Intelligent device navigation method, intelligent device, and computer-readable storage medium

Abstract

The application relates to the technical field of navigation, and particularly provides an intelligent device navigation method, an intelligent device and a computer readable storage medium, aiming to solve the problem that existing intelligent and reliable navigation cannot be performed in the scene of no prior map, invalid map or lack of high-level semantic information of the map. For this purpose, the intelligent device navigation method comprises the following steps: acquiring an identification result of a navigation mark in a space environment and a road topology map, obtaining a reference road and a target road corresponding to the navigation mark from the road topology map based on the identification result, and navigating the intelligent device by using the reference road and the target road.

Description

Technical Field

[0001] This application relates to the field of navigation technology, specifically providing a navigation method for smart devices, a smart device, and a computer-readable storage medium. Background Technology

[0002] Currently, the operation of intelligent driving and robot navigation systems on structured roads relies heavily on prior environmental information provided by maps. However, relying on maps has significant limitations: on the one hand, the production, updating, and maintenance of maps are costly and difficult to cover all areas (such as underground parking garages, remote parks, indoor venues, and newly built roads); on the other hand, when the environment undergoes dynamic changes (such as road diversions or temporary traffic control), the pre-stored maps quickly become invalid, leading to system "blindness".

[0003] Related technologies also provide map-free navigation solutions. Existing map-free navigation solutions mainly rely on simultaneous localization and mapping (SLAM) technology. However, the maps generated by such methods usually lack high-level semantic information. The system cannot understand natural human guidance signs such as "exit" and "Area A". Therefore, it lacks the ability to make high-level decisions ("which road should I take to my destination") at complex intersections. In scenarios where there is no prior map, the map is invalid, or the map lacks high-level semantic information, it cannot provide intelligent and reliable navigation. Summary of the Invention

[0004] This application aims to solve the aforementioned technical problem, namely, to address the issue that existing navigation methods cannot perform intelligently and reliably in scenarios where there is no prior map, the map is invalid, or the map lacks high-level semantic information.

[0005] In a first aspect, this application provides a navigation method for a smart device, comprising:

[0006] Obtain the recognition results of navigation signs in the spatial environment and the road topology map of the spatial environment;

[0007] Based on the recognition results, the reference road and the target road corresponding to the navigation sign are determined from the road topology map respectively; the navigation sign is used to provide guidance for smart devices on the reference road, and the target road is the road pointed to by the navigation sign;

[0008] Intelligent device navigation is performed based on the reference road and the target road.

[0009] In some embodiments, obtaining the recognition results of navigation signs in the space environment includes:

[0010] Acquire environmental images of the space environment;

[0011] The navigation sign is identified from the environmental image using a large visual language model, and the identification result of the navigation sign is obtained.

[0012] In some embodiments, the identification result includes location information and semantic information. Based on the identification result, the reference road and target road corresponding to the navigation identifier are determined from the road topology map, including:

[0013] Based on the location information of the navigation markers, a reference road is determined from the road topology map;

[0014] Based on the semantic information of the navigation identifier, the target road corresponding to the reference road is determined from the road topology map.

[0015] In some embodiments, the road topology map includes multiple roads and their respective position information in a target coordinate system; determining a reference road from the road topology map based on the position information of the navigation marker includes:

[0016] Based on the location information of the navigation sign, the navigation sign is projected onto the target coordinate system corresponding to the road topology map to obtain the projection location information of the navigation sign;

[0017] Based on the projection position information and the position information of each of the multiple roads, the roads whose overlap or spacing meets the preset conditions are determined from the multiple roads and used as the reference roads.

[0018] In some embodiments, the roads whose interval distances satisfy a preset condition are determined from the plurality of roads, and are included as the reference roads:

[0019] The road with the smallest interval distance among the multiple roads is selected as the reference road.

[0020] In some embodiments, the roads whose interval distances satisfy a preset condition are determined from the plurality of roads, and are included as the reference roads:

[0021] From the plurality of roads, roads with an interval distance less than a preset distance threshold are determined as reference roads, and the number of reference roads is greater than 1.

