METHOD AND APPARATUS FOR ESTIMATING POSITIONING ACCURACY, METHOD AND APPARATUS FOR TRAINING ACCURACY ESTIMATION MODEL, ELECTRONIC DEVICE, AND COMPUTER PROGRAM

A machine learning-based method for evaluating positioning accuracy errors in high-precision systems addresses the lack of effective evaluation methods, improving vehicle safety and control through accurate error estimation.

JP2024543959AActive Publication Date: 2024-11-26TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
JP2024532403
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2022-07-15
Filing Date
2023-04-27
Publication Date
2024-11-26
Estimated Expiration
2043-04-27

AI Technical Summary

Technical Problem

Current technologies lack effective means to evaluate accuracy errors in high-precision positioning, which is crucial for vehicle control, collision avoidance, and route planning in intelligent transportation systems.

Method used

A method and device for estimating positioning accuracy using a machine learning model trained on a training sample set that includes vehicle travel information and intermediate variables, allowing for the evaluation of positioning accuracy errors by inputting first position information and intermediate variables into a pre-trained accuracy estimation model.

Benefits of technology

The method effectively evaluates accuracy errors in high-precision positioning, improving accuracy estimation by predicting actual position information and estimating accuracy errors relative to the actual position, enhancing vehicle safety and control systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024543959000001_ABST
    Figure 2024543959000001_ABST
Patent Text Reader

Abstract

The embodiments of the present invention provide a method, device, electronic device, and storage medium for estimating positioning accuracy, and are applied to fields such as maps, navigation, autonomous driving, V2X (Vehicle-to-everything), ITS, and cloud computing. The method for estimating positioning accuracy includes the steps of acquiring first driving information of a vehicle from a sensor, acquiring first position information of the vehicle and a first intermediate variable used in a process of determining the first position information based on the first driving information, determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, and inputting the first position information and the first intermediate variable into the target accuracy estimation model to acquire second position information of the vehicle and an accuracy error of the first position information relative to the second position information. The embodiments of the present invention can effectively evaluate the accuracy error of high-precision positioning.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] This application claims priority to a Chinese patent application filed on July 15, 2022, bearing application number 202210837512.8 and entitled "Method, apparatus, electronic device and storage medium for estimating positioning accuracy," the entire contents of which are incorporated herein by reference.

[0002] The present invention relates to the field of Intelligent Transport Systems (ITS), and more particularly to a method, device, electronic device and storage medium for estimating positioning accuracy. [Background technology]

[0003] Intelligence is one of the main trends in the current development of automobiles. The process of automobile intelligence depends on the maturity of vehicle sensors, algorithm software and decision-making platforms. The combination of high-precision positioning and high-precision maps provides vehicles with accurate absolute position information, which, combined with the relative position information of sensors, improves the safety of smart driving. As automobiles become more intelligent, the importance of high-precision positioning becomes increasingly prominent.

[0004] Accuracy error evaluation is one of the important parts of high-precision positioning. The accuracy error of one positioning, i.e., the magnitude of the error that may be included in the current positioning, is very useful for subsequent vehicle control, collision avoidance, smart vehicle speed control, route planning and behavior decision-making. However, currently there is no means to effectively evaluate the accuracy error of high-precision positioning. Summary of the Invention

[0005] The embodiments of the present invention provide a method, device, electronic device, and storage medium for estimating positioning accuracy, which are capable of effectively evaluating the accuracy error of high-precision positioning.

[0006] In a first aspect of an embodiment of the present invention, a method for estimating positioning accuracy is provided, comprising the steps of: acquiring first driving information of a vehicle from a sensor; acquiring first position information of the vehicle and first intermediate variables used in the process of determining the first position information based on the first driving information; determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variables, wherein the accuracy estimation model is obtained by training a machine learning model based on a training sample set, and training samples in the training sample set include position information of a sample vehicle and intermediate variables used in the process of determining the position information; and inputting the first position information and the first intermediate variables into the target accuracy estimation model, and acquiring second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

[0007] In a second aspect of an embodiment of the present invention, a method for training an accuracy estimation model is provided, the method including the steps of acquiring second driving information of a sample vehicle at a first time from a sensor, acquiring third position information of the sample vehicle and second intermediate variables used in the process of determining the third position information based on the second driving information, acquiring fourth position information of the sample vehicle at the first time from a positioning device, determining a training sample set including the third position information, the fourth position information and the second intermediate variables, and training the accuracy estimation model based on the training sample set.

[0008] In a third aspect of an embodiment of the present invention, a positioning accuracy estimation device is provided, comprising: an acquisition unit that acquires first driving information of a vehicle from a sensor; a processing unit that acquires first position information of the vehicle and first intermediate variables used in the process of determining the first position information based on the first driving information; and a determination unit that determines a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variables, wherein the accuracy estimation model is obtained by training a machine learning model based on a training sample set, and training samples in the training sample set include position information of a sample vehicle and intermediate variables used in the process of determining the position information, and the target accuracy estimation model receives the first position information and the first intermediate variables and acquires second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

[0009] In a fourth aspect of an embodiment of the present invention, a training device for an accuracy estimation model is provided, the device including: a first acquisition unit that acquires second driving information of a sample vehicle at a first time from a sensor; a processing unit that acquires third position information of the sample vehicle and second intermediate variables used in the process of determining the third position information based on the second driving information; a second acquisition unit that acquires fourth position information of the sample vehicle at the first time from a positioning device; a determination unit that determines a training sample set including the third position information, the fourth position information and the second intermediate variables; and a training unit that trains the accuracy estimation model based on the training sample set.

