A method, apparatus, device, and storage medium for determining an intersection path

Abstract

The present disclosure provides a method, apparatus, device, and storage medium for determining an intersection path, which relates to the field of computer technology, and particularly to fields such as artificial intelligence, autonomous driving, vehicle networking, and in-vehicle maps. The specific implementation solution is as follows: Obtain the entry and exit road group information of the first intersection. According to the entry and exit road group information, determine the predicted intersection path from the pre-configured intersection path association information. The intersection path association information is the association relationship information between the second intersection and the intersection path of the second intersection. The intersection path represents the connection line between different lanes in the intersection, and the first intersection is different from the second intersection. Based on the predicted intersection path, determine the intersection path of the first intersection. By using the pre-configured intersection path association information, the present disclosure predicts the possible intersection path of the current intersection. It can provide a safer and more human-driving-habit-compliant path recommendation, reducing the costs caused by the mismatch between the map and the actual use process.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular to fields such as artificial intelligence, autonomous driving, vehicle networking, and in-vehicle maps. Background Art

[0002] A high-precision map is map data with higher position accuracy and rich expression that serves autonomous driving and assisted driving. In terms of intelligent driving, it can represent the lane-level topological relationship. At the same time, path decision-making and planning can be carried out based on such topological relationships.

[0003] Currently, a general path planning strategy makes a decision on the passing behavior at a certain intersection by combining the topological connection of the high-precision map and the road condition perception during actual driving. Among them, the weight of the intersection connection method of the high-precision map is still relatively large. Summary of the Invention

[0004] The present disclosure provides a method, an apparatus, a device, and a storage medium for determining an intersection path.

[0005] According to a first aspect of the present disclosure, a method for determining an intersection path is provided. The method may include: obtaining the entry / exit road group information of a first intersection. The entry / exit road group information includes entry / exit road relationship information, and the entry / exit road relationship information represents the corresponding relationship between a first lane entering the first intersection and a second lane exiting the first intersection. Then, according to the entry / exit road group information, a predicted intersection path may be determined from the pre-configured intersection path association information. The intersection path association information is the association relationship information between a second intersection and the intersection path of the second intersection. The intersection path represents the connection line between different lanes in the intersection, and the first intersection is different from the second intersection. After that, based on the predicted intersection path, the intersection path of the first intersection may be determined. By using the pre-configured intersection path association information, the present disclosure predicts the possible intersection path of the current intersection, which can provide a safer and more human-driving-habit-compliant path recommendation and reduce the cost caused by the mismatch between the map and the actual use process.

[0006] According to a second aspect of the present disclosure, there is provided an apparatus for determining an intersection path, including: an acquisition module configured to acquire entry / exit road group information of a first intersection, where the entry / exit road group information includes entry / exit road relationship information, and the entry / exit road relationship information represents the corresponding relationship between a first lane entering the first intersection and a second lane exiting the first intersection; a determination module configured to determine a predicted intersection path from pre-configured intersection path association information according to the entry / exit road group information, where the intersection path association information is the association relationship information between a second intersection and the intersection path of the second intersection, and the intersection path represents the connection line between different lanes in the intersection, and the first intersection is different from the second intersection; the determination module is further configured to determine the intersection path of the first intersection based on the predicted intersection path. The present disclosure predicts possible intersection paths of the current intersection by using pre-configured intersection path association information. It can provide safer and more human-driving-habit-compliant path recommendations, reducing the costs caused by the mismatch between the map and the actual use process.

[0007] According to a third aspect of the present disclosure, there is provided a device for determining an intersection path, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any one of the methods in the first aspect above.

[0008] According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, where the computer instructions are used to cause a computer to execute any one of the methods in the first aspect above.

[0009] According to a fifth aspect of the present disclosure, there is provided a computer program product including a computer program, and the computer program implements any one of the methods in the first aspect above when executed by a processor.

[0010] A method, apparatus, device, and storage medium for determining an intersection path provided by the present disclosure predict possible intersection paths of the current intersection by using pre-configured intersection path association information. It can provide safer and more human-driving-habit-compliant path recommendations, reducing the costs caused by the mismatch between the map and the actual use process.

[0011] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understandable through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them:

[0013] Figure 1 It is a schematic diagram of lane connection at an intersection according to an embodiment of the present disclosure;

[0014] Figure 2 It is a flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0015] Figure 3 It is another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0016] Figure 4 It is yet another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0017] Figure 5 It is still another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0018] Figure 6 It is another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0019] Figure 7 It is yet another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0020] Figure 8 It is a schematic flowchart of a method for determining an intersection path according to an embodiment of the present disclosure;

[0021] Figure 9 It is a schematic diagram of a device for determining an intersection path according to an embodiment of the present disclosure;

[0022] Figure 10 It is a schematic diagram of a device for determining an intersection path according to an embodiment of the present disclosure. Detailed implementation manners

[0023] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following.

[0024] The main application scenarios of the present disclosure can be, for example, scenarios where path prediction is performed for a certain intersection when using a map application. For example, it can be in the process of map application production, where path prediction is performed for some intersections in the map where vehicle driving trajectories have not been collected, or newly built intersections. Of course, in other scenarios, it can also be when the user is at a certain intersection and performs path prediction for the upcoming intersection.

[0025] One of the major problems at present is the selection of intersection paths. In related technologies, the connection method of intersections in high-precision maps with a relatively large weight is usually used to predict paths. However, the intersection connection method of high-precision maps cannot be uniformly applied to variable intersection scenarios according to a certain set of rules. With the differences in intersection size, shape, number of lanes, alignment method, and lane turning arrangement. A more human-driving-habit-compliant way of passing through intersections has always been a problem difficult to solve in intelligent driving.

