if the degree of match is poor, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is poor, and it needs to be corrected. During implementation, it is only necessary to input the trajectory verification information and the trajectory points into the predetermined machine learning model to verify whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches the predetermined matching degree; and if the predetermined matching degree is not reached, it is determined that an offset has occurred.

[0048] Of course, using the requester location information uploaded by the requester mobile terminal or the network location information uploaded by an external standard positioning device as the matching basis of the predetermined machine learning model is only an example. Those skilled in the art can set the predetermined machine learning model according to needs. For example, matching the driving trajectory corresponding to the trajectory point with the angle information output by the gyroscope. If the gyroscope recognizes a right turn at the intersection, but the trajectory is still straight, it can be determined that there is a problem with the driving trajectory, and the trajectory points on the driving trajectory need to be corrected.

[0049] The above-mentioned predetermined machine learning model can be created according to the training method of the existing machine learning model, that is, to create some labeled data sets, which include correct trajectory verification information, correct trajectory points uploaded by the service party's mobile terminal without offset, and trajectory points uploaded by the service party's mobile terminal with offset. By training a basic model with these data sets, the required predetermined machine learning model can be obtained. Of course, based on different requirements for the predetermined machine learning model, the data content labeled in the labeled data set can be transformed according to the needs. Those skilled in the art can refer to the existing methods for implementation, which will not be repeated here.

[0050] S203, in the event of an offset, determining the observation probability matrix of the HMM model based on the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information, and the position information of the trajectory point.

[0051] In the process of correcting the trajectory points, since the trajectory verification information is a reference for verifying the driving trajectory, the trajectory verification information is also used for correction during the correction process to maximize the information utilization. For example, in the event of an offset, the location information of the deviated trajectory point in the driving trajectory of the offset section can be directly replaced with the requesting party location information and / or network location information, thereby enabling rapid correction.

[0052] However, although it is feasible to directly use the requesting party location information or the network location information, there is still a situation where the requesting party location information or the network location information is also offset. Therefore, the embodiment of the present disclosure adopts the HMM model to correct the offset trajectory points, that is, converting the deviation correction problem of the driving trajectory into a road matching problem, making the road matching more accurate through multi-source information, and then correcting the offset trajectory points to the correct road direction.

[0053] In map road matching, the observation probability is mainly used to measure the probability of observing a trajectory point on a certain implicit state (road), that is, the degree of association between the trajectory point and a certain road (Link). It mainly considers two factors: the distance between the trajectory point and the candidate Link, and the angle between the forward direction of the trajectory point and the direction of the candidate Link itself.

[0054] Based on the above factors, the trajectory verification information is used again to perform calculations when determining the observation probability matrix of the HMM model to obtain the observation probability matrix.

[0055] S204, determining a state transition probability matrix of the HMM model according to the relative change of the posture information, the real road network information and the position information of the trajectory points.

[0056] The state transition probability refers to the possibility of transitioning from the previous state to the current state. Specifically in the map road matching task, it refers to the probability of transitioning from the link matched by the previous trajectory point to the link matched by the current trajectory point. For example, the posture information of this embodiment is the angle information output by the gyroscope, and the relative change can be the angle change and the heading change. Then, when the service provider transfers from this intersection to the next three intersections, the state transition probability can be determined by the angle and heading changes calculated by the angle information output by the gyroscope, the trajectory point and the road surface distance of the trajectory point calculated by the real road network information, and other parameters.

[0057] In the feature selection of probability calculation, considering that the task of map road matching is to examine the possibility of a series of observation points moving along a certain road, the difference between the earth surface distance between the two trajectory points and the actual path distance is selected as the main feature. That is, it can be considered that the closer the straight-line distance between these two points is to the actual road distance from point A to point B, the greater the possibility of its occurrence.

[0058] Based on the above factors, the trajectory verification information is used again to perform calculations when determining the state transition probability matrix of the HMM model to obtain the state transition probability matrix.

[0059] S205, performing Viterbi decoding on the observation probability matrix and the state transition probability matrix, determining the real trajectory that best matches the vehicle's moving process according to the Viterbi decoding result, and correcting the deviation trajectory points in the driving trajectory through the real trajectory to correct the driving trajectory.

