Intelligent driving method and device for cross-layer scene and storage medium

Abstract

The invention discloses a cross-layer scene intelligent driving method and device and a storage medium, and relates to the field of intelligent driving, and the method comprises the steps: carrying out the semantic recognition according to a real-time machine vision image, and determining the semantic anchor point information; navigation is carried out according to the semantic anchor point information and the destination information until the vehicle arrives at a cross-layer ramp; when a cross-layer ramp is reached, ramp information is determined according to a real-time machine vision image; and motion compensation is carried out according to the ramp information and the pose information of the vehicle, and automatic driving is carried out according to a compensation result. According to the method, a navigation path obtained based on destination information is optimized according to semantic anchor point information, so that a vehicle can automatically drive to a cross-layer ramp according to the identified semantic anchor point information under the condition of not depending on positioning information. And when the vehicle arrives at the cross-layer ramp, the ramp information obtained through machine vision and the pose information of the vehicle are used for controlling the vehicle, and reliable cross-layer automatic driving is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent driving, and in particular to an intelligent driving method and device for a cross-layer scene and a storage medium. BACKGROUND

[0002] With the continuous evolution of automobile intelligence, automatic driving is widely used.

[0003] However, for scenarios in which navigation software cannot provide accurate maps, such as cross-layer automatic driving in a multi-layer garage, the vehicle's position cannot be accurately located, resulting in the inability to complete automatic driving. How to achieve reliable automatic driving in a cross-layer automatic driving scenario has become a problem to be solved. SUMMARY

[0004] The present application provides an intelligent driving method and device for a cross-layer scene and a storage medium to solve the problem of poor reliability of current cross-layer automatic driving.

[0005] According to an aspect of the present application, an intelligent driving method for a cross-layer scene is provided, comprising:

[0006] According to the real-time machine vision image, semantic recognition is performed to determine semantic anchor point information;

[0007] According to the semantic anchor point information and destination information, navigation is performed until the vehicle reaches a cross-layer ramp;

[0008] When the cross-layer ramp is reached, the ramp information is determined according to the real-time machine vision image; motion compensation is performed according to the ramp information and the vehicle's pose information, and automatic driving is performed according to the compensation result.

[0009] According to another aspect of the present application, an intelligent driving device for a cross-layer scene is provided, comprising:

[0010] An anchor point recognition module is configured to perform semantic recognition according to a real-time machine vision image to determine semantic anchor point information;

[0011] An anchor point navigation module is configured to perform navigation according to the semantic anchor point information and destination information until the vehicle reaches a cross-layer ramp;

[0012] A ramp driving module is configured to, when the cross-layer ramp is reached, determine the ramp information according to the real-time machine vision image; perform motion compensation according to the ramp information and the vehicle's pose information, and perform automatic driving according to the compensation result.

[0013] According to another aspect of the present application, an electronic device is provided, comprising:

[0014] at least one processor; and

[0015] a memory in communication with the at least one processor; wherein

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the intelligent driving method for cross-layer scenarios according to any one of the embodiments of the application.

[0017] According to another aspect of the application, a computer readable storage medium is provided, which stores computer instructions for enabling a processor to implement the intelligent driving method for cross-layer scenarios according to any one of the embodiments of the application when executed by the processor.

[0018] The technical solution of the embodiment of the application performs semantic recognition on real-time machine vision images to determine semantic anchor point information, performs navigation according to the semantic anchor point information and destination information until the vehicle reaches the cross-layer ramp, determines ramp information according to real-time machine vision images when the cross-layer ramp is reached, performs motion compensation according to the ramp information and the pose information of the vehicle, and performs automatic driving according to the compensation result. Compared with the poor reliability of current cross-layer automatic driving, the technical solution of the embodiment of the application can perform semantic recognition on road signs in real-time machine vision images through machine vision technology to obtain semantic anchor point information for navigation. The navigation path obtained based on the destination information is optimized according to the semantic anchor point information, so that the vehicle can perform automatic driving according to the recognized semantic anchor point information without relying on positioning information, and drive to the cross-layer ramp. When the cross-layer ramp is reached, the vehicle is controlled using the ramp information obtained through machine vision and the pose information of the vehicle, to achieve reliable cross-layer automatic driving.

[0019] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the application, nor is it used to limit the scope of the application. Other features of the application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort.