[0026] In some related technologies, in the current static map data of high-precision maps, virtual lane geometry is established between the actual lane before entering the intersection and the actual lane after entering the intersection to connect the front and rear lanes. At the same time, the connection between lanes is carried out according to a unified alignment principle. For example, the left alignment principle is adopted. For example Figure 1 As shown in a schematic diagram of intersection lane connection, it can be seen that the connection line ① between the entering lane 101 and the exiting lane 102 represents the intersection connection path (or intersection path, lane connection line). There may be a virtual intersection path between each entering lane and exiting lane in the intersection. For some intersections, when the number of entering lanes and exiting lanes is inconsistent, the left alignment principle can be used for lane connection. For example, if there are 3 entering lanes and 2 exiting lanes, the first entering lane from the left is connected to the first exiting lane from the left, the second entering lane from the left is connected to the second exiting lane from the left, and the third entering lane from the left is connected to the last exiting lane from the left (i.e., the second exiting lane from the left). Of course, assuming there are 4 entering lanes, the fourth entering lane from the left is still connected to the last exiting lane from the left (i.e., the second exiting lane from the left). That is to say, when the number of entering lanes and exiting lanes is inconsistent, the extra lanes are connected to the last lane from the left in the corresponding entering / exiting lanes. Of course, the right alignment principle is the opposite, that is, connect each lane in turn from the right.

[0027] Continuing back to Figure 1 , in related technologies, the intersection paths obtained by lane connection do not differ in data attributes, that is, there is no difference in path weights. For example Figure 1 The 4 intersection paths shown in. For the straight-ahead direction, both intersection path ① and intersection path ② can pass through this intersection. However, usually people tend to choose intersection path ① during driving. Another example is the left-turn direction. Both intersection path ③ and intersection path ④ can reach the same exiting lane. However, people's driving habits usually feel that the trajectory of intersection path ④ is more comfortable.

[0028] In actual vehicle driving applications, when a vehicle passes through an intersection using a high-precision map, it will obtain the lane connection of the intersection (i.e., the intersection path) before passing through the intersection, and perform path planning at the local lane level. For example, it will give priority to the existing intersection path in the high-precision map. If there are obstacles or obstructing vehicles in the environment, it will change lanes or bypass obstacles according to the existing intersection connection conditions.

[0029] However, in the above-mentioned related technologies, the intersection connection rules of high-precision maps are very simple and cannot be used for intersections of various types and shapes. At the same time, the lane connection rules are inconsistent with people's actual driving habits or safer path selection. Therefore, when performing autonomous driving, if it is used as prior information, it may bring very bad effects and experiences to users. For example, in the actual application of autonomous driving, the related technology will frequently require the map to modify the path according to the actual driving, which is very unfavorable to the mass production of high-precision maps and has poor use effect.

[0030] Therefore, the present disclosure provides a method for determining an intersection path, by using pre-configured intersection path association information to predict the possible intersection path of the current intersection. It can provide safer and more human driving habit-compliant path recommendations, reducing the cost caused by the mismatch between the map and the actual use process. It can provide a more accurate prior information for the perception module, planning and control (PnC) module and other modules of the autonomous driving, so that the vehicle can provide users with a safer and more human driving habit-compliant way of passing the intersection.

[0031] Next, the present disclosure will be described in detail with reference to the accompanying drawings.

[0032] Figure 2 It is a flow chart of a method for determining an intersection path according to an embodiment of the present disclosure.

[0033] like Figure 2As shown, the present disclosure also provides a method for determining an intersection path. This method can be applied to devices such as terminal devices or network devices. Among them, terminal devices can include, for example, but are not limited to, mobile phones, wearable devices, tablet computers, handheld computers, laptop computers, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), laptop computers, mobile computers, augmented reality (AR) devices, virtual reality (VR) devices, artificial intelligence (AI) devices, and / or in-vehicle devices, and any other terminal devices or portable terminal devices.

[0034] In some other examples, the network device can be, for example, a server or a server cluster. Of course, it can also be a server or a server cluster on a running virtual machine, and the present disclosure does not make any limitations.

[0035] The method involved in the present disclosure can include the following steps:

[0036] S201, obtain the entry and exit road group information of the first intersection.

[0037] In some examples, the device can obtain the entry and exit road group information of the first intersection. Among them, the first intersection can be an intersection that requires path planning. For example, it can be an intersection where vehicle driving trajectories were not collected during the map production process, or a newly built intersection; or for another example, it can be an intersection that the user is currently about to pass through in actual applications.

[0038] In some examples, the entry and exit road group information can include entry and exit road relationship information. The entry and exit road relationship information can represent the corresponding relationship between the first lane entering the first intersection and the second lane exiting the first intersection. It can be understood that the first intersection can be regarded as the entry intersection, and the second intersection can be regarded as the exit intersection.

[0039] Of course, in other examples, the entry / exit road group information may further include multiple entry / exit road relationship information. Each entry / exit road relationship information may include corresponding entry lanes, exit lanes, lane widths of each lane, lane angles corresponding to the entry / exit road relationship information, lane types of the entry lanes, lane types of the exit lanes, and other information. Therefore, the intersection characteristics corresponding to the intersection can be obtained through the entry / exit lane group information. For example, it may include the number of entry lanes at the first intersection, the number of exit lanes, the lane widths of each lane, the lane angles corresponding to each entry / exit road relationship information, the lane types of the entry lanes, the lane types of the exit lanes, and other information. Among them, the entry lane may be the lane corresponding to entering the intersection, such as the above-mentioned first lane; the exit lane may be the lane corresponding to exiting the intersection, such as the above-mentioned second lane.

[0040] It can be understood that the entry / exit road group information can be regarded as the intersection characteristics abstracted from the first intersection. Of course, in some examples, these characteristics may be vector geometric characteristics, such as including two-dimensional coordinate information and height information. Among them, the two-dimensional coordinate information may correspond to the map coordinates established in the map data, such as longitude and latitude coordinates, etc.

