Vehicle navigation method and device, vehicle and computer program product

Abstract

The invention discloses a vehicle navigation method and device, a vehicle and a computer program product, and relates to the technical field of vehicle navigation. The method comprises the steps that real-time navigation data and real-time environment data of a vehicle are acquired, and the real-time navigation data are generated based on cloud road network information corresponding to a current navigation layer of the vehicle in a multi-layer overpass; a pre-trained target recognition model is adopted to recognize the real-time environment data, a recognition result is obtained, and the recognition result is used for representing the current driving layer of vehicles in the multi-layer overpass; based on the real-time navigation data and the identification result, performing matching verification on the current navigation layer and the current driving layer to obtain a verification result; and according to a verification result, performing navigation reset updating on the real-time navigation data. The technical problem of low vehicle driving safety caused by low accuracy of a navigation system in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of vehicle navigation technology, and in particular, to a vehicle navigation method, device, vehicle and computer program product. Background Art

[0002] Since the multi-layer roads inside the overpass are visually similar, it is easy for drivers to mistakenly enter other driving layers when driving. In related technologies, navigation systems usually rely on global satellite navigation systems to provide real-time location information and path planning for vehicles. However, in the complex scene of the overpass, due to the lack of accurate height information, it is difficult for the navigation system to accurately determine the vertical position of the vehicle, resulting in low accuracy of navigation path planning, which in turn leads to low vehicle driving safety.

[0003] In addition, the navigation correction function mainly relies on map matching technology to match the vehicle position obtained by the global satellite navigation system with the pre-stored road network position. However, navigation correction can usually only be performed when there is a significant deviation between the vehicle position and the road network position. In the complex scene of the overpass, the paths of different bridge layers have small deviations in the road network position, making it difficult to perform navigation corrections in a timely and accurate manner.

[0004] As can be seen from the above, how to improve the accuracy of the navigation system and improve the driving safety of vehicles has become one of the important technical issues in the relevant technical field. Currently, no effective solution has been proposed for the above problems. Summary of the invention

[0005] The embodiments of the present application provide a vehicle navigation method, device, vehicle and computer program product to at least solve the technical problem in the related art that the navigation system has low accuracy, resulting in low vehicle driving safety.

[0006] According to one aspect of an embodiment of the present application, a vehicle navigation method is provided, comprising: acquiring real-time navigation data and real-time environmental data of the vehicle, wherein the real-time navigation data is generated based on cloud-based road network information corresponding to a current navigation layer of the vehicle in a multi-layer overpass; using a pre-trained target recognition model to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize a current driving layer of the vehicle in the multi-layer overpass; based on the real-time navigation data and the recognition result, matching and verifying the current navigation layer and the current driving layer to obtain a verification result; and resetting and updating the navigation of the real-time navigation data according to the verification result.

[0007] Furthermore, the target recognition model is obtained through machine learning training based on multiple groups of multimodal training data, each group of multimodal training data includes: road sample images and text annotations, wherein the text annotations are used to characterize the target driving layer corresponding to the road sample image in the multi-layer overpass.

[0008] Furthermore, real-time environmental data is acquired in real time by an image acquisition device configured for the vehicle, and the real-time environmental data includes at least a road image corresponding to a current driving layer; a target recognition model is used to recognize the real-time environmental data, and a recognition result is obtained, including: using a target recognition model to predict and recognize the current driving layer corresponding to the road image to obtain a model prediction result; and performing spatiotemporal filtering processing on the model prediction result to obtain a recognition result.

[0009] Further, the model prediction results are subjected to spatiotemporal filtering processing to obtain recognition results, including: based on the target spatiotemporal filtering index, the model prediction results are smoothed and filtered to obtain recognition results, wherein the target spatiotemporal filtering index includes a time continuity index and a driving distance continuity index.

[0010] Furthermore, obtaining real-time navigation data includes: obtaining a navigation sequence corresponding to a current trip of the vehicle; identifying at least a portion of the sequence in the navigation sequence corresponding to a multi-layer overpass; and annotating attributes of a current navigation layer corresponding to at least a portion of the sequence to obtain real-time navigation data.

[0011] Furthermore, based on the real-time navigation data and the recognition results, a matching check is performed on the current navigation layer and the current driving layer, and the verification result includes: analyzing based on the real-time navigation data and the recognition results to determine whether the current navigation layer and the current driving layer correspond to the same layer in the multi-layer overpass; in response to the current navigation layer and the current driving layer corresponding to the same layer in the multi-layer overpass, the verification result determines that the current navigation layer and the current driving layer match each other.

[0012] Further, according to the verification result, resetting and updating the navigation of the real-time navigation data includes: in response to determining according to the verification result that the current navigation layer and the current driving layer do not match each other, resetting and updating the navigation of the real-time navigation data.

[0013] According to another aspect of an embodiment of the present application, a vehicle navigation device is also provided, including: an acquisition module, used to acquire real-time navigation data and real-time environmental data of the vehicle, wherein the real-time navigation data is generated based on cloud-based road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass; an identification module, used to use a pre-trained target recognition model to identify the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; a verification module, used to match and verify the current navigation layer and the current driving layer based on the real-time navigation data and the recognition result to obtain a verification result; and an update module, used to reset and update the navigation of the real-time navigation data according to the verification result.

