Positioning of a vehicle in a multi-level road system

Abstract

The invention relates to positioning of a vehicle in a multilevel road system. The invention relates to a method performed by a road recognition system of a vehicle to position the vehicle in a multilevel road system. The road recognition system determines a position of the vehicle with support of a positioning system. The road recognition system also identifies a multilevel road system in which the vehicle is located based on the vehicle position with support of at least a first digital map. The road recognition system also obtains image data with support of one or more image capturing devices adapted to capture a surrounding environment of the vehicle. The road recognition system determines a road level in which the vehicle is located based on the fed image data by a neural network trained to classify road levels based on image content of the environment. The road recognition system identifies a road and / or a lane of the multilevel road system in which the vehicle is located based on the determined road level with support of at least the first digital map.

Description

Technical Field

[0001] This invention relates to the positioning of vehicles in multi-level road systems. Background Technology

[0002] Large cities typically have highway-like roads that pass through and / or encircle the city to alleviate traffic congestion from narrow local roads. Such cities can construct several levels of roads with different exit and merging lanes, i.e., multi-level road systems, which have roads that can travel in the same direction.

[0003] However, multi-level road systems can pose problems when it comes to determining which road level a vehicle is located within. That is, automatic positioning technologies are typically based on commercially available (often relatively low-level) positioning systems, such as GNSS or GPS. While a vehicle's geographic location (e.g., horizontal position) can be derived from this, it can present challenges (e.g., in the absence of sensors such as altimeters and / or atmospheric pressure sensors) in deriving the vehicle's vertical position. Summary of the Invention

[0004] Therefore, the purpose of the embodiments herein is to provide a method for identifying the road and / or lane where a vehicle is located in a multi-road system in an improved and / or alternative manner.

[0005] The above objectives can be achieved through the subject matter disclosed herein. Embodiments are set forth in the appended claims, the following description, and the accompanying drawings.

[0006] The disclosed subject matter relates to a method for locating a vehicle in a multi-level road system, performed by a road recognition system. The road recognition system determines the vehicle's position with the support of a positioning system. The road recognition system also identifies the multi-level road system in which the vehicle is located based on its position (with the support of at least a first digital map). Furthermore, the road recognition system acquires image data with the support of one or more image acquisition devices suitable for capturing the vehicle's surrounding environment. The road recognition system further determines the road level in which the vehicle is located based on the fed image data using a neural network trained to classify road levels based on the environment of the image content. Additionally, the road recognition system identifies (with the support of at least a first digital map) the road and / or lane of the multi-level road system in which the vehicle is located based on the determined road level.

[0007] The disclosed subject matter also relates to a road recognition system for locating vehicles in a multi-level road system. The road recognition system includes a location determination unit for (and / or adapted to) determining the vehicle's position with the support of a positioning system. The road recognition system also includes a multi-level system recognition unit for (and / or adapted to) identifying the multi-level road system in which the vehicle is located based on its position, supported by at least a first digital map. Furthermore, the road recognition system includes an image data acquisition unit for (and / or adapted to) acquiring image data with the support of one or more image acquisition devices adapted to acquire the vehicle's surrounding environment. Additionally, the road recognition system includes a level determination unit for (and / or adapted to) determining the road level in which the vehicle is located based on the fed image data using a neural network trained to classify road levels based on the environment of the image content. Moreover, the road recognition system includes a road / lane recognition unit for (and / or adapted to) identifying the road and / or lane in the multi-level road system in which the vehicle is located based on the determined road level, supported by at least a first digital map.

[0008] Furthermore, the disclosed subject matter relates to a vehicle that includes the road recognition system described herein.

[0009] Furthermore, the disclosed subject matter relates to a computer program product comprising a computer program containing computer program code means arranged to cause a computer or processor to perform the steps of the road recognition system described herein, the computer program being stored on a computer-readable medium or carrier wave.

[0010] The disclosed subject matter also relates to a non-volatile computer-readable storage medium having the computer program product stored thereon.