[0022] In some embodiments, the semantic information includes directional information, and determining the target road corresponding to the reference road from the road topology map based on the semantic information of the navigation identifier includes:

[0023] Based on the directional information of the navigation marker, a road that is connected to the reference road and whose relative orientation and directional information match the reference road are determined from the road topology map, and is taken as the target road.

[0024] In some embodiments, the semantic information includes text information and directional information, and determining the target road corresponding to the reference road from the road topology map based on the semantic information of the navigation identifier includes:

[0025] Based on the text and direction information of the navigation identifier, a road that is connected to the reference road, matches the text information, and matches the relative orientation and direction information of the reference road is determined from the road topology map and is taken as the target road.

[0026] In some embodiments, obtaining the recognition result of navigation identifiers in the spatial environment includes: obtaining the navigation identifier of at least one navigation destination in the spatial environment and obtaining the recognition result of the navigation identifier corresponding to each navigation destination, so as to obtain the reference road and target road corresponding to the navigation identifier of each navigation destination;

[0027] Before performing intelligent device navigation based on the reference road and the target road, the method further includes:

[0028] Obtain navigation instructions and the current location information of the smart device, wherein the navigation instructions include a preset navigation destination;

[0029] Based on the preset navigation destination, a matching target navigation identifier is determined from the navigation identifiers of at least one navigation destination, so as to perform smart device navigation based on the current location information, the reference road corresponding to the target navigation identifier, and the target road.

[0030] In some embodiments, obtaining navigation identifiers for at least one navigation destination in the spatial environment includes: obtaining multiple navigation identifiers for the same navigation destination in the spatial environment to obtain reference roads and target roads corresponding to the multiple navigation identifiers for the same navigation destination.

[0031] In a second aspect, this application provides a smart device comprising:

[0032] At least one processor;

[0033] And, a memory communicatively connected to the at least one processor;

[0034] The memory stores a computer program, which, when executed by the at least one processor, implements the intelligent device navigation method described above.

[0035] In a third aspect, this application provides a computer-readable storage medium storing a plurality of program codes, characterized in that the program codes are adapted to be loaded and run by a processor to perform the intelligent device navigation method described in any of the preceding claims.

[0036] By adopting the above technical solution, this application can obtain the recognition results of navigation signs in the spatial environment and a road topology map, and based on the recognition results, obtain the reference road and target road corresponding to the navigation sign from the road topology map, and use the reference road and target road to navigate the intelligent device. This solution can bind the semantic information of navigation signs with the physical path, realize multi-step spatial reasoning based on navigation signs, improve the ability of intelligent navigation decision-making, eliminate the dependence on pre-stored high-precision maps, and overcome the problem that existing technologies cannot perform intelligent and reliable navigation in scenarios without prior maps, map failure scenarios, or maps lacking high-level semantic information. Attached Figure Description

[0037] The preferred embodiments of this application are described below with reference to the accompanying drawings, in which:

[0038] Figure 1 This is a schematic flowchart of the main steps of the intelligent device navigation method provided in the embodiments of this application;

[0039] Figure 2 This is a flowchart illustrating a smart device navigation method provided in an embodiment of this application;

[0040] Figure 3 This is a flowchart illustrating a smart device navigation method provided in another embodiment of this application;

[0041] Figure 4 This is a schematic diagram of the spatial environment provided in a specific example of this application;

[0042] Figure 5 This is a road topology map of the spatial environment provided in this application as a specific example;

[0043] Figure 6 This is a partial flowchart of the smart device navigation method provided in a specific example of this application. Detailed Implementation

[0044] Some embodiments of this application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of this application and are not intended to limit the scope of protection of this application.

[0045] Figure 1 This is a schematic flowchart of the main steps of the intelligent device navigation method provided in the embodiments of this application, which includes:

[0046] Step S11: Obtain the recognition results of navigation signs in the spatial environment and the road topology map of the spatial environment.

[0047] Step S12: Based on the recognition results, determine the reference road and target road corresponding to the navigation sign from the road topology map.

[0048] Among them, navigation signs are used to provide guidance for smart devices on reference roads, and the target road is the road that the navigation sign points to.