[0014] According to another aspect of an embodiment of the present application, a vehicle is also provided, including an on-board memory and an on-board processor, wherein a computer program is stored in the on-board memory, and the on-board processor is configured to run the computer program to execute any one of the above-mentioned vehicle navigation methods.

[0015] According to another aspect of an embodiment of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the vehicle navigation method according to any one of the above items is implemented.

[0016] In an embodiment of the present application, real-time navigation data and real-time environmental data of a vehicle are obtained, wherein the real-time navigation data is generated based on cloud-based road network information corresponding to a current navigation layer of the vehicle in a multi-layer overpass; a pre-trained target recognition model is used to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; based on the real-time navigation data and the recognition result, a match verification is performed on the current navigation layer and the current driving layer to obtain a verification result; and according to the verification result, the navigation of the real-time navigation data is reset and updated.

[0017] It is easy to notice that the embodiment of the present application generates real-time navigation data of the vehicle by using cloud road network information, uses a pre-trained target recognition model to recognize the environmental data collected by the vehicle in real time, and combines the recognition result with the real-time navigation data to verify the current navigation layer and the current driving layer. Therefore, the embodiment of the present application achieves the purpose of providing a more accurate navigation route, thereby achieving the technical effect of improving the accuracy of vehicle navigation and the safety of vehicle driving, and further solves the technical problem of low accuracy of the navigation system in the related art, resulting in low vehicle driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and the related descriptions of the embodiments are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 is a hardware structure block diagram of an optional computing terminal for implementing a vehicle navigation method according to an embodiment of the present application;

[0020] Figure 2 is a flow chart of a vehicle navigation method according to an embodiment of the present application;

[0021] Figure 3 is a block diagram of a system structure for vehicle navigation during driving on a multi-layer overpass section according to an embodiment of the present application;

[0022] Figure 4is a flow chart of a bridge layer attribute identification process according to an embodiment of the present application;

[0023] Figure 5 is a flow chart of a vehicle navigation update process according to an embodiment of the present application;

[0024] Figure 6 It is a structural block diagram of a vehicle navigation device according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments only include a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] According to an embodiment of the present application, a method embodiment of a vehicle navigation method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0028] First, the operating environment of the above method embodiment is exemplarily described. Figure 1 is a hardware structure block diagram of an optional computing terminal for implementing a vehicle navigation method according to an embodiment of the present application, such as Figure 1As shown, a computing terminal 10 (for example, a computer terminal, a mobile smart terminal, a vehicle terminal, or a cloud computing virtual terminal, etc.) may include: one or more processors 102, a memory 104 for storing data, and a transmission device 106 for implementing a communication function, wherein each processor 102 may include but is not limited to a processing component such as a microprocessor (Microcontroller Unit, referred to as MCU) or a programmable logic device (Field Programmable Gate Array, referred to as FPGA).

[0029] The computing terminal 10 may further include: a display device 110, an input / output device 108, a Universal Serial Bus (USB) port (the USB port may be used as one of the ports of a computer bus, not shown in the figure), a network interface (not shown in the figure), a power supply (not shown in the figure) and a camera (not shown in the figure). A person skilled in the art may understand that Figure 1 The structure of the computing terminal 10 shown is only for illustration and does not impose a strict limitation on the structure of the computing terminal 10. For example, the computing terminal 10 may also include Figure 1 More or fewer components than those shown in FIG. 1 , or computing terminal 10 may have the same components as those shown in FIG. Figure 1 Different categories of components are shown.

[0030] It should be noted that the one or more processors 102 and / or other data processing circuits in the computing terminal 10 may be embodied in whole or in part as software, hardware, firmware or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the vehicle terminal 10 (or mobile device).

[0031] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the vehicle navigation method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned vehicle navigation method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the vehicle terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0032] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the vehicle terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0033] In the above operating environment, an embodiment of the present application provides a vehicle navigation method. The vehicle navigation method can be used to provide an autonomous driving vehicle navigation function for a preset application scenario. The above preset application scenarios may include the following scenarios in the field of household cars: commuting autonomous driving scenarios, artificial intelligence (ie, Artificial Intelligence, referred to as AI) driving scenarios for household cars, automatic parking assist (Automatic Parking Assist, referred to as APA) scenarios, and intelligent navigation assistance (Navigation Guided Pilot, referred to as NGP) scenarios in urban areas or high-speed areas. In addition, the above preset application scenarios may also include, but are not limited to: transportation vehicle navigation scenarios of smart driving trucks or unmanned trucks in the field of logistics and transportation, operation route navigation scenarios of autonomous driving agricultural vehicles in the field of agricultural machinery, flight route navigation scenarios of drones, and mobile path navigation scenarios of intelligent robots (such as cleaning robots, service robots, delivery robots, etc.).