[0021] Figure 1 The flowchart of the intelligent driving method for cross-layer scenarios provided by the embodiment of the application is shown in Fig. 1;

[0022] Figure 2A flowchart of the intelligent driving method of the cross-layer scene provided by the embodiment of the present application is shown in Figure 2;

[0023] Figure 3 A flowchart of the intelligent driving method of the cross-layer scene provided by the embodiment of the present application is shown in Figure 3;

[0024] Figure 4 A structure diagram of the intelligent driving device of the cross-layer scene provided by the embodiment of the present application is shown in Figure 4. Figure 1 ;

[0025] Figure 5 A structure diagram of the intelligent driving device of the cross-layer scene provided by the embodiment of the present application is shown in Figure 5. Figure 2 ;

[0026] Figure 6 A structure diagram of the electronic device for implementing the intelligent driving method of the cross-layer scene of the embodiment of the present application is shown in Figure 6. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0028] It should be noted that the terms "include" and "have" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device containing a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0029] The inventor found that with the continuous evolution of automobile intelligence, automatic driving is widely used. However, for the scene where the navigation software cannot provide accurate maps, such as the scene of cross-layer automatic driving in a multi-layer garage, due to the inability to accurately locate the position of the vehicle, the automatic driving cannot be completed.

[0030] The core limitation of current parking lot intelligent driving technology is the dual challenge of environmental adaptability and system robustness. Existing solutions generally have positioning continuity defects in poor satellite signal environments: the inertial navigation system in the cross-layer scene produces uncontrollable drift, and visual positioning is easily disturbed by temporary occlusion or sudden changes in light; the dynamic obstacle response mechanism has significant perception lag for lateral escape targets and low-speed moving bodies, and conservative braking is frequently triggered in narrow lane scenarios; and it is highly dependent on pre-acquired high-precision maps or fixed infrastructure, resulting in high deployment costs and difficulty in adapting to dynamic environmental changes such as temporary construction and parking space occupation, resulting in insufficient overall system robustness. How to achieve reliable autonomous driving in cross-layer autonomous driving scenarios is a problem that needs to be solved.

[0031] Figure 1 is a flowchart of an intelligent driving method for a cross-layer scene provided by an embodiment of the present application. The embodiment can be applied to the case where a vehicle performs autonomous driving cross-layer navigation in an underground parking lot or an above-ground parking building. The method can be executed by a cross-layer scene intelligent driving device, which can be realized in the form of hardware and / or software. The cross-layer scene intelligent driving device can be configured in an automobile controller or other electronic device. As shown in Figure 1 , it includes:

[0032] Step S101: performing semantic recognition according to real-time machine vision images to determine semantic anchor point information.

[0033] The arrow and logo in the machine vision image can be recognized. For example, the driving direction arrow on the ground, the exit word on the top of the floor, and the graphic logo related to the exit word in the machine vision image are recognized, and when the information related to the anchor points such as the exit, the entrance, the way to the upper layer, and the way to the lower layer is found, the semantic anchor point information is recognized. The semantic anchor point information can be the road or intersection leading to the exit of the current layer.

[0034] Step S102: navigating according to the semantic anchor point information and the destination information until the vehicle reaches the cross-layer ramp.

[0035] Optionally, navigating according to the semantic anchor point information and the destination information can be implemented as:

[0036] determining an initial navigation route according to the semantic map and the destination information; optimizing the initial navigation route according to the semantic anchor point information to obtain a target navigation route, and navigating according to the target navigation route.

[0037] The semantic map contains the positions corresponding to different semantic anchor point information in each floor. The semantic anchor point information obtained in step S101 is matched with the semantic map to determine the current position of the vehicle.

[0038] The above embodiment can determine an initial navigation route based on destination information, and then optimize the initial navigation route in combination with semantic anchor point information and a semantic map to obtain a target navigation route.

[0039] Further, after optimizing the initial navigation route according to the semantic anchor point information, the method further comprises:

[0040] Supplementing the semantic map according to the recognized semantic anchor point information.

[0041] The above embodiment can supplement the semantic map according to the real-time recognized semantic anchor point information, and realize optimization of the semantic map in the process of autonomous driving. With an increase in the number of navigation, the content of the semantic map of the parking lot of the layer is continuously enriched, and the accuracy of the target navigation path is improved.

[0042] In step S103, when reaching the cross-layer ramp, ramp information is determined according to real-time machine vision images; motion compensation is performed according to the ramp information and vehicle pose information, and autonomous driving is performed according to the compensation result.