[0060] Viterbi decoding is to determine the best match from all candidate matching road segment sequences, that is, the actual trajectory of the final map road matching, and then correct the trajectory points uploaded by the service provider's mobile terminal through the position information in the matched actual trajectory to obtain the corrected actual trajectory of the vehicle.

[0061] The above process from S203 to S205 is a process of correcting the deviated trajectory points through the trajectory verification information to correct the driving trajectory.

[0062] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience.

[0063] The above process is described in detail below with reference to the accompanying drawings and specific examples.

[0064] In order to solve the problem that the service-side mobile terminal cannot accurately locate itself in certain circumstances, the disclosed embodiments provide a method for correcting the driving trajectory. The method uses the requesting party's mobile phone or an external standard positioning device to correct the service-side positioning information, thereby solving the problem of inaccurate positioning information of trajectory points uploaded by the service-side mobile terminal due to problems such as mobile phone quality, improving the accuracy of the driving trajectory and thus enhancing the user experience.

[0065] In the disclosed embodiment, after the service provider's mobile terminal uploads the driving trajectory of the positioning information of the preliminary trajectory point, the cloud backend uses the machine learning model to judge the quality of the trajectory. If the trajectory is offset, the position information uploaded by the requesting party's mobile terminal is used to correct the deviation of the trajectory part, or the trajectory verification information uploaded by the external device is used to calculate the high-precision position information, and transmit it to the service provider's mobile terminal via Bluetooth to correct the deviation of the trajectory part. Among them, the above-mentioned trajectory point is the GPS position, and the external device may include GPS, accelerometer, gyroscope, barometer, Wifi, etc. The external standard positioning device may be GPS or Wifi, etc.

[0066] Of course, if the position information uploaded by the requesting party's mobile phone terminal can be used to correct the deviated track part, the network position information uploaded by the external standard positioning device can also be directly used to correct the deviated track part without calculating the high-precision position information. Those skilled in the art can choose any method to correct the deviated track part according to actual needs, which is not limited here.

[0067] The exemplary process of the present disclosure embodiment is as follows:

[0068] Step 1: Receive the location information uploaded by the service provider and the requester's mobile terminal, as well as the trajectory verification information from the external device.

[0069] After the requesting party gets on the vehicle, the cloud service backend receives the location information of the service party's mobile terminal and the requesting party's mobile terminal respectively through the Internet or Bluetooth.

[0070] Normally, when an order is completed, the cloud backend calculates the fee to be paid for the order based on the driving trajectory uploaded by the service party terminal (i.e., the summary of location information during driving).

[0071] Step 2: Determine the quality of the trajectory uploaded by the service provider's mobile terminal.

[0072] Since the trajectory points of the vehicle are known, the trajectory quality can be determined based on the trajectory points.

[0073] The GPS position of the service provider is likely to be deviated when it is under a tall building or an overpass. Figure 3 The lower row of trajectory points is the original GPS output. However, the actual driving route of the vehicle is the route above. When the deviation is far, the road matching result is usually wrong. Therefore, the embodiment of the present disclosure uses a predetermined machine learning model to judge the quality of the driving trajectory uploaded by the service party's mobile terminal.

[0074] The model can match the location information uploaded by the service-side mobile terminal with the location information uploaded by the requesting terminal or the high-precision location information calculated by an external standard positioning device. If the matching degree is high, it means that the trajectory uploaded by the service-side mobile terminal is of good quality, and the driving trajectory can be directly used for subsequent cost calculation and other processes; if the matching degree is poor, it means that the trajectory uploaded by the service-side terminal is of poor quality and needs to be corrected.

[0075] Step 3: If the quality of the trajectory uploaded by the service provider's mobile terminal is poor, the driving trajectory is corrected. The process of correcting the driving trajectory includes the following two methods: (1) and (2).

[0076] (1) Use the location information uploaded by the requesting mobile terminal to correct the driving trajectory.

[0077] When it is determined that the quality of the trajectory uploaded by the service-side mobile terminal is poor, the cloud backend can use the location information uploaded by the requesting terminal received in step 1 to correct the trajectory, and replace the location points with large deviations in the trajectory with the location point information uploaded by the requesting terminal, thereby correcting the driving trajectory uploaded by the service-side mobile terminal.