[0010] In a fifth aspect of an embodiment of the present invention, there is provided an electronic device comprising a processor capable of executing computer instructions and a memory having computer instructions stored therein, the computer instructions being configured to be loaded by the processor to perform a method according to the first or second aspect above.

[0011] In a sixth aspect of an embodiment of the present invention, there is provided a computer readable storage medium having stored thereon computer instructions which, when loaded and executed by a processor of a computing device, cause the computing device to perform a method according to the first or second aspect above.

[0012] In a seventh aspect of an embodiment of the present invention, there is provided a computer program product or computer program comprising computer instructions, the computer instructions being read and executed by a processor of a computing device from a computer readable storage medium to cause the computing device to perform a method according to the first or second aspect above.

[0013] According to the above technical means, an embodiment of the present invention can effectively evaluate the accuracy error of high-precision positioning by obtaining first position information of a vehicle and a first intermediate variable used in the process of determining the first position information based on the first driving information, determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, inputting the first position information and the first intermediate variable into the target accuracy estimation model, and obtaining second position information of a vehicle and an accuracy error of the first position information relative to the second position information. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a schematic diagram of the architecture of a system according to an embodiment of the present invention. [Diagram 2] 1 is a schematic flowchart of a method for training a model according to an embodiment of the present invention. [Diagram 3] FIG. 1 is a schematic diagram of a network architecture according to an embodiment of the present invention. [Figure 4] 1 is a schematic flowchart of a method for obtaining training sample data according to an embodiment of the present invention; [Diagram 5]2 is a schematic flowchart of a method for estimating positioning accuracy according to an embodiment of the present invention. [Figure 6] FIG. 2 is a diagram showing a specific example of a method for estimating positioning accuracy according to an embodiment of the present invention. [Figure 7] 1 is a schematic block diagram of a positioning accuracy estimation device according to an embodiment of the present invention; [Figure 8] FIG. 1 is a schematic block diagram of a model training device according to an embodiment of the present invention. [Figure 9] 1 is a schematic block diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] The following clearly and completely describes the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and are not all of the embodiments. All other embodiments that can be obtained by those skilled in the art based on the embodiments of the present invention without creative work are within the scope of the present invention.

[0016] In the embodiment of the present invention, "B corresponding to A" means that B is related to A. In one aspect, B may be determined based on A. In addition, determining B based on A does not mean determining B simply based on A, and B may be determined based on A and / or other information.

[0017] In the description of this application, unless otherwise specified, "at least one" means one or more, and "multiple" means two or more. In addition, "and / or" describes a related relationship of related objects, and means that there can be three types of relationship, for example, A and / or B may mean that only A exists, that A and B exist simultaneously, or that only B exists, where A and B may be singular or plural. The symbol " / " usually means an "or" relationship of the related objects before and after. "At least one of the following" or similar expressions means any combination of these items, including any combination of singular (individual) or plural (individual). For example, at least one of a, b, or c may mean a, b, c, ab, ac, bc, or abc, where a, b, and c may be singular or plural.

[0018] In addition, descriptions such as "first" and "second" in the embodiments of the present invention are merely intended to distinguish the objects of explanation, and do not distinguish between sequential orders, and do not indicate any particular limitations on the number of devices in the embodiments of the present invention, and do not limit the embodiments of the present invention.

[0019] It should be understood that any particular feature, structure or characteristic associated with any embodiment herein is included in at least one embodiment of the present invention, and that the particular feature, structure or characteristic described herein may be combined in any suitable manner in one or more embodiments.

[0020] Additionally, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server comprising a series of steps or units is not necessarily limited to the steps or units explicitly listed, but may include those not explicitly listed or other steps or units inherent to those processes, methods, products, or apparatus.

[0021] The embodiments of the present invention are applied in the field of artificial intelligence.

[0022] Here, artificial intelligence (AI) refers to theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, evolve, and extend human intelligence, sense the environment, acquire knowledge, and use the knowledge to obtain optimal results. In other words, artificial intelligence is a comprehensive technology of computer science that seeks to understand the substance of intelligence and create new intelligent machines that respond in a manner similar to human intelligence. Artificial intelligence is the study of the design principles and realization methods of various intelligent machines, and the endowment of machines with the functions of sensing, reasoning, and decision-making.

[0023] Artificial intelligence technology is a comprehensive field that covers a wide range of fields, including both hardware and software technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interactive systems, mechatronics, etc. The software technologies of artificial intelligence mainly include computer vision technology, voice processing technology, natural language processing technology, and machine learning / deep learning.

