WO2021155685A1 - 一种更新地图的方法、装置和设备 - Google Patents

一种更新地图的方法、装置和设备 Download PDF

Info

Publication number
WO2021155685A1
WO2021155685A1 PCT/CN2020/125615 CN2020125615W WO2021155685A1 WO 2021155685 A1 WO2021155685 A1 WO 2021155685A1 CN 2020125615 W CN2020125615 W CN 2020125615W WO 2021155685 A1 WO2021155685 A1 WO 2021155685A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
abnormal scene
sensing data
map
abnormal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/125615
Other languages
English (en)
French (fr)
Inventor
丁涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP20917388.9A priority Critical patent/EP4089659A4/en
Publication of WO2021155685A1 publication Critical patent/WO2021155685A1/zh
Priority to US17/879,252 priority patent/US20220373353A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
    • G08G1/0969—Systems involving transmission of navigation instructions to the vehicle having a display in the form of a map
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3807—Creation or updating of map data characterised by the type of data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G01C21/30—Map- or contour-matching
    • G01C21/32—Structuring or formatting of map data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3833—Creation or updating of map data characterised by the source of data
    • G01C21/3848—Data obtained from both position sensors and additional sensors
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/251—Fusion techniques of input or preprocessed data
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0112—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0116—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from roadside infrastructure, e.g. beacons
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0133—Traffic data processing for classifying traffic situation
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G08G1/0141—Measuring and analyzing of parameters relative to traffic conditions for specific applications for traffic information dissemination
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/04—Detecting movement of traffic to be counted or controlled using optical or ultrasonic detectors
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G08G1/096708—Systems involving transmission of highway information, e.g. weather, speed limits where the received information might be used to generate an automatic action on the vehicle control
    • G08G1/096725—Systems involving transmission of highway information, e.g. weather, speed limits where the received information might be used to generate an automatic action on the vehicle control where the received information generates an automatic action on the vehicle control
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G08G1/096766—Systems involving transmission of highway information, e.g. weather, speed limits where the system is characterised by the origin of the information transmission
    • G08G1/096775—Systems involving transmission of highway information, e.g. weather, speed limits where the system is characterised by the origin of the information transmission where the origin of the information is a central station