[0011] Therefore, a method is introduced that can identify the correct road and / or lane of a multi-level road system in which a vehicle is located. That is, since the vehicle's position is determined (with the support of a positioning system), the vehicle's geographic location, such as its horizontal position, is established. Furthermore, since the multi-level road system in which the vehicle is located is identified based on the vehicle's position (with the support of at least a first digital map), the multi-level road system can be identified in the digital maps by mapping the vehicle's position to one or more digital maps, whereby, considering the digital maps, the vehicle is considered to be located in or near the multi-level road system. Moreover, since image data is acquired with the support of one or more image acquisition devices suitable for acquiring the vehicle's surrounding environment, data is derived from one or more acquired images showing the vehicle's surrounding environment. Furthermore, since the road level in which the vehicle is located is determined by a neural network based on the fed image data, this neural network is trained to classify road levels based on the background of the image content. By utilizing a pre-trained neural network, the current road level can be established, which is adapted to identify and subsequently classify road levels based on clues, cues, and / or indications of the scene, setting, and / or situation of the vehicle's surrounding environment from the acquired image data. In other words, image data is processed by a neural network pre-trained (based on visual and environmental cues) to classify what road level an image might indicate. For example, the neural network can be trained to recognize and / or classify an image showing the sky with no structure and / or road structure above it as indicating the highest road level and / or a road without an upper road level, while an image showing road structure above the vehicle and vulnerable road users at the same level as the vehicle can be recognized and / or classified by the neural network as a ground road level and / or a lower road level. Moreover, that is, since the road and / or lane of the multi-level road system in which the vehicle is located is identified based on the determined road level with the support of at least a first digital map, by processing the image data through a neural network and establishing potential, feasible, and / or possible road levels associated with the current vehicle position in the face of a digital map, the corresponding road level of the multi-level road system can be identified, and the corresponding road and / or lane of the road level associated with the current vehicle position can then be derived. Accordingly, through the introduced concept, the correct road level can be found from multiple road levels of the multi-level road system, and the correct road and / or lane in which the vehicle is located can then be determined.

[0012] For this reason, a method is provided for identifying the road and / or lane in which a vehicle is located in a multi-road system in an improved and / or alternative manner.

[0013] The technical features and advantages of the above methods will be discussed in more detail below. Attached Figure Description

[0014] Various aspects of non-limiting embodiments, including specific features and advantages, will be readily understood from the following detailed description and accompanying drawings, in which:

[0015] Figure 1 A schematic diagram illustrating an exemplary road recognition system for a vehicle according to an embodiment of the present invention is shown;

[0016] Figure 2 This is a schematic block diagram illustrating an exemplary road recognition system according to an embodiment of the present invention; and

[0017] Figure 3 This is a flowchart illustrating an exemplary method performed by a road recognition system according to an embodiment of the present invention. Detailed Implementation

[0018] Non-limiting embodiments of the invention will now be described more fully below with reference to the accompanying drawings, in which presently preferred embodiments of the invention are illustrated. However, the invention may be presented in many different forms and should not be construed as limited to the embodiments set forth herein. The same reference numerals always refer to the same elements. Dashed lines in some of the boxes in the figures indicate that these elements or actions are optional rather than mandatory.

[0019] In the following, based on embodiments relating to vehicle positioning in a multi-level road system, a method will be disclosed that can identify the correct road and / or lane in which a vehicle is located in a multi-level road system.

[0020] Now, referring to the attached diagram, Figure 1 The diagram shows an exemplary road recognition system 1 for a vehicle 2 according to an embodiment of the present invention, while... Figure 2 A schematic block diagram of an exemplary road recognition system 1 according to an embodiment of the present invention is shown. The road recognition system 1 is adapted to locate a vehicle 2 in a multi-level road system 3.

[0021] The exemplary vehicle 2 can be represented by any (e.g., known) manned or unmanned vehicle, such as an engine-driven or electric vehicle, like a car, truck, van, minivan, bus, tractor, and / or motorcycle. Furthermore, according to the example, vehicle 2 can optionally refer to "autonomous and / or at least partially autonomous vehicle," "unmanned and / or at least partially unmanned vehicle," and / or "autonomous and / or at least partially autonomous vehicle."

[0022] A multi-level road system 3 can be represented by an arbitrary road structure with arbitrary dimensions, having multiple road levels 4, for example, having two or more road levels 4 extending at least to some extent in the same or substantially the same road direction. In an exemplary embodiment... Figure 1The diagram illustrates two road levels 4: a first ground-level road level 41 and a second upper-level road level 42. It can be noted that, although not shown here, the multi-level road system 3 may optionally include one or more underground road levels 4.