[0043] Optionally, the motion compensation according to the ramp information and the vehicle pose information can be implemented as:

[0044] Obtaining pose information determined by an inertial navigation system; optimizing the ramp information according to inertial information obtained by the inertial navigation system; and performing motion compensation according to the optimized ramp information and the pose information.

[0045] The inertial navigation system can provide current pose information of the vehicle, such as an angle of the vehicle with the horizontal direction, a vehicle travel resistance, a vehicle travel direction, etc.

[0046] The above embodiment can quickly determine the pose information in combination with the inertial navigation system, and improve the timeliness and accuracy of dynamic compensation.

[0047] The embodiment of the application fuses visual semantic anchor point recognition and inertial dynamic compensation to construct a double-redundancy positioning architecture. The vehicle captures positioning features such as floor identification and ramp texture in real time by using a surround-view camera, matches the positioning features with a lightweight semantic map, and then determines the floor and position of the vehicle according to the semantic map. In addition, an inertial navigation-odometer fusion system is activated in the ramp area to automatically correct a motion model according to the slope angle. The scheme can solve the drift problem of traditional positioning methods in a cross-layer scene, realize global continuous and stable positioning in an environment with poor satellite signals, and significantly improve the pose perception ability of the vehicle in a complex parking lot environment.

[0048] The intelligent driving method for cross-layer scenes provided by the embodiment of the present application can perform semantic recognition on real-time machine vision images to determine semantic anchor point information, perform navigation according to the semantic anchor point information and destination information until the vehicle reaches a cross-layer ramp, determine ramp information according to real-time machine vision images when the cross-layer ramp is reached, perform motion compensation according to the ramp information and the pose information of the vehicle, and perform automatic driving according to the compensation result. Compared with the poor reliability of current cross-layer automatic driving, the intelligent driving method for cross-layer scenes provided by the embodiment of the present application can perform semantic recognition on road signs in real-time machine vision images through machine vision technology to obtain semantic anchor point information for navigation. The navigation path obtained based on the destination information is optimized according to the semantic anchor point information, so that the vehicle can perform automatic driving according to the recognized semantic anchor point information without relying on positioning information, and drive to the cross-layer ramp. When the cross-layer ramp is reached, the vehicle is controlled using the ramp information obtained through machine vision and the pose information of the vehicle, to achieve reliable cross-layer automatic driving.

[0049] Figure 2 is the flow of the intelligent driving method for cross-layer scenes provided by the embodiment of the present application Figure 2 As a further description of the above embodiment, the intelligent driving method for cross-layer scenes provided by the embodiment of the present application can recognize obstacles around the vehicle and adjust automatic driving according to the threat type of the recognized obstacles. Specifically, after the automatic driving according to the compensation result, the method further includes: performing obstacle recognition according to real-time machine vision images to determine the threat type of the obstacles; and adjusting the navigation path according to the threat type of the obstacles and a preset risk avoidance strategy. Correspondingly, as shown in Figure 3 The intelligent driving method for cross-layer scenes can be implemented through the following steps:

[0050] Step S201: performing semantic recognition according to real-time machine vision images to determine semantic anchor point information.

[0051] Step S202: performing navigation according to the semantic anchor point information and destination information until the vehicle reaches a cross-layer ramp.

[0052] Step S203: when the cross-layer ramp is reached, determining ramp information according to real-time machine vision images; performing motion compensation according to the ramp information and the pose information of the vehicle, and performing automatic driving according to the compensation result.

[0053] Step S204: performing obstacle recognition according to real-time machine vision images to determine the threat type of the obstacles.

[0054] Optionally, the obstacle recognition according to real-time machine vision images to determine the threat type of the obstacles can be implemented as:

[0055] According to the real-time machine vision image, an obstacle in a radial direction of a vehicle travel direction is identified, and a type of the obstacle is determined; and a threat type of the obstacle is determined according to the type of the obstacle.

[0056] The above embodiment can identify the obstacle in the radial direction of the vehicle travel direction, and since only the obstacle in the radial direction has a high collision risk, by identifying the obstacle moving in the radial direction, the calculation amount of identification can be reduced, the obstacle can be quickly identified, and the obstacle identification efficiency is improved.

[0057] In step S205, the navigation path is adjusted according to the threat type of the obstacle and a preset risk avoidance strategy.