[0078] (2) Use high-precision position information calculated by external standard positioning equipment to correct the driving trajectory.

[0079] When it is determined that the quality of the trajectory uploaded by the service party's mobile terminal is poor, the cloud background can also use the high-precision location information obtained by the external standard positioning device to replace the location information of the trajectory points with large deviations in the driving trajectory, thereby correcting the driving trajectory uploaded by the service party's mobile terminal.

[0080] by Figure 3 Take the offset trajectory in as an example, Figure 3 The GPS points uploaded by the service terminal (the lower row of track points in the figure) have been offset. Figure 4 As shown, WIFI positioning ( Figure 4 The points in the figure that are close to or overlap with the previous route are correct and can reflect the actual route taken by the user. At the same time, the gyroscope output angle is accurate.

[0081] Based on the HMM model, the network positioning, the angle information output by the gyroscope, and the position of the requesting party are integrated to assist in real-time correction of the offset GPS points. The correction problem is converted into a road matching problem, that is, the road matching is made more accurate through multi-source information, and then the offset GPS points are corrected to the correct road direction.

[0082] The HMM model is used in road matching, where the HMM model is divided into three parts: observation probability matrix, state transition probability matrix and Viterbi decoding.

[0083] In the process of constructing the observation probability matrix, a virtual road network is constructed using WIFI, and the observation probability matrix is ​​calculated by combining the virtual road network information, the real road network information, the heading angle calculated by the gyroscope, and the GPS information. The observation probability of matching the GPS point to the virtual road network (that is, the probability that a GPS point belongs to a specific j-th link) is calculated as follows:

[0084]

[0085] proj_dist_prob=beta_projection*e -(point2linkdist / theta) ;

[0086]

[0087]

[0088] The above "proj_dist_prob" refers to By analogy, to avoid viewing inconvenience, third-level or second-level subscripts may appear in the above formula or the following formula in the form of underscores, and will not be repeated later.

[0089] Among them, A is the preset weight coefficient, beta_projection indicates whether it is the weight of vertical projection. If the GPS projection is on the extension line of the straight road, the weight is low, and if the projection point is on the road, the weight is high. factor is the speed influencing factor, point2linkdist is the distance from the point to the virtual link, loc_diff and bearing_diff are the angular difference between the two points and the heading angle difference, beta and theta are both normalization coefficients, which adjust the sensitivity of the angle change amplitude to the observation probability.

[0090] In calculating the state transition probability matrix, the relative angle change output by the gyroscope and the road distance are used to obtain a more accurate state transition probability from the previous link i to the current link j.

[0091] a ij =θ*prob dist +(1-θ)prob angle ;

[0092]

[0093]

[0094] Among them, route_dist is the road distance between two points (the road distance projected from the GPS point to several roads), gps_dist is the straight-line distance between two points (the straight-line distance of the original GPS points), the smaller the difference, the greater the probability, bearing_diff is the heading angle difference between the two points, gyro_diff is the angle change output by the sensor, and θ refers to a fixed weight coefficient, which means the proportion of the influence of prob_dist and prob_angle on the transfer probability.

[0095] Finally, Viterbi decoding is used to find the best matching real road. After obtaining the matching road, the offset GPS point is appropriately offset to the matching real road direction, such as Figure 5 The middle row of trajectory points in .

[0096] By using the disclosed embodiments, when the service provider's driving trajectory deviates from the road network, if there is no other reference trajectory, the above-mentioned deviation correction method can be used for correction. If the correction result falls on the road network, the correction result can be used. Furthermore, because satellite position, geographical location, weather, terrain, etc. may affect the GPS positioning accuracy, the above-mentioned factors can also be used in the trajectory quality judgment process to increase the accuracy of the judgment.