[0023] The phrase “road recognition system” can refer to “road assignment system,” “road and / or lane recognition system,” and / or “road hierarchy assessment system,” while “of” the vehicle can mean “included in” the vehicle and / or “mounted” on the vehicle. On the other hand, “for” locating a vehicle can mean “suitable for” locating a vehicle, and “for locating said vehicle” can mean “for locating said vehicle,” “for road and / or lane assignment of said vehicle,” “for road hierarchy assessment,” and / or “for assessing the road hierarchy of said vehicle.” Furthermore, the phrase “multi-level road system” can always refer to “multi-level road system,” “road system comprising at least two road levels,” and / or “multi-level road structure and / or network.”

[0024] The road recognition system 1 (e.g., via location determination unit 101) is adapted and / or configured to determine the location 20 of the vehicle 2 with the support of the positioning system 21. Thus, the geographic location 20 of the vehicle 2, such as its horizontal position 20, is established.

[0025] Determining the location 20 of vehicle 2 with the support of positioning system 21 can be achieved in any (e.g., known) manner, potentially through additional support from dead reckoning calculations and / or similar methods. Similarly, positioning system 21 can be represented by any (e.g., known) sensors and / or functions suitable for sensing and / or determining (e.g., the vehicle's) whereabouts and / or geographic location (e.g., via GNSS, such as GPS). Thus, positioning system 21 can be at least partially included (and / or mounted) in vehicle 2, for example, associated with vehicle 2's (e.g., known) optional navigation system and / or (e.g., known) optional perception system, optional advanced driver assistance system (ADAS), and / or optional automatic driving (AD) system.

[0026] The phrase “determine the location of…” can mean “derive the location of…”, while “location” can mean “geographical location” and / or “horizontal location”. On the other hand, the phrase “with the support of a positioning system” can mean “through input from a positioning system”, “from a positioning system”, “with the support of at least a positioning system” and / or “with the support of a positioning system including and / or mounted in the vehicle”, while “the location of the vehicle” can mean “vehicle location”.

[0027] The road recognition system 1 (e.g., via multi-layer system recognition unit 102) is adapted and / or configured to identify the multi-layer road system 3 in which the vehicle 2 is located based on the vehicle location 20, with the support of at least a first digital map 22. Thus, by mapping the vehicle location 20 to one or more digital maps 22, the multi-layer road system 3 can be identified in the digital map, for example... Figure 1 In the example of the multi-level road system, considering digital map 22, vehicle 2 is considered to be located in or near multi-level road system 3.

[0028] Identifying the multi-level road system 3 in which vehicle 2 is located based on vehicle location 20 can be accomplished in any manner (e.g., known) based on digital map 22. For example, the multi-level road system 3 can be identified and / or marked in digital map 22, and / or it can be derived from providing multiple digital maps 22 for a specific geographic area and / or road segment (e.g., corresponding digital maps covering the corresponding road levels) if the multi-level road structure 3 is applicable to said specific geographic area and / or road segment. At least the first digital map 22 can be represented by any (e.g., known) digital map, such as a high-definition (HD) map and / or its equivalents and / or successor maps. Moreover, the digital map 22 can be at least partially included (and / or incorporated into) vehicle 2, for example, associated with vehicle 2's (e.g., known) navigation system and / or (e.g., known) optional perception system, optional ADAS, and / or optional AD system.

[0029] The phrase “identify…multi-level road system” can mean “derive…multi-level road system” and / or “determine…multi-level road system”, while “identify based on the vehicle location with the support of at least a first digital map” can mean “identify by mapping the vehicle location to at least a first digital map”, “identify based on the vehicle location with respect to at least a first digital map”, and / or “based on comparing the vehicle location with at least a first digital map”. Furthermore, the phrase “at least a first digital map” can mean “one or more digital maps”, and, by example, further means only “digital map”. The phrase “multi-level road system in which the vehicle is located” can mean “determine and / or assume the multi-level road system in which the vehicle is located” and / or “the multi-level road system in which the vehicle is located or within a predetermined distance of the vehicle location”.

[0030] The road recognition system 1 (e.g., via image data acquisition unit 104) is adapted and / or configured to acquire image data 5 with the support of one or more image acquisition devices 23 adapted to acquire the surrounding environment of the vehicle 2. Thus, data 5 is derived from one or more images of the vehicle's surrounding environment acquired by the image acquisition devices 23.