[0058] Optionally, the navigation path is adjusted according to the threat type of the obstacle and the preset risk avoidance strategy, which can be implemented as:

[0059] If the threat type of the obstacle is a first threat type, the vehicle is stopped. If the threat type of the obstacle is a second threat type, an alternative lane is identified, and a target lane is selected according to a vehicle condition of the alternative lane and a turning radius of the vehicle.

[0060] The danger degree of the first threat type is greater than that of the second threat type.

[0061] The above embodiment can perform different automatic driving processing on obstacles of different threat types, and further improve safety.

[0062] In view of the dynamic obstacle perception lag problem, a threat evaluation model based on a radial velocity component is established. The relative velocity component of the approaching direction of the obstacle is calculated in real time, and a multi-level response decision is made in combination with the target semantic type. An emergency avoidance strategy is triggered for a high-threat target, and path re-planning or speed adjustment is started for a medium-low-threat target. This mechanism significantly optimizes the response efficiency to dangerous scenes such as children suddenly rushing out and stroller lateral movement, and avoids the risk of sudden braking or collision caused by response lag in traditional schemes.

[0063] A hierarchical fusion cost map planning strategy is adopted to decouple traffic efficiency, safety constraints and environmental factors into independent decision layers. The system dynamically weights and fuses three layers of cost according to real-time perception data: preferentially selects paths with sufficient turning radius, actively avoids risk areas such as children activity areas, and optimizes lane selection based on lighting information. At the cross-layer junction, a map preloading mechanism is used to realize navigation seamless switching. This scheme makes the planned path have safety and traffic efficiency under the premise of completely breaking away from the dependence on high-precision maps, and supports temporary construction, occupancy obstacles and other dynamic scenes.

[0064] The above embodiment can identify obstacles around the vehicle during automatic driving of the vehicle, especially when driving on a cross-layer slope, thereby improving driving safety of pedestrians around the vehicle and the vehicle itself.

[0065] Figure 4 A flowchart of an intelligent driving method for a cross-layer scene is provided for the embodiment of the present application, which is further used to illustrate the above embodiment. The method comprises the following steps.

[0066] In step S301, semantic recognition is performed according to real-time machine vision images to determine semantic anchor point information.

[0067] In step S302, an initial navigation route is determined according to a semantic map and destination information.

[0068] In step S303, the initial navigation route is optimized according to the semantic anchor point information to obtain a target navigation route, and navigation is performed according to the target navigation route until the vehicle reaches a cross-layer slope.

[0069] In step S304, the semantic map is supplemented according to the recognized semantic anchor point information.

[0070] In step S305, when the cross-layer slope is reached, slope information is determined according to real-time machine vision images.

[0071] In step S306, pose information determined by an inertial navigation system is obtained, and the slope information is optimized according to inertial information obtained by the inertial navigation system.

[0072] In step S307, motion compensation is performed according to the optimized slope information and the pose information, and automatic driving is performed according to a compensation result.

[0073] In step S308, obstacles in a radial direction of a vehicle travel direction are identified according to real-time machine vision images, and an obstacle type is determined, and a threat type of the obstacle is determined according to the obstacle type.

[0074] In step S309, if the threat type of the obstacle is a first threat type, the vehicle is stopped.

[0075] In step S310, if the threat type of the obstacle is a second threat type, an alternative lane is identified, and a target lane is selected according to a vehicle condition of the alternative lane and a turning radius of the vehicle.

[0076] The intelligent driving method for cross-layer scenarios provided in this invention can accurately locate and improve the confidence of semantic anchor point recognition, dynamically assess threats and correct cost map weights in real time, and output smooth control commands through hierarchical planning. The lightweight architecture significantly reduces storage and computing power requirements, enabling the system to be deployed on in-vehicle embedded platforms. It eliminates the decision-making lag encountered by traditional solutions when connecting ramps or navigating narrow roads. It achieves strong robustness to temporary obstacles and environmental changes, enabling vehicles to maintain smooth and safe autonomous driving capabilities across all scenarios.