[0034] When the above-mentioned preset application scenarios are scenarios in fields other than the field of household cars, those skilled in the art should be able to understand that the target vehicle in the above-mentioned vehicle navigation method is replaced with other objects (such as agricultural machinery, drones, robots, etc.), and correspondingly, the target turning section is replaced with the area to be traveled related to other objects, and the data related to other vehicles is replaced with data related to other objects (for example, the body size data of the vehicle is replaced with the body size data of the drone, etc.). On this basis, the specific implementation method of the above-mentioned vehicle navigation method is exemplified in the embodiment of the present application, taking the field of household cars as an example.

[0035] Under the above operating environment, the embodiment of the present application provides the following Figure 2 The vehicle navigation method shown, Figure 2 is a flow chart of a vehicle navigation method according to an embodiment of the present application, such as Figure 2 As shown, the method includes the following steps S201 to S204.

[0036] Step S201, obtaining real-time navigation data and real-time environment data of the vehicle, wherein the real-time navigation data is generated based on cloud road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass.

[0037] The above-mentioned real-time navigation data may be information such as the path, position, direction, etc. updated in real time during the driving process of the vehicle. The above-mentioned real-time environmental data may include but are not limited to: roads, buildings, traffic signs, pedestrians, other vehicles, temperature, etc. The real-time environmental data can be obtained through a variety of sensors on the vehicle (such as cameras, radar sensors, temperature sensors). The above-mentioned current navigation layer may be a specific bridge layer in a multi-layer overpass where the vehicle should be located according to the real-time navigation path. The above-mentioned cloud road network information may be road network data stored in a cloud server, and the cloud road network information may include but are not limited to the location, direction, height, bridge layer, etc. of the road, and the cloud road network information can be updated in real time.

[0038] The specific implementation method of acquiring real-time environmental data may be: multiple sensors on the vehicle collect the vehicle's surrounding environment data in real time and transmit it to the vehicle navigation system.

[0039] The specific implementation method of generating real-time navigation data based on the cloud road network information corresponding to the current navigation layer can be: first, the navigation system identifies the section of the navigation route that the vehicle is about to travel that contains multi-layer overpasses according to the current navigation route, and determines the current navigation layer from the multi-layer overpasses; secondly, after the navigation system determines the current navigation layer, the navigation system will obtain the road network data corresponding to the current navigation layer from the cloud; finally, the precise location data in the cloud road network information is used to locate the vehicle in real time, obtain the location information, calculate the navigation path, and generate real-time navigation data. The above-mentioned location information may include but is not limited to the bridge layer of the vehicle in the multi-layer overpass and the position of the vehicle on the road.

[0040] It is easy to understand that in the embodiments of the present application, by utilizing the vehicle's real-time navigation data and real-time environmental data, the vehicle's navigation path can be dynamically adjusted according to the real-time vehicle driving conditions and environmental changes around the vehicle during vehicle driving, thereby improving the real-time and accuracy of the vehicle's navigation solution.

[0041] Step S202, using a pre-trained target recognition model to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass.

[0042] The target recognition model may be a machine learning model trained with a large amount of data, and the target recognition model may be used to identify the current driving layer of a vehicle in a multi-layer overpass. The current driving layer may be a specific bridge layer in a multi-layer overpass where the vehicle is actually located during the current driving process.

[0043] Exemplarily, current driving images captured by multiple vehicle-mounted cameras are used as input, and feature extraction and scene classification are performed on the current driving images through a deep learning algorithm (such as a convolutional neural network) to identify the current driving layer of the vehicle during driving on a multi-layer overpass.

[0044] The specific implementation method of using the pre-trained target recognition model to recognize the real-time environmental data and obtain the recognition result can be: during the vehicle driving process, the environmental data is collected in real time, the real-time environmental data is used as input, and the real-time environmental data is processed using the trained target recognition model to identify the structural features related to the multi-layer overpass in the environmental data (such as the distribution of bridge piers, the connection method of the bridge deck, etc.), thereby determining the current driving layer of the vehicle.

[0045] It is easy to understand that in the embodiment of the present application, by utilizing the target recognition model to identify real-time environmental data, the information of the vehicle's current driving layer can be identified in a timely manner, ensuring that the navigation system can respond quickly to environmental changes and improve driving safety.

[0046] Step S203, based on the real-time navigation data and the recognition result, the current navigation layer and the current driving layer are matched and verified to obtain a verification result.

[0047] The above verification results can be used to indicate the compatibility of the current navigation layer and the current driving layer.

[0048] Based on the real-time navigation data and the recognition result, the current navigation layer and the current driving layer are matched and verified. The specific implementation method for obtaining the verification result can be: the navigation system matches and verifies the information of the current navigation layer in the real-time navigation data with the information of the current driving layer in the recognition result, determines whether the current navigation layer and the current driving layer match, and obtains the verification result.

[0049] It is easy to understand that in the embodiment of the present application, based on the real-time navigation data and recognition results, it is possible to verify whether the navigation path information provided by the navigation system matches the current actual driving path information of the vehicle, especially in a complex multi-layer overpass driving environment, the current navigation layer and the current driving layer can be matched and verified. Compared with the navigation system in the related art that only relies on the global satellite navigation system, the embodiment of the present application can promptly discover the situation where the vehicle driving path does not match the navigation path, improve the accuracy and reliability of the navigation system, and thus improve the safety of vehicle driving.