[0097] The third embodiment of the present disclosure provides a device for correcting a driving trajectory, the structure of which is shown in FIG. Figure 6As shown, including:

[0098] A receiving module 10 is used for receiving trajectory verification information and trajectory points uploaded by a mobile terminal of a service provider; a judging module 20 is coupled with the receiving module 10 and is used for inputting the trajectory verification information and trajectory points into a predetermined machine learning model to determine whether a driving trajectory corresponding to the trajectory points is offset; a correcting module 30 is coupled with the judging module 20 and is used for correcting the deviated trajectory points by means of trajectory verification information in case of offset so as to correct the driving trajectory; wherein the trajectory verification information at least includes: posture information output by the mobile terminal of the service provider, real road network information and auxiliary position information, wherein the posture information output by the mobile terminal may be angle information output by a gyroscope in the mobile terminal, etc.; the auxiliary position information at least includes: position information uploaded by the mobile terminal of the requester and / or network position information uploaded by an external positioning device.

[0099] The above-mentioned track points are usually GPS points obtained by the service provider's mobile terminal. GPS can determine the location of the vehicle in real time. Of course, the uploaded track points may be inaccurate, so verification is required.

[0100] The track verification information may be information related to the vehicle where the service provider is located. To ensure the accuracy of the verification, as much other information as possible may be uploaded to verify whether the uploaded track points are correct. In specific implementation, the network location information may be WIFI positioning information uploaded by a WIFI device.

[0101] When using the above track verification information, more information can be inferred based on the track verification information that has been uploaded, thereby increasing the accuracy of the verification.

[0102] The above-mentioned predetermined machine learning model is used to determine the quality of the driving trajectory uploaded by the service provider's mobile terminal, that is, to determine whether there are any offset trajectory points.

[0103] The predetermined machine learning model can first determine the driving trajectory based on the trajectory points, and then match the location information of the trajectory points uploaded by the service-side mobile terminal with the requester location information uploaded by the requester mobile terminal or the high-precision network location information calculated using an external standard positioning device. If the degree of match is high, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is good, and the driving trajectory formed by the trajectory points can be directly used for subsequent cost calculation and other processes; if the degree of match is poor, it means that the quality of the trajectory points uploaded by the service-side mobile terminal is poor, and it needs to be corrected. Therefore, the above-mentioned judgment module is specifically used to: determine the driving trajectory based on the trajectory points, and verify whether the matching degree of the parameter information of the driving trajectory and the trajectory verification information reaches the predetermined matching degree; if the predetermined matching degree is not reached, determine that an offset has occurred.

[0104] Of course, using the requester location information uploaded by the requester mobile terminal or the network location information uploaded by an external standard positioning device as the matching basis of the predetermined machine learning model is only an example. Those skilled in the art can set the predetermined machine learning model according to needs. For example, matching the driving trajectory corresponding to the trajectory point with the angle information output by the gyroscope. If the gyroscope recognizes a right turn at the intersection, but the trajectory is still straight, it can be determined that there is a problem with the driving trajectory, and the trajectory points on the driving trajectory need to be corrected.

[0105] The above-mentioned predetermined machine learning model can be created according to the training method of the existing machine learning model, that is, to create some labeled data sets, which include correct trajectory verification information, correct trajectory points uploaded by the service party's mobile terminal without offset, and trajectory points uploaded by the service party's mobile terminal with offset. By training a basic model with these data sets, the required predetermined machine learning model can be obtained. Of course, based on different requirements for the predetermined machine learning model, the data content labeled in the labeled data set can be transformed according to the needs. Those skilled in the art can refer to the existing methods for implementation, which will not be repeated here.

[0106] In the process of correcting the trajectory points, since the trajectory verification information is a reference for verifying the driving trajectory, the trajectory verification information is also used for correction in the correction process to maximize the information utilization.

[0107] The correction module may be specifically used to: determine a deviated track point in a deviated driving track, and replace the position information of the deviated track point with the requesting party position information and / or network position information.

[0108] The above-mentioned correction module can also be specifically used for: determining the observation probability matrix of the HMM model based on the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information and the position information of the trajectory points; determining the state transition probability matrix of the HMM model based on the relative change of the posture information, the real road network information and the position information of the trajectory points; performing Viterbi decoding on the observation probability matrix and the state transition probability matrix, determining the real trajectory that best matches the vehicle's travel process based on the results of Viterbi decoding, and correcting the deviated trajectory points in the driving trajectory through the real trajectory.