Definitions

  • This application relates to the field of intelligent driving technology, and in particular to a map update method, device and equipment.
  • the map used by the vehicle is one of the most important components.
  • autonomous vehicles must accurately know their position on the road, such as the distance to the side of the road shoulder, the distance to the lane line, and so on. Therefore, the absolute accuracy of maps used by autonomous vehicles generally needs to reach the decimeter level or even the centimeter level.
  • the map also needs to store various traffic elements in the traffic scene, including the road network data, lane lines, and traffic signs of traditional maps, as well as the slope, curvature, heading, elevation, and roll of the lane. Data such as degree.
  • More than 95% of the current autonomous driving schemes rely on high-precision maps to obtain current road information and environmental information. If the map used by the autonomous vehicle cannot be updated in time, and the map does not match the current actual road, then It will be a very dangerous thing for autonomous driving. Therefore, in order to create a safe driving environment, it is essential to update the maps on which autonomous vehicles rely in real time, accurately and quickly.
  • the present application provides a map update method, device, and equipment, which can improve the efficiency of map update, ensure the safety of the autonomous driving environment, save computing power, and reduce update costs.
  • the first aspect of the present application provides a map update method.
  • the method includes: when an abnormal scene occurs, acquiring abnormal scene sensing data, and calculating a minimum safety margin based on the abnormal scene sensing data.
  • the minimum safety boundary refers to the minimum impact range of the abnormal scene marked on the map on the traffic.
  • the map is updated according to the minimum safety margin obtained by calculation.
  • the above method triggers the map update program when an abnormal scene occurs, uses the abnormal scene sensing data to determine the minimum safety margin, and updates the map according to the calculated minimum safety margin, which improves the real-time of map refreshing, thereby ensuring the safety of the autonomous driving environment sex.
  • calculating the minimum safety margin based on the abnormal scene sensing data includes: obtaining the vehicle drivable area according to the abnormal scene sensing data, and then calculating the minimum safety margin of the abnormal scene on the map according to the vehicle drivable area.
  • the vehicle-driving area refers to the area where the vehicle can pass safely determined from the driving perspective. This application first calculates the drivable area, and then calculates the minimum safety margin through the drivable area, which improves the accuracy of the calculation result.
  • the abnormal scene sensing data includes vehicle-mounted sensor sensing data.
  • the method for calculating the driving area of the vehicle according to the abnormal scene sensing data includes: inputting the sensor data of the vehicle-mounted sensor to a pre-trained neural network to obtain the driving area of the vehicle.
  • the pre-trained neural network model is used to calculate the vehicle's travelable area, which improves the calculation speed and accuracy of the calculation results, and ensures the accuracy and effectiveness of the map.
  • the method for calculating the driving area of the vehicle includes: inputting various types of sensing data obtained by the on-board sensors into corresponding multiple types of pre-trained neural networks to obtain the driving area of the vehicle. Multiple estimates. Then, multiple estimates of the vehicle's drivable area are merged to obtain the merged vehicle's drivable area.
  • the invention integrates the travelable area obtained by calculating multiple sensor data, improves the accuracy of estimating the vehicle travelable area, and ensures the accuracy and effectiveness of map update.
  • the abnormal scene sensing data includes vehicle sensor sensing data and road monitoring sensing data.
  • the method for acquiring road monitoring sensing data includes: determining the location of nearby abnormal scenes according to the abnormal scene location information in the vehicle sensor sensing data.
  • the road monitoring cameras are assembled, and then the road monitoring data before and after the abnormal scenes collected by these camera sets are obtained.
  • This application considers road monitoring data in combination with real scene conditions. Road surveillance cameras have fixed positions, clear orientations, and the quality of surveillance images is becoming more and more perfect. Therefore, the use of road surveillance data can more accurately calculate the vehicle's travelable area, fully utilize existing surveillance resources, and achieve reasonable allocation of resources.
  • the method of calculating the travelable area of the vehicle according to the abnormal scene sensing data includes: comparing the road monitoring data before and after the occurrence of the abnormal scene to obtain the travelable area of the vehicle.
  • the image processing method of computer vision is used to calculate the travelable area of the vehicle, which provides another option for calculating the travelable area of the vehicle.
  • the method of calculating the minimum safety margin according to the vehicle's drivable area includes calculating the minimum safety margin according to the location information of the abnormal scene and the vehicle's drivable area.
  • calculating the minimum safety margin based on the location information of the abnormal scene and the driveable area of the vehicle includes: determining the coordinates of the driveable area of the vehicle in the vehicle coordinate system using the location information of the abnormal scene as a reference point. Then, according to the mapping relationship between the vehicle coordinate system and the global coordinate system adopted by the map, the coordinates of the vehicle's travelable area are converted to coordinates in the global coordinate system to obtain the minimum safety margin of the abnormal scene.
  • the driveable area of the vehicle can be directly mapped to the map, and the method is simple and accurate.
  • the method for acquiring abnormal scene sensing data includes: when an abnormal scene is detected, the vehicle-mounted communication device triggers a vehicle-mounted sensor to acquire the abnormal scene sensing data.
  • the in-vehicle communication device actively triggers the map correction program after acquiring the abnormal scene sensing data, which greatly improves the real-time response of the system and guarantees the personal safety of the passengers of the self-driving vehicle.
  • this proactive approach to the law clarifies the updated location, which greatly saves computing resources.
  • the vehicle-mounted sensor sensing data includes obstacle information/point cloud information collected by the vehicle-mounted radar, pictures and videos collected by the vehicle-mounted camera, and location information acquired by the vehicle-mounted satellite positioning receiving system.
  • a car is often equipped with a variety of on-board sensors. The combination of multiple types of sensor data can ensure the accuracy of map updates.
  • the abnormal scenario includes: a traffic accident, road construction, or vehicle failure.
  • the abnormal scene mainly refers to the situation where the road is abnormally occupied.
  • the vehicle and the construction vehicle in the traffic accident or breakdown trigger the on-board communication device to obtain the abnormal scene sensing data to perform the update of the map.
  • the update method provided by the present invention triggers the map update operation by ordinary social vehicles. Compared with the prior art, the map acquisition vehicle or crowd-sourcing vehicle is used to acquire and trigger the map update, which reduces the cost of map update.
  • the map update method provided by the present application enables the abnormal vehicle to obtain the abnormal scene sensing data in the first time when the abnormal scene occurs.
  • This method clarifies the updated road sections, saves computing power, but also improves the response speed of the map system to abnormal situations and improves the efficiency of map refreshing.
  • the vehicle that triggers the update is an ordinary social vehicle, and there is no need for a map service company to send a collection vehicle or hire a crowdsourced vehicle to collect, which reduces the cost of map refresh.
  • the introduction of a variety of sensor data to calculate the vehicle's travelable area has improved the accuracy of map updates.
  • the present application provides a map update device, the update device includes: an acquisition module, a processing module, and an update module;
  • the acquisition module is used to obtain abnormal scene sensing data; the processing module is used to calculate the minimum safety margin of the abnormal scene based on the abnormal scene sensing data, and the minimum safety margin is used to mark the abnormal scene pair on the map.
  • the minimum impact range of traffic; the update module is used to update the map according to the minimum safety boundary obtained by calculation.
  • the processing module is configured to: obtain a vehicle-drivable area of an abnormal scene according to the abnormal scene sensing data, where the vehicle-drivable area refers to a vehicle safe-passable area determined from a driving perspective; Then, the minimum safety margin of the abnormal scene is calculated according to the travelable area of the vehicle.
  • the abnormal scene sensing data includes vehicle-mounted sensor sensing data
  • the processing module is specifically configured to input the vehicle-mounted sensor sensing data into a pre-trained neural network to obtain the vehicle's travelable area.
  • the processing module is also used to input various types of sensing data obtained by the on-board sensors into corresponding various types of pre-trained neural networks to obtain multiple estimates of the vehicle's travelable area ; Combine these multiple estimates, and calculate the fused vehicle driving area.
  • the abnormal scene sensing data includes vehicle sensor sensing data and road monitoring sensing data.
  • the specific method for the acquisition module to obtain the road monitoring sensing data is as follows: according to the abnormal scene location information contained in the vehicle sensor sensing data, the road monitoring camera set near the abnormal scene is determined. Then, the road monitoring sensing data collected by the road monitoring camera collection is acquired, and the road monitoring sensing data includes the road monitoring data before the occurrence of the abnormal scene and the road monitoring data after the occurrence of the abnormal scene.
  • the processing module is further configured to compare the road monitoring data collected by the road monitoring camera set before and after the occurrence of the abnormal scene to obtain the travelable area of the vehicle.
  • the processing module is further configured to calculate the minimum safety margin of the abnormal scene according to the location information of the abnormal scene and the driveable area of the vehicle.
  • the processing module calculates the minimum safety margin based on the location of the abnormal scene and the vehicle's driveable area, which specifically includes: taking the location information of the abnormal scene as a reference point, and acquiring the vehicle's driveable area based on the sensor data of the vehicle.
  • the coordinates of the vehicle's coordinate system then, according to the mapping relationship between the vehicle's coordinate system and the global coordinate system adopted by the map, the coordinates of the vehicle's travelable area are converted to the coordinates of the global coordinate system to obtain the smallest abnormal scene Security boundary.
  • the acquiring module is specifically configured to acquire the abnormal scene sensing data when an abnormal scene is detected.
  • the vehicle-mounted sensor sensing data includes obstacle information/point cloud information collected by the vehicle-mounted radar, pictures and videos collected by the vehicle-mounted camera, and location information acquired by the vehicle-mounted satellite positioning receiving system.
  • the abnormal scenario includes: a traffic accident, road construction, or vehicle failure.
  • this application provides a map update system, which specifically includes a vehicle-mounted communication device and a cloud server.
  • the vehicle-mounted communication device is used to obtain abnormal scene sensing data when an abnormal scene is detected, and send the abnormal scene sensing data to the cloud system to trigger the cloud system to perform an update.
  • the cloud server is configured to calculate a minimum safety margin based on the acquired abnormal scene sensing data, and update the map. The minimum safety margin is used to mark the minimum impact range of the abnormal scene on traffic on the map.
  • this application provides another map update system, which specifically includes a vehicle-mounted communication device and a cloud server.
  • the vehicle-mounted communication device is used to obtain abnormal scene sensing data when an abnormal scene is detected; calculate a minimum safety margin based on the obtained abnormal scene sensing data, and the minimum safety margin is used to identify the abnormality on a map The minimum impact range of the scene on the traffic; and the minimum safety boundary is sent to the cloud system to trigger the cloud system to perform an update.
  • the cloud server is configured to update the map according to the obtained minimum security boundary.
  • the present application provides a vehicle-mounted communication box.
  • the vehicle-mounted communication box includes a processor, a memory, a communication interface, and a bus.
  • Communication the memory is used to store computer-executed instructions, and when the on-board communication box is running, the processor executes the computer-executed instructions in the memory to use the hardware resources in the on-board communication box to perform all the above aspects or All aspects are the operation steps of the method implemented by the vehicle-mounted communication device in the method described in any possible implementation manner.
  • the present application provides a networked vehicle, characterized in that the networked vehicle includes a vehicle-mounted communication device and a vehicle-mounted sensor.
  • the vehicle-mounted sensor acquires abnormal scene sensing data, and calculates based on the abnormal scene sensing data
  • the minimum safety boundary is used to identify the minimum impact range of the abnormal scene on the traffic on the map, and send the minimum safety boundary to the cloud system to trigger the cloud system to update the map.
  • the present application provides another connected vehicle, which is characterized in that the connected vehicle includes a vehicle-mounted communication device and a vehicle-mounted sensor.
  • the vehicle-mounted sensor acquires abnormal scene sensing data and senses the abnormal scene
  • the data is sent to the cloud system, which triggers the cloud system to perform an update.
  • the present application provides a cloud server, which is characterized in that when an abnormal scene occurs, the cloud server obtains abnormal scene sensing data, and calculates a minimum safety margin based on the abnormal scene sensing data to update the map.
  • the present application provides a computer-readable storage medium that stores instructions in the computer-readable storage medium, which when run on a computer, causes the computer to execute the above-mentioned first aspect and possible implementations. method.
  • this application provides a computer program product containing instructions, which when run on a computer, enables the computer to execute the method described in the first aspect and possible implementations.
  • FIG. 1A is a schematic diagram of an application scenario of an existing map update method provided by an embodiment of the present application.
  • FIG. 1B is a schematic diagram of an application scenario of the map update method of the present invention provided by an embodiment of the present application.