[0024] With the research and progress of artificial intelligence technology, artificial intelligence technology is being researched and applied in many fields, such as smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, autonomous driving, drones, robots, smart medical care, smart customer service, etc. With the development of technology, it is expected that artificial intelligence technology will be applied in more and more fields and will play an increasingly important role.

[0025] The embodiment of the present invention relates to an automatic driving technology in artificial intelligence technology. The automatic driving technology is a technology in which a computer can automatically and safely operate a vehicle without any human voluntary operation through the cooperation of artificial intelligence, computer vision, radar, monitoring equipment and global positioning system. The automatic driving technology usually includes technologies such as high-precision maps, environment sensing, action decision, route planning, motion control, etc. The automatic driving technology has been widely applied. Specifically, the technical means according to the embodiment of the present invention relates to a technology for evaluating the accuracy error of positioning, and can be specifically applied to the evaluation of the positioning accuracy of automatic driving.

[0026] An embodiment of the present invention may also relate to Machine Learning (ML) in artificial intelligence technology. ML is a multidisciplinary field that includes probability theory, statistics, approximation theory, convex analysis, algorithmic complexity theory, and other fields. It is a specialized study of how computers simulate or realize human learning behavior, acquire new knowledge or skills, and reorganize existing knowledge structures to constantly improve their own performance. Machine learning is at the heart of artificial intelligence and is a fundamental method for endowing computers with intelligence, and its applications span various fields of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, trust networks, reinforcement learning, transition learning, inductive learning, and teacher learning.

[0027] 1 is a schematic diagram of a system architecture according to an embodiment of the present invention. As shown in FIG. 1, the system architecture may include a user device 101, a data collection device 102, a training device 103, an execution device 104, a database 105, and a content library 106.

[0028] Here, the data collection device 102 reads out the training data from the content library 106 and stores the read out training data in the database 105. The training data according to the embodiment of the present invention includes location information #1, location information #2, and intermediate variables used in the process of determining the location information #1. Here, the location information #1 is determined based on the driving information of the sample vehicle acquired by a sensor, and the location information #2 is the location information of the sample vehicle collected by a positioning device.

[0029] The training device 103 trains the machine learning model based on the training data held in the database 105, so that the machine learning model after training can effectively evaluate the positioning accuracy. The machine learning model obtained by the training device 103 can be applied to different systems or devices.

[0030] 1, the execution device 104 includes an I / O interface 107 for data interaction with an external device. For example, vehicle travel information collected by a sensor and sent by the user device 101 is received through the I / O interface. A calculation module 109 in the execution device 104 obtains the vehicle positioning accuracy based on the travel information through a trained machine model. The machine learning model may send a corresponding result to the user device 101 through the I / O interface.

[0031] Here, the user device 101 may include a smart car, an in-vehicle terminal, a mobile phone, a tablet computer, a notebook computer, a palmtop computer, a mobile internet device (MID), or other terminal devices.

[0032] The execution device 104 may be a server.

[0033] As an example, the server may be a computing device such as a rack server, a blade server, a tower server, or a cabinet server, etc. The server may be a standalone test server or a test server cluster consisting of multiple test servers.

[0034] The server may be one or more servers, where there are at least two servers for providing different services and / or at least two servers for providing the same service, such as in a load balancing manner, but embodiments of the present invention are not limited thereto.

[0035] Here, the server may be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. The server may be a node of a blockchain.

[0036] In this embodiment, the execution device 104 is connected to the user device 101 through a network. The network may be a wireless or wired network, such as a corporate Intranet, the Internet, a Global System of Mobile communication (GSM), a Wideband Code Division Multiple Access (WCDMA), a 4G network, a 5G network, Bluetooth, Wi-Fi, a telephone network, etc.

[0037] 1 is merely a schematic diagram of the architecture of a system according to an embodiment of the present invention, and is not limited to the positional relationships between the illustrated apparatuses, devices, modules, etc. In some aspects, the data collection apparatus 102 may be the same apparatus as the user apparatus 101, the training apparatus 103, and the execution apparatus 104. The database 105 may be distributed on one server or on multiple servers, and the content library 106 may be distributed on one server or on multiple servers.

[0038] The following describes terms relevant to embodiments of the present invention.

[0039] High-precision positioning: Usually means positioning with an accuracy of decimeter or centimeter level or higher, which can provide vehicles with relatively high-precision positioning results and is one of the essential core technologies in safe driving such as autonomous driving and remote driving. High-precision positioning plays an important role in accurate lateral / longitudinal positioning of vehicles, obstacle detection and collision avoidance, smart vehicle speed control, route planning and behavior decision-making, etc.

[0040] High-precision positioning allows the vehicle to accurately determine its own absolute position through a high-precision absolute reference system. The high-precision absolute reference system may be, for example, a high-precision map. The map layer of the high-precision map includes many road attribute elements with centimeter-level accuracy, including but not limited to information on road edges, lane edges, centerlines, etc. The vehicle can perform accurate navigation based on the information of the high-precision map while traveling.

[0041] Multi-source fusion positioning: A technology that fuses multiple positioning technologies based on information fusion policies, and can fuse related positioning methods such as satellite positioning, wireless communication signal positioning, and sensor positioning, thereby obtaining better fusion positioning results than single positioning methods. Multi-source fusion positioning can achieve high-precision positioning.