[0041] S202. According to the entry / exit road group information, determine the predicted intersection path from the pre-configured intersection path association information.

[0042] In some examples, the device may determine the predicted intersection path corresponding to the first intersection from the pre-configured intersection path association information according to the entry / exit road group information of the first intersection obtained in S201. Among them, the intersection path association information is the association relationship information between the second intersection and the intersection path of the second intersection. The first intersection and the second intersection are different intersections. The intersection path is represented as the connection line between different entry lanes and exit lanes in an intersection.

[0043] In some examples, a correlation relationship table between intersections and intersection paths may be pre-constructed. Each table corresponds to an intersection and can be uniquely identified by the above-mentioned intersection characteristics. All intersection paths corresponding to the intersection can be stored in each table. Of course, in some examples, height information may also be included. For example, for some intersections with overlapping two-dimensional coordinates of some roads, such as some overpasses, loop intersections, viaducts, etc. The device can match the entry / exit road group information with the correlation relationship table. Determine the correlation relationship table of the second intersection similar to the current first intersection. And use the entry / exit road relationship information in the entry / exit road group information to determine all possible predicted intersection paths. That is to say, the device uses the correlation relationship table of the second intersection similar to the first intersection to predict the possible paths of the first intersection according to the entry / exit road relationship information of the first intersection, that is, determine the predicted intersection path corresponding to the first intersection.

[0044] Of course, the intersection paths stored in the above association table can be pre-configured intersection paths that conform to human driving habits.

[0045] In some other examples, an association model between intersections and intersection paths, that is, an intersection connection model, can be pre-constructed. The intersection connection model can be the intersection connection model corresponding to the second intersection. The device determines the predicted intersection path corresponding to the first intersection based on the entry / exit road relationship information of the first intersection and through the intersection connection model corresponding to the second intersection.

[0046] It can be understood that any equivalent method can also be used to construct the association relationship between intersections and intersection paths, and based on this relationship, the predicted intersection path corresponding to the first intersection is determined using the entry / exit road relationship information of the first intersection. The present disclosure does not make any limitations.

[0047] In some other examples, the first intersection can also be the same as the second intersection, that is, path prediction is performed for the intersection where the vehicle driving trajectory has been collected. For example, it can be applied to the scenario where the user drives to an intersection where the association relationship between intersections and intersection paths has been constructed and re-performs path prediction for this intersection.

[0048] S203. Determine the intersection path of the first intersection based on the predicted intersection path.

[0049] In some examples, the device can use the predicted intersection path determined in S202 as the intersection path of the first intersection. For example, the device directly uses the predicted intersection path corresponding to the first intersection determined in S203 as the intersection path of the first intersection. And it can be used for path recommendation during subsequent autonomous driving.

[0050] The present disclosure predicts possible intersection paths of the current intersection by using pre-configured intersection path association information. It can provide safer and more human-driving-habit-compliant path recommendations, reducing the costs caused by the mismatch between the map and the actual usage process.

[0051] In some embodiments, in order to better simulate human driving habits and more effectively incorporate the height information of the road, that is, path prediction can be performed for roads with the same two-dimensional coordinates but different heights. For the intersection path association information in S202, an intersection connection model can be used. The intersection connection model is pre-trained based on the association relationship between the second intersection and the intersection paths of the second intersection. Figure 3 It is another flowchart of the method for determining the intersection path according to the embodiments of the present disclosure. As Figure 3 shown, S202 determining the predicted intersection path from the pre-configured intersection path association information according to the entry / exit road group information may include the following steps:

[0052] S301. Determine an intersection connection model that matches the entry / exit road group information based on the entry / exit road group information.

[0053] In some examples, the device can determine an intersection connection model that matches the entry / exit road group information. Among them, each intersection connection model can be considered to correspond to an intersection. It can be understood that the intersection connection model is pre-trained based on the association relationship between an intersection and the intersection paths of that intersection. It can be understood that the intersection connection model corresponds to a second intersection.

[0054] For example, the driving trajectories of each intersection and the corresponding intersections can be collected in advance. For instance, it can include the driving trajectory of the host vehicle for data collection, as well as the driving trajectories of other vehicles (which can be called obstacle vehicles) collected by the host vehicle. In the training phase, the network architecture of a neural network can be used for training. For example, the entry / exit road group information corresponding to each path is used as the input of the model, and the real trajectories of the vehicles are used as labels to perform supervised learning training on the neural network. Thus, the intersection connection model corresponding to the corresponding intersection is constructed. Obviously, different intersections correspond to different constructed intersection connection models.

[0055] It can be understood that the vehicle driving trajectories can be obtained through existing trajectory fitting methods, such as using an aggregation method, which will not be elaborated in this disclosure.

[0056] In some examples, an intersection connection model library can be constructed for the intersection connection models corresponding to different intersections. Each intersection connection model can be uniquely identified by the entry / exit road group information corresponding to its respective intersection. Since the road conditions of each intersection vary more or less, the unique intersection can be determined through the entry / exit road group information corresponding to its respective intersection, and thus the intersection connection model corresponding to that intersection can be determined.

[0057] In some examples, the entry / exit road group information of the first intersection can be used to match different intersection connection models in the intersection connection model library. If the entry / exit road group information corresponding to a certain intersection connection model has the highest similarity, then this intersection connection model can be used as the matching intersection connection model. If there are multiple intersection connection models whose corresponding entry / exit road group information has the highest and same similarity with the entry / exit road group information of the first intersection. Then, one can be selected from the multiple intersection connection models as the matching intersection connection model according to a preset method. For example, the preset method can be random selection or setting weights for different intersection connection models and selecting the intersection connection model with a higher weight, etc. The specific method of selecting one from multiple intersection connection models as the matching intersection connection model can be arbitrarily selected according to the actual situation, and this disclosure does not make a limitation.