  • Fig. 2 is a schematic diagram of the logical architecture of the map update system provided by an embodiment of the present application.
  • Fig. 3 is a schematic flowchart of a map update method provided by an embodiment of the present application.
  • FIG. 4 is a schematic flowchart of another map update method provided by an embodiment of the present application.
  • Fig. 5 is a schematic diagram of a process for calculating a drivable area provided by an embodiment of the present application.
  • Fig. 6 is a schematic diagram of a drivable area calculated by using a neural network according to an embodiment of the present application.
  • FIG. 7 is a schematic diagram of multi-layer fusion of multiple drivable areas according to an embodiment of the present application.
  • Fig. 8 is a schematic diagram of another process for calculating a drivable area provided by an embodiment of the present application.
  • Fig. 9 is a schematic structural diagram of a map updating device provided by an embodiment of the present application.
  • Fig. 10 is a schematic structural diagram of a map updating device provided by an embodiment of the present application.
  • the map update product is mounted on the vehicle through the front-mounted installation.
  • the computing platform can estimate the relative position information of the vehicle in real time.
  • the real-time sensed results are compared with the current high-precision map.
  • the system finds that the real-time perception result does not match the map, it is reported to the map processing system in the cloud.
  • the cloud system processes the data uploaded by the vehicles to form map update information, and distributes it to each vehicle to complete the closed loop of the entire map update.
  • the current update method using collection vehicles/crowdsourced vehicles is difficult to quickly match the updated road section. It can only be updated according to pre-determined routes or random driving routes that rely on crowdsourced vehicles, which cannot guarantee the real-time validity of the map.
  • FIG. 1A when a car 101 has a traffic accident and the road is blocked, if the set trajectory of the collection car/crowd-sourcing car 102 (car A) is driving along the direction of A1 or A2 , The map will not be updated in time. When other self-driving cars pass through this abnormal road section, since their own maps cannot be updated in time, dangerous scenes of self-driving will appear.
  • the collection vehicle/crowd-sourcing vehicle needs to follow a preset route or a random route, compare the scanned scene with the existing map in real time, and then report to the cloud when an abnormality is found.
  • Such an update method has extremely high requirements for computing power and communication bandwidth.
  • each collection vehicle/crowdsourced vehicle needs to be equipped with a unique set of sensors and processing and calculation units, the number of collection vehicles/crowdsourced vehicles is very limited, which is suitable for large-scale, large-scale high-precision maps , Obviously can not meet the requirements of fast refresh.
  • Fig. 1B is an application scenario of an embodiment of the present invention.
  • the vehicle 101 with an abnormal traffic is equipped with a TBOX (Telecommunications Box), a positioning module, and other on-board sensors.
  • TBOX Telecommunications Box
  • the cloud system will also obtain the data collected by the road camera based on the positioning information.
  • a minimum safety margin (as shown by the dashed frame around the vehicle 101 in FIG. 1B) is calculated to update the high-precision map.
  • the update method provided by the present invention can actively report the abnormal scene data at the first time when the abnormal scene occurs, trigger the emergency correction procedure of the cloud system, and improve the response speed of map refresh.
  • the calculation results of multiple sensor data can be fused with each other, which improves the accuracy of map updates and ensures the safety of the autonomous driving environment.
  • FIG 2 is a schematic diagram of a network architecture of a map update system provided by this application.
  • the system includes a cloud system 201, a network 202, a road monitoring camera 203, and a vehicle 204 that triggers an abnormal scene.
  • the cloud system 201 communicates with the road monitoring camera 203 or the vehicle 204 that triggers the abnormal scene through the network 202.
  • the cloud system 201 can process a large amount of sensor data, calculate the minimum drivable area, and then project it to the map according to the positioning to obtain the minimum safety margin, and finally update the map.
  • the network 202 is a medium for realizing the transmission of data from on-board sensors and road surveillance cameras to the cloud system.
  • the network 202 includes a wired and/or wireless transmission method, where the wired transmission method includes data transmission in the form of Ethernet, optical fiber, and the like.
  • Wireless transmission methods include broadband cellular network transmission methods such as 3G, 4G (Fourth generation), or 5G (fifth generation).
  • the road monitoring cameras 203 have a fixed position and a clear orientation, and are distributed on both sides of the road. It has a networking function and can upload a fixed-view video stream of abnormal scenes to the cloud to calculate the drivable area, thereby updating a map with the smallest safety margin.
  • the vehicle 204 that triggers the abnormal scene refers to a vehicle that has a traffic accident, a faulty vehicle, a construction vehicle, or the like that causes an abnormal lane state.
  • the vehicle 204 that triggers the abnormal scene includes a Telecommunications Box (TBOX) 2041, a central gateway 2042, a body control module (BCM) 2043, a human-computer interaction controller 2044, a vehicle sensor 2045, a black box device 2046, etc.
  • Devices or devices can communicate through a Controller Area Network (CAN) or In-car Ethernet (In-car Ethernet), which is not limited in this application.
  • the communication box 2041 is used to implement communication between the vehicle 204 that triggers the abnormal scene and the cloud system 201.
  • the body controller 2043 is used to control basic hardware devices of the automobile such as the door 20431 and the window 20432.
  • the human-machine interaction controller 2044 includes in-vehicle entertainment (In-Vehicle Infotainment, IVI) and/or hardware monitor interface (Hardware Monitor Interface, HMI) and other in-vehicle entertainment control systems.
  • the vehicle-mounted sensor 2045 includes a radar 20451, a camera 20452, and a positioning module 20453. The data of the vehicle-mounted sensor 2045 will be uploaded to the cloud through the communication box 2041.
  • the radar 20451 may use radio signals to sense objects in the surrounding environment of the vehicle.
  • the radar 20451 may be a lidar (Lidar) to provide point cloud information of the surrounding environment.
  • the camera (camera or camera group) 20452 can be used to capture multiple images of the surrounding environment of the vehicle.
  • the camera 20452 may be a still camera or a video camera.
  • at least one camera (camera or camera group) 20452 may be installed on the front and rear bumpers, side mirrors, and windshield of the vehicle, respectively.
  • the positioning module 20453 may adopt a global positioning system (Global Positioning System, GPS), a Beidou system or other systems, and can output global positioning information with a certain accuracy.
  • the black box device 2046 is used to record the body data of the smart car in an emergency.
  • the car in addition to communicating with the outside world through the communication box, the car can also be implemented through other devices.
  • the management system shown in FIG. 2 may not include the central gateway 2042, and each controller and sensor may be directly connected to the communication box 2041.
  • FIG. 2 is only to better illustrate the method for updating the high-precision map provided by this application, and does not constitute a limitation to the embodiments of this application.
  • Figure 3 is the overall solution process of the present invention. The method includes the following steps:
  • S301 The abnormal vehicle sends the sensor data of the on-board sensor.
  • the vehicle that triggers the abnormal scene actively acquires the sensor data of its own vehicle and sends it to the cloud system within the first time.
  • the sensing data of on-board sensors mainly includes obstacle information/point cloud information collected by radar, pictures and videos collected by on-board cameras (cameras or camera groups), and position information fed back by the global positioning system (GPS) or Beidou navigation system .
  • Vehicle sensor data is actively sent to the cloud system by the vehicle that triggers the abnormal scene as soon as the abnormality occurs. This method of actively triggering the map update program improves the efficiency of map update. Improve the safety of self-driving cars.
  • the vehicle that triggers the abnormal scene in this embodiment may be equipped with an advanced driver assistance system (ADAS) to realize the function of automatic driving, or may not have the function of automatic driving.
  • ADAS advanced driver assistance system
  • the vehicle may be a vehicle that has been involved in a traffic accident, a vehicle that has broken down, or a construction vehicle.
  • abnormal vehicles will also push alarm information to a collection of autonomous vehicles that meet preset conditions to remind vehicles or passengers to pay attention to the information matching between the map and the actual road, and to modify the path planning strategy. Reduce the priority of vehicles passing this abnormal road section and change to detour other roads.
  • the set of vehicles that meet the preset conditions may be a set of autonomous vehicles within 500 meters from the location of the abnormal scene, which is not specifically limited in the present invention.
  • the alarm information can include the location of the abnormal scene, the reason for the abnormality, and the expected duration of the abnormality.
  • S302 The cloud system receives the abnormal scene sensing data.
  • the abnormal scene sensing data may only include the data of the vehicle-mounted sensor, or may also include road monitoring data determined according to the location data provided by the vehicle-mounted sensor.
  • the cloud system first receives the on-board sensor sensing data uploaded by the vehicle that triggered the abnormal scene.
  • the vehicle-mounted sensor sensing data includes the location information of the abnormal scene, the point cloud information collected by the vehicle-mounted radar, and the video data collected by the vehicle-mounted camera.
  • the cloud system can locate a collection of road monitoring cameras near the abnormal scene to obtain road monitoring data before and after the abnormal scene occurs.
  • the road camera collection has a fixed position and a fixed orientation.
  • the cloud system combined with the position information provided by the on-board sensor can quickly locate the road surveillance camera collection near the abnormal scene.
  • the relationship between the road camera and the abnormal scene meets the preset condition.
  • the preset condition may be that the distance is less than a certain threshold, and the orientation of the camera set can capture an abnormal scene.
  • S303 The cloud system calculates the drivable area.
  • the cloud system calculates the minimum drivable area based on the acquired sensing data of the abnormal scene.
  • the so-called drivable area is considered from the perspective of vehicle driving, that is, the area on the road where vehicles can pass safely.
  • the abnormal scene sensing data can come from on-board sensors or from a collection of road monitoring cameras. According to different data, the cloud system can use different methods to calculate the vehicle's travelable area.
  • S304 The cloud system calculates the minimum safety margin to update the map.
  • the so-called minimum safety margin is relative to the map.
  • a range that has the least impact on the traffic road that is, the vehicle can pass smoothly
  • the cloud system projects the calculated drivable area on the map according to the positioning information, and calculates the minimum safety margin to update the map.
  • the method of projecting the drivable area of the vehicle onto the map is as follows: First, the position of the abnormal scene is used as the origin of the coordinates to establish the own vehicle coordinate system. Combined with the data provided by the on-board sensors, determine the coordinates of the driveable area in the own vehicle coordinate system. Secondly, it is necessary to determine the mapping relationship between the vehicle coordinate system and the global coordinate system adopted by the map.
  • the scale or mapping relationship between the two coordinate systems can be determined according to the distance between the fixed road element such as surveillance camera or road sign and the position of the abnormal scene, and the distance between the fixed road element and the position of the abnormal scene on the map .
  • the cloud system when the drivable area is mapped, the cloud system also needs to combine the road information on the high-precision map to optimize the map level. Exemplarily, smoothing or regularizing the mapped area is performed.
  • steps S302-S304 the calculation of the vehicle's travelable area and the minimum safety margin does not necessarily have to be performed by the cloud system.
  • the abnormal vehicle itself is equipped with a high-performance computing center, it can also be performed by the on-board computing center.
  • the vehicle can drive area and the minimum safety margin are calculated, and then the vehicle-mounted communication device uploads the calculation result of the minimum safety margin to the cloud, and the cloud performs map update.
  • the cloud system after the cloud system updates the map according to the minimum safety boundary, it can also use the data of the collected vehicle/crowd-sourcing vehicle to assist in the correction of the update result.
  • S305 Collect the collected data of the collection vehicle/crowd-sourcing vehicle for map correction.
  • the cloud system can directly collect the data collected by the collection vehicles/crowdsourcing vehicles that are about to pass the road section where the abnormal scene is located, or call idle collection vehicles/crowdsourcing vehicles to collect data on the abnormal road sections. Recalculate the drivable area according to the data of the collected vehicle/crowd-sourcing vehicle, project the calculated drivable area onto the map and calculate the new minimum safety margin, revise and adjust the minimum safety margin previously calculated through sensor data, and update the high-precision map.
  • S306 Determine that the abnormal scene has ended, and cancel the abnormal mark.
  • the vehicle that triggered the abnormal scene can actively report the abnormal end information.
  • the construction truck actively sends a message to remind the cloud abnormality has ended.
  • the abnormal end information can also be reported to the cloud by the collection vehicle/crowdsourcing vehicle.
  • the cloud cancels the abnormal mark on the previous map after confirming that the abnormal information has ended.
  • the vehicle that triggers the abnormal scene actively reports the abnormal data, and initiates the emergency correction program of the cloud map.
  • the efficiency of this update method is greatly improved.
  • the map refresh operation can be completed 1-5 minutes after the occurrence of the abnormal scene, and the map with the minimum safety boundary can be updated without the need for collection vehicles/crowd-sourced vehicles to pass by, which reduces the refresh cost.
  • the method of the present invention greatly saves computing power.
  • the cloud system can not only calculate the drivable area through the data of the on-board sensors, but also calculate the road monitoring video stream data obtained by the positioning.
  • the two minimum safety margins calculated by the two methods can also be fused and adjusted, which improves the accuracy of map update.
  • the foregoing step S303 calculating the drivable area can be implemented in the following three ways.
  • Implementation method 1 The cloud system calculates the driveable area based on the visual data collected by the on-board camera (camera/camera group). With reference to Figure 5, the steps of the above method are as follows:
  • S401 Pre-train the Multinet network using the data set marked with the drivable area.
  • the so-called MultiNet is composed of multiple network structures, which can perform three tasks at the same time: classification, detection, and semantic segmentation.
  • this network When this network performs three tasks, it generally shares an encoder for feature extraction.
  • the structure of the encoder can be VGG16 or other network architectures, such as GoogleNet and ResNet101.
  • three decoders After extracting the features, three decoders are used to perform the three tasks of classification, detection, and semantic segmentation. Use a large number of data sets to train the Multinet network, and use the validation set to verify the accuracy of the model until the accuracy of the model meets the preset requirements.
  • S402 Input the visual data collected by the vehicle-mounted camera to the aforementioned Multinet neural network.
  • Fig. 6 is a schematic diagram of using the MultiNet network to identify the drivable area.
  • the irregular area 501 filled with black dots is directly calculated by the Multinet network to represent the safe passage of the autonomous vehicle, that is, the vehicle drivable area.
  • Implementation method 2 The cloud system calculates the drivable area based on the point cloud data collected by the on-board radar.
  • the second implementation is similar to the first implementation, except that the data type is converted from two-dimensional ordinary visual data to a three-dimensional point cloud array.
  • a point cloud is a set of 3D points that carry depth information.
  • the point cloud can also have the RGB value of each point to form a colorful point cloud.
  • a point cloud neural network such as PointNet
  • the point cloud travelable area picture 603 calculated by using the millimeter wave radar sensor data is rasterized, and the size of the raster is not limited.
  • Each grid in 601, 602, and 603 has a confidence level in the previous calculation process.
  • the so-called confidence refers to the probability that the grid is recognized as a vehicle-drivable area.
  • the three confidence levels corresponding to the same grid are summed proportionally. If the result is greater than the preset threshold, the grid is a drivable area, otherwise it is an undrivable area.
  • the three drivable areas 601, 602, and 603 are fused and adjusted to improve the accuracy of the calculation of the vehicle drivable area.
  • Implementation method 3 The cloud system calculates the drivable area based on the road monitoring sensor data. With reference to Figure 8, the steps of the above implementation are as follows:
  • the cloud system obtains the historical video stream of the abnormal scene collected by the road surveillance camera.
  • the cloud system locates a collection of surveillance cameras near the abnormal scene based on the location information provided by the vehicle-mounted sensor. Then, the cloud system can obtain the video stream and historical video stream of the abnormal scene from the database of the storage device, or directly from the road surveillance camera terminal.
  • the embodiment of the present application does not limit the manner of obtaining the current abnormal video stream and the historical video stream, and may be determined according to actual conditions.
  • S702 The cloud system extracts a vehicle-free pedestrian picture from the historical video stream.
  • a picture of pedestrians without vehicles means that there are no vehicles, pedestrians, construction or any obstacles on the road in the picture, and it only contains a complete and clean road. This step can be completed by selecting a suitable neural network, or by using traditional image processing methods.
  • S703 Use traditional computer vision methods to continuously compare the pictures of pedestrians without vehicles with the video streams of the abnormal scenes to obtain the abnormal areas.
  • S704 Calculate the drivable area in reverse. Using methods such as the expansion threshold, reverse calculation is performed on the abnormal area obtained in step 703 to obtain the vehicle travelable area.
  • abnormal road monitoring sensing data in addition to using traditional computer vision methods to compare historical data and abnormal data, abnormal road monitoring sensing data can also be directly input to the neural network to obtain the vehicle's travelable area.
  • the above-mentioned calculated drivable areas are all safe areas for driving vehicles. Calculating the drivable area is the prerequisite for updating the high-precision map with the smallest safety margin.
  • the calculation data sources of implementation mode 1 and implementation mode 2 are both on-board sensors, while the calculation data sources of implementation mode 3 are road surveillance cameras. Both types of data can calculate the driving area of the vehicle.
  • the minimum safety margins calculated from the projections of these two types of drivable areas can also be adjusted mutually to improve the accuracy of the minimum safety margins.
  • the map update method provided in this application calculates the minimum drivable area of the vehicle's view angle according to the data of the vehicle-mounted sensor uploaded by the vehicle-mounted communication device or the located road monitoring data, and then projects it to the corresponding position on the map to refresh the minimum safety margin map.
  • the vehicle that triggers the abnormal scene actively starts the map correction program, which greatly improves the map refresh efficiency, and the high-precision map refresh operation can be completed 1-5 minutes after the abnormal scene occurs. There is no need to wait for collection vehicles/crowdsourcing vehicles, which reduces the cost of refreshing. And the updated road section is clear, which reduces the huge computing power requirements and processing time caused by the high-precision map correction in the cloud.
  • the vehicle-mounted communication device of the present application may be a smart vehicle with communication capabilities.
  • FIG. 9 is a map update device 800 provided by an embodiment of the application.
  • the update device 800 may include: an acquisition module 801, an alarm module 802, a processing module 803, an update module 804, and a verification module 805.
  • the obtaining module 801 is used to obtain sensing data of the abnormal scene when the abnormal scene occurs;
  • the alarm module 802 (optional) is used to push alarm information to a collection of vehicles meeting preset conditions;
  • the processing module 803 is configured to calculate the minimum safety boundary of the abnormal scene according to the acquired abnormal scene sensing data, and the minimum safety boundary is used to mark the minimum impact range of the abnormal scene on the traffic on the map;
  • the update module 804 is configured to update the map according to the minimum safety boundary obtained by calculation.
  • the verification module 805 (optional) is used to call/collect information collected by the collection vehicle or crowdsourced vehicle on the road section where the abnormal scene is located, correct the minimum safety margin, and cancel the abnormality mark when it is determined that the abnormality has ended.
  • the processing module 803 is specifically configured to: obtain the vehicle's drivable area of the abnormal scene according to the abnormal scene sensing data, where the vehicle's drivable area refers to the safe passable area of the vehicle determined from the driving perspective; according to the vehicle's drivable area Calculate the minimum safety margin of the abnormal scene.
  • the abnormal scene sensing data includes vehicle-mounted sensor sensing data
  • the processing module 803 is further configured to input the vehicle-mounted sensor sensing data into a pre-trained neural network to obtain the vehicle's travelable area.
  • the processing module 803 is further configured to input various types of vehicle sensor sensing data obtained into corresponding multiple types of pre-trained neural networks to obtain multiple estimates of the vehicle's travelable area; The two estimates are fused, and the fused vehicle's travelable area is calculated.
  • the abnormal scene sensing data includes vehicle sensor sensing data and road monitoring sensing data.
  • the specific method for the acquisition module 801 to obtain road monitoring sensing data is as follows: acquiring the abnormal scene location information contained in the vehicle sensor sensing data; The location information of the scene, the road monitoring camera set near the abnormal scene is determined; the road monitoring sensing data collected by the road monitoring camera set is obtained, and the road monitoring sensing data includes the road monitoring data before the abnormal scene and the abnormality Road monitoring data after the scene occurs.
  • the processing module 803 is further configured to compare the road monitoring data collected by the road monitoring camera set before and after the occurrence of the abnormal scene to obtain the travelable area of the vehicle.
  • the processing module 803 is further configured to calculate the minimum safety margin based on the location information of the abnormal scene and the vehicle's drivable area.
  • the processing module 803 calculates the minimum safety margin of the abnormal scene according to the location information of the abnormal scene and the vehicle's drivable area specifically includes: taking the position information of the abnormal scene as a reference point, and acquiring according to sensor data of on-board sensors The coordinates of the vehicle's travelable area in the vehicle's coordinate system; according to the mapping relationship between the vehicle's coordinate system and the global coordinate system adopted by the map, the coordinates of the vehicle's travelable area are converted to coordinates in the global coordinate system to obtain The minimum safety margin for abnormal scenarios.
  • the obtaining module 801 is specifically configured to: when an abnormal scene is detected, obtain the abnormal scene sensing data.
  • the vehicle-mounted sensor sensing data includes: obstacle information/point cloud information collected by the vehicle-mounted radar, pictures and videos collected by the vehicle-mounted camera, and location information acquired by the vehicle-mounted satellite positioning receiving system.
  • the abnormal scene includes: a traffic accident, road construction, or vehicle failure.
  • this application also provides a map update system.
  • the map update system includes a vehicle-mounted communication device and a cloud server.
  • the vehicle-mounted communication device is used to obtain the abnormal scene sensing data when an abnormal scene is detected, and send the abnormal scene sensing data to the cloud system to trigger the cloud system to perform the update;
  • the cloud server is used to calculate according to the acquired abnormal scene sensing data The minimum safety margin is to update the map.
  • this application also provides another map update system.
  • the map update system includes a vehicle-mounted communication device and a cloud server.
  • the vehicle-mounted communication device is used to obtain abnormal scene sensing data when an abnormal scene is detected; calculate the minimum safety margin based on the obtained abnormal scene sensing data; then send the minimum safety margin to the cloud system to trigger the cloud system to perform an update.
  • the cloud server is used to update the map according to the obtained minimum security boundary.
  • this application also provides a networked vehicle.
  • the networked vehicle includes on-board communication devices and on-board sensors.
  • the vehicle-mounted sensor acquires the abnormal scene sensing data, calculates the minimum safety margin of the abnormal scene according to the abnormal scene sensing data, and sends the minimum safety margin to the cloud system to trigger the cloud system to update the map.
  • this application also provides another connected vehicle.
  • the networked vehicle includes on-board communication devices and on-board sensors.
  • the vehicle-mounted sensor acquires the abnormal scene sensing data, and sends the abnormal scene sensing data to the cloud system to trigger the cloud system to perform an update.
  • modules of the above device is only a division of logical functions, and may be fully or partially integrated into a physical entity during actual implementation, or may be physically separated.
  • modules can all be implemented in the form of software called by processing elements; they can also be implemented in the form of hardware; some modules can be implemented in the form of calling software by processing elements, and some of the modules can be implemented in the form of hardware.
  • each step of the above method or each of the above modules can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.
  • the above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more application specific integrated circuits (ASIC), or one or more microprocessors (digital signal processor, DSP), or, one or more field programmable gate arrays (FPGA), etc.
  • ASIC application specific integrated circuit
  • DSP digital signal processor
  • FPGA field programmable gate arrays
  • the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program codes.
  • these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).
  • SOC system-on-a-chip
  • the computer program product includes one or more computer instructions.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website, computer, server, or data center.
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)).
  • FIG. 10 is a schematic structural diagram 900 of a map updating device provided by an embodiment of the application.
  • the device may include a processor 901, a communication interface 902, a memory 903, and a system bus 904.
  • the memory 903 and the communication interface 902 are connected to the processor 901 through the system bus 904, and communicate with each other.
  • the memory 903 is used to store computer execution instructions
  • the communication interface 902 is used to communicate with other devices
  • the processor 901 executes computer instructions to implement the solutions shown in the foregoing method embodiments.
  • the system bus mentioned in FIG. 10 may be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) bus, etc.
  • PCI peripheral component interconnect standard
  • EISA extended industry standard architecture
  • the system bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
  • the communication interface is used to realize the communication between the database access device and other devices (such as the client, the read-write library and the read-only library).
  • the memory may include random access memory (RAM), and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
  • the above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other Programming logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • CPU central processing unit
  • NP network processor
  • DSP digital signal processor
  • ASIC application-specific integrated circuit
  • FPGA field programmable gate array
  • an embodiment of the present application further provides a storage medium, in which instructions are stored in the storage medium, which when run on a computer, cause the computer to execute the method shown in the foregoing method embodiment.
  • an embodiment of the present application further provides a chip for executing instructions, and the chip is configured to execute the method shown in the foregoing method embodiment.
  • the size of the sequence numbers of the above-mentioned processes does not mean the order of execution.
  • the execution order of the processes should be determined by their functions and internal logic, and should not be implemented in this application.
  • the implementation process of the example constitutes any limitation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Automation & Control Theory (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Atmospheric Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种地图更新方法、装置(800)、设备(900)，可应用于自动驾驶(Automated driving)或者智能驾驶(Intelligent Driving)等领域。其中，地图更新方法包括：在异常场景发生时，获取各种类型的异常场景感应数据，根据异常场景感应数据计算最小安全边界。最后，根据计算获得的最小安全边界，对地图进行更新。地图更新方法在异常场景发生时就能触发地图更新程序，无需等待采集车/众包车，提高了地图刷新的实时性，从而保障了自动驾驶环境的安全性。