[0042] Here, satellite positioning is a technology that uses satellites (e.g., GNSS) to perform positioning. Wireless communication signal positioning is a technology that uses wireless communication signals (e.g., WiFi signals, Ultra Wideband (UWB) signals) to perform positioning. Sensor positioning is a technology that uses information collected by sensors (e.g., visual sensors, vehicle speed sensors, Inertial Measurement Unit (IMU) sensors) to perform positioning.

[0043] Visual positioning: A technology that uses information collected by visual sensors to perform positioning. Visual positioning may use visual sensors to recognize road attribute elements in a high-precision map positioning layer and estimate vehicle position information using visual algorithms. Visual positioning has significant cost advantages because it can highly reuse high-precision maps and sensors such as cameras and does not require additional hardware.

[0044] Inertial positioning: A technology that uses information collected by an IMU sensor to perform positioning. Inertial positioning uses an IMU sensor to measure the angular velocity and acceleration information of a vehicle, and automatically estimates the instantaneous speed and position information of the carrier using Newton's laws of motion. It does not rely on external information, does not radiate energy to the outside world, is not disturbed, and has good concealment properties.

[0045] Global Navigation Satellite System (GNSS): refers generally to a satellite navigation system, including the global, regional and augmented satellite navigation systems, such as the Global Positioning System (GPS), Glonass, the European Galileo, the Beidou satellite navigation system, and associated augmentation systems, such as the Wide Area Augmentation System (WAAS), the European Geostationary Navigation System (EGNOS) and the Multi-Function Transport Satellite Augmentation System (MSAS), and may include other satellite navigation systems under construction or to be constructed in the future. The international GNSS system is a complex combination of a multi-system, multi-level and multi-mode system.

[0046] Typically, typical satellite positioning achieves positioning by simultaneously receiving multiple satellite signals. However, certain fluctuations occur when the satellite signals pass through the ionosphere and troposphere, resulting in errors and resulting in a positioning accuracy of only meters. A ground augmented base station can improve positioning accuracy by calculating satellite positioning errors using real-time kinematic (RTK) carrier phase difference technology and further correcting the position. For example, the accuracy of satellite positioning can be increased from the meter level to the centimeter level.

[0047] Satellite positioning is also prone to weak or lost signals when there are ground obstructions, in which case the vehicle can use inertial or visual positioning, or other positioning capabilities to ensure the navigation system continues to work.

[0048] RTK carrier phase differential technology: A differential method that processes the carrier phase observation measurements of two measurement stations in real time, and calculates the coordinates by obtaining the difference by transmitting the carrier phase collected by the reference station to the user receiver. RTK carrier phase differential technology uses the carrier phase dynamic real-time differential method to achieve centimeter-level positioning accuracy in real time outdoors, providing new measurement principles and methods for construction sampling, topographical measurement, and various control measurements, and improving work efficiency.

[0049] Accuracy error evaluation is one of the important parts of high-precision positioning. The accuracy error of one positioning, i.e., the magnitude of the error that may be included in the current positioning, is very useful for subsequent vehicle control, collision avoidance, smart vehicle speed control, route planning and behavior decision-making. However, currently there is no means to effectively evaluate the accuracy error of high-precision positioning.

[0050] In consideration of the above problems, embodiments of the present invention provide a method, device, electronic device, and storage medium for estimating positioning accuracy that can effectively evaluate the accuracy error of high-precision positioning.

[0051] Specifically, an embodiment of the present invention may acquire first driving information of a vehicle from a sensor, acquire first position information of the vehicle and a first intermediate variable used in the process of determining the first position information based on the first driving information, determine a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, input the first position information and the first intermediate variable into the target accuracy estimation model, and acquire second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

[0052] Here, the accuracy estimation model is obtained by training a machine learning model based on a training sample set, and the training sample in the training sample set includes location information of a sample vehicle and intermediate variables used in the process of determining the location information. As an example, the training sample set may include third location information, fourth location information, and a second intermediate variable used in the process of determining the third location information. The third location information is determined based on second driving information of the sample vehicle collected by a sensor at a first time, and the fourth location information includes location information of the sample vehicle collected by a positioning device at the first time.

[0053] Therefore, an embodiment of the present invention can effectively evaluate the accuracy error of high-precision positioning by obtaining first position information of a vehicle and a first intermediate variable used in the process of determining the first position information based on the first driving information, determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, inputting the first position information and the first intermediate variable into the target accuracy estimation model, and obtaining second position information of a vehicle and an accuracy error of the first position information relative to the second position information.

[0054] In an embodiment of the present invention, the accuracy estimation model can realize prediction of actual position information of the vehicle collected by a positioning device and estimation of accuracy error of the positioning information for the actual position information based on vehicle positioning information determined based on vehicle driving information collected by a sensor and intermediate variables used in the process of determining the positioning information.

[0055] The method for estimating the positioning accuracy of the embodiment of the present invention may be divided into two stages, namely, an offline model construction stage and an online estimation stage. The offline model construction stage may be trained to obtain an accuracy estimation model based on collected data of a sample vehicle. The online estimation stage may collect vehicle driving information in real time, and estimate the vehicle positioning accuracy based on the driving information and the accuracy estimation model constructed in the offline model construction stage.