[0077] This invention addresses three core pain points in intelligent parking system driving: positioning gaps, response lag, and high cost dependence. It creatively proposes a multimodal collaborative control architecture. Through a fusion positioning mechanism combining visual semantic anchors and real-time inertial navigation compensation, it completely solves the positioning drift problem in cross-layer scenarios, achieving centimeter-level continuous positioning even in environments with poor satellite signals. Combined with a radial velocity component dynamic threat assessment model and a hierarchical cost map planning algorithm, it significantly improves response speed and safety in the face of lateral obstacles and complex scenarios. Simultaneously, it employs a lightweight semantic map architecture, completely eliminating reliance on high-precision maps and fixed infrastructure, and supporting real-time adaptation to dynamic environments such as temporary construction and parking space occupancy. Ultimately, it achieves seamless autonomous parking across all scenarios, resulting in a simultaneous leap in system robustness, traffic efficiency, and deployment economy.

[0078] This invention uses a coupled positioning mechanism of visual semantic anchors and inertial navigation real-time compensation to solve cross-layer positioning drift; establishes a dynamic threat assessment model based on radial velocity components to achieve rapid response; and designs a lightweight layered cost map planning algorithm to ensure path safety and efficiency without relying on high-precision maps.

[0079] Figure 4 This is a schematic diagram of the structure of an intelligent driving device for cross-level scenarios provided by an embodiment of the present invention. This embodiment is applicable to situations where vehicles perform autonomous driving navigation across levels in underground parking lots or above-ground parking garages. This device can be executed by an intelligent driving device for cross-level scenarios, which can be implemented in hardware and / or software. This intelligent driving device for cross-level scenarios can be configured in electronic devices such as vehicle controllers, where the vehicle controller can be an in-vehicle infotainment system, an intelligent driving domain controller, or a vehicle controller. Figure 5 As shown, the device includes: an anchor point recognition module 41, an anchor point navigation module 42, and a ramp driving module 43.

[0080] Anchor point recognition module 41 is used to perform semantic recognition based on real-time machine vision images and determine semantic anchor point information;

[0081] Anchor point navigation module 42 is used to navigate according to the semantic anchor point information and destination information until the vehicle reaches the cross-level ramp;

[0082] The ramp driving module 43 is configured to determine ramp information according to real-time machine vision images when a cross-layer ramp is reached, perform motion compensation according to the ramp information and vehicle pose information, and perform automatic driving according to a compensation result.

[0083] On the basis of the above embodiment, the anchor navigation module 42 is configured to:

[0084] determine an initial navigation route according to the semantic map and destination information;

[0085] optimize the initial navigation route according to the semantic anchor information to obtain a target navigation route, and perform navigation according to the target navigation route.

[0086] On the basis of the above embodiment, the anchor navigation module 42 is configured to: Figure 6 as shown in the figure, further comprising a map supplement module 44, which is configured to supplement the semantic map according to the recognized semantic anchor information after optimizing the initial navigation route according to the semantic anchor information.

[0087] On the basis of the above embodiment, the ramp driving module 43 is configured to perform motion compensation according to the ramp information and vehicle pose information, and the motion compensation comprises:

[0088] obtaining pose information determined by an inertial navigation system, and optimizing the ramp information according to inertial information obtained by the inertial navigation system;

[0089] performing motion compensation according to the optimized ramp information and pose information.

[0090] On the basis of the above embodiment, the anchor navigation module 42 further comprises an obstacle identification module 45 and an obstacle avoidance module 46.

[0091] The obstacle identification module 45 is configured to identify obstacles according to real-time machine vision images after automatic driving according to the compensation result, and determine a threat type of the obstacles.

[0092] The obstacle avoidance module 46 is configured to adjust a navigation path according to the threat type of the obstacles and a preset safety strategy.

[0093] On the basis of the above embodiment, the obstacle identification module 45 is configured to:

[0094] identify obstacles in a radial direction of a vehicle travel direction according to real-time machine vision images, and determine an obstacle type;

[0095] determine a threat type of the obstacles according to the obstacle type.

[0096] On the basis of the above embodiment, the obstacle avoidance module 46 is configured to:

[0097] if the threat type of the obstacle is the first threat type, stopping the vehicle;

[0098] if the threat type of the obstacle is the second threat type, identifying an alternative lane, and selecting a target lane according to a traffic condition of the alternative lane and a turning radius of the vehicle.