Description

一种更新地图的方法、装置和设备 技术领域
本申请涉及智能驾驶技术领域，尤其涉及一种地图更新方法、装置和设备。
背景技术
近年来，智能汽车(Smart/intelligent car)已成为车辆发展的新趋势，越来越多的汽车采用了辅助驾驶系统(Advanced Driving Assistance System)和自动驾驶(Automated Driving)系统。在自动驾驶领域，车辆所使用的地图是最为重要的组件之一。在运行过程中，自动驾驶汽车必须精确地知道自己在路面上的位置，例如与旁边路肩的距离，与车道线的距离等等。所以，自动驾驶车辆所使用的地图的绝对精度一般需要达到分米级，甚至厘米级。除了高精度的坐标之外，地图还需要存储交通场景中的各种交通要素，包括传统地图的道路网数据、车道线以及交通标志，还会含有车道的坡度、曲率、航向、高程、侧倾程度等数据。
目前的自动驾驶方案中，95％以上都依赖于高精地图来获取当前道路信息和环境信息，如果自动驾驶车辆所使用的地图无法及时更新，出现了地图和当前实际道路不匹配的情况，那么对于自动驾驶来说将会是一件非常危险的事情。因此，为了营造一个安全的行车环境，对自动驾驶车辆所依赖的地图进行实时、准确、快速地更新是必不可少的。
发明内容
本申请提供一种地图更新方法、装置和设备，可以提高地图刷新的效率，保障了自动驾驶环境的安全性，同时节约算力，降低刷新成本。
本申请第一方面提供一种地图更新方法，该方法包括：在异常场景发生时，获取异常场景感应数据，根据异常场景感应数据计算最小安全边界。所述最小安全边界指的是在地图上标识出的异常场景对交通的最小影响范围。然后，根据计算获得的最小安全边界，对地图进行更新。上述方法在异常场景发生时，触发地图更新程序，利用异常场景感应数据确定最小安全边界，并根据计算出的最小安全边界更新地图，提高了地图刷新的实时性，从而保障了自动驾驶环境的安全性。
在一种可能的实现方式中，根据异常场景感应数据计算最小安全边界，包括：根据异常场景感应数据获得车辆可行驶区域，然后再根据车辆可行驶区域计算异常场景在地图上的最小安全边界。其中，车辆可行驶区域是指以行车视角确定的车辆可安全通行区域。本申请先计算可行驶区域，再通过可行驶区域计算最小安全边界，提高了计算结果的准确性。
在另一种可能的实现方式中，异常场景感应数据包括车载传感器感应数据。根据异常场景感应数据计算车辆可行驶区域的方法包括：将车载传感器感应数据输入至预训练的神经网络以获取车辆可行驶区域。根据车载传感器感应数据，利用预训练的神经网络模型计算车辆可行驶区域，提高了计算速度以及计算结果的准确度，保障了地图的准确有效性。
在另一种可能的实现方式中，计算车辆可行驶区域的方法包括：将车载传感器获得的多种类型的感应数据分别输入至对应的多种类型的预训练的神经网络以获取车辆可行驶区域的多个估计。然后，将车辆可行驶区域的多个估计进行融合，得到融合后的车辆可行驶区域。本发明将多个传感器数据计算得到可行驶区域进行融合，提高了车辆可行驶区域估算的准确度，保证了地图更新的准确性、有效性。
在另一种可能的实现方式中，异常场景感应数据包括车载传感器感应数据和道路监控感应数据，其中，道路监控感应数据的获取方法包括：根据车载传感器感应数据中的异常场景位置信息确定附近的道路监控摄像头集合，然后，获取这些摄像头集合采集到的异常场景发生前后的道路监控数据。本申请结合真实的场景情况，考虑了道路监控数据。道路监控摄像头位置固定，朝向明确，且监控图像的质量日益完善，因此，利用道路监控数据可以更准确地计算出车辆可行驶区域，充分调用了现有的监控资源，实现资源的合理配置。
在另一种可能的实现方式中，根据异常场景感应数据计算车辆可行驶区域的方法包括：比对异常场景发生前后的道路监控数据，获得车辆可行驶区域。使用了计算机视觉的图像处理方法计算车辆可行驶区域，提供了计算车辆可行驶区域的另一种选择。
在另一种可能的实现方式中，根据车辆可行驶区域计算最小安全边界的方法包括根据异常场景的位置信息和车辆可行驶区域计算最小安全边界。
在另一种可能的实现方式中，根据异常场景的位置信息和车辆可行驶区域计算最小安全边界包括：以异常场景的位置信息作为参考点确定车辆可行驶区域在自车坐标系下的坐标。然后根据自车坐标系和地图采用的全局坐标系之间的映射关系，将车辆可行驶区域的坐标转换为全局坐标系下的坐标，获得所述异常场景的最小安全边界。通过坐标系之间的对应关系，可以直接将车辆可行驶区域映射至地图，方法简洁且准确。
在另一种可能的实现方式中，获取异常场景感应数据的方法包括：在检测到发生异常场景时，车载通信装置触发车载传感器获取所述异常场景感应数据。在异常场景发生时，车载通信装置获取异常场景感应数据之后主动触发地图修正程序，大大提高了系统响应的实时性，保障了自动驾驶车辆乘客的人身安全。此外，这种主动触法的方式明确了更新的地段，极大程度上节约了计算资源。
在另一种可能的实现方式中，车载传感器感应数据包括：车载雷达所采集到的障碍物信息/点云信息、车载相机所采集到的图片和视频、车载卫星定位接收系统获取的位置信息。为了确保车辆的行车安全，一辆车往往配有多种车载传感器。多种类型的传感器数据相结合，可以保障地图更新的准确度。
在另一种可能的实现方式中，异常场景包括：交通事故、道路施工或者车辆故障。异常场景主要是指道路被异常占用的情况，发生交通事故或者故障的车辆以及施工车在上述异常场景发生时，触发车载通信装置获取异常场景感应数据以执行地图的更新。本发明提供的更新方法由普通社会车辆触发地图更新操作，相对比现有技术采用地图采集车或者众包车去采集并触发地图更新的方式，降低了地图刷新的成本。
通过上述描述，本申请提供的地图更新方法使得异常车辆在发生异常场景时在第一时间内获取异常场景感应数据。这样的方法在明确了更新路段，节约算力的同时也提高了地图系统对异常情况的响应速度，提高地图刷新效率。且触发更新的车辆为普通社会车辆，无需地图服务公司派使采集车或者雇佣众包车去采集，降低了地图刷新的成本。此外，引入多种传感器数据计算车辆可行驶区域，提高了地图更新的准确度。
第二方面，本申请提供一种地图的更新装置，所述更新装置，包括：获取模块、处理模块和更新模块；
所述获取模块，用于获取异常场景感应数据；所述处理模块，用于根据异常场景感应数据，计算异常场景的最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；所述更新模块，用于根据计算获得的最小安全边界，对地图进行更新。
在一种可能的实现方式中，处理模块用于：根据所述异常场景感应数据，获得异常场景的车辆可行驶区域，所述车辆可行驶区域是指以行车视角确定的车辆可安全通行区域；然后根据的得到的车辆可行驶区域计算异常场景的最小安全边界。
在另一种可能的实现方式中，异常场景感应数据包括车载传感器感应数据，处理模块具体用于：将车载传感器感应数据输入预训练的神经网络，获得车辆可行驶区域。
在另一种可能的实现方式中，处理模块还用于将车载传感器获得的多种类型的感应数据分别输入至对应的多种类型的预训练的神经网络，获取车辆可行驶区域的多个估计；将这多个估计进行融合，计算获得融合后的车辆可行驶区域。
在另一种可能的实现方式中，异常场景感应数据包括车载传感器感应数据和道路监控感应数据。获取模块获取道路监控感应数据的具体方式如下：根据车载传感器感应数据包含的异常场景位置信息，确定异常场景附近的道路监控摄像头集合。然后再获取道路监控摄像头集合采集的道路监控感应数据，所述道路监控感应数据包括所述异常场景发生前的道路监控数据和所述异常场景发生后的道路监控数据。
在另一种可能的实现方式中，处理模块还用于比对所述道路监控摄像头集合采集到所述异常场景发生前和发生后的道路监控数据，获得所述车辆可行驶区域。
在另一种可能的实现方式中，处理模块还用于根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界。
在另一种可能的实现方式中，处理模块根据异常场景的位置和车辆可行驶区域计算最小安全边界具体包括：以异常场景的位置信息作为参考点，根据车载传感器感应数据获取车辆可行驶区域在自车坐标系下的坐标；然后，根据自车坐标系和地图所采用的全局坐标系之间的映射关系，将车辆可行驶区域的坐标转换为全局坐标系下的坐标，获得异常场景的最小安全边界。
在另一种可能的实现方式中，获取模块具体用于：在检测到发生异常场景时，获取所述异常场景感应数据。
在另一种可能的实现方式中，车载传感器感应数据包括：车载雷达所采集到的障碍物信息/点云信息、车载相机所采集到的图片和视频、车载卫星定位接收系统获取的位置信息。
在另一种可能的实现方式中，异常场景包括：交通事故、道路施工或者车辆故障。
第二方面提供的地图更新装置及可能的实现方式所能够达到的技术效果，与根据第一方面地图更新方法及可能的实现方式所能够达到的技术效果相同，在此不再赘述。
第三方面，本申请提供一种地图更新系统，具体包括车载通信装置、云端服务器。所述车载通信装置，用于在检测到发生异常场景时，获取异常场景感应数据，并将异常场景感应数据发送至云端系统，触发云端系统执行更新。所述云端服务器，用于根据获取到的所述异常场景感应数据，计算最小安全边界，对地图进行更新，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围。
第四方面，本申请提供另外一种地图更新系统，具体包括车载通信装置、云端服务器。所述车载通信装置，用于在检测到发生异常场景时，获取异常场景感应数据；根据获取到的异常场景感应数据，计算最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；并将所述最小安全边界发送至云端系统，触发所述云端系统执行更新。所述云端服务器，用于根据获取到的所述最小安全边界，对地图进行更新。
第五方面，本申请提供一种车载通信盒子，所述车载通信盒子包括处理器、存储器、通 信接口、总线，其中，所述处理器、存储器和通信接口之间通过总线连接并完成相互间的通信，所述存储器中用于存储计算机执行指令，所述车载通信盒子运行时，所述处理器执行所述存储器中的计算机执行指令以利用所述车载通信盒子中的硬件资源执行上述所有方面或所有方面任一种可能实现方式中所述方法中车载通信装置所实现的方法的操作步骤。
第六方面，本申请提供一种联网车辆，其特征在于所述联网车辆包括车载通信装置和车载传感器，当检测异常场景发生时，所述车载传感器获取异常场景感应数据，根据异常场景感应数据计算最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围，并将最小安全边界发送至云端系统，触发云端系统执行地图的更新。
第七方面，本申请提供另一种联网车辆，其特征在于所述联网车辆包括车载通信装置和车载传感器，当检测异常场景发生时，所述车载传感器获取异常场景感应数据，并将异常场景感应数据发送至云端系统，触发云端系统执行更新。
第八方面，本申请提供一种云端服务器，其特征在于，当异常场景发生时，所述云端服务器获取异常场景感应数据，并且根据异常场景感应数据计算最小安全边界对地图进行更新。
第九方面，本申请提供一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，当其在计算机上运行时，使得计算机执行上述第一方面以及可能实现方式中所述的方法。
第十方面，本申请提供一种包含指令的计算机程序产品，当其在计算机上运行时，使得计算机执行上述第一方面以及可能实现方式中所述的方法。
附图说明
图1A是本申请实施例提供的现有地图更新方法的应用场景示意图。
图1B是本申请实施例提供的本发明地图更新方法的应用场景示意图。
图2是本申请实施例提供的地图更新系统的逻辑架构示意图。
图3是本申请实施例提供的地图更新方法的流程示意图。
图4是本申请实施例提供的另一种地图更新方法的流程示意图。
图5是本申请实施例提供的一种计算可行驶区域的流程示意图。
图6是本申请实施例提供的利用神经网络计算出来的可行驶区域示意图。
图7是本申请实施例提供的将多种可行驶区域进行多层融合的示意图。
图8是本申请实施例提供的另一种计算可行驶区域的流程示意图。
图9是本申请实施例提供的一种地图更新装置的结构示意图。
图10是本申请实施例提供的一种地图更新设备的结构示意图。
具体实施方式
在现有技术中，将地图更新产品通过前装搭载上车，基于采集车/众包车上的传感器，计算平台可以实时估计车辆的相对位置信息。再结合深度学习和SLAM(Simultaneous Localization And Mapping)算法感知道路元素(例如红绿灯、路灯等)将实时感知到的结果和当前的高精地图进行对比。当系统发现实时感知结果和地图不匹配时，就上报给云端的地图处理系统。云端系统对车辆上传的数据进行处理，形成地图更新信息，并且下发到各个车辆，完成整个地图更新的闭环。
但是，当前利用采集车/众包车进行更新的方法，难以迅速匹配到更新路段，只能按照事 先定好的路线，或依赖于众包车的随机行驶路线进行更新，无法保证地图的实时有效性。示例性的，如图1A所示，当一辆车101发生了交通事故而导致道路不通时，如果采集车/众包车102(车A)的设定轨迹是沿着A1方向或者A2方向行驶时，将无法及时更新地图。当有其他自动驾驶汽车经过此异常路段时，由于其自身的地图无法得到及时的更新，将会出现自动驾驶危险场景。
此外，采集车/众包车需要沿着预设的路线或者随机路线，将扫描的场景与现有的地图进行实时比对，等发现异常时，再上报云端。这样的更新方法对算力和通讯带宽的要求极高。此外，由于每个采集车/众包车都需要配备加装一套独有的传感器和处理计算单元，所以采集车/众包车的数量十分有限，这对于大规模、大范围的高精地图来说，显然无法达到快速刷新的要求。
基于此，本申请提供一种快速更新地图的方法。图1B是本发明实施例的应用场景。发生交通异常的车辆101自车配有TBOX(Telecommunications Box，通信盒子)、定位模块以及其他的车载传感器，当道路上发生异常交通场景时，立刻通过TBOX上传车载传感器的数据。除此之外，云端系统还会根据定位信息来获取道路摄像头采集的数据。根据收集到的数据，计算出最小安全边界(如图1B中车辆101周围的虚线框所示)以更新高精地图。之后再调用采集车/众包车102在此路段采集的信息，来完成高精地图的完整修正。本发明提供的更新方法，可以在异常场景发生的第一时间主动上报异常场景数据，触发云端系统的紧急修正程序，提高了地图刷新的响应速度。除此之外，多种传感器数据的计算结果可以相互融合，提高了地图更新的准确度，确保了自动驾驶环境的安全性。
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然所描述的实施例是本申请一部分实施例而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本申请保护的范围。
图2为本申请提供的一种地图更新系统的组网架构示意图。如图2所示，该系统包括云端系统201、网络202、道路监控摄像头203、触发异常场景的车辆204。云端系统201通过网络202和道路监控摄像头203或者触发异常场景的车辆204进行通信。其中，云端系统201能够对大量传感器数据进行处理，计算出最小可行驶区域，然后根据定位投影至地图以获取最小安全边界，最后更新地图。网络202是实现将车载传感器的数据以及道路监控摄像头的数据传输至云端系统的媒介。网络202包括以有线和/或无线传输的方式，其中，有线的传输方式包括利用以太、光纤等形式进行数据传输。无线传输方式包括3G、4G(Fourth generation)、或5G(fifth generation)等宽带蜂窝网络传输方式。
道路监控摄像头203位置固定，朝向明确，分布在道路两旁。其具有联网功能，可以上传异常场景固定视角的视频流给云端以计算可行驶区域，从而更新出具有最小安全边界的地图。
触发异常场景的车辆204是指发生交通事故的车辆、故障车辆、施工车辆等导致车道状态异常的车辆。触发异常场景的车辆204包括通信盒子(Telecommunications Box，TBOX)2041、中央网关2042、车身控制器(BodyControl Module，BCM)2043，人机交互控制器2044、车载传感器2045、黑匣子设备2046等，上述各个器件或者设备可以通过控制器局域网络(ControllerArea Network，CAN)或者车内以太网(In-car Ethernet)进行通信，本申请对此不做任何限定。其中，通信盒子2041用于实现触发异常场景的车辆204和云端系统201 之间的通信。车身控制器2043用于控制车门20431、车窗20432等汽车的基础硬件设备。人机交互控制器2044包括车载娱乐(In-VehicleInfotainment，IVI)和/或硬件监视器接口(HardwareMonitor Interface，HMI)等车载娱乐控制系统，负责支持人和车辆的交互，通常用于管理仪表20441、中控显示20442等设备。车载传感器2045则包括雷达20451、相机20452、定位模块20453，车载传感器2045的数据将通过通信盒子2041上传至云端。雷达20451可利用无线电信号来感测车辆的周边环境内的物体，在一些实施例中，雷达20451可以是激光雷达(Lidar)，提供周边环境的点云信息。相机(摄像头或者摄像头组)20452可用于捕捉车辆的周边环境的多个图像。相机20452可以是静态相机或视频相机。作为示例，可以在车辆的前后保险杠、侧视镜、挡风玻璃上分别安装至少一个相机(摄像头或者摄像头组)20452。定位模块20453可以采用全球定位系统(GlobalPositioning System，GPS)，也可以采用北斗系统或者其他系统，能够输出一定精度的全局定位信息。黑匣子设备2046用于在紧急情况下记录智能汽车的车身数据。
可选的，汽车除了通过通信盒子与外界进行通信外，还可以通过其他设备实现。可选的，图2所示的管理系统中也可以不包括中央网关2042，各个控制器、传感器可以直接与通信盒子2041相连。
值得说明的是，图2所示的系统架构仅仅是为了更好的说明本申请所提供的高精地图的更新方法，并不构成对本申请实施例的限定。
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。
图3是本发明的整体解决方案流程，方法包括如下步骤：
S301：异常车辆发送车载传感器感应数据。
在交通事故、车辆故障、道路施工等异常场景发生时，触发该异常场景的车辆主动获取自车的车载传感器感应数据并在第一时间内将其发送至云端系统。车载传感器的感应数据主要包括雷达所采集到的障碍物信息/点云信息、车载相机(摄像头或者摄像头组)采集到的图片和视频、全球定位系统(GPS)或者北斗导航系统反馈出来的位置信息。车载传感器感应数据由触发异常场景的车辆在发生异常的第一时间主动发送给云端系统，这种主动触发地图更新程序的方法提高了地图更新的效率，无需等待采集车/众包车发现异常，确保了自动驾驶汽车行驶的安全性。