[0049] Step S13: Perform intelligent device navigation based on the reference road and the target road.

[0050] The identification results include at least the location information and semantic information of the navigation sign.

[0051] A road topology map can include the topological edges corresponding to multiple roads in the spatial environment, as well as the position information of the roads corresponding to the topological edges in the target coordinate system.

[0052] In some embodiments, the acquisition of the navigation sign recognition result and the road topology map in step S11 can be performed simultaneously or separately, as detailed below. Figure 2 Description of the embodiments.

[0053] In some embodiments, step S12 may specifically involve determining the road to which the navigation sign belongs from the road topology map based on the location information in the identification result. That is, the navigation sign is used to provide guidance for smart devices on the road, and the road is used as a reference road. For details, please refer to the description below.

[0054] In some embodiments, step S12 may specifically involve determining the target road corresponding to the navigation sign from the road topology map based on the semantic information in the recognition result, as described below.

[0055] In some embodiments, step S13 may specifically be: generating a navigation trajectory or control command based on the reference road and the target road to enable navigation of the smart device in the current spatial environment.

[0056] The above describes a smart device navigation method provided in this application. It obtains the recognition results of navigation markers in the spatial environment and a road topology map. Based on the recognition results, it obtains the reference road and target road corresponding to the navigation marker from the road topology map, and then uses the reference road and target road to navigate the smart device. This solution can bind the semantic information of navigation markers with the physical path, enabling multi-step spatial reasoning based on navigation markers, improving the ability of intelligent navigation decision-making, eliminating the dependence on pre-stored high-precision maps, and overcoming the problem that existing technologies cannot perform intelligent and reliable navigation in scenarios without prior maps, map failure scenarios, or scenarios where maps lack high-level semantic information.

[0057] Figure 2 This is a flowchart illustrating a smart device navigation method provided in an embodiment of this application, which includes:

[0058] Step S21: Obtain the recognition results of navigation signs in the spatial environment. The recognition results include location information and semantic information.

[0059] Step S22: Obtain the road topology map of the spatial environment.

[0060] Step S23: Based on the location information of the navigation markers, determine the reference road from the road topology map.

[0061] Step S24: Based on the semantic information of the navigation identifier, determine the target road corresponding to the reference road from the road topology map.

[0062] Step S25: Perform intelligent device navigation based on the reference road and the target road.

[0063] In some embodiments, obtaining the recognition result of navigation marks in the space environment in step S21 includes:

[0064] Acquire environmental images of the spatial environment; use a large visual language model to identify navigation signs from the environmental images and obtain the recognition results of the navigation signs.

[0065] The recognition results include the location information and semantic information of the navigation sign. The location information can be obtained by using a large visual language model to recognize the environment image and output the detection box corresponding to the navigation sign, based on the position of the detection box in the environment image.

[0066] In some embodiments, navigation identifiers may include pointing icons, such as a straight arrow icon or a left turn arrow icon, and correspondingly, semantic information may include pointing information such as the direction of the arrow corresponding to the pointing icon; in other embodiments, navigation identifiers may also include text, such as the text "exit" used to identify the type of destination, or the text "500 meters" used to identify the distance, and correspondingly, semantic information may also include the text information corresponding to the text.

[0067] It should be noted that there may be one or more navigation identifiers. Multiple navigation identifiers can be identifiers targeting the same navigation destination but located in different positions within the spatial environment. In other embodiments, multiple navigation identifiers can also be identifiers targeting different navigation destinations; see [link to relevant documentation] for details. Figure 3 Description of the corresponding embodiments.

[0068] In some embodiments, step S22 may specifically involve directly receiving a road topology map of the spatial environment provided externally.

[0069] Specifically, a road topology map of the spatial environment can be acquired using a perception system that utilizes point clouds, vision, or other sensors. This road topology map displays multiple feasible roads in the spatial environment, the connectivity between these roads, and the location information of each road within a target coordinate system. For example, the target coordinate system can be a three-dimensional coordinate system in the physical world, and the road topology map can include the location information of each road within that three-dimensional coordinate system.