[0056] Each step will be described in detail below with reference to the drawings.

[0057] First, the offline model construction stage will be described.

[0058] 2 is a schematic flow chart of a method 200 for training a model according to an embodiment of the present invention. The method 200 may be executed by any electronic device having data processing capability. For example, the electronic device may be implemented as a server or a terminal device, and for example, the electronic device may be implemented as the training device 103 in FIG. 1, but the present invention is not limited thereto.

[0059] 3 is a schematic diagram of a network architecture according to an embodiment of the present invention. The network architecture includes a vehicle information collection module 301, a real-time positioning module 302, an information statistics module 303, a true value collection module 304, a model construction module 305, an information judgment module 306, and an accuracy estimation model 307. Here, the vehicle information collection module 301, the real-time positioning module 302, the information statistics module 303, the true value collection module 304, and the model construction module 305 may be used for model training in the offline model construction stage. The model training method 200 will be described below with reference to FIG. 3.

[0060] As shown in FIG. 2, the method 200 includes steps 210-250.

[0061] Step 210: Obtain second driving information of the sample vehicle at a first time from a sensor.

[0062] As an example, the vehicle information collection module 301 in Fig. 3 may obtain second driving information of the sample vehicle collected by a sensor at a first time. The first time may be, but is not limited to, the current and previous periods. After obtaining the second driving information, the vehicle information collection module 301 may send the second driving information to the real-time positioning module 302. Preferably, the second driving information may be sent to the information statistics module 303.

[0063] In some aspects, the second driving information includes sample vehicle driving information collected by at least one of an inertial measurement unit (IMU) sensor, a vehicle speed sensor, and a visual sensor.

[0064] As an example, the sensor may be various sensors for collecting vehicle driving information installed on a sample vehicle, such as a GNSS sensor, an IMU sensor, a vehicle speed sensor, a visual sensor, etc., or may be other sensor equipment.

[0065] The driving information of the sample vehicle collected by the GNSS sensor may include, for example, the longitude, latitude, DOP value, accuracy value, whether the current positioning is valid, whether the current positioning is fixed, etc. of the currently positioned sample vehicle.

[0066] Here, when there is no obstruction or the obstruction is weak (e.g., when driving normally on a road), the GNSS sensor can receive the GNSS signal normally, so the current positioning is valid, and in this case, the GNSS sensor may also be referred to as being valid. When there is a relatively large amount of obstruction (e.g., a tunnel, an overpass, or a high mountain road area), the GNSS sensor cannot receive the GNSS signal normally, so the current positioning is invalid, and in this case, the GNSS sensor may also be referred to as being invalid or ineffective.

[0067] The solution of the GNSS sensor obtained by adopting the RTK carrier phase difference technique may be referred to as a fixed solution. When the GNSS signal of the GNSS sensor is blocked (for example, in a high mountain road area), when the RTK carrier phase difference technique cannot be adopted, other types of solutions, such as floating-point solutions, may be obtained. Whether the current positioning is fixed or not means whether the current positioning adopts the RTK carrier phase difference technique to obtain a fixed solution. Here, the accuracy of the fixed solution is the highest, reaching the centimeter level.

[0068] The driving information of the sample vehicle collected by the IMU sensor may include, for example, acceleration values ​​in the x-, y-, and z-axis directions of an accelerometer, and angular velocities in the x-, y-, and z-axis directions of a gyro.

[0069] The travel information of the sample vehicle collected by the vehicle speed sensor may include, for example, the speed direction and size of the sample vehicle.

[0070] The driving information of the sample vehicle collected by the visual sensor may include, for example, the equation of the lane in which the sample vehicle is located, the line type of the lane (e.g., solid, dashed, white or yellow line, etc.), and the coordinates of obstacles.

[0071] Step 220: Obtain third location information of the sample vehicle based on the second driving information, and a second intermediate variable used in the process of determining the third location information.

[0072] 3 may obtain the third location information of the sample vehicle and the second intermediate variables used in the process of determining the third location information based on the above-mentioned second driving information. After obtaining the third location information and the second intermediate variables, the real-time positioning module 302 may transmit the third location information and the second intermediate variables to the model construction module 305.

[0073] In some aspects, the real-time positioning module 302 may acquire second driving information of the sample vehicle collected by each sensor from the vehicle information collection module 301, and may acquire third position information of the sample vehicle by referring to other information. In some aspects, the other information may include map information such as a high-precision map.

[0074] In some aspects, the third location information may include the longitude, latitude and vehicle heading angle of the sample vehicle.

[0075] Here, the direction angle of the vehicle may be an angle at which the front of the vehicle is offset from a predetermined direction. For example, the direction angle of the vehicle may be an angle at which the front of the vehicle is offset clockwise from due north.

[0076] In the process of determining the third location information of the sample vehicle based on the above second driving information, some intermediate variables, ie, the above second intermediate variables, may be obtained.

[0077] In some aspects, third position information and second intermediate variables may be obtained based on the second travel information and the map information, where the second intermediate variables may include error information between lane information determined based on the second travel information and lane information of the sample vehicle in the map information.