[0031] Image data 5 can be acquired in any (e.g., known) manner with the support of at least the first image acquisition device 23, for example, derived from one or more images acquired by said image acquisition device 23 (e.g., continuously and / or intermittently). Similarly, one or more image acquisition devices 23, which may include and / or be mounted in the vehicle 2 and distributed in any feasible manner, can be represented by any sensors, functions, and / or systems (e.g., one or more cameras) suitable for acquiring the surrounding environment of the vehicle 2. According to an example, the image acquisition device 23 can be configured in association with optional perception systems, optional ADAS, and / or optional AD systems of the vehicle 2 (e.g., known). Furthermore, image data 5 can be represented by any feasible data derived from the image acquisition device 23 and can also have any feasible size and / or format. In addition, image data 5 can cover any portion of the vehicle's surrounding environment in any direction of the vehicle 2 (e.g., in at least the forward direction of the vehicle 2).

[0032] The phrase "acquiring image data" can refer to "deriving and / or acquiring image data," while "image data" can refer to "image data of at least a portion of the vehicle's surrounding environment" and / or "one or more images." Furthermore, according to the example, "acquiring image data with the support of..." can refer to "acquiring image data from one or more images derived with the support of...". On the other hand, the phrase "suitable for acquiring the vehicle's surrounding environment" can refer to "suitable for acquiring at least a portion of the vehicle's surrounding environment."

[0033] Road recognition system 1 (e.g., via hierarchy determination unit 105) is adapted and / or configured to determine the road hierarchy in which vehicle 2 is located based on fed image data 5 using a neural network trained to classify road hierarchies based on the environment of the image content. Thus, by utilizing a pre-trained neural network adapted to identify and subsequently classify road hierarchies based on clues, cues, and / or indications of the scene, setting, and / or situation of the vehicle 2's surrounding environment from the acquired image data 5, a current road hierarchy can be established. That is, image data 5 passes through a neural network pre-trained to classify (based on visual and environmental cues) what road hierarchy the image might indicate. For example, the neural network can be trained to identify and / or classify an image showing the sky with no structure and / or road structure above it as indicating the highest road hierarchy and / or a road without an upper road hierarchy, while an image showing the road structure above the image and vulnerable road users at the same level as the vehicle can be identified and / or classified by the neural network as a ground road hierarchy and / or a lower road hierarchy. In an exemplary manner... Figure 1In the image data 5, the image data includes and / or displays, in an exemplary manner, the road structure (including the upper road level 42) above the vehicle 2 and the environment of the bases of buildings, pillars and trees in the same horizontal plane as the vehicle 2, so that the neural network can classify the road level associated with it as a ground road level based on the specific environment.

[0034] Optionally, the environment of the image content may refer to (and / or include) features of one or more road structures, such as road ramps. Additionally or alternatively, the environment of the image content may refer to features of one or more static objects, such as buildings and / or pillars (e.g., their bases and / or tops) and / or landscapes and / or topography. Furthermore, additionally or alternatively, the environment of the image content may refer to features of one or more other road users, such as the presence or absence of other road users (e.g., vulnerable road users). Furthermore, additionally or alternatively, the environment of the image data may refer to the extent of sky views and / or side views, such as those obscured (and / or unobstructed) by one or more elevated structures and / or underground walls. Furthermore, additionally or alternatively, the environment of the image data may refer to the extent of ambient light, such as sunlight, direct sunlight, and / or direct moonlight.

[0035] Alternatively, determining the road level where vehicle 2 is located may include feeding image data 5 through a neural network trained to classify road levels based on the environment of the image content and additionally based on geographic location. Thus, the environment considering the image content can vary and / or have different meanings based on geographic location, such as country, state, and / or region, which can therefore affect the classification of road levels.

[0036] The neural network can be represented by any feasible neural network (e.g., known ones), such as an artificial neural network (ANN). Optionally, the neural network can include and / or be represented as a deep neural network (DNN), such as a convolutional neural network (CNN). Moreover, according to examples, the neural network can include and / or be represented as a CNN of a type of scene classifier known in the art, and / or its equivalents and / or successors. The neural network can be trained and / or has been trained based on a dataset (e.g., a labeled dataset). For example, the input data of the dataset can be (e.g., in a target city where multi-level road systems are common) collected (and / or has been collected) by one or more image acquisition vehicles, and then (for at least the first known vehicle location) labeled (e.g., manually) based on the known corresponding road level, which is derived from knowledge of the road level and / or number of road levels at the vehicle location, for example, from a digital map, such as a digital map of a type corresponding to and / or similar to the aforementioned first digital map 22.