可选的，本实施例中的触发异常场景的车辆可以配备高级辅助驾驶系统(ADAS)以实现自动驾驶的功能，也可以不具备自动驾驶的功能。具体的，该车辆可以是发生交通事故的车辆、发生故障的车辆或者是施工车辆。
此外，除了发送车载传感器感应数据给云端系统之外，异常车辆也会推送告警信息给满足预设条件的自动驾驶车辆集合以提示车辆或者乘客注意地图和现实道路的信息匹配，修改路径规划策略，降低车辆通行此异常路段的优先级，改为绕道其他道路。示例性的，满足预设条件的车辆集合可以是距离异常场景位置500米范围内的自动驾驶车辆集合，本发明对此不做具体限定。告警信息可以包含异常场景的位置，异常原因以及异常预计持续的时间。
S302：云端系统接收异常场景感应数据。
异常场景感应数据可以仅仅包括车载传感器的数据，也可以还包括根据车载传感器提供的位置数据而确定的道路监控数据。
云端系统首先接收触发异常场景的车辆上传的车载传感器感应数据。其中，车载传感器感应数据包括了异常场景的位置信息、车载雷达采集的点云信息、车载相机采集的视频数据。
可选的，云端系统在接收车载传感器提供的位置数据之后，可以定位异常场景附近的道路监控摄像头集合，以获取异常场景发生前后的道路监控数据。道路摄像头集合位置固定，朝向也固定，云端系统结合车载传感器提供的位置信息可以快速定位到异常场景附近的道路监控摄像头集合。道路摄像头与异常场景之间的关系满足预设条件。示例性的，预设条件可以是距离小于一定的阈值，且摄像头集合的朝向能拍摄到异常场景。
S303：云端系统计算可行驶区域。
云端系统根据获取到的异常场景感应数据，计算最小可行驶区域。所谓可行驶区域，是从车辆行车视角来考虑的，即道路上车辆可以安全通行的区域。异常场景感应数据可以来源于车载传感器也可以来源于道路监控摄像头集合，根据不同的数据，云端系统可以采用不同的方法计算车辆可行驶区域。
S304：云端系统计算最小安全边界以更新地图。
所谓最小安全边界是相对于地图而言的。示例性的，以异常车辆或者施工区域为对象，在地图上圈出对交通道路影响最小(即车辆能顺利通行)的范围，这范围的边界就是最小安全边界。云端系统根据定位信息，将计算好的可行驶区域投影至地图上，计算出最小安全边界以更新地图。
将车辆的可行驶区域投影至地图的方法具体如下：首先，以异常场景的位置为坐标原点，建立自车坐标系。结合车载传感器提供的数据，确定可行驶区域在自车坐标系下的坐标。其次，需要确定自车坐标系和地图所采用的全局坐标系之间的映射关系。可以根据监控摄像机或者路标等固定的道路元素与异常场景位置之间的距离，以及在地图上所述固定道路元素和异常场景位置之间的距离来确定两种坐标系之间的比例尺或者映射关系。然后，再根据两种坐标系之间的映射关系获取车辆可行驶区域在全局坐标系下的坐标，从而在地图上标出最小安全边界。可选的，当可行驶区域被映射上去时，云端系统还需结合高精地图上的道路信息，进行地图层面上的优化。示例性的，对映射出来的区域进行平滑处理或者规则化处理。
可选的，在更新地图的同时在异常场景还需在所处的位置打上异常标记，提醒过往自动驾驶车辆注意高精地图和现实道路的信息匹配，修改路径规划策略，降低车辆通行此异常路段的优先级，改为绕道其他道路。
需要说明的是，在步骤S302-S304中，车辆可行驶区域以及最小安全边界的计算不一定必须由云端系统执行，当异常车辆自身配有高性能的计算中心时，也可以由车载计算中心执行车辆可行驶区域和最小安全边界的计算，然后车载通信装置将最小安全边界的计算结果上传至云端，由云端执行地图更新。
在另一种可选的实现方式中，结合图4，当云端系统根据最小安全边界更新地图之后，还可以结合采集车/众包车的数据对更新结果进行辅助修正。
S305：收集采集车/众包车的采集数据进行地图修正。
在更新完具有最小安全边界的地图之后，云端系统可以直接收集即将路经异常场景所在路段的采集车/众包车采集的数据，或者调用空闲的采集车/众包车去异常路段采集数据。根据采集车/众包车的数据重新计算可行驶区域，将计算出来的可行驶区域投影至地图算出新的最小安全边界之后，对之前通过传感器数据计算出来的最小安全边界进行修正调整，更新高精地图。
S306：判断异常场景已经结束，取消异常标记。
当异常场景结束时，可以由触发异常场景的车辆主动上报异常结束信息。示例性的，施 工结束之后，施工车主动发送信息提醒云端异常已经结束。可选的，异常结束信息还可以由采集车/众包车上报给云端。可选的，云端在确认异常信息已经结束之后，取消之前地图上的异常标记。
本申请实施例提供的地图更新方法，触发异常场景的车辆主动上报异常数据，启动云端地图紧急修正程序。这样的更新方法效率极大提升，在异常场景出现后1-5分钟就可以完成地图的刷新操作，无需采集车/众包车路过就能更新出具有最小安全边界地图，降低了刷新成本。且与之前的实时扫描比对方法相比，本发明方法极大地节约了算力。而且云端系统除了可以通过车载传感器的数据计算可行驶区域之外，还能根据定位获取到的道路监控视频流数据进行计算。两种方式计算出来的两个最小安全边界也可以进行融合调整，提高了地图更新的准确度。
示例性的，在上述实施例的基础上，结合图5-8，上述步骤S303计算可行驶区域可以通过以下三种方式实现。
实现方式一：云端系统根据车载相机(摄像头/摄像头组)采集的视觉数据计算可行驶区域。结合图5，上述方法的步骤如下所示：
S401：使用标注过可行驶区域的数据集对Multinet网络进行预训练。
所谓MultiNet，顾名思义是由多个网络结构组成，可以同时进行三项任务：分类、检测、语义分割。此网络进行三个任务时，一般共用一个编码器进行特征提取。编码器的结构可以采用VGG16，也可以是别的网络架构，比如GoogleNet、ResNet101。提取特征之后，再分别使用三个解码器进行分类、检测、语义分割这三个任务。使用大量的数据集对Multinet网络进行训练，并采用验证集验证模型的准确度，直至模型的准确度达到预设要求。
S402：将车载相机采集的视觉数据输入至上述Multinet神经网络。
S403：神经网络直接输出带可行驶区域的图片数据
比起传统的计算机图像处理方法，利用预训练好的神经网络模型获取可行驶区域，不仅加快了计算速度，也提高了计算的准确度。图6是使用MultiNet网络识别可行驶区域的示意图，用黑点填充的不规则区域501是由Multinet网络直接计算得出的代表自动驾驶车辆可以安全通行的区，即车辆可行驶区域。
实现方式二：云端系统根据车载雷达采集的点云数据计算可行驶区域。
第二实现方式与第一实现方式类似，只是数据类型由二维的普通视觉数据转换成了三维的点云阵列。点云是一组3D点，携带深度信息，点云也可以具有每个点的RGB值，形成一个彩色的点云。通过点云神经网络(例如PointNet)对场景数据进行分割，可以获得点云版的可行驶区域。
进一步的，当车辆拥有多种传感器时，可以通过多种传感器的数据分别计算出多个可行驶区域结果，再将它们进一步融合来获得置信度最佳的可行驶区域。多种传感器虽然在同一辆车上，但是采集的视角仍然具有不同的偏差。所以在融合之前，需要将这些可行驶区域的视角进行统一，也就是将这些可行驶区域都转换成自车坐标系下的坐标，然后，再进行可行驶区域的融合，得到最佳的车辆可行驶区域。示例性的，融合的方式如图7所示，将转换后的利用车载相机(摄像头或摄像头组)的数据计算得到的可行驶区域图片601、利用激光雷达传感器数据计算的点云可行驶区域图片602、利用毫米波雷达传感器数据计算的点云可行驶区域图片603栅格化，栅格的大小不予限定。601、602、603中每一个栅格在之前计算的过程中都产生了一个置信度，所谓置信度指的就是该栅格确认为车辆可行驶区域的概率。将 同一个栅格对应的三个置信度按比例求和，若结果大于预设阈值，则该栅格为可行驶区域，否则为不可行驶区域。按照上述方法将601、602、603三个可行驶区域进行融合调整，提高车辆可行驶区域计算的准确度。
实现方式三：云端系统根据道路监控感应数据计算可行驶区域。结合图8，上述实现方式的步骤如下所示：
S701：云端系统获取道路监控摄像头所采集的异常场景历史视频流。
云端系统根据车载传感器提供的位置信息，定位异常场景附近的监控摄像头集合。然后，云端系统可以从存储设备的数据库中获取异常场景的视频流和历史视频流，也可以直接从道路监控摄像头终端获取。本申请实施例不对获取当前异常视频流以及历史视频流的方式进行限定，可以根据实际情况确定。
S702：云端系统从历史视频流中提取无车辆行人图片。
无车辆行人的图片是指该图片中道路上没有车辆、行人，也没有施工或者任何障碍物，仅仅包含完整干净的道路。此步骤可以选择合适的神经网络完成，也可以使用传统的图像处理方法完成。
S703：利用传统计算机视觉方法将无车辆行人的图片与异常场景的视频流持续进行比对，以获取异常区域。
S704：反向计算可行驶区域。利用膨胀阈值等方法，对步骤703取得的异常区域进行反向计算，从而得到车辆可行驶区域。
值得说明的是，对于道路监控感应数据，除了采用传统计算机视觉方法比对历史数据和异常数据以外，亦可以直接将异常的道路监控感应数据输入至神经网络以获取车辆可行驶区域。
上述计算出来的可行驶区域都是针对行驶车辆而言的可行驶的安全区域。计算可行驶区域是更新出具有最小安全边界的高精地图的前提。实现方式一和实现方式二的计算数据来源都是车载传感器，而实现方式三的计算数据来源是道路监控摄像头，这两类数据都可以计算出车辆的可行驶区域。根据这两类可行驶区域投影计算出来的最小安全边界也可以互相进行一个辅助调整，提高最小安全边界的准确度。
本申请提供的地图更新方法，根据车载通信装置上传车载传感器的数据或者根据定位到的道路监控数据计算车辆视角的最小可行驶区域，然后投影至地图的相应位置，以刷新出具有最小安全边界的地图。在该技术方案中，触发异常场景的车辆主动启动地图修正程序，极大提升了地图刷新效率，可在异常场景出现后1-5分钟完成高精地图刷新操作。无需等待采集车/众包车，降低了刷新的成本。且更新路段明确，降低云端因为高精地图修正而导致的超大算力要求和处理时间。
在本申请的另一种可能设计中，本申请的车载通信装置可以是具有通信能力的智能车辆。
值得说明的是，对于上述方法实施例，为了简单描述，故将其都表述为一系列的动作组合，但是本领域的技术人员应该知悉，本申请并不受所描述的动作顺序的限制。
本领域的技术人员根据以上描述的内容，能够想到的其他合理步骤组合，也属于本申请的保护范围内。其次，本领域技术人员也应该熟悉，说明书中所描述的实施例均属于优选实施例，所涉及的动作并不一定是本申请所必须的。
上文结合图2至图8，详细描述了根据本申请实施例所提供的地图更新方法，下面将结合图9-10，介绍本申请的地图更新装置，车载通信盒子。对于本申请装置实施例中未披露的 细节，请参照本申请方法实施例。
图9为本申请实施例提供的一种地图更新装置800，该更新装置800可以包括：获取模块801、告警模块802、处理模块803、更新模块804、验证模块805。
获取模块801，用于在异常场景发生时，获取异常场景感应数据；
告警模块802(可选的)，用于向满足预设条件的车辆集合推送告警信息；
处理模块803，用于根据获取到的异常场景感应数据，计算异常场景的最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；
更新模块804，用于根据计算获得的所述最小安全边界，对地图进行更新。
验证模块805(可选的)，用于调用/收集采集车或者众包车在异常场景所在路段采集的信息，对最小安全边界进行修正，且当判断异常已经结束时，取消异常标记。
可选的，处理模块803具体用于：根据异常场景感应数据，获得异常场景的车辆可行驶区域，所述车辆可行驶区域是指以行车视角确定的车辆可安全通行区域；根据车辆可行驶区域计算所述异常场景的最小安全边界。
可选的，异常场景感应数据包括车载传感器感应数据，所述处理模块803还用于将车载传感器感应数据输入预训练的神经网络，获取所述车辆可行驶区域。
可选的，处理模块803还用于将获得的多种类型的车载传感器感应数据分别输入至对应的多种类型的预训练的神经网络，获取车辆可行驶区域的多个估计；然后，将多个估计进行融合，计算获得融合后的车辆可行驶区域。
可选的，异常场景感应数据包括车载传感器感应数据和道路监控感应数据，所述获取模块801获取道路监控感应数据的具体方式如下：获取所述车载传感器感应数据包含的异常场景位置信息；根据异常场景的位置信息，确定所述异常场景附近的道路监控摄像头集合；获取道路监控摄像头集合采集的道路监控感应数据，所述道路监控感应数据包括所述异常场景发生前的道路监控数据和所述异常场景发生后的道路监控数据。
可选的，处理模块803还用于比对所述道路监控摄像头集合采集到所述异常场景发生前和发生后的道路监控数据，获得所述车辆可行驶区域。
可选的，处理模块803还用于根据异常场景的位置信息和车辆可行驶区域计算最小安全边界。
可选的，处理模块803根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界具体包括：以异常场景的位置信息作为参考点，根据车载传感器感应数据获取车辆可行驶区域在自车坐标系下的坐标；根据自车坐标系和所述地图采用的全局坐标系之间的映射关系，将车辆可行驶区域的坐标转换为全局坐标系下的坐标，获得异常场景的最小安全边界。
可选的，获取模块801具体用于：在检测到发生异常场景时，获取所述异常场景感应数据。
可选的，车载传感器感应数据包括：车载雷达所采集到的障碍物信息/点云信息、车载相机所采集到的图片和视频、车载卫星定位接收系统获取的位置信息。
可选的，异常场景包括：交通事故、道路施工或者车辆故障。
作为另一种可能的实施例，本申请还提供一种地图更新系统。该地图更新系统包括车载通信装置和云端服务器。车载通信装置用于在检测到发生异常场景时，获取异常场景感应数据，并将异常场景感应数据发送至云端系统，触发云端系统执行更新；云端服务器用于根据 获取到的异常场景感应数据，计算最小安全边界，对地图进行更新。
作为另一种可能的实施例，本申请还提供另一种地图更新系统。该地图更新系统包括车载通信装置和云端服务器。车载通信装置，用于在检测到发生异常场景时，获取异常场景感应数据；根据获取到的异常场景感应数据，计算最小安全边界；然后将最小安全边界发送至云端系统，触发云端系统执行更新。云端服务器，用于根据获取到的最小安全边界，对地图进行更新。
作为另一种可能的实施例，本申请还提供一种联网车辆。该联网车辆包括车载通信装置和车载传感器。当检测异常场景发生时，车载传感器获取异常场景感应数据，根据异常场景感应数据计算异常场景的最小安全边界，并将最小安全边界发送至云端系统，触发所述云端系统执行地图的更新。
作为另一种可能的实施例，本申请还提供一种另一种联网车辆。该联网车辆包括车载通信装置和车载传感器。当检测到异常场景发生时，车载传感器获取异常场景感应数据，并将异常场景感应数据发送至云端系统，触发云端系统执行更新。
需要说明的是，应理解以上装置的各个模块的划分仅仅是一种逻辑功能的划分，实际实现时可以全部或部分集成到一个物理实体上，也可以物理上分开。且这些模块可以全部以软件通过处理元件调用的形式实现；也可以全部以硬件的形式实现；还可以部分模块通过处理元件调用软件的形式实现，部分模块通过硬件的形式实现。
在实现过程中，上述方法的各步骤或以上各个模块可以通过处理器元件中的硬件的集成逻辑电路或者软件形式的指令完成。例如，以上这些模块可以是被配置成实施以上方法的一个或多个集成电路，例如：一个或多个特定集成电路(application specific integrated circuit，ASIC)，或，一个或多个微处理器(digital signal processor，DSP)，或，一个或者多个现场可编程门阵列(field programmable gate array，FPGA)等。再如，当以上某个模块通过处理元件调度程序代码的形式实现时，该处理元件可以是通用处理器，例如中央处理器(central processing unit，CPU)或其它可以调用程序代码的处理器。再如，这些模块可以集成在一起，以片上系统(system-on-a-chip，SOC)的形式实现。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘solid state disk(SSD))等。
图10为本申请实施例提供的地图更新设备的结构示意图900。如图10所示，该装置可以包括处理器901、通信接口902、存储器903和系统总线904。存储器903和通信接口902通过系统总线904和处理器901连接，并完成相互间的通信。存储器903用于存储计算机执行 指令，通信接口902用于和其他设备进行通信，处理器901执行计算机指令实现上述方法实施例所示的方案。
图10中提到的系统总线可以是外设部件互连标准(peripheral component interconnect，PCI)总线或扩展工业标准结构(extended industry standard architecture，EISA)总线等。所述系统总线可以分为地址总线、数据总线、控制总线等。为便于表示，图中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。通信接口用于实现数据库访问装置与其他设备(例如客户端、读写库和只读库)之间的通信。存储器可能包含随机存取存储器(random access memory，RAM)，也可能还包括非易失性存储器(non-volatile memory)，例如至少一个磁盘存储器。
上述的处理器可以是通用处理器，包括中央处理器CPU、网络处理器(network processor，NP)等；还可以是数字信号处理器DSP、专用集成电路ASIC、现场可编程门阵列FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。
可选的，本申请实施例还提供一种存储介质，所述存储介质中存储有指令，当其在计算机上运行时，使得计算机执行如上述方法实施例所示的方法。
可选的，本申请实施例还提供一种运行指令的芯片，所述芯片用于执行如上述方法实施例所示的方法。
可以理解的是，在本申请的实施例中涉及的各种数字编号仅为描述方便进行的区分，并不用来限制本申请的实施例的范围。
可以理解的是，在本申请的实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请的实施例的实施过程构成任何限定。
最后应说明的是：以上各实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述各实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (26)