[0070] In some embodiments, step S23 may specifically be:

[0071] Based on the location information of the navigation sign, the navigation sign is projected onto the target coordinate system corresponding to the road topology map to obtain the projected location information of the navigation sign; based on the projected location information and the location information of multiple roads, roads that overlap or have a spacing distance that meets the preset conditions are determined from the multiple roads and used as reference roads.

[0072] The location information of navigation signs can be represented by the two-dimensional coordinates of the detection box. When the target coordinate system is the three-dimensional coordinate system of the physical world, the navigation signs can be projected onto the target coordinate system corresponding to the road topology map. This can be done by using known intrinsic parameters of the image acquisition device, such as focal length, and extrinsic parameters, such as the relative position of the image acquisition device to the smart device and the pose of the smart device, to transform the location information of the navigation signs into the three-dimensional coordinate system of the physical world, so as to determine the installation position of the navigation signs in the physical world.

[0073] In some embodiments, projecting navigation signs onto the target coordinate system corresponding to the road topology map based on the location information of the navigation signs can be achieved by transforming the local location information of the navigation signs, such as the location information of a certain edge of a detection box or the bottom center point of the detection box, into the target coordinate system corresponding to the road topology map. This effectively reduces the amount of computation and is beneficial for improving subsequent decision-making efficiency. Accordingly, the location information obtained after transforming the local location information into a coordinate system can be used to characterize the projected location information of the navigation signs.

[0074] In some embodiments, when a navigation sign is used to provide guidance to a smart device on a road, the navigation sign corresponds to a reference road, and the installation location of the navigation sign overlaps with or is minimally spaced from the reference road.

[0075] From multiple roads, roads whose interval distances meet preset conditions are identified as reference roads, including:

[0076] The road with the smallest interval among multiple roads is selected as the reference road.

[0077] In other embodiments, navigation markers can be used to provide guidance to smart devices on multiple roads simultaneously, with each navigation marker corresponding to multiple reference roads. Roads whose intervals meet preset conditions are determined from the multiple roads and used as reference roads include:

[0078] From multiple roads, roads with an interval distance less than a preset distance threshold are identified as reference roads, and the number of reference roads is greater than 1.

[0079] The preset distance threshold can be flexibly set based on the actual road layout, such as the distance between the installation location of the navigation sign and the road with the furthest interval indicated by it.

[0080] In some embodiments, semantic information includes pointing information, and step S24 may specifically be as follows:

[0081] Based on the directional information of the navigation signs, the road that is connected to the reference road and whose relative orientation and directional information match the reference road is determined from the road topology map and is taken as the target road.

[0082] If the navigation sign is a right-turn arrow, the direction information can be "turn right". Starting from the reference road corresponding to the right-turn arrow, the road that is connected to the reference road and located to the right of the reference road and extends from the road topology map is determined as the target road. The target road is the road that the navigation points to based on the navigation sign in the road topology map.

[0083] In some embodiments, determining the relative orientation with respect to a reference road can be achieved by: determining the relative orientation with respect to the reference road based on the angle differences between each road connected to the reference road and the current reference road. Furthermore, the angle differences with the current reference road can be used to determine whether it matches the directional information. For example, when the aforementioned instruction information is "turn right," a road connected to the reference road with an angle of approximately 90° to the extension direction of the reference road can be determined from the road topology map as the target road; when the instruction information is "go straight," a road connected to the reference road with an angle of approximately 0° to the extension direction of the reference road can be determined from the road topology map as the target road.

[0084] In other embodiments, semantic information may include text information and directional information, and step S24 may specifically be as follows:

[0085] Based on the text and direction information of the navigation signs, the road that is connected to the reference road, matches the text information, and matches the relative orientation and direction information of the reference road is determined from the road topology map and is taken as the target road.

[0086] As an example, the text information can be text used to identify distance, such as "500 meters". Based on the example above where the directional information is "turn right", the road that is connected to the current reference road, located 500 meters ahead and to the right of the reference road can be determined from the road topology map as the target road.