[0058] It can be understood that the device can access the intersection connection model library in a wired or wireless manner. Among them, the wireless methods involved in the present disclosure may include wireless communication solutions such as 2G / 3G / 4G / 5G / 6G. Or it includes wireless communication solutions such as wireless local area networks (WLAN), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), ZigBee, and infrared technology (IR). Among them, WLAN can be, for example, a wireless fidelity (Wi-Fi) network.

[0059] S302. Obtain a predicted intersection path according to the entry and exit road relationship information and the intersection connection model.

[0060] In some examples, the device can determine a predicted intersection path for the first intersection according to the entry and exit road relationship information of the first intersection and the corresponding intersection connection model of the second intersection determined in S301.

[0061] For example, input the entry and exit road relationship information of the first intersection into the intersection connection model corresponding to the second intersection, so as to obtain the corresponding predicted intersection path of the first intersection. It can be understood that since the entry and exit road relationship information of the second intersection matches that of the first intersection, it can be considered that the second intersection may be very similar to the first intersection. Therefore, the intersection connection model of the second intersection can be used to predict the intersection path of the first intersection. Based on the actual entry and exit road relationship information of the first intersection and the intersection connection model of the second intersection, the possible intersection path of the first intersection is speculated.

[0062] The present disclosure uses the intersection connection model to determine and predict the intersection path of the first intersection, which can ensure that the model better learns human driving habits, so as to more accurately predict the intersection path, thereby improving safety and driving experience.

[0063] In some embodiments, Figure 4 is another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure. As Figure 4 shown, determining the intersection path of the first intersection based on the predicted intersection path in S203 may further include the following steps:

[0064] S401. Determine the path selection probability corresponding to the predicted intersection path according to the pre-configured path selection probability database.

[0065] In some examples, after the device determines the predicted intersection path, it can also determine the corresponding path selection probability for each predicted intersection path according to the pre-configured path selection probability database.

[0066] For example, the actual driving trajectories collected corresponding to the second intersection can be used to determine the user selection situation of each intersection path at the second intersection. For example, 75% of the vehicles drive on intersection path A at the second intersection, 10% of the vehicles drive on intersection path B, and 15% of the vehicles drive on intersection path C. Correspondingly, the path selection probability of intersection path A can be 75%, the path selection probability of intersection path B can be 10%, and the path selection probability of intersection path C can be 15%.

[0067] In other examples, the corresponding path selection probabilities can also be configured for different intersection paths according to principles such as the curvature change or the shortest path of different intersection paths.

[0068] In the pre-configured path selection probability database, the path selection probabilities of each intersection path corresponding to the second intersection are stored. Since the predicted intersection path is determined through the intersection path association information corresponding to the second intersection. Therefore, the device can determine the path selection probability corresponding to the predicted intersection path through the path selection probability database.

[0069] S402. Determine the intersection path of the first intersection with a path selection probability based on the predicted intersection path and the path selection probability corresponding to the predicted intersection path.

[0070] In some examples, the device can determine the intersection path of the first intersection with a path selection probability based on the path selection probability corresponding to the predicted intersection path determined in S401 and the corresponding predicted intersection path. That is to say, when the device determines the intersection path of the first intersection, it determines the path selection probability of the corresponding intersection path at the same time.

[0071] In some examples, for multiple predicted intersection paths with the same entry intersection, the one with the highest path selection probability can also be used as the intersection path of the first intersection. For example, if there are two predicted intersection paths, where predicted intersection path 1 corresponds to entry intersection 1 and exit intersection 1, and predicted intersection path 2 corresponds to entry intersection 1 and exit intersection 2. And the path selection probability corresponding to predicted intersection path 1 is 80%, and the path selection probability corresponding to predicted intersection path 2 is 20%. Then it can be found that the path selection probability corresponding to predicted intersection path 1 is higher. Therefore, predicted intersection path 1 is used as the intersection path of the first intersection, and predicted intersection path 2 is discarded.

[0072] When determining the intersection path of the first intersection, the present disclosure simultaneously determines the path selection probability of the corresponding intersection path, so as to provide users with recommended choices of different paths during the autonomous driving stage, which is beneficial for users to select a safer and more accurate driving path based on different selection probabilities.

[0073] In some embodiments, Figure 5 is another flowchart of the method for determining the intersection path according to the embodiment of the present disclosure. As Figure 5 shown, the method may further include the following steps:

[0074] S501, Obtain vehicle driving trajectory information.

[0075] In some examples, the device may also obtain vehicle driving trajectory information. Among them, the vehicle driving trajectory information corresponds to the intersection path of the first intersection. For example, the device may obtain the vehicle driving trajectory information collected by the vehicle currently passing through the first intersection. For example, a vehicle may drive through the first intersection, collect its own vehicle driving trajectory information, and send it to the device.

[0076] S502, According to the vehicle driving trajectory information, determine the path selection update probability corresponding to the intersection path of the first intersection.

[0077] In some examples, the device may determine the path selection update probability corresponding to the intersection path of the first intersection according to the vehicle driving trajectory information obtained in S501. For example, assume that the intersection paths recommended by the device for the first intersection include intersection path A and intersection path B. When actually applied, the user selects intersection path A and actually drives along intersection path A. Then, the device determines the path selection update probability corresponding to intersection path A of the first intersection through the vehicle driving trajectory information fed back by the vehicle. It can be understood that the path selection update probability is the new selection probability of this intersection path.

[0078] S503, Update the path selection probability corresponding to the corresponding intersection path in the path selection probability database based on the path selection update probability.

[0079] In some examples, the device may update the path selection probability corresponding to the corresponding intersection path in the path selection probability database based on the path selection update probability determined in S502.

[0080] For example, the intersection path A of the first intersection is predicted based on the intersection path A' of the second intersection. After the device determines the path selection update probability corresponding to intersection path A of the first intersection, it can update the path selection probability corresponding to intersection path A' of the second intersection in the path selection probability database.