  1. 一种地图的更新方法，其特征在于，所述方法包括：
    获取异常场景感应数据；
    根据所述异常场景感应数据，计算所述异常场景的最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；
    根据计算获得的所述最小安全边界，对地图进行更新。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述异常场景感应数据，计算所述异常场景的最小安全边界包括：
    根据所述异常场景感应数据，获得所述异常场景的车辆可行驶区域，所述车辆可行驶区域是指以行车视角确定的车辆可安全通行区域；
    根据所述车辆可行驶区域计算所述异常场景的最小安全边界。
  3. 根据权利要求2所述的方法，其特征在于，所述异常场景感应数据包括车载传感器感应数据，
    所述根据所述异常场景感应数据，获得所述异常场景的车辆可行驶区域包括：
    将所述车载传感器感应数据输入预训练的神经网络，获取所述车辆可行驶区域。
  4. 根据权利要求3所述的方法，其特征在于，所述将所述车载传感器感应数据输入预训练的神经网络，获取所述车辆可行驶区域包括：
    将所述车载传感器获得的多种类型的感应数据分别输入至对应的多种类型的预训练的神经网络，获取车辆可行驶区域的多个估计；
    将所述车辆可行驶区域的多个估计进行融合，计算获得融合后的车辆可行驶区域。
  5. 根据权利要求2所述的方法，其特征在于，所述异常场景感应数据包括车载传感器感应数据和道路监控感应数据，所述道路监控感应数据通过如下方式获取：
    获取所述车载传感器感应数据包含的所述异常场景的位置信息；
    根据所述异常场景的位置信息，确定所述异常场景附近的道路监控摄像头集合；
    获取所述道路监控摄像头集合采集的道路监控感应数据，所述道路监控感应数据包括所述异常场景发生前的道路监控数据和所述异常场景发生后的道路监控数据。
  6. 根据权利要求5所述的方法，其特征在于，所述根据所述异常场景感应数据，计算所述异常场景的车辆可行驶区域包括：
    比对所述道路监控摄像头集合采集到所述异常场景发生前和发生后的道路监控数据，获得所述车辆可行驶区域。
  7. 根据权利要求3或6所述的方法，所述根据所述车辆可行驶区域计算所述异常场景的最小安全边界包括：
    根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界。
  8. 根据权利要求7所述的方法，其特征在于，所述根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界包括：
    以所述异常场景的位置信息作为参考点，根据所述车载传感器感应数据获取所述车辆可行驶区域在自车坐标系下的坐标；
    根据所述自车坐标系和所述地图采用的全局坐标系之间的映射关系，将所述车辆可行驶区域的坐标转换为所述全局坐标系下的坐标，获得所述异常场景的最小安全边界。
  9. 根据权利要求1-8任一所述的方法，其特征在于，所述获取异常场景感应数据包括：
    在检测到发生异常场景时，车载通信装置触发车载传感器获取所述异常场景感应数据。
  10. 根据权利要求3-9任一所述的方法，其特征于，所述车载传感器感应数据包括：
    车载雷达所采集到的障碍物信息/点云信息、车载相机所采集到的图片和视频、车载卫星定位接收系统获取的位置信息。
  11. 根据权利要求1-10任一所述的方法，其特征在于，所述异常场景包括：交通事故、道路施工或者车辆故障。
  12. 一种地图更新装置，其特征在于，包括：获取模块、处理模块和更新模块；
    所述获取模块，用于获取异常场景感应数据；
    所述处理模块，用于根据所述异常场景感应数据，计算所述异常场景的最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；
    所述更新模块，用于根据计算获得的所述最小安全边界，对地图进行更新。
  13. 根据权利要求12所述的装置，其特征在于，所述处理模块具体用于：
    根据所述异常场景感应数据，获得所述异常场景的车辆可行驶区域，所述车辆可行驶区域是指以行车视角确定的车辆可安全通行区域；
    根据所述车辆可行驶区域计算所述异常场景的最小安全边界。
  14. 根据权利要求13所述的装置，其特征在于，所述异常场景感应数据包括车载传感器感应数据，所述处理模块还用于将所述车载传感器感应数据输入预训练的神经网络，获取所述车辆可行驶区域。
  15. 根据权利要求13所述的装置，其特征在于，所述处理模块还用于将所述车载传感器获得的多种类型的感应数据分别输入至对应的多种类型的预训练的神经网络，获取车辆可行驶区域的多个估计；将所述车辆可行驶区域的多个估计进行融合，计算获得融合后的车辆可行驶区域。
  16. 根据权利要求13所述的装置，其特征在于所述异常场景感应数据包括车载传感器感应数据和道路监控感应数据，所述获取模块获取道路监控感应数据的具体方式如下：
    获取所述车载传感器感应数据包含的所述异常场景的位置信息；
    根据所述异常场景的位置信息，确定所述异常场景附近的道路监控摄像头集合；
    获取所述道路监控摄像头集合采集的道路监控感应数据，所述道路监控感应数据包括所述异常场景发生前的道路监控数据和所述异常场景发生后的道路监控数据。
  17. 根据权利要求16所述的装置，其特征在于，所述处理模块还用于比对所述道路监控摄像头集合采集到所述异常场景发生前和发生后的道路监控数据，获得所述车辆可行驶区域。
  18. 根据权利要求14或17所述的装置，其特征在于，所述处理模块还用于根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界。
  19. 根据权利要求18所述的装置，其特征在于，所述处理模块根据所述异常场景的位置信息和所述车辆可行驶区域计算所述异常场景的最小安全边界具体包括：
    以所述异常场景的位置信息作为参考点，根据所述车载传感器感应数据获取所述车辆可行驶区域在自车坐标系下的坐标；
    根据所述自车坐标系和所述地图采用的全局坐标系之间的映射关系，将所述车辆可行驶区域的坐标转换为所述全局坐标系下的坐标，获得所述异常场景的最小安全边界。
  20. 根据权利要求12-19任一所述的装置，其特征在于，所述获取模块具体用于：
    在检测到发生异常场景时，获取所述异常场景感应数据。
  21. 根据权利要求14-20任一所述的装置，其特征在于，所述车载传感器感应数据包括：
    车载雷达所采集到的障碍物信息/点云信息、车载相机所采集到的图片和视频、车载卫星定位接收系统获取的位置信息。
  22. 根据权利要求12-21所述的装置，其特征在于，所述异常场景包括：交通事故、道路施工或者车辆故障。
  23. 一种地图更新系统，其特征在于，所述系统包括车载通信装置、云端服务器；
    所述车载通信装置，用于在检测到发生异常场景时，获取异常场景感应数据，并将所述异常场景感应数据发送至云端系统，触发所述云端系统执行更新；
    所述云端服务器，用于根据获取到的所述异常场景感应数据，计算最小安全边界，对地图进行更新，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围。
  24. 一种地图更新系统，其特征在于，所述系统包括车载通信装置、云端服务器；
    所述车载通信装置，用于在检测到发生异常场景时，获取异常场景感应数据；根据获取到的异常场景感应数据，计算最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围；并将所述最小安全边界发送至云端系统，触发所述云端系统执行更新；
    所述云端服务器，用于根据获取到的所述最小安全边界，对地图进行更新。
  25. 一种联网车辆，其特征在于所述联网车辆包括车载通信装置和车载传感器，当检测异常场景发生时，所述车载传感器获取异常场景感应数据，根据所述异常场景感应数据计算所述异常场景的最小安全边界，所述最小安全边界用来在地图上标识所述异常场景对交通的最小影响范围，并将所述最小安全边界发送至云端系统，触发所述云端系统执行地图的更新。
  26. 一种联网车辆，其特征在于所述联网车辆包括车载通信装置和车载传感器，当检测异常场景发生时，所述车载传感器获取异常场景感应数据，并将所述异常场景感应数据发送至云端系统，触发所述云端系统执行更新。
PCT/CN2020/125615 2020-02-04 2020-10-31 一种更新地图的方法、装置和设备 Ceased WO2021155685A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP20917388.9A EP4089659A4 (en) 2020-02-04 2020-10-31 MAP UPDATE METHOD, DEVICE AND DEVICE
US17/879,252 US20220373353A1 (en) 2020-02-04 2022-08-02 Map Updating Method and Apparatus, and Device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010079832.2A CN113223317B (zh) 2020-02-04 2020-02-04 一种更新地图的方法、装置和设备
CN202010079832.2 2020-02-04