[0037] The phrase "determine road hierarchy" can refer to "identify and / or classify road hierarchy," while "where the vehicle is located" can refer to "where the vehicle is believed to be located," "where the vehicle is positioned," and / or "where the vehicle is believed to be positioned." Furthermore, the phrase "based on feeding the image data through a neural network" can refer to "based on feeding the image data into a neural network," "based on passing the image data through a neural network," "by feeding the image data through a neural network," "at least based on feeding the image data through a neural network," and / or "based on applying a neural network to the image data." Moreover, "neural network" can refer to "artificial neural network" and / or "at least a first neural network." On the other hand, the phrase "trained to classify" can refer to "trained to identify and / or classify" and / or "pre-trained to classify," while "based on the environment of the image content" can refer to "based on the scene, setting, and / or situation in the image content," "based on clues, hints, and / or indications in the image content," and / or "based on clues, hints, and / or indications of the scene, setting, and / or situation in the image content."

[0038] Optionally, the road recognition system 1 may (e.g., via an optional layer determination unit 103) be adapted and / or configured to identify the number of road layers 4 in the multi-level road system 3, supported by at least a first digital map 22, thereby determining the road layers based on the fed image data 5 via a neural network trained to classify road layers based on the environment of the image content. Thus, the road layers obtained from the image data 5 via the neural network can be narrowed down to, matched with, and / or compared to the number of road layers 4 in the multi-level road system 3, and accordingly, the obtained road layers may be adjusted accordingly. Therefore, if the road layers obtained by the neural network potentially do not exist in the multi-level road system 3 (e.g., exemplary non-existent underground road layers and / or road layers higher than those set by the multi-level road system 3), then in order to determine the road layers, the available number of road layers 4 in the multi-level road system 3 can be considered to adjust the road layers obtained by the neural network.

[0039] The road recognition system 1 (e.g., via road / lane recognition unit 106) is adapted and / or configured to identify the road 31 and / or lane 311 of the multi-level road system 3 in which the vehicle 2 is located, based on a determined road hierarchy, with the support of at least a first digital map 22. Thus, by processing the image data 5 through a neural network and establishing potential, feasible, and / or possible road hierarchies associated with the current vehicle position 20 in relation to the digital map 22, the corresponding road hierarchy of the multi-level road system 3 (in...) can be identified. Figure 1(represented by ground road level 41) and then derive the corresponding road 31 and / or lane 311 associated with that road level 41 of the current vehicle position 20. Accordingly, through the introduced concept, the correct road level can be found from the multiple road levels 4 of the multi-level road system 3, and the correct road 31 and / or lane 311 in which the vehicle 2 is located can then be determined.

[0040] The phrase "identify based on the determined road level" can mean "determine and / or derive based on the determined road level," and further means "identify by using, selecting, and / or considering the determined road level." Moreover, the phrase "identify based on the determined road level with the support of the at least first digital map" can mean "identify based on the determined road level while considering the at least first digital map" and / or "identify by selecting the road level of a multi-level road system indicated by the determined road level." On the other hand, the phrase "lane of the multi-level road system" can mean "lane of the road in the multi-level road system," and "where the vehicle is located" can mean "along the path the vehicle is positioned along," "where the vehicle is located," and / or "where the vehicle believes it is located."

[0041] Optionally, the obtained image data 5 can be adapted for use as training data for a neural network. Thus, image data 5, for example, labeled with vehicle location 20, multi-level road system 3, the optional number of levels of multi-level road system 3, the determined number of road levels, and / or the identified road 31 and / or lane 311 of the multi-level road system 3 in which vehicle 2 is located, can be used to train a neural network.