[0050] Step S204: Reset and update the real-time navigation data according to the verification result.

[0051] According to the verification result, a specific implementation method of resetting and updating the real-time navigation data may be: the navigation system recalculates the navigation path based on the verification result to generate new real-time navigation data.

[0052] It is easy to understand that in the embodiment of the present application, the verification result is used to reset and update the real-time navigation data, and the erroneous navigation path can be automatically detected and corrected according to the verification result, thereby ensuring the real-time nature of the navigation data, improving the accuracy of the navigation data, and improving the robustness of the navigation system, which can effectively avoid driving risks caused by navigation path errors.

[0053] Through the technical solutions provided by the above steps S201 to S204, the embodiment of the present application obtains the real-time navigation data and real-time environmental data of the vehicle, wherein the real-time navigation data is generated based on the cloud road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass; the real-time environmental data is identified by using a pre-trained target recognition model to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; based on the real-time navigation data and the recognition result, the current navigation layer and the current driving layer are matched and verified to obtain a verification result; according to the verification result, the real-time navigation data is reset and updated. It is easy to notice that the embodiment of the present application generates the real-time navigation data of the vehicle by using the cloud road network information, identifies the environmental data collected by the vehicle in real time by using the pre-trained target recognition model, and verifies the current navigation layer and the current driving layer in combination with the recognition result and the real-time navigation data. Therefore, the embodiment of the present application achieves the purpose of providing a more accurate navigation route, thereby achieving the technical effect of improving the accuracy of vehicle navigation and improving the safety of vehicle driving, thereby solving the technical problem of low accuracy of the navigation system in the related technology, resulting in low vehicle driving safety.

[0054] Other optional implementations of the above-mentioned vehicle navigation method of the embodiment of the present application are further described below.

[0055] As an optional implementation, in the above-mentioned vehicle navigation method, the target recognition model is obtained through machine learning training based on multiple groups of multimodal training data, and each group of multimodal training data includes: road sample images and text annotations, wherein the text annotations are used to characterize the target driving layer corresponding to the road sample image in a multi-layer overpass.

[0056] The road sample images may be actual road images of vehicles traveling on different bridge layers in a multi-layer overpass environment, and the road sample images may include but are not limited to bridge deck features, bridge pier distribution, road markings, traffic signs, etc. The text annotations may be text descriptions of scene information in each road sample image.

[0057] In order to improve the accuracy of the target recognition model in identifying the driving layer, after obtaining multiple road sample images, corresponding detailed text annotations are added to each road sample image to obtain multiple groups of multimodal training data. The target recognition model (such as the visual-language model based on the Transformer architecture) is trained using the multiple groups of multimodal training data. The parameters of the model are adjusted through the back propagation algorithm so that the model can extract features related to the bridge layer from the image and match them with the text annotations. After multiple iterations, the accuracy of the target recognition model in identifying the driving layer is gradually improved.

[0058] Through the technical solution provided above, in the embodiment of the present application, the target recognition model is trained using multiple groups of multimodal training data, which enables the target recognition model to learn the image features of each bridge layer in the multi-layer overpass, thereby accurately identifying the specific bridge layer where the vehicle is currently traveling in actual applications, thereby improving the accuracy of the target recognition model in identifying the driving layer.

[0059] As an optional implementation, the real-time environment data is acquired in real time by an image acquisition device configured by the vehicle, and the real-time environment data at least includes a road image corresponding to the current driving layer; in the above step S202, the real-time environment data is recognized by using a target recognition model, and obtaining the recognition result may also include the following execution steps:

[0060] Step S2021, using the target recognition model to predict and recognize the current driving layer corresponding to the road image to obtain a model prediction result;

[0061] Step S2022, performing spatiotemporal filtering processing on the model prediction result to obtain a recognition result.

[0062] The above model prediction result can be a preliminary judgment based on the target recognition model to identify the current driving layer in the road image. The above spatiotemporal filtering process can be used to remove noise in the model prediction result and smooth the model prediction result. The above noise can be noise caused by the influence of time or driving distance.

[0063] By using the target recognition model to predict and identify the current driving layer, it is possible to actively perceive and understand the vertical structure of the current road environment, ensuring that the navigation system can timely and accurately update the vehicle position and driving layer information. Compared with relying on global satellite navigation system positioning, the embodiment of the present application can provide more detailed and reliable information on the vertical positioning of multi-layer overpasses. Even when the global satellite navigation system signal is poor or inaccurate, it can provide correct navigation guidance based on environmental characteristics, avoiding navigation errors caused by positioning errors, and improving the intelligence and robustness of the navigation system.

[0064] By performing spatiotemporal filtering on the model prediction results, the target recognition model can more accurately determine the current driving layer of the vehicle in the multi-layer overpass, enhance the accuracy of the model prediction, and reduce the uncertainty of single image prediction.