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/879,252 Continuation US20220373353A1 (en) 2020-02-04 2022-08-02 Map Updating Method and Apparatus, and Device

Publications (1)

Publication Number Publication Date
WO2021155685A1 true WO2021155685A1 (zh) 2021-08-12

Family

ID=77085413

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/125615 Ceased WO2021155685A1 (zh) 2020-02-04 2020-10-31 一种更新地图的方法、装置和设备

Country Status (4)

Country Link
US (1) US20220373353A1 (zh)
EP (1) EP4089659A4 (zh)
CN (1) CN113223317B (zh)
WO (1) WO2021155685A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114252087A (zh) * 2021-12-22 2022-03-29 广州小鹏自动驾驶科技有限公司 一种地图数据的处理方法及装置、车辆和存储介质
CN114332384A (zh) * 2021-11-19 2022-04-12 清华大学 一种车载高清地图数据源内容分发方法及装置
CN115113253A (zh) * 2022-06-21 2022-09-27 一汽奔腾轿车有限公司 一种结合5g的高精度车辆定位方法
CN115752446A (zh) * 2022-09-16 2023-03-07 北京四维远见信息技术有限公司 一种车载激光扫描数据同步方法、装置、电子设备及介质

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3882813B1 (en) 2020-03-20 2025-05-07 Aptiv Technologies AG Method for generating a dynamic occupancy grid
EP3888988B1 (en) 2020-03-30 2024-09-04 Aptiv Technologies AG Method and system for determining a usable distance in front of a vehicle
EP3905105A1 (en) 2020-04-27 2021-11-03 Aptiv Technologies Limited Method for determining a collision free space
EP3905106A1 (en) * 2020-04-27 2021-11-03 Aptiv Technologies Limited Method for determining a drivable area
CN114136307B (zh) * 2021-12-07 2024-01-26 上汽大众汽车有限公司 一种车载导航系统地图全自动更新方法
CN114676212B (zh) * 2022-03-17 2025-09-05 北京百度网讯科技有限公司 地图数据的处理方法、装置、电子设备和存储介质
CN114689036B (zh) * 2022-03-29 2025-06-17 深圳海星智驾科技有限公司 地图更新方法、自动驾驶方法、电子设备及存储介质
CN115329024B (zh) * 2022-08-18 2023-09-26 北京百度网讯科技有限公司 一种地图数据更新方法、装置、电子设备及存储介质
CN116259028A (zh) * 2023-05-06 2023-06-13 杭州宏景智驾科技有限公司 用于激光雷达的异常场景检测方法、电子设备和存储介质
CN117009365B (zh) * 2023-08-08 2026-04-10 浙江吉利控股集团有限公司 高精地图更新方法、装置、设备及存储介质
CN119728398B (zh) * 2024-12-20 2026-01-30 国电投河南新能源有限公司 基于物联网与人工智能的新能源场站巡检与故障诊断方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101414408A (zh) * 2008-10-03 2009-04-22 邓湘 触发事件区域代码化智能交通系统
US20130226442A1 (en) * 2012-02-29 2013-08-29 James D. Lynch Three-dimensional traffic flow presentation
US20150179066A1 (en) * 2013-12-24 2015-06-25 Tomer RIDER Road hazard communication
CN106164997A (zh) * 2014-02-27 2016-11-23 通腾运输公司 用于将危险与数字地图的地区相关联的方法
CN108021625A (zh) * 2017-11-21 2018-05-11 深圳广联赛讯有限公司 车辆异常聚集地监控方法及系统、计算机可读存储介质
CN108088455A (zh) * 2017-12-14 2018-05-29 山东中图软件技术有限公司 一种导航方法
CN109808709A (zh) * 2019-01-15 2019-05-28 北京百度网讯科技有限公司 车辆行驶保障方法、装置、设备及可读存储介质
CN110047270A (zh) * 2019-04-09 2019-07-23 南京锦和佳鑫信息科技有限公司 自动驾驶专用车道上应急管理和道路救援的方法

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103500503B (zh) * 2013-09-17 2016-09-07 北京中广睛彩导航科技有限公司 一种基于众包模式的精准路况分析方法及系统
CN105139657B (zh) * 2015-10-21 2017-12-12 招商局重庆交通科研设计院有限公司 一种基于v2i的道路边界与事故黑点的提取方法及系统
CN105644567A (zh) * 2015-12-29 2016-06-08 大陆汽车投资（上海）有限公司 基于adas的驾驶辅助系统
US20170327035A1 (en) * 2016-05-10 2017-11-16 Ford Global Technologies, Llc Methods and systems for beyond-the-horizon threat indication for vehicles
US10522038B2 (en) * 2018-04-19 2019-12-31 Micron Technology, Inc. Systems and methods for automatically warning nearby vehicles of potential hazards
CN108646752B (zh) * 2018-06-22 2021-12-28 奇瑞汽车股份有限公司 自动驾驶系统的控制方法及装置
CN109461321A (zh) * 2018-12-26 2019-03-12 爱驰汽车有限公司 自动驾驶电子围栏更新方法、系统、设备及存储介质
CN109766405B (zh) * 2019-03-06 2022-07-12 路特迩科技(杭州)有限公司 基于电子地图的交通和旅行信息服务系统及方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101414408A (zh) * 2008-10-03 2009-04-22 邓湘 触发事件区域代码化智能交通系统
US20130226442A1 (en) * 2012-02-29 2013-08-29 James D. Lynch Three-dimensional traffic flow presentation
US20150179066A1 (en) * 2013-12-24 2015-06-25 Tomer RIDER Road hazard communication
CN106164997A (zh) * 2014-02-27 2016-11-23 通腾运输公司 用于将危险与数字地图的地区相关联的方法
CN108021625A (zh) * 2017-11-21 2018-05-11 深圳广联赛讯有限公司 车辆异常聚集地监控方法及系统、计算机可读存储介质
CN108088455A (zh) * 2017-12-14 2018-05-29 山东中图软件技术有限公司 一种导航方法
CN109808709A (zh) * 2019-01-15 2019-05-28 北京百度网讯科技有限公司 车辆行驶保障方法、装置、设备及可读存储介质
CN110047270A (zh) * 2019-04-09 2019-07-23 南京锦和佳鑫信息科技有限公司 自动驾驶专用车道上应急管理和道路救援的方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4089659A4 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114332384A (zh) * 2021-11-19 2022-04-12 清华大学 一种车载高清地图数据源内容分发方法及装置
CN114252087A (zh) * 2021-12-22 2022-03-29 广州小鹏自动驾驶科技有限公司 一种地图数据的处理方法及装置、车辆和存储介质
CN114252087B (zh) * 2021-12-22 2022-07-01 广州小鹏自动驾驶科技有限公司 一种地图数据的处理方法及装置、车辆和存储介质
CN115113253A (zh) * 2022-06-21 2022-09-27 一汽奔腾轿车有限公司 一种结合5g的高精度车辆定位方法
CN115752446A (zh) * 2022-09-16 2023-03-07 北京四维远见信息技术有限公司 一种车载激光扫描数据同步方法、装置、电子设备及介质

Also Published As

Publication number Publication date
CN113223317B (zh) 2022-06-10
EP4089659A1 (en) 2022-11-16
CN113223317A (zh) 2021-08-06
US20220373353A1 (en) 2022-11-24
EP4089659A4 (en) 2023-07-12

Similar Documents

Publication Publication Date Title
CN113223317B (zh) 一种更新地图的方法、装置和设备
CN109920246B (zh) 一种基于v2x通信与双目视觉的协同局部路径规划方法
CN112712717B (zh) 一种信息融合的方法、装置和设备
US11520331B2 (en) Methods and apparatus to update autonomous vehicle perspectives
EP4152204A1 (en) Lane line detection method, and related apparatus
US11887324B2 (en) Cross-modality active learning for object detection
CN110920611B (zh) 基于邻车的车辆控制方法和装置
US20260022942A1 (en) Systems and methods for deriving path-prior data using collected trajectories
WO2022001618A1 (zh) 一种车辆的车道保持控制方法、装置及系统
CN105793669B (zh) 车辆位置推定系统、装置、方法以及照相机装置
US11403947B2 (en) Systems and methods for identifying available parking spaces using connected vehicles
CN112149550A (zh) 一种基于多传感器融合的自动驾驶车辆3d目标检测方法
CN110942038B (zh) 基于视觉的交通场景识别方法、装置、介质及电子设备
WO2019077999A1 (ja) 撮像装置、画像処理装置、及び、画像処理方法
CN115691213B (zh) 预警方法、装置、设备和可读存储介质
CN114216469B (zh) 一种对高精地图进行更新的方法、智慧基站及存储介质
US20240328822A1 (en) Operational design domain management for vehicles having automated driving systems
US12123734B2 (en) Automatic annotation of drivable road segments
DE102023109040A1 (de) Vereinheitlichtes framework und werkzeuge zur kommentierung von fahrspurbegrenzungen
CN111709354B (zh) 识别目标区域的方法、装置、电子设备和路侧设备
CN112305499B (zh) 一种根据光源进行定位的方法及装置
CN119600563A (zh) 用于收集数据以用于对象检测模型的后续训练的方法
CN119152456A (zh) 车辆周围的环境感知方法、装置、设备及存储介质
US12187316B2 (en) Camera calibration for underexposed cameras using traffic signal targets
Alrousan et al. Multi-sensor fusion in slow lanes for lane keep assist system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20917388

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020917388

Country of ref document: EP

Effective date: 20220812

NENP Non-entry into the national phase

Ref country code: DE