[0087] This solution can effectively identify the semantic information of navigation signs, achieve understanding of high-level semantic information, and accurately and effectively bind semantic information with physical roads, which is beneficial for the subsequent realization of intelligent navigation.

[0088] In some embodiments, step S25 may specifically involve generating a navigation trajectory or control command based on the reference road and the target road to enable navigation of the smart device in the current spatial environment.

[0089] In another embodiment, a target navigation identifier can be determined based on the user's navigation needs, and then navigation can be performed based on the reference road and target road corresponding to the target navigation identifier. For details, please refer to the description of the following embodiments.

[0090] The above is an embodiment of a smart device navigation method provided in this application, which can achieve [unclear meaning - possibly related to navigation]. Figure 1 It achieves the same beneficial effects as the corresponding embodiments, and can effectively identify the semantic information of navigation signs and realize the effective binding of navigation signs with reference roads and target roads.

[0091] See Figure 3 , Figure 3 This is a flowchart illustrating a smart device navigation method according to another embodiment of this application, which may include:

[0092] Step S31: Obtain the navigation identifier of at least one navigation destination in the spatial environment and obtain the recognition result of the corresponding navigation identifier for each navigation destination. The recognition result includes location information and semantic information.

[0093] Step S32: Obtain the road topology map of the spatial environment.

[0094] Step S33: For each navigation destination, determine the reference road from the road topology map based on the location information of the navigation sign; the navigation sign is used to provide guidance for smart devices on the reference road.

[0095] Step S34: For each navigation destination, determine the target road corresponding to the reference road from the road topology map based on the semantic information of the navigation identifier.

[0096] Step S35: Obtain navigation instructions and the current location information of the smart device. The navigation instructions include a preset navigation destination.

[0097] Step S36: Based on the preset navigation destination, determine the matching target navigation identifier from the navigation identifiers of at least one navigation destination.

[0098] Step S37: Perform intelligent device navigation based on the current location information, the reference road corresponding to the target navigation sign, and the target road.

[0099] Step S32 can be implemented in the same way as step S22 above. Steps S33 and S34 can be adapted to be executed for each navigation destination, and implemented in the same way as steps S23 and S24.

[0100] In some embodiments, obtaining the navigation identifier of at least one navigation destination in the spatial environment in step S31 can be obtaining the navigation identifiers of one or more navigation destinations in the spatial environment and obtaining the identification result of the navigation identifier corresponding to each navigation destination.

[0101] As an example, navigation destinations may include exits, shopping malls, and parking lots. Navigation signs may include text indicating the navigation destination; correspondingly, semantic information may include text information related to the navigation destination.

[0102] In some embodiments, obtaining the navigation identifier of at least one navigation destination in the spatial environment in step S31 may include: obtaining multiple navigation identifiers of the same navigation destination in the spatial environment to obtain the reference road and the target road corresponding to the multiple navigation identifiers of the same navigation destination respectively.

[0103] Among these, navigation markers for the same navigation destination can be located in different positions within the current spatial environment. Accordingly, in step S33, for the same navigation destination, reference roads can be determined from the road topology map based on the location information of each navigation marker. In step S34, for the same navigation destination, the target roads corresponding to the reference roads can be determined from the road topology map based on the semantic information of each navigation marker, thereby obtaining the reference roads and target roads corresponding to multiple navigation markers for the same navigation destination.

[0104] In some embodiments, obtaining the navigation instruction in step S35 can be obtaining the voice instruction "navigate to the exit", where "exit" is the preset navigation destination, which can be flexibly set according to the user's needs.

[0105] In some embodiments, step S36 may specifically involve comparing the semantic information of a preset navigation destination and a navigation identifier, wherein the semantic information includes text information used to indicate the navigation destination, and when the text information matches the preset navigation destination, the current navigation identifier is used as the target navigation identifier.

[0106] In some embodiments, multiple target navigation identifiers can be identified, and these multiple target navigation identifiers can correspond to different location information, different reference roads, and target roads.

[0107] In some embodiments, step S37 may specifically involve generating a navigation trajectory or control command based on the reference road and target road corresponding to the target navigation identifier and the current location information, and performing smart device navigation.