[0099] The intelligent driving device for cross-layer scenarios provided by the embodiment of the present application comprises an anchor point identification module 41, a navigation anchor point module 42, and a ramp driving module 43. The anchor point identification module 41 is used for performing semantic identification according to real-time machine vision images to determine semantic anchor point information. The navigation anchor point module 42 is used for performing navigation according to the semantic anchor point information and destination information until the vehicle reaches a cross-layer ramp. The ramp driving module 43 is used for determining ramp information according to real-time machine vision images when the cross-layer ramp is reached, performing motion compensation according to the ramp information and vehicle pose information, and performing automatic driving according to the compensation result. Compared with the poor reliability of the current cross-layer automatic driving, the intelligent driving device for cross-layer scenarios provided by the embodiment of the present application can perform semantic identification on road signs in real-time machine vision images through machine vision technology to obtain semantic anchor point information for navigation. The navigation path obtained based on the destination information is optimized according to the semantic anchor point information, so that the vehicle can perform automatic driving according to the identified semantic anchor point information without relying on positioning information and drive to the cross-layer ramp. When the cross-layer ramp is reached, the vehicle is controlled using the ramp information obtained through machine vision and the vehicle pose information, and reliable cross-layer automatic driving is achieved.

[0100] The intelligent driving device for cross-layer scenarios provided by the embodiment of the present application can execute the intelligent driving method for cross-layer scenarios provided by any embodiment of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0101] Figure 6 is a structural schematic diagram of an electronic device provided by the embodiment of the present application. The electronic device 10 is intended to represent various forms of electronic devices such as a car controller, such as a vehicle information entertainment system controller, an intelligent driving domain controller, or a whole vehicle controller. The components shown herein, their connections and relationships, and their functions are merely examples, and are not intended to limit the implementation of the present application described and / or claimed herein.

[0102] As ​As shown, the electronic device 10 includes at least one processor 11, and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., communicatively connected to the at least one processor 11, where the memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer programs stored in the read-only memory (ROM) 12 or loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0103] Various components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a touch screen, physical buttons, etc., an output unit 17, such as various types of displays, a speaker, etc., a storage unit 18, such as a magnetic disk, an optical disk, etc., and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0104] The processor 11 can be various general and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs various methods and processes described above, such as the intelligent driving method across layers of scenarios.

[0105] In some embodiments, the intelligent driving method across layers of scenarios can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the intelligent driving method across layers of scenarios described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to perform the intelligent driving method across layers of scenarios by any other appropriate means, such as by means of firmware.

[0106] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system on a chip systems (SOCs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0107] Computer programs used to implement the intelligent driving method for cross-layer scenarios of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the computer program running on the processor implements the functions / operations specified in the flow diagrams and / or block diagrams. The computer program can execute entirely on a machine, partly on a machine, partly on a machine as a stand-alone software package, partly on a machine and partly on a remote machine or entirely on a remote machine or server.

[0108] The embodiments of the application also provide a computer readable storage medium, which stores computer instructions for causing a processor to execute an intelligent driving method for cross-layer scenarios, comprising:

[0109] According to the real-time machine vision image, semantic recognition is performed to determine semantic anchor point information;

[0110] According to the semantic anchor point information and destination information, navigation is performed until the vehicle reaches the cross-layer ramp;

[0111] When the cross-layer ramp is reached, ramp information is determined according to the real-time machine vision image; motion compensation is performed according to the ramp information and the pose information of the vehicle, and automatic driving is performed according to the compensation result.

[0112] On the basis of the above-mentioned embodiments, optionally, according to the semantic anchor point information and destination information, navigation is performed, comprising:

[0113] According to the semantic map and the destination information, an initial navigation route is determined;

[0114] According to the semantic anchor point information, the initial navigation route is optimized to obtain a target navigation route, and navigation is performed according to the target navigation route.

[0115] On the basis of the above-mentioned embodiments, optionally, after the initial navigation route is optimized according to the semantic anchor point information, the method further comprises:

[0116] According to the identified semantic anchor point information, the semantic map is supplemented.

[0117] On the basis of the above-mentioned embodiments, optionally, the motion compensation is performed according to the slope information and the pose information of the vehicle, and the method comprises:

[0118] Obtaining the pose information determined by the inertial navigation system; and optimizing the slope information according to the inertial information obtained by the inertial navigation system;

[0119] Performing the motion compensation according to the optimized slope information and the pose information.

[0120] On the basis of the above-mentioned embodiments, optionally, after the automatic driving is performed according to the compensation result, the method further comprises:

[0121] Performing obstacle identification according to real-time machine vision images to determine a threat type of the obstacle;

[0122] Adjusting the navigation path according to the threat type of the obstacle and a preset risk avoidance strategy.