[0065] In an exemplary application scenario, a vehicle is traveling on a two-level overpass section. The vehicle traveling on the first level overpass section is recorded as under the bridge, and the vehicle traveling on the second level overpass section is recorded as on the bridge. Figure 3 As shown, in the target recognition model module, a multimodal vision-language large model is used to predict and recognize the road image corresponding to the current driving layer, determine whether the vehicle is currently on or under the bridge, and obtain the model prediction result.

[0066] Through the technical solutions provided by the above steps S2021 to S2022, the embodiment of the present application adopts a target recognition model to predict and identify the current driving layer of the road image, and performs spatiotemporal filtering on the model prediction results. This not only improves the accuracy of the navigation system's vehicle positioning and the accuracy of the navigation route in complex traffic environments, but also enhances the safety of vehicle driving.

[0067] As an optional implementation, in the above step S2022, performing spatiotemporal filtering on the model prediction result to obtain the recognition result may also include the following execution steps:

[0068] Step S20221, based on the target spatiotemporal filtering index, the model prediction result is smoothed and filtered to obtain the recognition result, wherein the target spatiotemporal filtering index includes a time continuity index and a driving distance continuity index.

[0069] The target spatiotemporal filtering index can be a signal processing index that combines the time dimension and the space dimension, and the target spatiotemporal filtering index can be used to filter and optimize the model prediction results. The time continuity index can be used to indicate the continuity and consistency of the model prediction results in the time series. The driving distance continuity index can be used to indicate the consistency between the vehicle driving path and the model prediction results.

[0070] Based on the target spatiotemporal filtering index, the model prediction results are smoothed and filtered. The specific implementation steps to obtain the recognition results can be: according to the model prediction results of the target recognition model within a period of time, time series analysis techniques (such as sliding window averaging, dynamic time warping, etc.) are used to identify and eliminate abnormal model prediction results; according to the vehicle's driving path, the model prediction results are spatially smoothed to ensure that the model prediction results match the actual driving path of the vehicle.

[0071] By introducing the time continuity indicator, the navigation system can perform consistency analysis on the model's continuous prediction results over a period of time, eliminate abnormal prediction results caused by external interference (such as light changes, occlusion, etc.), and thus improve the stability of the recognition results.

[0072] By introducing the driving distance continuity indicator, the navigation system can combine the vehicle's driving trajectory and perform spatial smoothing on the model's prediction results. Even in the case of occasional misjudgment of the model, the navigation system can correct the recognition results based on the continuity of the vehicle's actual driving path, thereby ensuring the accuracy of the positioning information and reducing navigation errors.

[0073] In an exemplary application scenario, Figure 3 As shown, the spatiotemporal filtering technology is used to perform consistency analysis on the continuous prediction results of the model within a period of time, and the prediction results of the model are spatially smoothed in combination with the vehicle's driving trajectory, and the false detection or missed detection prediction results of the model are filtered to obtain more accurate recognition results.

[0074] Through the technical solution provided by the above step S20221, the embodiment of the present application performs smooth filtering on the model prediction results based on the target spatiotemporal filtering index, which not only improves the stability and accuracy of the recognition results, reduces the false alarm rate, but also enhances the robustness and adaptability of the navigation system.

[0075] As an optional implementation, in the above step S201, acquiring real-time navigation data may further include the following execution steps:

[0076] Step S2011, obtaining the navigation sequence corresponding to the current trip of the vehicle;

[0077] Step S2012, identifying at least a portion of the sequence in the navigation sequence corresponding to the multi-layer overpass;

[0078] Step S2013: annotate the attributes of the current navigation layer corresponding to at least part of the sequence to obtain real-time navigation data.

[0079] The navigation sequence of the current trip may be sequence information of the entire navigation route of the vehicle from the starting point to the destination, and the navigation sequence of the current trip may include data such as global satellite navigation system coordinate points, road types, and driving directions. At least part of the sequence corresponding to the multi-layer overpass may include location information such as the entrance, exit, and turning point of the multi-layer overpass. The attribute annotation may be used to determine the attribute information of the bridge layer.

[0080] By analyzing the complete navigation sequence of the vehicle's journey, the navigation system can accurately identify multi-layer overpass sections and then label the attributes of the current driving layer, ensuring that the navigation system can fully consider the bridge layer on which the vehicle is actually driving, avoiding path planning errors caused by inaccurate recognition of bridge layer information, and improving the accuracy of path planning.

[0081] In an exemplary application scenario, Figure 3 As shown, in the cloud map module, the navigation sequence of the current trip corresponding to the vehicle is obtained, the navigation sequence of the current trip is judged, the navigation sequences corresponding to multiple bridge layers are automatically identified and the attributes of the sequences are marked to prevent the navigation system from erroneously triggering navigation switching due to the influence of other bridge layers.

[0082] The above process of identifying the attributes of the bridge layer is as follows: Figure 4 As shown, in the cloud map module, the recognition results obtained in the target recognition model module and the real-time navigation data are obtained to determine whether the navigation data is consistent with the recognition results. If the navigation data is consistent with the recognition results, the on-bridge attributes / under-bridge attributes are output; if the navigation information is inconsistent with the recognition results, the unknown attributes are output.