[0108] It can generate navigation trajectories or control commands from the current location, through reference roads, to the target road.

[0109] When multiple target navigation markers are identified, the target navigation marker that is closest to the current location can be determined based on the current location information and the location information of the multiple target navigation markers in the driving direction toward the preset navigation destination. The reference road corresponding to the closest target navigation marker is then used as the starting road. The reference road corresponding to the current target navigation marker is used as the starting road, and a continuous navigation trajectory is generated together with the reference roads corresponding to other target navigation markers and the target road. For details, please refer to the description of the specific embodiments below.

[0110] The above is another embodiment of the intelligent device navigation method provided in this application, which can achieve [the following]: Figure 2 The same beneficial effects as the corresponding embodiments can be achieved by first obtaining the reference road and target road corresponding to the navigation identifier of at least one navigation destination, and then matching the target navigation identifier of the preset navigation destination according to the navigation instruction, thereby facilitating the quick acquisition of the reference road and target road corresponding to the target navigation identifier and improving the efficiency of navigation decision-making.

[0111] This application provides a specific example of a navigation method for intelligent devices, applied to vehicles. See also... Figures 4 to 6 As shown, Figure 4 This is a schematic diagram of the spatial environment provided in a specific example of this application, with the vehicle located at parking point P1. Figure 5 This is a road topology diagram of the spatial environment provided in a specific example of this application. Each topology edge can correspond to a road, and each node can correspond to an intersection. It shows parking point P1, road E1, road E2, passage E3 and road N1, as well as intersection J1 between road E1 and road E2, and intersection J2 where road E2, road E3 and road N1 intersect.

[0112] Figure 6 This is a flowchart of some steps in the smart device navigation method provided in the specific example of the application.

[0113] Step 1: Environmental perception and data input.

[0114] Environmental perception can include identifying navigational markers and road topology maps of the spatial environment.

[0115] The system uses images of the current spatial environment captured by the vehicle's cameras to identify the visual language model (VLM). Figure 4 The two navigation icons in the text.

[0116] Among them, can Figure 4 The two navigation icons are denoted as A and B. Icon A's text content is "Exit," pointing to an icon labeled "Straight Arrow." The corresponding semantic information in the recognition result is: {Text information: "Exit," Direction information: "Straight"}. Icon B's text content is "Exit," pointing to an icon labeled "Right Turn Arrow." The corresponding semantic information in the recognition result is: {Text information: "Exit," Direction information: "Right Turn"}. Simultaneously, VLM can also obtain two-dimensional (2D) bounding boxes for the navigation icons, and the recognition result also includes the location information of the navigation icons obtained from the 2D bounding boxes.

[0117] Among them, the road topology map G, which can receive the spatial environment provided by the outside, is as follows: Figure 5 As shown.

[0118] The data input may include externally provided road topology map G and VLM recognition results.

[0119] Step 2: Core associative reasoning, which includes two stages.

[0120] Phase 1: Binding navigation identifiers to reference roads E_ref.

[0121] The goal of Phase One is to determine which road in the road topology map each navigation sign is physically attached to, i.e., which road the navigation sign is used to guide smart devices, and to identify the corresponding road as the reference road.

[0122] The inputs for this phase one are: the 2D detection box of the navigation sign, the vehicle's current precise pose (from the vehicle's positioning module), and the externally provided road topology map G (containing the 3D spatial coordinates of the roads corresponding to each topological edge in the physical world 3D coordinate system).

[0123] This stage of the process includes:

[0124] Spatial Projection and Matching: For each navigation sign's detection frame, the installation position of the sign in the physical world is calculated using known camera parameters and vehicle pose. For example, when a navigation sign is installed on a wall or the lower part of a column, the 3D projection position information of the navigation sign can be obtained by projecting the bottom center point of the detection frame into 3D space, and the topological edge that overlaps with or is closest to the navigation sign's projection can be found from the road topology map.

[0125] Binding decision: based on Figure 4 and Figure 5 As shown, the projected position of marker A overlaps with or is closest to the topological edge E1 in space. Therefore, marker A is bound to E1, and E1 is used as its reference road, meaning that marker A provides information for vehicles traveling on E1.