[0078] As an example, the second intermediate variable may include two types of information:

[0079] (1) Intermediate variables obtained based on the information collected by the visual sensor and the high-precision map, such as the difference between the sensor lane width and the map lane width, lane optimization distance, initial optimization error, and final optimization error. (2) Based on intermediate variables obtained from the information collected by the GNSS sensor, the visual sensor, and the high-precision map, such as the probability that the positioning point is in each lane, the lane with the highest probability of the positioning point, whether the lane information provided by the GNSS sensor matches the lane information provided by the visual sensor, etc. In some aspects, the second intermediate variables may further include at least one of a covariance of an optimization algorithm used to determine the third position information and a statistical value of the second travel information during the second period of time.

[0080] As an example, the second intermediate variable may further include the following two kinds of information:

[0081] (3) Covariances of the optimization algorithm, such as the covariance of the velocity, the covariance of the IMU sensor deviations, and the covariance of the estimated attitude (e.g., the estimated vehicle longitude, latitude, and vehicle heading angle). (4) Statistics of parameters collected by a statistical sensor within a certain period of time, for example, statistics of information collected by a GNSS sensor within 20 seconds, the average value and variance of the vehicle speed within 1 second, the average value and variance of the acceleration values ​​of the accelerometer in the x-, y-, and z-axis directions within 10 seconds, the average value and variance of the angular velocity values ​​of the gyroscope in the x-, y-, and z-axis directions, the distance between the position of the last positioning point obtained by the GNSS sensor and the positioning result, etc. Here, in the above (1), the lane optimization distance may mean a distance obtained by optimizing the lane obtained by visual positioning when the lane marking of the lane where the vehicle is located obtained by visual positioning is aligned with the lane marking of the lane where the vehicle is located on the high-precision map. The initial optimization error may mean an error between the lane marking of the lane where the vehicle is located obtained by visual positioning before the lane marking of the lane where the vehicle is located on the high-precision map is aligned. The final optimization error may mean an error between the lane marking of the lane where the vehicle is located obtained by visual positioning after the lane marking of the lane where the vehicle is located on the high-precision map is aligned.

[0082] In the above (2), the GNSS sensor may perform highly accurate positioning using an RTK carrier phase difference technique, obtain the probability of the vehicle's positioning point being in each lane, and the lane with the highest probability of the positioning point may be determined as the lane in which the vehicle is located.The visual sensor may further obtain the probability of the vehicle's positioning point being in each lane based on the optimization result in the above (1), and determine the lane with the highest probability of the positioning point as the lane in which the vehicle is located.

[0083] In the above (3), different optimization algorithms and different optimization goals may be adopted, and the present invention is not limited thereto. The above optimization algorithm may be, for example, an optimization algorithm based on a form of global optimization, such as optimizing information collected by all sensors within a certain period of time, or may be, for example, an optimization algorithm based on a form of filtering, such as optimizing using information collected by sensors at previous and subsequent times.

[0084] In (4) above, the statistical values ​​of the information collected by the GNSS sensor within 20 seconds include, but are not limited to, the valid ratio, the fixed ratio, the average and variance of the accuracy value, the average and variance of the DOP value, the accuracy of the last GNSS positioning point, the DOP value, whether it is fixed or not, whether it is valid or not, etc.

[0085] Here, the valid ratio may mean a ratio of valid positioning of the GNSS sensor to all positioning of the GNSS sensor, and the fixed ratio may mean a ratio of fixed solutions to all positioning results among the positioning results acquired by the GNSS sensor.

[0086] 3 may tally up the statistical values ​​of the parameters provided by the vehicle information collection module 301 within a certain period of time to obtain the statistical value (4) above, but the present invention is not limited to this. After obtaining the statistical values, the information statistics module 303 may transmit the statistical values ​​to the real-time positioning module 302 or transmit the statistical values ​​to the model construction module 305.

[0087] Step 230: Obtain fourth position information of the sample vehicle collected at the first time by a positioning device.

[0088] As an example, the positioning device may be the true value collection module 304 in Fig. 3 and may be installed in the sample vehicle. Therefore, the fourth position information may be referred to as true value position information, that is, the fourth position information may be regarded as the actual position of the sample vehicle, but the present invention is not limited thereto.

[0089] The positioning equipment may have a high positioning accuracy, for example equipment such as SPAN-ISA-100C or SPAN-CPT. The higher the accuracy of the positioning equipment used, the more effective the estimation of the positioning accuracy.

[0090] In some aspects, the fourth location information may include the longitude, latitude, and vehicle heading angle of the sample vehicle.

[0091] Step 240: Determine a training sample set including the third position information, the fourth position information, and the second intermediate variable.

[0092] As an example, the model building module 305 in Fig. 3 may determine a training sample set. The training sample set may include a number of training samples, and each training sample may include the third position information obtained in step 220 above, the second intermediate variable, and the fourth position information in step 230.