[0083] Through the technical solutions provided by the above steps S2011 to S2013, the embodiment of the present application generates a technical solution of real-time navigation data by identifying the part related to the multi-layer overpass in the navigation sequence and annotating the attributes of the current navigation layer. This can not only improve the accuracy and real-time performance of the navigation path planning of the navigation system, but also provide the driver with more accurate and intuitive navigation information, thereby enhancing the safety of vehicle driving.

[0084] As an optional implementation, in the above step S203, based on the real-time navigation data and the recognition result, the current navigation layer and the current driving layer are matched and verified, and obtaining the verification result may also include the following execution steps:

[0085] Step S2031, analyzing based on the real-time navigation data and the recognition result, determining whether the current navigation layer and the current driving layer correspond to the same layer in the multi-layer overpass;

[0086] Step S2032, in response to the current navigation layer and the current driving layer corresponding to the same layer in the multi-layer overpass, the verification result determines that the current navigation layer and the current driving layer match each other.

[0087] In order to ensure that the path information provided by the navigation system corresponds to the bridge layer on which the vehicle is actually traveling, and to improve the accuracy and reliability of navigation, the system compares and analyzes the current navigation layer information contained in the real-time navigation data with the current driving layer identification result to determine whether the current navigation layer and the current driving layer correspond to the same layer in a multi-layer overpass.

[0088] In an exemplary application scenario, Figure 3 As shown, based on the real-time navigation data and recognition results, analysis is performed to determine whether the current navigation layer and the current driving layer correspond to the same layer in a multi-layer overpass. If the current navigation layer and the current driving layer correspond to the same layer in a multi-layer overpass, the verification result determines that the current navigation layer and the current driving layer match, and the navigation data is not updated.

[0089] Through the technical solutions provided by the above steps S2031 to S2032, the navigation system of the present application can determine in real time whether the bridge layer currently actually driven by the vehicle is consistent with the bridge layer displayed by the navigation planning route through comparative analysis of the real-time navigation data and the recognition results. This real-time verification mechanism avoids the situation where the navigation route is wrong due to confusion of bridge layer information, and improves the accuracy of navigation information.

[0090] As an optional implementation, in the above step S204, according to the verification result, resetting and updating the real-time navigation data may further include the following execution steps:

[0091] Step S2041, in response to determining according to the verification result that the current navigation layer and the current driving layer do not match each other, the real-time navigation data is reset and updated.

[0092] In response to determining that the current navigation layer and the current driving layer do not match each other according to the verification result, the real-time navigation data is reset and updated: the navigation system determines whether the real-time navigation data needs to be updated based on the verification result. If the verification result shows that the current navigation layer and the current driving layer do not match each other, the navigation system will recalculate the navigation path based on the accurate information of the current driving layer and generate new real-time navigation data.

[0093] It should be noted that if the verification result shows that the current navigation layer matches the current driving layer, it means that the navigation data is accurate and there is no need to update the navigation data.

[0094] In an exemplary application scenario, Figure 3 As shown, in response to determining that the current navigation layer and the current driving layer do not match each other according to the verification result, the navigation data is updated, a "switch navigation route" button is generated on the navigation display interface, and the navigation route is switched according to the button pressing result. In addition, in response to determining that the current navigation layer and the current driving layer do not match each other according to the verification result, the navigation system can also automatically switch the navigation route.

[0095] The above process of switching navigation routes according to button pressing results is as follows: Figure 5As shown in the figure, when the verification result is that the current navigation layer does not match the current driving layer, the flag bit (recorded as hw_flag) is set according to the recognition result output by the target recognition model module. When the user chooses not to switch navigation, hw_flag is set to 0; when the user chooses to switch navigation and the target recognition model module outputs the result as on the bridge, hw_flag is set to 1; when the user chooses to switch navigation and the target recognition model module outputs the result as under the bridge, hw_flag is set to 2.

[0096] When hw_flag=0 in the autonomous driving controller, the central domain controller does not process; when hw_flag=1 or hw_flag=2 in the autonomous driving controller, the central domain controller responds to the route on the bridge switching to the route under the bridge or the route under the bridge switching to the route on the bridge, the navigation system re-plans the navigation route, switches the navigation route, and the autonomous driving controller records the information of navigation route deviation and continues to monitor whether the current navigation layer matches the current driving layer.

[0097] Through the technical solution provided by the above step S2041, in the embodiment of the present application, when the system detects that the current navigation layer information does not match the actual driving layer of the vehicle, the real-time navigation data is reset and updated to ensure that the navigation system can quickly respond to the actual driving conditions and can correct the navigation route in time when deviations occur in the navigation route, thereby improving the accuracy and real-time performance of navigation.

[0098] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0099] It should be noted that, for the above method embodiments, for the sake of simplicity, the technical solutions in the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of actions in the described action combinations, because according to the present application, some of the above steps can be performed in other orders or simultaneously. In addition, those skilled in the art should also be aware that the embodiments described in the specification of this application belong to preferred embodiments, and the actions and modules involved are not necessarily necessary to implement the technical solutions of the present application.