[0109] In map road matching, the observation probability is mainly used to measure the probability of observing a trajectory point on a certain implicit state (road), that is, the degree of association between the trajectory point and a certain road (Link). It mainly considers two factors: the distance between the trajectory point and the candidate Link, and the angle between the forward direction of the trajectory point and the direction of the candidate Link itself.

[0110] Based on the above factors, the trajectory verification information is used again to perform calculations when determining the observation probability matrix of the HMM model to obtain the observation probability matrix.

[0111] The state transition probability refers to the possibility of transitioning from the previous state to the current state. Specifically in the map road matching task, it refers to the probability of transitioning from the link matched by the previous trajectory point to the link matched by the current trajectory point. For example, the posture information of this embodiment is the angle information output by the gyroscope, and the relative change can be the angle change and the heading change. Then, when the service provider transfers from this intersection to the next three intersections, the state transition probability can be determined by the angle and heading changes calculated by the angle information output by the gyroscope, the trajectory point and the road surface distance of the trajectory point calculated by the real road network information, and other parameters.

[0112] In the feature selection of probability calculation, considering that the task of map road matching is to examine the possibility of a series of observation points moving along a certain road, the difference between the earth surface distance between the two trajectory points and the actual path distance is selected as the main feature. That is, it can be considered that the closer the straight-line distance between these two points is to the actual road distance from point A to point B, the greater the possibility of its occurrence.

[0113] Based on the above factors, the trajectory verification information is used again to perform calculations when determining the state transition probability matrix of the HMM model to obtain the state transition probability matrix.

[0114] Viterbi decoding is to determine the best match from all candidate matching road segment sequences, that is, the actual trajectory of the final map road matching, and then correct the trajectory points uploaded by the service provider's mobile terminal through the position information in the matched actual trajectory to obtain the corrected actual trajectory of the vehicle.

[0115] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience.

[0116] A fourth embodiment of the present disclosure provides a storage medium, which is a computer-readable medium and stores a computer program. When the computer program is executed by a processor, the method provided by any embodiment of the present disclosure is implemented, including the following steps S11 to S13:

[0117] S11, receiving trajectory verification information and trajectory points uploaded by the service provider's mobile terminal;

[0118] S12, inputting the trajectory verification information and the trajectory point into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory point is offset;

[0119] S13, in the event of deviation, correcting the deviated trajectory point through trajectory verification information to correct the driving trajectory;

[0120] The trajectory verification information includes at least: the posture information output by the service provider's mobile terminal, the real road network information and the auxiliary location information; the auxiliary location information includes at least: the location information uploaded by the requester's mobile terminal and / or the network location information uploaded by the external positioning device.

[0121] When the computer program is executed by the processor to input the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points has deviated, the processor specifically executes the following steps: determining the driving trajectory based on the trajectory points; verifying whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; and determining that a deviation has occurred if the predetermined matching degree has not been reached.

[0122] When the computer program is executed by the processor to correct the deviated trajectory point through trajectory verification information, the processor specifically executes the following steps: determining the deviated trajectory point in the deviated driving trajectory; replacing the position information of the deviated trajectory point with the requesting party position information and / or network position information.

[0123] When the computer program is executed by the processor to correct the deviated trajectory points through trajectory verification information, the processor specifically executes the following steps: determining the observation probability matrix of the HMM model based on the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information and the position information of the trajectory points; determining the state transition probability matrix of the HMM model based on the relative change of the posture information, the real road network information and the position information of the trajectory points; performing Viterbi decoding on the observation probability matrix and the state transition probability matrix, determining the real trajectory that best matches the vehicle's travel process based on the results of the Viterbi decoding, and correcting the deviated trajectory points in the driving trajectory through the real trajectory.

[0124] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience.

[0125] Optionally, in this embodiment, the above-mentioned storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk. Optionally, in this embodiment, the processor executes the method steps recorded in the above-mentioned embodiment according to the program code stored in the storage medium. Optionally, the specific examples in this embodiment can refer to the examples described in the above-mentioned embodiment and the optional implementation mode, and this embodiment will not be repeated here. Obviously, it should be understood by those skilled in the art that the above-mentioned modules or steps of the present disclosure can be implemented with a general computing device, which can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented with executable program codes of computing devices, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that here, or they can be made into individual integrated circuit modules respectively, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the present disclosure is not limited to any specific combination of hardware and software.