[0081] The present disclosure can receive vehicle driving trajectory information during the application stage, continuously update the path selection probabilities corresponding to the paths of the corresponding intersections in the path selection probability database, so as to provide more accurate path selection probabilities for the paths of the corresponding intersections. Thus, during the autonomous driving stage, more intersection paths that conform to human driving habits can be provided for users.

[0082] In some embodiments, Figure 6 is another flowchart of a method for determining an intersection path according to an embodiment of the present disclosure. As Figure 6 shown, the method may further include the following steps:

[0083] S601, obtain the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

[0084] In some examples, the device can obtain the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection. Among them, the intersection characteristics of the first intersection include the entry-exit road relationship information of the first intersection. It can be understood that the intersection characteristics of the first intersection may include the number of entry lanes, the number of exit lanes, the lane widths of each lane, the lane angles corresponding to each entry-exit road relationship information, the lane types of the entry lanes, the lane types of the exit lanes, and other information.

[0085] In some examples, the intersection characteristics of the first intersection can be newly collected, for example, collected by a large number of real vehicles during driving. Of course, in other examples, it can also be obtained based on the entry-exit road group information obtained in S201, and the present disclosure does not make a limitation.

[0086] It can be understood that generally, the vehicle driving trajectory information of the first intersection can be collected by real vehicles during driving, and the intersection characteristics of the first intersection can be collected at the same time.

[0087] S602, optimize the intersection connection model with the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

[0088] In some examples, the device can use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection obtained in S601 to optimize the intersection connection model, that is, optimize and train the intersection connection model corresponding to the second intersection.

[0089] It can be understood that since the intersection connection model of the second intersection matches the entry-exit road relationship information of the first intersection, it can be considered that the first intersection is similar to the second intersection. Therefore, the intersection connection model of the second intersection can be optimized by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection. Thus, a more perfect intersection connection model can be obtained.

[0090] The present disclosure optimizes and trains a matching intersection connection model by using the intersection characteristics of a first intersection and the vehicle driving trajectory information of the first intersection, so as to obtain a more perfect intersection connection model, and improve the accuracy when subsequently using the intersection connection model to predict paths for corresponding intersections.

[0091] In some embodiments, corresponding conditions can be preset for optimizing the intersection connection model. Therefore, Figure 7 is a flowchart of another method for determining an intersection path according to an embodiment of the present disclosure. As Figure 7 shown, the method may further include the following steps:

[0092] S701, determine the matching degree between the entry / exit road group information and the intersection connection model.

[0093] In some examples, the device can determine the matching degree between the entry / exit road group information corresponding to the first intersection and the intersection connection model corresponding to the second intersection. For example, when matching the entry / exit road group information corresponding to the first intersection with each model in the intersection connection model library, the matching degree between the entry / exit road group information corresponding to the first intersection and the corresponding intersection connection model can be determined. For example, when determining the intersection model of the matching second intersection, the matching degree between the entry / exit road group information corresponding to the first intersection and the matching one can be determined simultaneously. For example, the matching degree reaches 90%.

[0094] S702, if the matching degree meets the preset condition, use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection to generate an intersection connection model corresponding to the first intersection.

[0095] In some examples, when the device determines that the matching degree between the entry / exit road group information corresponding to the first intersection and the intersection connection model corresponding to the second intersection in S701 meets the preset condition, the device can use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection to generate an intersection connection model corresponding to the first intersection.

[0096] For example, a matching degree threshold can be preset in advance. The preset condition can be that when the matching degree is less than the matching degree threshold, use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection to generate an intersection connection model corresponding to the first intersection. That is to say, if the matching degree between the entry / exit road group information of the first intersection and the intersection connection model of the second intersection does not reach the preset matching degree threshold, it can be considered that the entry / exit road group information of the first intersection and the intersection connection model of the second intersection are not very well matched. Therefore, the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection can be used to generate a corresponding intersection connection model for the first intersection.

[0097] For another example, if the matching degree between the entry / exit road group information of the first intersection and the intersection connection model of the second intersection reaches a preset matching degree threshold, that is, the matching degree is greater than or equal to the matching degree threshold. Then it can be considered that the entry / exit road group information of the first intersection is basically matched with the intersection connection model of the second intersection. It can be considered that the first intersection and the second intersection are the same intersection or identical intersections. Then the steps in Figure 6 can be executed to optimize the intersection connection model of the second intersection.

[0098] Of course, in some other examples, the matching degree may not be considered, and the intersection connection model corresponding to the first intersection can be directly generated by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection. The present disclosure does not make any limitations.

[0099] By generating a new intersection connection model for the first intersection when the first intersection and the second intersection do not match well, the present disclosure enriches the intersection connection model library, which is beneficial to subsequent path prediction for other intersections and improves the accuracy of the predicted path.

[0100] Figure 8 is a schematic flowchart of a process for determining an intersection path according to an embodiment of the present disclosure. As Figure 8 shown, the present disclosure provides a schematic flowchart of a process for determining an intersection path. The process can generally include three parts, namely, the part of establishing an intersection connection model library, the part of map annotation, and the part of vehicle-end application. It can be understood that usually, the establishment of the model library can be applied to a network device, the map annotation can be applied to a network device or a terminal device, and the vehicle-end application can be applied to a terminal device. Of course, this method can be applied to any device such as a network device or a terminal device according to the actual situation, and the present disclosure does not make any limitations.

[0101] Among them, the part of establishing an intersection connection model library can include the following steps:

[0102] S801, sample feature extraction.

[0103] In some examples, the device extracts the features of each sample from a large number of intersection samples obtained. For example, the intersection characteristics corresponding to each intersection and the possible vehicle driving trajectories corresponding to each intersection are extracted.

[0104] S802, supervised training of the intersection connection model.