[0100] According to the embodiment of the present application, a device embodiment of a vehicle navigation device is also provided, and the vehicle navigation device is used to implement the above method embodiment and various optional implementation methods of the method embodiment. The technical contents described above are not repeated in the device embodiment. It should be noted that in the following related descriptions of the device embodiment, "module" can be software, hardware, or a combination of software and hardware for implementing a specified function. Figure 6 As shown, the device comprises:

[0101] An acquisition module 601 is used to acquire real-time navigation data and real-time environment data of a vehicle, wherein the real-time navigation data is generated based on cloud road network information corresponding to a current navigation layer of the vehicle in a multi-layer overpass;

[0102] The recognition module 602 is used to recognize the real-time environment data using a pre-trained target recognition model to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass;

[0103] A verification module 603 is used to perform a matching verification on the current navigation layer and the current driving layer based on the real-time navigation data and the recognition result to obtain a verification result;

[0104] The updating module 604 is used to perform navigation reset and update on the real-time navigation data according to the verification result.

[0105] It should be noted that the above modules 601 to 604 correspond to steps S201 to S204 in the method embodiment. The four modules are the same as the instances and application scenarios implemented by the corresponding steps, but are not limited to the contents disclosed in the above method embodiment.

[0106] Optionally, in the above-mentioned vehicle navigation device, the target recognition model is obtained through machine learning training based on multiple groups of multimodal training data, and each group of multimodal training data includes: road sample images and text annotations, wherein the text annotations are used to characterize the target driving layer corresponding to the road sample image in the multi-layer overpass.

[0107] Optionally, in the above-mentioned vehicle navigation device, the real-time environmental data is acquired in real time by an image acquisition device configured for the vehicle, and the real-time environmental data includes at least a road image corresponding to the current driving layer; the above-mentioned recognition module 602 is also used to: use the target recognition model to predict and recognize the current driving layer corresponding to the road image to obtain a model prediction result; perform spatiotemporal filtering on the model prediction result to obtain the recognition result.

[0108] Optionally, the above-mentioned recognition module 602 is also used to: based on the target spatiotemporal filtering index, smooth and filter the model prediction result to obtain the recognition result, wherein the target spatiotemporal filtering index includes a time continuity index and a driving distance continuity index.

[0109] Optionally, the acquisition module 601 is further used to: acquire a navigation sequence corresponding to the current trip of the vehicle; identify at least a portion of the sequence in the navigation sequence corresponding to the multi-layer overpass; and perform attribute annotation on the current navigation layer corresponding to at least a portion of the sequence to obtain the real-time navigation data.

[0110] Optionally, the verification module 603 is also used to: perform analysis based on the real-time navigation data and the recognition result to determine whether the current navigation layer and the current driving layer correspond to the same layer in the multi-layer overpass; in response to the current navigation layer and the current driving layer corresponding to the same layer in the multi-layer overpass, the verification result determines that the current navigation layer and the current driving layer match each other.

[0111] Optionally, the updating module 604 is further configured to: in response to determining according to the verification result that the current navigation layer and the current driving layer do not match each other, perform navigation reset and update on the real-time navigation data.

[0112] It should be noted that the various modules mentioned in the above device embodiments can be implemented by software, hardware, or a combination of software and hardware. For example, when hardware is used to implement the above modules, the various modules can be set in the same processor, or the various modules can be set in different processors in any combination. For another example, the above modules can be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n), and the above modules can also be run as part of the device on a processor such as Figure 1 In the computing terminal 10 shown.

[0113] Through the device technical solutions corresponding to the above-mentioned acquisition module, identification module, verification module, and update module, the embodiment of the present application acquires the real-time navigation data and real-time environmental data of the vehicle through the acquisition module, wherein the real-time navigation data is generated based on the cloud road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass; the recognition module uses a pre-trained target recognition model to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; the verification module matches and verifies the current navigation layer and the current driving layer based on the real-time navigation data and the recognition result to obtain a verification result; the update module performs navigation reset and update on the real-time navigation data according to the verification result. As a result, the embodiment of the present application achieves the purpose of providing a more accurate navigation route, thereby realizing the technical effect of improving the accuracy of vehicle navigation and improving the safety of vehicle driving, thereby solving the technical problem of low accuracy of the navigation system in the related technology, which leads to low vehicle driving safety.

[0114] According to an embodiment of the present application, a vehicle embodiment is also provided, which includes an on-board memory and an on-board processor, wherein a computer program is stored in the on-board memory, and the on-board processor is configured to run the computer program to implement the above-mentioned vehicle navigation method.

[0115] According to an embodiment of the present application, a computer program product is also provided. The computer program product includes a computer program, and when the computer program is executed by a processor, the above-mentioned vehicle navigation method can be implemented.

[0116] Optionally, the computer program product may provide vehicle navigation services based on the vehicle navigation method.

[0117] Optionally, in this embodiment, the computer program product may be a set of instructions and codes pre-written according to the vehicle navigation method. The computer program product may run on various computer platforms, including personal computers, servers, mobile devices, etc.