[0126] according to Figure 4 and Figure 5 As shown, the projection position of the identifier B is also most strongly associated with the spatial relationship of E2. Therefore, the identifier B is bound to the topological edge E2, and E2 is used as its reference road.

[0127] This stage can output: a binding list [(Identifier A, Reference Road: E1), (Identifier B, Reference Road: E2)].

[0128] Phase Two: Derivation of the Target Road E_target Based on Navigation Identifiers

[0129] The goal of Phase Two is to deduce the target road that the vehicle should eventually enter from the road topology map, based on the semantic information (textual information and directional information) of the navigation markers, starting from the reference road.

[0130] The inputs for Phase 2 are: the binding list output from Phase 1, the semantic information of the navigation identifiers, and the complete road topology map G.

[0131] For identifier A, the processing steps in this stage include:

[0132] Instruction Analysis: The semantic information of identifier A is "exit" + "go straight", and the corresponding logic is: "Proceed along the current reference road direction to reach the exit."

[0133] Topological reasoning: Starting from the reference road E1, the vehicle is traveling from P1 to J1. At node J1 (the end of E1), we find all connected topological edges originating from J1 in the road topology graph G to find the road connected to the reference road.

[0134] It then determines which topological edge best matches the "straight ahead" directional information. This can be done by calculating the angle difference between each connected topological edge and the direction the vehicle enters J1. If the angle between the direction of E2 and the extension direction of E1 is the smallest, and the angle difference is less than 15 degrees, then E2 is determined to match the "straight ahead" direction.

[0135] Based on the semantic information of identifier A, starting from E1 in the road topology graph G, the connected topology edge E2 that best fits the semantics of "straight ahead" is determined, and the target road is finally derived.

[0136] For identifier B, the processing steps in this stage include:

[0137] Instruction analysis: The semantic information of identifier B is "exit" + "turn right", and the corresponding logic is: "Turn right at the end of the reference road to reach the exit."

[0138] Topological reasoning: The starting point is E2 and the ending point is J2. At node J2, find a connected topological edge that starts from J2 and whose direction makes an angle of approximately 90° with the entering direction.

[0139] Derivation complete: The target road identified by B is E3.

[0140] Output: Updated mapping table [(Identifier A, Reference Road: E1, Target Road: E2), (Identifier B, Reference Road: E2, Target Road: E3)].

[0141] Step 3: Navigation Decision and Execution.

[0142] The vehicle is located at parking point P1. The system receives the user's navigation command, such as "navigate to the exit," matches the identifier in the updated mapping table, and provides the final planned route. For example, it extracts reference road E1 for identifier A and target road E2; it extracts reference road E2 for identifier B and target road E3.

[0143] Based on the reference road and target road corresponding to each of the signs A and B, a navigation trajectory is generated, that is, starting from P1, passing through E1→E2→E3 to reach the exit, and the vehicle is controlled to execute the operation based on this navigation trajectory.

[0144] Those skilled in the art will understand that all or part of the processes in the method of the above-described embodiment can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0145] Another aspect of this application provides a computer-readable storage medium.

[0146] In one embodiment of a computer-readable storage medium according to this application, the computer-readable storage medium can be configured to store a program that performs the intelligent device navigation method of the above-described method embodiments. This program can be loaded and run by a processor to implement the intelligent device navigation method. For ease of explanation, only the parts related to the embodiments of this application are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of this application. The computer-readable storage medium can be a storage device comprising various electronic devices. Optionally, in the embodiments of this application, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0147] Another aspect of this application provides a smart device.

[0148] In one embodiment of a smart device according to this application, the smart device may include at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program, which, when executed by the at least one processor, implements the method described in any of the above embodiments. The smart device described in this application may include driving equipment, smart vehicles, robots, and other devices.

[0149] In some embodiments of this application, the smart device may further include at least one sensor for sensing information. The sensor is communicatively connected to any type of processor mentioned in this application. Optionally, the smart device may further include an intelligent driving system for guiding the smart device to drive autonomously or providing driving assistance. The processor communicates with the sensor and / or the intelligent driving system to perform the methods described in any of the above embodiments.