[0105] In some examples, the device can perform supervised training on a neural network model based on the intersection characteristics extracted in S801 and the vehicle driving trajectories of the corresponding intersections, so as to train the intersection connection models corresponding to each intersection.

[0106] In some examples, after obtaining the intersection connection models corresponding to different intersections, an intersection connection model library can be constructed based on each intersection connection model.

[0107] The map annotation part can include the following steps:

[0108] S803. Obtain the information of the entry / exit road group.

[0109] In some examples, the device obtains the information of the entry / exit road group of the first intersection. The information of the entry / exit road group can include the entry / exit road relationship information. The entry / exit road relationship information can represent the corresponding relationship between the first lane entering the first intersection and the second lane exiting the first intersection.

[0110] In some examples, it can be to select and mark the entry lanes and exit lanes of a certain intersection, so as to determine the information of the entry / exit road group of the first intersection.

[0111] S804. Match with the intersection connection model library.

[0112] In some examples, the device can match the obtained information of the entry / exit road group in S804 with the constructed intersection connection model library, so as to determine the matching intersection connection model.

[0113] It can be understood that if multiple intersection connection models are matched, the intersection connection model with the highest matching degree can be selected as the successfully matched intersection connection model. Of course, in some examples, if there are multiple intersection connection models with the highest matching degree, one of the intersection connection models can be selected according to a preset method. For example, a random method or any equivalent method can be adopted, which is not limited in this disclosure.

[0114] S805. Perform map annotation.

[0115] In some examples, the device can annotate the map through the intersection connection model. For example, all possible intersection paths of the first intersection are determined through the information of the entry / exit road group of the first intersection and the matched intersection connection model. And all possible intersection paths of the first intersection are annotated at the position of the first intersection on the map.

[0116] S806. Determine whether the intersection path is the first annotation.

[0117] In some examples, the device can determine whether the annotation of the first intersection on the map this time is the first annotation. If so, continue to execute S807, otherwise, S808 can be executed.

[0118] S807. Determine the initial path selection probability.

[0119] In some examples, if it is determined that this annotation of the first intersection in the map is the first annotation, the device can determine the initial path selection probabilities pre-configured for each intersection path corresponding to the first intersection.

[0120] S808, update the path selection probabilities.

[0121] In some examples, if it is determined that this annotation of the first intersection in the map is not the first annotation, the device can determine the path selection update probabilities for each intersection path corresponding to the first intersection from the path selection probability database. Then, use the path selection update probabilities to annotate each intersection path of the first intersection.

[0122] In some examples, the initial path selection probabilities can also be pre-stored in the path selection probability database, and the initial path selection probabilities in S807 can also be determined through the path selection probability database. The present disclosure does not make any limitations.

[0123] The vehicle-end application part can include the following steps:

[0124] S809, obtain the map.

[0125] In some examples, the device can obtain the annotated map data. It can be understood that the annotated map data contains the recommended intersection paths of the corresponding intersections and the path selection probabilities of each intersection path.

[0126] In some examples, the annotated map data can be obtained through the PnC module on the device.

[0127] S810, the device makes a path decision based on the intersection paths of the first intersection and based on preset conditions, and controls the vehicle to drive according to the path decision result.

[0128] In some examples, the device can determine the intersection paths to be executed according to the preset decision scheme based on the annotated map obtained in S809, and control the vehicle to drive according to the decision result.

[0129] For example, when the vehicle travels to the first intersection, the vehicle determines through the annotated map obtained in S809 that there are two intersection paths available for driving at present, namely intersection path XX and intersection path YY. Assume that the path selection probability of intersection path XX is x%, and the path selection probability of intersection path YY is y%. The device can select an intersection path as the upcoming intersection path according to the pre-configured decision scheme. The pre-configured decision scheme can, for example, preferentially select the intersection path with the highest path selection probability. Or based on the current environment around the vehicle, such as detecting obstacles in some lanes, etc., and comprehensively make a decision to determine a more suitable intersection path for driving. It can be understood that the present disclosure does not make any limitations on how to make a decision.

[0130] In some examples, after the vehicle is controlled to travel along a certain intersection path, the vehicle travel trajectory can be obtained, and the path selection probability of the corresponding intersection path in the path selection probability database can be updated.

[0131] For example, if the vehicle chooses to travel along intersection path XX, the path selection probability corresponding to the intersection path of the corresponding intersection in the path selection probability database can be updated, for example, updated to x’%.

[0132] S811. Obtain environmental information.

[0133] In some examples, the device can also obtain the environmental information around the vehicle, such as the intersection characteristics of the first intersection, so as to store this data and the vehicle travel trajectory together as an intersection sample, which is convenient for subsequent periodic or aperiodic update of the intersection connection model, or generating an intersection connection model corresponding to the first intersection based on the newly stored intersection sample.

[0134] It can be understood that Figure 8 In the manner shown, first establish an intersection connection model, and then when labeling intersection paths on the map, by determining the matching intersection connection model, the intersection paths of each intersection can be predicted. For example, it includes intersections where vehicle travel trajectories have not been collected, or newly built intersections, or intersections where vehicle travel trajectories have been collected, etc.

[0135] The present disclosure predicts possible intersection paths of the current intersection by using pre-configured intersection path association information, which can provide safer and more human-driving-habit-compliant path recommendations, reduce the costs caused by the mismatch between the map and the actual use process, and improve the user experience.

[0136] Based on the same concept, the embodiments of the present disclosure also provide a device for determining intersection paths.

[0137] It can be understood that, in order to implement the above functions, a device for determining intersection paths provided by the embodiments of the present disclosure includes corresponding hardware structures and / or software modules for executing each function. Combining the units and algorithm steps of the various examples disclosed in the embodiments of the present disclosure, the embodiments of the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the way of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the technical solution of the embodiments of the present disclosure.