[0126] A fifth embodiment of the present disclosure provides an electronic device. The structural diagram of the electronic device can be as follows: Figure 7 As shown, at least a memory 901 and a processor 902 are included, the memory 901 stores a computer program, and the processor 902 implements the method provided by any embodiment of the present disclosure when executing the computer program on the memory 901. Exemplarily, the steps of the electronic device computer program are as follows S21 to S23:

[0127] S21, receiving trajectory verification information and trajectory points uploaded by the service provider's mobile terminal;

[0128] S22, inputting the trajectory verification information and the trajectory point into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory point is offset;

[0129] S23, in the event of deviation, correcting the deviated trajectory point through trajectory verification information to correct the driving trajectory;

[0130] The trajectory verification information includes at least: the posture information output by the service provider's mobile terminal, the real road network information and the auxiliary location information; the auxiliary location information includes at least: the location information uploaded by the requester's mobile terminal and / or the network location information uploaded by the external positioning device.

[0131] When the processor executes a computer program stored in the execution memory that inputs trajectory verification information and trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points has deviated, the processor specifically executes the following computer program: determining the driving trajectory based on the trajectory points; verifying whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; and determining that a deviation has occurred if the predetermined matching degree is not reached.

[0132] When the processor executes a computer program stored in the memory for correcting deviated trajectory points through trajectory verification information, the processor specifically executes the following computer program: determining the deviated trajectory points in the deviated driving trajectory; replacing the position information of the deviated trajectory points with the requesting party position information and / or network position information.

[0133] When the processor executes a computer program stored in the memory for correcting deviated trajectory points through trajectory verification information, the processor specifically executes the following computer program: determining the observation probability matrix of the HMM model based on the auxiliary position information, the azimuth information and the heading angle information inferred from the posture information, the real road network information and the position information of the trajectory points; determining the state transition probability matrix of the HMM model based on the relative change of the posture information, the real road network information and the position information of the trajectory points; performing Viterbi decoding on the observation probability matrix and the state transition probability matrix, determining the real trajectory that best matches the vehicle's travel process based on the results of the Viterbi decoding, and correcting the deviated trajectory points in the driving trajectory through the real trajectory.

[0134] The disclosed embodiment performs offset verification on the trajectory points of the service party's mobile terminal through the received trajectory verification information. There are many types of trajectory verification information and the verification accuracy is high. Once it is found that the driving trajectory has deviated, the trajectory verification information is used again to correct the trajectory points to ensure that the final driving trajectory is a corrected driving trajectory, and then accurate billing can be performed based on the corrected driving trajectory, which provides a better user experience.

[0135] In addition, although exemplary embodiments have been described herein, the scope includes any and all embodiments based on the present disclosure with equivalent elements, modifications, omissions, combinations (e.g., various embodiments intersecting schemes), adaptations or changes. The elements in the claims will be interpreted broadly based on the language adopted in the claims, and are not limited to the examples described in this specification or during the implementation of this application, and the examples will be interpreted as non-exclusive. Therefore, this specification and examples are intended to be considered as examples only, and the true scope and spirit are indicated by the following claims and the full scope of their equivalents.

[0136] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. For example, a person of ordinary skill in the art can use other embodiments when reading the above description. In addition, in the above-mentioned specific embodiments, various features can be grouped together to simplify the present disclosure. This should not be interpreted as an intention that a disclosed feature that is not required to be protected is necessary for any claim. On the contrary, the subject matter of the present disclosure may be less than all the features of a specific disclosed embodiment. Thus, the following claims are incorporated into the specific embodiments as examples or embodiments, wherein each claim is independently used as a separate embodiment, and it is considered that these embodiments can be combined with each other in various combinations or arrangements. The scope of the present disclosure should be determined with reference to the attached claims and the full scope of equivalent forms granted by these claims.