[0093] For example, an autonomous vehicle may have a control system designed for driving in an urban environment and a control system designed for driving on a highway, since different vehicles have different Operational Design Domains (ODDs). As a result, a large amount of collected data is required in different scenarios, such as at least having to cover highways, urban highways, urban general roads, highways or urban tunnel sections, etc.

[0094] The embodiment of the present invention may collect data generated by a large number of sample vehicles while driving in different scenarios to generate a training sample set. For example, in a sufficiently large number of sections such as expressways, urban expressways, and urban general roads, a sensor may collect driving information while the sample vehicles are driving, and a positioning device may collect position information while the sample vehicles are driving. For example, data of a distance of at least 5,000 kilometers, or 10,000 kilometers or more may be collected in each scenario.

[0095] In one preferred embodiment, when determining the training sample set, the positioning result (e.g., the third position information) obtained by the real-time positioning module 302 and the positioning result (e.g., the fourth position information) obtained by the true value collection module 304 may be time-aligned, and all intermediate variables of the real-time positioning module 302 may be extracted. As an example, one training sample may be generated at a certain time point. That is, each training sample may include the positioning result obtained by the real-time positioning module 302, the positioning result obtained by the true value collection module 304, and the intermediate variables of the real-time positioning module 302 at a corresponding time point.

[0096] In some aspects, for a tunnel section, a simulation of the tunnel section may be performed on data of a sample vehicle collected on a normal section (a section that does not include a tunnel) to obtain training sample data corresponding to the tunnel section.

[0097] Specifically, in tunnel sections, GNSS signals are blocked, so data collected by GNSS sensors cannot be obtained, and if one relies solely on other sensors, such as IMU sensors, estimating the generated cumulative errors cannot meet the accuracy requirements, and the accuracy of the location information collected by positioning equipment in the tunnel sections may be reduced.

[0098] Typically, in the technology of autonomous driving or remote driving, it is sufficient to require that a relatively high positioning accuracy can be maintained for a certain period or distance of entering the tunnel for the tunnel section. Therefore, a simulation of the tunnel section may be performed by artificially setting a part of the file collected by the GNSS sensor as invalid. Accordingly, the section corresponding to the part set as invalid in the file collected by the GNSS sensor may be regarded as the tunnel section. In this way, training sample data corresponding to the tunnel section is obtained, and the fourth position information of the sample vehicle collected by the positioning device in the training sample has a sufficiently high accuracy with respect to the position information collected by the positioning device in the actual tunnel.

[0099] 4 is a schematic flowchart of a method 400 for acquiring training sample data according to an embodiment of the present invention. In the method 400, the driving information of a sample vehicle collected by a GNSS sensor may be periodically set to invalid and valid in a time sequence to generate training sample data corresponding to a tunnel section.

[0100] It should be noted that, although Figure 4 illustrates steps or operations of a method for obtaining a training sample set corresponding to a tunnel section, these steps or operations are merely exemplary and embodiments of the invention may perform other operations or variations of the operations in the figure. Additionally, the steps in Figure 4 may be performed in an order different from that shown in the figure, and not all of the operations in the figure may be performed.

[0101] As shown in FIG. 4, the method 400 includes steps 401-409.

[0102] Step 401: Obtain all valid files collected by the GNSS sensor.

[0103] As an example, in the road measurement stage, travel information of a sample vehicle collected by a sensor in a road section in a scenario without a tunnel of the sample vehicle may be acquired, where the travel information of the vehicle collected by the GNSS sensor may be stored in a file format.

[0104] Step 402: It is determined whether or not a time longer than 200 seconds has elapsed since the start time of the file.

[0105] Here, the file start time may be the time when the sample vehicle started traveling, i.e., the time when the GNSS sensor started collecting vehicle traveling information. By determining whether a time longer than 200 s has elapsed since the file start time, the acquired training sample data for the simulation of the tunnel can include the traveling information of the sample vehicle collected by the GNSS sensor 200 s before entering the tunnel.

[0106] Step 403: It is determined whether the speed has ever exceeded 20 km / h.

[0107] Here, by determining whether the speed has ever exceeded 20 km / h, it is possible to ensure that the vehicle is in a normal driving state, and therefore to ensure that the vehicle driving information obtained by the sensor is useful.

[0108] Step 404: Setting the files collected by the GNSS sensor to invalid.

[0109] Preferably, when invalidating a file collected by a GNSS sensor, information in a file formed from vehicle driving information collected by sensors other than the GNSS sensor of the sample vehicle (e.g., an IMU sensor, a visual sensor, a speed sensor, etc.) may be read out, and the location information of the sample vehicle collected by a positioning device may be obtained.

[0110] Step 405: It is determined whether the invalid time exceeds 300 seconds.

[0111] Here, the invalid time is 300 seconds, which corresponds to the travel time for simulating the tunnel section being 300 seconds. If the invalid time exceeds 300 seconds, the next step 406 is executed.

[0112] Step 406: Set the file collected by the GNSS sensor as valid.

[0113] Note that setting the file collected by the GNSS sensor to be valid after the invalid time exceeds 300 s corresponds to a scenario simulating a vehicle traveling in a normal section after exiting a tunnel section.

[0114] Step 407: The file information is read.