[0138] As an exemplary implementation manner Figure 9It is a schematic diagram of a device for determining intersection paths shown in an exemplary embodiment of the present disclosure. Refer to Figure 9 As shown, a device 900 for determining intersection paths is provided. The device 900 can implement any of the methods involved in the above Figures 2 to 8 . The device 900 may include: an acquisition module 901, configured to acquire entry / exit road group information of a first intersection, where the entry / exit road group information includes entry / exit road relationship information, and the entry / exit road relationship information represents the corresponding relationship between a first lane entering the first intersection and a second lane exiting the first intersection; a determination module 902, configured to determine a predicted intersection path from pre-configured intersection path association information according to the entry / exit road group information, where the intersection path association information is association relationship information between a second intersection and the intersection path of the second intersection, and the intersection path represents the connection line between different lanes in the intersection, and the first intersection is different from the second intersection; the determination module 902 is further configured to determine the intersection path of the first intersection based on the predicted intersection path.

[0139] The present disclosure predicts possible intersection paths of the current intersection by using pre-configured intersection path association information. It can provide safer and more human-driving-habit-compliant path recommendations, reducing the costs caused by the mismatch between the map and the actual use process.

[0140] In some possible implementation manners, the intersection path association information is an intersection connection model, and the intersection connection model is pre-trained based on the association relationship between the intersection paths of the second intersection and the second intersection; the determination module 902 is further configured to: determine an intersection connection model that matches the entry / exit road group information according to the entry / exit road group information; obtain the predicted intersection path according to the entry / exit road relationship information and the intersection connection model.

[0141] The present disclosure uses the intersection connection model to determine the prediction of the intersection path of the first intersection, which can ensure that the model better learns human driving habits, so as to more accurately predict the intersection path, thereby improving safety and driving experience.

[0142] In some possible implementation manners, the determination module 902 is further configured to: determine the path selection probability corresponding to the predicted intersection path according to a pre-configured path selection probability database; determine the intersection path of the first intersection with the path selection probability based on the predicted intersection path and the path selection probability corresponding to the predicted intersection path.

[0143] When the present disclosure determines the intersection path of the first intersection, it simultaneously determines the path selection probability of the corresponding intersection path, so as to provide users with selection recommendations for different paths during the automatic driving stage, which is beneficial for users to select safer and more accurate driving paths based on different selection probabilities.

[0144] In some possible embodiments, the apparatus 900 further includes an updating module 903; the obtaining module 901 is further configured to obtain vehicle driving trajectory information, where the vehicle driving trajectory information corresponds to the intersection path of the first intersection; the determining module 902 is further configured to determine a path selection update probability corresponding to the intersection path of the first intersection according to the vehicle driving trajectory information; the updating module 903 is configured to update the path selection probability corresponding to the corresponding intersection path in the path selection probability database based on the path selection update probability.

[0145] The present disclosure can receive vehicle driving trajectory information in the application stage and continuously update the path selection probability corresponding to the corresponding intersection path in the path selection probability database, so as to provide a more accurate path selection probability for the corresponding intersection path. Thus, in the autonomous driving stage, a more human-driving-habit-compliant intersection path can be provided for users.

[0146] In some possible embodiments, the apparatus 900 further includes an optimization module 904; the obtaining module 901 is further configured to obtain the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection, where the intersection characteristics include the entering / exit road relationship information of the intersection; the optimization module 904 is configured to optimize the intersection connection model by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

[0147] The present disclosure optimizes and trains the matching intersection connection model by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection, so as to obtain a more perfect intersection connection model and improve the accuracy of subsequent path prediction for the corresponding intersection by using the intersection connection model.

[0148] In some possible embodiments, the apparatus 900 further includes a generating module 905; the determining module 902 is further configured to determine the matching degree between the entering / exit road group information and the intersection connection model; the generating module 905 is configured to, if the matching degree meets a preset condition, generate an intersection connection model corresponding to the first intersection by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

[0149] The present disclosure generates a new intersection connection model for the first intersection when the first intersection does not match well with the second intersection, thereby enriching the intersection connection model library, which is beneficial to subsequent path prediction for other intersections and improving the accuracy of the predicted path.

[0150] Regarding the apparatus involved in the above-mentioned present disclosure, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0151] In the technical solution of the present disclosure, the acquisition, storage, and application of the user's personal information all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0152] According to an embodiment of the present disclosure, the present disclosure further provides a device for determining an intersection path, a readable storage medium, and a computer program product.

[0153] Figure 10 FIG. shows a schematic block diagram of a device 1000 for determining an intersection path that can be used to implement an embodiment of the present disclosure. It can be understood that the device 1000 can be a terminal device or a network device. The device 1000 is intended to represent various forms of digital computers, such as, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, a server cluster, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0154] As Figure 10 shown, the device 1000 includes a computing unit 1001, which can execute various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0155] Multiple components in the device 1000 are connected to the I / O interface 1005, including: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a disk, an optical disc, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0156] The computing unit 1001 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1001 executes the various methods and processes described above, such as Figures 2 to 8 any of the methods described. For example, in some embodiments, Figures 2 to 8 any of the methods described can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 1008. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 1000 via the ROM 1002 and / or the communication unit 1009. When the computer program is loaded into the RAM 1003 and executed by the computing unit 1001, one or more steps of any of the methods described above can be executed. Alternatively, in other embodiments, the computing unit 1001 can be configured to execute the above Figures 2 to 8 any of the methods described by any other suitable means (e.g., by means of firmware).

[0157] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor, that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0158] The program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that the program codes, when executed by the processor or controller, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The program codes may be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0159] In the context of the present disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0160] In order to provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0161] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.

[0162] A computer system can include a client and a server. The client and the server are generally far from each other and usually interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, can also be a server of a distributed system, or a server incorporating a blockchain. Of course, in some examples, the server can also refer to a server cluster.