[0118] Optionally, in this embodiment, the instructions and codes corresponding to the computer program product are used to implement the following method steps: obtaining real-time navigation data and real-time environmental data of the vehicle, wherein the real-time navigation data is generated based on cloud-based road network information corresponding to a current navigation layer of the vehicle in a multi-layer overpass; using a pre-trained target recognition model to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; based on the real-time navigation data and the recognition result, matching and verifying the current navigation layer and the current driving layer to obtain a verification result; and resetting and updating the navigation of the real-time navigation data according to the verification result.

[0119] In the above-mentioned multiple embodiments of the present application, the description of each embodiment has different emphases. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0120] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of multiple modules can be a logical function division, and there can be any other possible division methods when actually implemented in the application scenario; for another example, multiple modules (or units or components in the modules) can be combined with each other and can be integrated into another system. For example, some features in the method embodiments described above can be ignored or skipped.

[0121] It should be noted that in the above-mentioned embodiments, the modules, components or units described as separate parts can be physically separated or physically integrated. The parts displayed as modules or units can be physical modules or physical units, or virtual modules or virtual units, that is, multiple modules or multiple units can be in the same position, or distributed in multiple positions or multiple spaces. In the application scenario, according to the actual needs of the scenario, some or all of the multiple modules or multiple units can be selected to implement the technical solutions of the embodiments of the present application, thereby achieving the corresponding technical objectives.

[0122] In particular, for an integrated functional module or functional unit, if it is implemented in the form of a software functional unit and sold or used as an independent product, the module or functional unit can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.

[0123] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A vehicle navigation method, characterized in that: include: Acquire real-time navigation data and real-time environment data of the vehicle, wherein the real-time navigation data is generated based on cloud road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass; Using a pre-trained target recognition model to recognize the real-time environmental data to obtain a recognition result, wherein the recognition result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; Based on the real-time navigation data and the recognition result, performing a matching check on the current navigation layer and the current driving layer to obtain a check result; According to the verification result, the real-time navigation data is reset and updated.

2. The vehicle navigation method according to claim 1, characterized in that: The target recognition model is obtained through machine learning training based on multiple groups of multimodal training data, each group of multimodal training data includes: road sample images and text annotations, wherein the text annotations are used to characterize the target driving layer corresponding to the road sample image in the multi-layer overpass.

3. The vehicle navigation method according to claim 1 or 2, characterized in that: The real-time environment data is acquired in real time by an image acquisition device configured by the vehicle, and the real-time environment data at least includes a road image corresponding to the current driving layer; The target recognition model is used to recognize the real-time environmental data, and the recognition result obtained includes: Using the target recognition model to predict and recognize the current driving layer corresponding to the road image to obtain a model prediction result; The model prediction result is subjected to spatiotemporal filtering processing to obtain the recognition result.

4. The vehicle navigation method according to claim 3, characterized in that: The model prediction result is subjected to spatiotemporal filtering to obtain the recognition result including: Based on the target spatiotemporal filtering index, the model prediction result is smoothed and filtered to obtain the recognition result, wherein the target spatiotemporal filtering index includes a time continuity index and a driving distance continuity index.

5. The vehicle navigation method according to claim 1, characterized in that: Acquiring the real-time navigation data includes: Obtaining a navigation sequence corresponding to a current trip of the vehicle; identifying at least a portion of the navigation sequence corresponding to the multi-level overpass; Attribute annotation is performed on the current navigation layer corresponding to the at least partial sequence to obtain the real-time navigation data.

6. The vehicle navigation method according to claim 1, characterized in that: Based on the real-time navigation data and the recognition result, the current navigation layer and the current driving layer are matched and verified, and the verification result obtained includes: Performing analysis based on the real-time navigation data and the recognition result to determine whether the current navigation layer and the current driving layer correspond to the same layer in the multi-layer overpass; In response to the current navigation layer and the current driving layer corresponding to the same layer in the multi-layer overpass, the verification result is determined as the current navigation layer and the current driving layer match each other.

7. The vehicle navigation method according to claim 1, characterized in that: According to the verification result, resetting and updating the real-time navigation data includes: In response to determining, according to the verification result, that the current navigation layer is not consistent with the current driving layer, a navigation reset and update is performed on the real-time navigation data.

8. A vehicle navigation device, characterized in that: include: An acquisition module, used to acquire real-time navigation data and real-time environment data of the vehicle, wherein the real-time navigation data is generated based on cloud road network information corresponding to the current navigation layer of the vehicle in the multi-layer overpass; an identification module, used to identify the real-time environmental data using a pre-trained target recognition model to obtain an identification result, wherein the identification result is used to characterize the current driving layer of the vehicle in the multi-layer overpass; A verification module, configured to perform a matching verification on the current navigation layer and the current driving layer based on the real-time navigation data and the recognition result to obtain a verification result; An updating module is used to perform navigation reset and update on the real-time navigation data according to the verification result.

9. A vehicle, characterized in that: The vehicle navigation method comprises an on-board memory and an on-board processor, wherein the on-board memory stores a computer program, and the on-board processor is configured to run the computer program to execute the vehicle navigation method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the vehicle navigation method according to any one of claims 1 to 7.

Images