[0150] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. Without departing from the principles of this application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of this application.

Claims

1. A navigation method for intelligent devices, characterized in that, include: Obtain the recognition results of navigation signs in the spatial environment and the road topology map of the spatial environment; Based on the recognition results, the reference road and the target road corresponding to the navigation sign are determined from the road topology map respectively; the navigation sign is used to provide guidance for smart devices on the reference road, and the target road is the road pointed to by the navigation sign; Intelligent device navigation is performed based on the reference road and the target road; The identification result includes location information. Determining the reference road corresponding to the navigation sign from the road topology map based on the identification result includes: determining the reference road from the road topology map based on the location information of the navigation sign; the road topology map includes multiple roads and their respective location information in the target coordinate system; determining the reference road from the road topology map based on the location information of the navigation sign includes: Based on the location information of the navigation sign, the navigation sign is projected onto the target coordinate system corresponding to the road topology map to obtain the projection location information of the navigation sign; Based on the projection position information and the position information of each of the multiple roads, the roads whose overlap or spacing meets the preset conditions are determined from the multiple roads and used as the reference roads.

2. The method according to claim 1, characterized in that, Obtaining the recognition results of navigation signs in the spatial environment includes: Acquire environmental images of the space environment; The navigation sign is identified from the environmental image using a large visual language model, and the identification result of the navigation sign is obtained.

3. The method according to claim 1, characterized in that, The recognition result also includes semantic information. Based on the recognition result, the target road corresponding to the navigation sign is determined from the road topology map, including: Based on the semantic information of the navigation identifier, the target road corresponding to the reference road is determined from the road topology map.

4. The method according to claim 1, characterized in that, The roads whose intervals meet preset conditions are determined from the plurality of roads, and these roads are used as reference roads, including: The road with the smallest interval distance among the multiple roads is selected as the reference road.

5. The method according to claim 1, characterized in that, The roads whose intervals meet preset conditions are determined from the plurality of roads, and these roads are used as reference roads, including: From the plurality of roads, roads with an interval distance less than a preset distance threshold are determined as reference roads, and the number of reference roads is greater than 1.

6. The method according to claim 3, characterized in that, The semantic information includes directional information, and determining the target road corresponding to the reference road from the road topology map based on the semantic information of the navigation identifier includes: Based on the directional information of the navigation marker, a road that is connected to the reference road and whose relative orientation and directional information match the reference road are determined from the road topology map, and is taken as the target road.

7. The method according to claim 3, characterized in that, The semantic information includes text information and directional information. Determining the target road corresponding to the reference road from the road topology map based on the semantic information of the navigation identifier includes: Based on the text information and directional information of the navigation identifier, a road that is connected to the reference road, matches the text information, and matches the relative orientation and directional information of the reference road is determined from the road topology map and is taken as the target road.

8. The method according to any one of claims 1 to 5, characterized in that, The identification results of spatial environment navigation marks include: obtaining the navigation marks of at least one navigation destination in the spatial environment and obtaining the identification results of the navigation marks corresponding to each navigation destination, so as to obtain the reference road and target road of the navigation mark corresponding to each navigation destination; Before performing intelligent device navigation based on the reference road and the target road, the method further includes: Obtain navigation instructions and the current location information of the smart device, wherein the navigation instructions include a preset navigation destination; Based on the preset navigation destination, a matching target navigation identifier is determined from the navigation identifiers of at least one navigation destination, so as to perform smart device navigation based on the current location information, the reference road corresponding to the target navigation identifier, and the target road.

9. The method according to claim 8, characterized in that, Obtaining navigation identifiers for at least one navigation destination in the spatial environment includes: obtaining multiple navigation identifiers for the same navigation destination in the spatial environment to obtain reference roads and target roads corresponding to the multiple navigation identifiers for the same navigation destination.

10. A smart device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program, which, when executed by the at least one processor, implements the intelligent device navigation method according to any one of claims 1 to 9.

11. A computer-readable storage medium storing a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by a processor to perform the smart device navigation method according to any one of claims 1 to 9.

Images