[0163] The present disclosure first makes a more reasonable improvement to the way of marking intersection connections, uses machine learning to abstract models and refine rules for different types of intersection connections, and automatically matches the most suitable connection model in production. Additionally, during use, the map model is continuously strengthened and improved based on actual vehicle operation data.

[0164] The solution involved in the present disclosure can be better applied to the planning and decision-making of autonomous vehicles at intersections in an urban scenario. Compared with the current situation where autonomous driving still relies on the information of high-precision maps in complex road conditions. After building an intersection map through the intersection model established by this solution, the vehicle will be more in line with human driving habits when passing through the intersection, improving safety and the driving experience.

[0165] At the same time, for the mass production of high-precision maps, the solution involved in the present disclosure greatly improves the automation degree of intersection mapping, improves efficiency, and reduces the costs caused by the mismatch between the map and the actual usage situation.

[0166] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved. There is no limitation herein.

[0167] The above specific embodiments do not limit the scope of protection of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present disclosure shall be included within the scope of protection of the present disclosure.

Claims

1. A method for determining an intersection path, the method comprises: Obtain the entry / exit road group information of a first intersection, wherein the entry / exit road group information includes entry / exit road relationship information, and the entry / exit road relationship information represents the corresponding relationship between a first lane entering the first intersection and a second lane exiting the first intersection; According to the entry / exit road group information, determine a predicted intersection path from the pre-configured intersection path association information, wherein the intersection path association information is the association relationship information between a second intersection and the intersection path of the second intersection, the intersection path represents the connection line between different lanes in the intersection, and the first intersection is different from the second intersection; Based on the predicted intersection path, determine the intersection path of the first intersection; Wherein, the determining the intersection path of the first intersection based on the predicted intersection path includes: According to the pre-configured path selection probability database, determine the path selection probability corresponding to the predicted intersection path; Based on the predicted intersection path and the path selection probability corresponding to the predicted intersection path, determine the intersection path of the first intersection with a path selection probability.

2. The method according to claim 1, wherein, The intersection path association information is an intersection connection model, and the intersection connection model is pre-trained based on the association relationship between a second intersection and the intersection path of the second intersection; The determining a predicted intersection path from the pre-configured intersection path association information according to the entry / exit road group information includes: According to the entry / exit road group information, determine the intersection connection model that matches the entry / exit road group information; According to the entry / exit road relationship information and the intersection connection model, obtain the predicted intersection path.

3. The method according to claim 1, wherein, The method further comprises: Obtain vehicle driving trajectory information, and the vehicle driving trajectory information corresponds to the intersection path of the first intersection; According to the vehicle driving trajectory information, determine the path selection update probability corresponding to the intersection path of the first intersection; Based on the path selection update probability, update the path selection probability corresponding to the corresponding intersection path in the path selection probability database.

4. The method according to any one of claims 1-3, wherein, The method further comprises: Obtain the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection, and the intersection characteristics include the entry / exit road relationship information of the intersection; Use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection to optimize the intersection connection model.

5. The method according to claim 4, wherein, The method further comprises: Determine the matching degree between the entry / exit road group information and the intersection connection model; If the matching degree meets the preset condition, use the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection to generate the intersection connection model corresponding to the first intersection.

6. An apparatus for determining an intersection path, comprises: An acquisition module, configured to acquire entry / exit road group information of a first intersection, where the entry / exit road group information includes entry / exit road relationship information, and the entry / exit road relationship information represents the corresponding relationship between a first lane entering the first intersection and a second lane exiting the first intersection; A determination module, configured to determine a predicted intersection path from pre-configured intersection path association information according to the entry / exit road group information, where the intersection path association information is association relationship information between a second intersection and an intersection path of the second intersection, the intersection path represents a connection line between different lanes in the intersection, and the first intersection is different from the second intersection; The determination module is further configured to determine the intersection path of the first intersection based on the predicted intersection path; Wherein, the determination module is further configured to: Determine the path selection probability corresponding to the predicted intersection path according to a pre-configured path selection probability database; Determine the intersection path of the first intersection with a path selection probability based on the predicted intersection path and the path selection probability corresponding to the predicted intersection path.

7. The apparatus according to claim 6, Wherein, The intersection path association information is an intersection connection model, and the intersection connection model is pre-trained based on the association relationship between the second intersection and the intersection path of the second intersection; The determination module is further configured to: Determine the intersection connection model that matches the entry / exit road group information according to the entry / exit road group information; Obtain the predicted intersection path according to the entry / exit road relationship information and the intersection connection model.

8. The apparatus according to claim 6, Wherein, The apparatus further includes an update module; The acquisition module is further configured to acquire vehicle driving trajectory information, and the vehicle driving trajectory information corresponds to the intersection path of the first intersection; The determination module is further configured to determine the path selection update probability corresponding to the intersection path of the first intersection according to the vehicle driving trajectory information; The update module is configured to update the path selection probability corresponding to the corresponding intersection path in the path selection probability database based on the path selection update probability.

9. The apparatus according to any one of claims 6-8, Wherein, The apparatus further includes an optimization module; The acquisition module is further configured to acquire the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection, and the intersection characteristics include the entry / exit road relationship information of the intersection; The optimization module is configured to optimize the intersection connection model with the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

10. The apparatus according to claim 9, Wherein, The apparatus further includes a generation module; The determination module is further configured to determine the matching degree between the entry / exit road group information and the intersection connection model; The generation module is configured to, if the matching degree meets a preset condition, generate an intersection connection model corresponding to the first intersection by using the intersection characteristics of the first intersection and the vehicle driving trajectory information of the first intersection.

11. An apparatus for determining a path at an intersection, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions, when executed by the at least one processor, enable the at least one processor to perform the method according to any one of claims 1-5.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein, the computer instructions are for causing the computer to perform the method according to any one of claims 1-5.

13. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-5.

Images