[0042] like Figure 2As further shown, the road recognition system 1 includes a location determination unit 101, a multi-layer system recognition unit 102, an optional layer number determination unit 103, an image data acquisition unit 104, a layer determination unit 105, and a road / lane recognition unit 106, all of which have been described in more detail above. Furthermore, the embodiments described herein for locating a vehicle 2 in a multi-layer road system 3 can be implemented by one or more processors (e.g., processor 107, referred to herein as GPU as an abbreviation for graphics processing unit) and computer program code for performing the functions and actions of the embodiments herein. The program code can also be provided as a computer program product, for example, in the form of a data carrier carrying computer program code for performing the embodiments herein when loaded into the road recognition system 1. One such carrier can be in the form of a CD-ROM and / or a hard disk drive, but other data carriers are also feasible. Furthermore, the computer program code can be provided as plain program code on a server and downloaded to the road recognition system 1. The road recognition system 1 may also include a memory 108, which includes one or more storage units. The memory 108 may be arranged to store, for example, information, and further, data, configurations, schedules, and applications to execute the methods described herein when executed in the road recognition system 1. For example, computer program code may be implemented in firmware, stored in the flash memory 108 of the embedded processor 107, and / or wirelessly downloaded, for example, from an external server. Furthermore, the location determination unit 101, the multi-layer system recognition unit 102, the optional layer number determination unit 103, the image data acquisition unit 104, the layer determination unit 105, the road / lane recognition unit 106, the optional processor 107, and / or the optional memory 108 may be at least partially included in one or more nodes 109 of the vehicle 2, such as an ECU. Those skilled in the art will also understand that the aforementioned units 101, 102, 103, 104, 105, 106, and any other units, interfaces, systems, controllers, modules, devices, elements, features, etc., described herein can refer to, include, contain, and / or be implemented as or a combination of the following: analog and digital circuitry, and / or one or more processors configured with software and / or firmware, the software and / or firmware being stored, for example, in memory (e.g., memory 108), and executed as described herein when executed by one or more processors (e.g., processor 107). One or more of these processors, along with other digital hardware, may be included in a single application-specific integrated circuit (ASIC), or several processors and various digital hardware may be distributed across several separate components (either individually packaged or assembled into a system-on-a-chip (SoC)).

[0043] Figure 2The image further shows the positioning system 21, digital map 22, image acquisition device 23, and image data 5, all of which have been discussed in more detail above.

[0044] Figure 3 This is a flowchart illustrating an exemplary method performed by a road recognition system 1 according to an embodiment of the present invention. The method is used to locate a vehicle 2 in a multi-level road system 3. Exemplary methods that can be repeated continuously include... Figures 1-2 With the support of [the relevant authority], one or more of the following actions are discussed. Moreover, where applicable, actions may be taken in any suitable order and / or one or more actions may be performed simultaneously and / or in an alternating order. For example, action 1004 and / or action 1005 may be performed simultaneously and / or before action 1001, action 1002 and / or action 1003.

[0045] Action 1001

[0046] In action 1001, the road recognition system 1 (e.g., with the support of the location determination unit 101) determines the location 20 of the vehicle 2 with the support of the positioning system 21.

[0047] Action 1002

[0048] In action 1002, the road recognition system 1 (e.g., with the support of the multi-layer system recognition unit 102) identifies the multi-layer road system 3 in which the vehicle 2 is located based on the vehicle location 20, with the support of at least the first digital map 22.

[0049] Action 1003

[0050] In optional action 1003, the road recognition system 1 can (e.g., with the support of optional layer number determination unit 103) identify the number of road layers 4 of the multi-layer road system 3 with the support of at least the first digital map 22.

[0051] Action 1004

[0052] In action 1004, the road recognition system 1 (e.g., with the support of the image data acquisition unit 104) acquires image data 5 with the support of one or more image acquisition devices 23 adapted to acquire the surrounding environment of the vehicle 2.

[0053] Action 1005

[0054] In action 1005, the road recognition system 1 (e.g., with the support of the hierarchy determination unit 105) determines the road hierarchy in which the vehicle 2 is located by means of a neural network based on the feed image data 5. The neural network is trained to classify road hierarchies based on the environment of the image content.

[0055] Alternatively, the neural network may include a deep neural network (DNN).

[0056] Further optionally, the environment of the image content may involve features of one or more road structures, features of one or more static objects, features of one or more other road users, the extent of an obscured sky view and / or side view, and / or the extent of ambient lighting.

[0057] Furthermore, optionally, the action 1005 of determining the road level where vehicle 2 is located may include (and / or the level determination unit 105 may be adapted) feeding image data 5 through a neural network trained to classify road levels based on the environment of the image content and additionally based on geographical location.

[0058] Further optionally, if optional action 1003 precedes action 1005, then action 1005 for determining the road level may include (and / or the level determination unit 105 may be adapted) additionally determining the road level based on the number of road levels 4.

[0059] Furthermore, the obtained image data 5 can optionally be used as training data for a neural network.

[0060] Action 1006

[0061] In action 1006, the road recognition system 1 (e.g., with the support of the road / lane recognition unit 106) identifies the road 31 and / or lane 331 of the multi-level road system 3 in which the vehicle 2 is located based on the determined road hierarchy, with the support of at least the first digital map 22.