[0137] Multiple embodiments of the present disclosure are described in detail above, but the present disclosure is not limited to these specific embodiments. Those skilled in the art can make various variations and modifications to the embodiments based on the concepts of the present disclosure, and these variations and modifications should all fall within the scope of protection required by the present disclosure.

Claims

1. A method for correcting a driving trajectory, characterized in that: include: Receive trajectory verification information and trajectory points uploaded by the service provider's mobile terminal; Inputting the trajectory verification information and the trajectory points into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory points is offset; In the event of deviation, the deviated track points are corrected by using the track verification information to correct the driving track; The trajectory verification information at least includes: the posture information output by the service party's mobile terminal, the real road network information and the auxiliary location information, and the auxiliary location information at least includes: the location information uploaded by the requesting party's mobile terminal and / or the network location information uploaded by the external positioning device, and The HMM model is used to correct the deviation trajectory points. The correcting the deviated trajectory point by using the trajectory verification information includes: Determine an observation probability matrix of the HMM model according to the auxiliary position information, the direction angle information and the heading angle information calculated from the posture information, the real road network information and the position information of the trajectory point; Determine the state transition probability matrix of the HMM model according to the relative change of the posture information, the real road network information and the position information of the trajectory point; Viterbi decoding is performed on the observation probability matrix and the state transition probability matrix, and a real trajectory that best matches the vehicle's travel process is determined according to the result of the Viterbi decoding, and the deviated trajectory points in the driving trajectory are corrected by the real trajectory.

2. The method according to claim 1, characterized in that The step of inputting the trajectory verification information and the trajectory point into a predetermined machine learning model to determine whether the driving trajectory corresponding to the trajectory point is offset comprises: Based on the trajectory points, determining a driving trajectory; Verifying whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; If the predetermined matching degree is not reached, it is determined that a deviation occurs.

3. The method according to claim 1 or 2, characterized in that The correcting the deviated trajectory point by using the trajectory verification information includes: Determining a deviated trajectory point in the driving trajectory where the deviation occurs; The location information of the deviated trajectory point is replaced with the requesting party location information and / or the network location information.

4. A device for correcting a driving trajectory, characterized in that: include: A receiving module, used to receive trajectory verification information and trajectory points uploaded by the service party's mobile terminal; A judgment module, used for inputting the trajectory verification information and the trajectory point into a predetermined machine learning model to judge whether the driving trajectory corresponding to the trajectory point is offset; A correction module, used for correcting the deviated trajectory points by using the trajectory verification information in case of deviation, so as to correct the driving trajectory; The trajectory verification information at least includes: the posture information output by the service party's mobile terminal, the real road network information and the auxiliary location information, and the auxiliary location information at least includes: the location information uploaded by the requesting party's mobile terminal and / or the network location information uploaded by the external positioning device, and The HMM model is used to correct the deviation trajectory points. The correction module is also used to determine the observation probability matrix of the HMM model according to the auxiliary position information, the direction angle information and the heading angle information calculated by the posture information, the real road network information and the position information of the trajectory point; Determine the state transition probability matrix of the HMM model according to the relative change of the posture information, the real road network information and the position information of the trajectory point; Viterbi decoding is performed on the observation probability matrix and the state transition probability matrix, and a real trajectory that best matches the vehicle's travel process is determined according to the result of the Viterbi decoding, and the deviated trajectory points in the driving trajectory are corrected by the real trajectory.

5. The device according to claim 4, characterized in that The judgment module is specifically used for: Based on the trajectory points, determining a driving trajectory; Verifying whether the matching degree between the parameter information of the driving trajectory and the trajectory verification information reaches a predetermined matching degree; In the event that the predetermined degree of matching is not achieved, it is determined that a deviation has occurred.

6. The device according to claim 4 or 5, characterized in that The correction module is specifically used for: Determining a deviated trajectory point in the driving trajectory where the deviation occurs; The location information of the deviated trajectory point is replaced with the requesting party location information and / or the network location information.

7. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

8. An electronic device comprising at least a memory and a processor, wherein a computer program is stored in the memory, wherein: The processor implements the steps of the method of any one of claims 1 to 3 when executing the computer program on the memory.