[0123] On the basis of the above-mentioned embodiments, optionally, the obstacle identification is performed according to real-time machine vision images to determine a threat type of the obstacle, and the method comprises:

[0124] Identifying the obstacle in a radial direction of a traveling direction of the vehicle according to real-time machine vision images to determine an obstacle type;

[0125] Determining the threat type of the obstacle according to the obstacle type.

[0126] On the basis of the above-mentioned embodiments, optionally, the navigation path is adjusted according to the threat type of the obstacle and a preset risk avoidance strategy, and the method comprises:

[0127] If the threat type of the obstacle is a first threat type, stopping the vehicle;

[0128] If the threat type of the obstacle is a second threat type, identifying an alternative lane, and selecting a target lane according to a lane condition of the alternative lane and a turning radius of the vehicle.

[0129] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer-readable storage medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of a machine-readable storage medium will include a one or more lines of a computer program code, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0130] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0131] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0132] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0133] It should be understood that the various forms of flow illustrated above can be used to reorder, add or delete steps. For example, the steps described in the present application can be performed in parallel, in series, in a different order, or any combination thereof, as long as the desired results of the technical solutions of the present application can be achieved, which are not limited herein.

[0134] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for intelligent driving across layers of scenarios, the method comprising: The method comprises the following steps: According to the real-time machine vision image, the semantic recognition is carried out, and the semantic anchor point information is determined; According to the semantic anchor point information and the destination information, the navigation is carried out until the vehicle reaches the cross-layer ramp; When the cross-layer ramp is reached, the ramp information is determined according to the real-time machine vision image; the motion compensation is carried out according to the ramp information and the pose information of the vehicle, and the automatic driving is carried out according to the compensation result.

2. The method of claim 1, wherein, According to the semantic anchor point information and the destination information, the navigation is carried out, which comprises: According to the semantic map and the destination information, the initial navigation route is determined; According to the semantic anchor point information, the initial navigation route is optimized to obtain the target navigation route, and the navigation is carried out according to the target navigation route.

3. The method of claim 2, wherein, After the initial navigation route is optimized according to the semantic anchor point information, the method further comprises the following steps: According to the recognized semantic anchor point information, the semantic map is supplemented.

4. The method of claim 1, wherein, According to the ramp information and the pose information of the vehicle, the motion compensation is carried out, which comprises: The pose information determined by the inertial navigation system is obtained; the ramp information is optimized according to the inertial information obtained by the inertial navigation system; According to the optimized ramp information and the pose information, the motion compensation is carried out.

5. The method of claim 1, wherein, After the automatic driving is carried out according to the compensation result, the method further comprises the following steps: According to the real-time machine vision image, the obstacle recognition is carried out, and the threat type of the obstacle is determined; According to the threat type of the obstacle and the preset risk avoidance strategy, the navigation path is adjusted.

6. The method of claim 5, wherein, According to the real-time machine vision image, the obstacle recognition is carried out, and the threat type of the obstacle is determined, which comprises: According to the real-time machine vision image, the obstacle in the radial direction of the vehicle travel direction is recognized, and the obstacle type is determined; According to the obstacle type, the threat type of the obstacle is determined.

7. The method of claim 5, wherein, According to the threat type of the obstacle and the preset risk avoidance strategy, the navigation path is adjusted, which comprises: If the threat type of the obstacle is the first threat type, the vehicle is stopped; If the threat type of the obstacle is the second threat type, the alternative lane is recognized, and the target lane is selected according to the lane condition of the alternative lane and the turning radius of the vehicle.

8. An intelligent driving apparatus for cross-layer scenarios, characterized by, The method comprises the following steps: An anchor point recognition module is configured to determine semantic anchor point information by performing semantic recognition on real-time machine vision images; An anchor point navigation module is configured to navigate according to the semantic anchor point information and destination information until the vehicle reaches a cross-layer ramp; A ramp driving module is configured to determine ramp information from real-time machine vision images when the cross-layer ramp is reached; perform motion compensation according to the ramp information and the pose information of the vehicle; and perform automatic driving according to the compensation result.

9. An electronic device, comprising: The electronic device comprises: At least one processor; and A memory connected in communication with the at least one processor; wherein The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the intelligent driving method for cross-layer scenarios according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to implement the intelligent driving method for cross-layer scenarios according to any one of claims 1-7 when executed.

Images