[0062] Those skilled in the art will recognize that the present invention is by no means limited to the preferred embodiments described above. Rather, many modifications and variations are possible within the scope of the appended claims. Furthermore, it should be noted that the drawings are not necessarily drawn to scale, and the dimensions of certain features may be enlarged for clarity. Instead, the focus is on illustrating the principles of the embodiments described herein. Additionally, in the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude a plurality.

Claims

1. A method for locating a vehicle (2) in a multi-level road system (3) by a road identification system (1) of the vehicle (2), the method comprising: The location (20) of the vehicle (2) is determined with the support of the positioning system (21); The multi-level road system (3) in which the vehicle (2) is located is identified based on the location (20) of the vehicle, supported by at least a first digital map (22). Image data (5) is obtained with the support of one or more image acquisition devices (23) used to acquire the surrounding environment of the vehicle (2); Based on the image data (5), the road level (6) where the vehicle (2) is located is determined by a neural network, which is trained to classify the road level based on the environment of the image content; and With the support of the at least first digital map (22), the road (31) and / or lane (311) of the multi-level road system (3) in which the vehicle (2) is located is identified based on the determined road level (6), thereby locating the vehicle (2) in the multi-level road system (3).

2. The method according to claim 1, further comprising: The number of road levels (4) of the multi-level road system (3) is identified with the support of at least the first digital map (22); The step of determining the road level (6) is based on the number of road levels (4).

3. The method according to claim 1 or 2, wherein, The environment in question involves: One or more road construction features; Features of one or more static objects; and / or Characteristics of one or more other road users.

4. The method according to claim 1 or 2, wherein, The step of determining the road level (6) where the vehicle (2) is located includes feeding the image data (5) through a neural network trained to classify road levels based on the environment of the image content and based on the geographical location.

5. The method according to claim 1 or 2, wherein, The neural network includes a deep neural network (DNN).

6. The method according to claim 1 or 2, wherein, The obtained image data (5) is used as training data for the neural network.

7. A road identification system (1) for a vehicle (2) for locating the vehicle (2) in a multi-level road system (3), the road identification system (1) comprising: A location determination unit (101) is used to determine the location (20) of the vehicle (2) with the support of a positioning system (21). Multi-layer system identification unit (102) is used to identify the multi-layer road system (3) in which the vehicle (2) is located based on the vehicle's location (20) with the support of at least a first digital map (22). Image data acquisition unit (104) is used to acquire image data (5) with the support of one or more image acquisition devices (23) for acquiring the surrounding environment of the vehicle (2); A hierarchy determination unit (105) is used to determine the road hierarchy (6) where the vehicle (2) is located based on the image data (5) fed in by a neural network, the neural network being trained to classify road hierarchy based on the environment of the image content; and A road / lane recognition unit is used to identify the road (31) and / or lane (311) of the multi-level road system (3) in which the vehicle (2) is located based on the determined road level (6) with the support of the at least first digital map (22), thereby locating the vehicle (2) in the multi-level road system (3).

8. The road recognition system (1) according to claim 7 further includes: Layer number determination unit (103) is used to identify the number of road layers (4) of the multi-level road system (3) with the support of the at least first digital map (22); The hierarchy determination unit (105) determines the road hierarchy (6) based on the number of road hierarchy (4).

9. The road recognition system (1) according to claim 7 or 8, wherein, The environment in question involves: One or more road construction features; Features of one or more static objects; and / or Characteristics of one or more other road users.

10. The road recognition system (1) according to claim 7 or 8, wherein, The hierarchy determination unit (105) is used to feed the image data (5) through a neural network, which is trained to classify road hierarchy based on the environment of the image content and on the geographical location.

11. The road recognition system (1) according to claim 7 or 8, wherein, The neural network includes a deep neural network (DNN).

12. The road recognition system (1) according to claim 7 or 8, wherein, The obtained image data (5) is used as training data for the neural network.

13. A vehicle (2) comprising the road identification system (1) according to claim 7 or 8.

14. A computer program product comprising a computer program containing computer program code means arranged to cause a computer or processor to perform the steps of the method according to any one of claims 1 to 6, the computer program being stored on a computer-readable medium or carrier wave.

15. A non-volatile computer-readable storage medium having a computer program product according to claim 14 stored thereon.

Images