[0115] Here, in response to a scenario in which the sample vehicle travels along a normal section, information in a file formed from driving information of the sample vehicle collected by the GNSS sensor and other sensors on the vehicle, and location information of the sample vehicle collected by a positioning device may be read out.

[0116] Step 408: It is determined whether the end of the file has been reached.

[0117] If the end of the file has been reached, the process ends. If the end of the file has not been reached, the next step 409 is executed.

[0118] Step 409: It is determined whether the valid time exceeds 200 seconds.

[0119] If the valid time exceeds 200 s, the next step 404 is executed, that is, the file collected by the GNSS sensor continues to be set as invalid. In this way, by periodically (i.e., alternately) setting the file of the GNSS sensor to invalid for 300 s and to valid for 200 s in accordance with the time sequence, data of the sample vehicle in the tunnel section can be generated based on data of the sample vehicle collected in the normal section.

[0120] In normal cases, when a vehicle enters a tunnel section in an autonomous driving mode, the driver may control the autonomous vehicle. Therefore, the most interesting period in the tunnel section is the vehicle's driving information for a predetermined period (e.g., 200 s) before entering the tunnel or a predetermined distance collected by the distance sensor, so that the data of a sample vehicle in the tunnel section can be generated based on this by periodically setting the file of the GNSS sensor to be disabled for 300 s and enabled for 200 s.

[0121] It should be noted that the above times of 200 s, 300 s, and speed of 20 km / h are specific examples given to facilitate understanding of the means of the embodiments of the present invention, and each of the above time or speed values ​​may be replaced with other numerical values, and the embodiments of the present invention are not limited to these.

[0122] Therefore, the embodiment of the present invention can realize the simulation of the tunnel section by periodically setting the file collected by the GNSS sensor to invalid and valid in a time sequence, and generate the data of the sample vehicle corresponding to the tunnel section. That is, the section corresponding to the part of the file collected by the GNSS sensor that is set to invalid may be regarded as the tunnel section. In this way, a training sample set corresponding to the tunnel section can be obtained, and the accuracy of the position information collected by the positioning device in the training sample set can be guaranteed to be sufficiently high.

[0123] Step 250: Train an accuracy estimation model based on the training sample set.

[0124] 3 may train the accuracy estimation model 307 based on a training sample set. For example, the model construction module 305 may input the training samples to the accuracy estimation model 307 to update parameters of the accuracy estimation model 307.

[0125] In some aspects, different accuracy estimation models may be trained for different scenarios, respectively, so as to obtain the best accuracy error estimation effect in each scenario. As an example, the following three scenarios may be included:

[0126] Scenario 1: Non-tunnel scenario + scenario with map Scenario 2: Tunnel scenario + scenario with map Scenario 3: Non-tunnel scenario + no map scenario As an example, the above map may be a high-precision map, but the present invention is not limited thereto.

[0127] In the above Scenario 1, the GNSS sensor is effective, and real-time positioning may be performed by combining vehicle travel information collected by the sensor with map information.

[0128] In Scenario 1, the accuracy estimation model may specifically include a first accuracy estimation model. The third position information and the second intermediate variable in the training sample corresponding to the first accuracy estimation model are determined based on the second travel information and map information (e.g., a high-precision map). The second travel information includes travel information of the sample vehicle collected by a GNSS sensor, and the second intermediate variable includes error information between lane information determined based on the second travel information and lane information of the sample vehicle in the map information.

[0129] As a specific example, in the training sample of the first accuracy estimation model, the second intermediate variables may include four types of intermediate variables in the above step 220, such as (1) intermediate variables obtained based on information collected by a visual sensor and a high-precision map, (2) intermediate variables obtained based on information collected by a GNSS sensor, a visual sensor, and a high-precision map, (3) covariance of the optimization algorithm, and (4) statistical values ​​of parameters collected within a certain period of time by a statistical sensor.

[0130] In the above Scenario 2, the GNSS sensor may be disabled, and real-time positioning may be performed by combining vehicle travel information collected by sensors other than the GNSS sensor with map information.

[0131] In the scenario 2, the accuracy estimation model may specifically be a second accuracy estimation model. The third position information and the second intermediate variable in the training sample corresponding to the second accuracy estimation model are determined based on a valid portion of the second travel information and map information (e.g., a high-precision map). The travel information of the sample vehicle collected by the GNSS sensor in the second travel information is set to be partially invalid. The second intermediate variable includes error information between the lane information determined based on the second travel information and the lane information of the sample vehicle in the map information.

[0132] Preferably, the travel information of the sample vehicle collected by the GNSS sensor in the second travel information is periodically set to invalid and valid in chronological order.

[0133] Specifically, the method for invalidating the travel information of the sample vehicle collected by the GNSS sensor may refer to the description of step 240 above, and the description thereof will be omitted here.

[0134] As a specific example, in the training sample of the second accuracy estimation model, the second intermediate variables may include four types of intermediate variables, such as (1) intermediate variables obtained based on information collected by a visual sensor and a high-precision map, (3) covariance of an optimization algorithm, and (4) statistical values ​​of parameters collected by a statistical sensor within a certain period of time (parameters collected by a GNSS sensor are not included) in the above step 220. For example, the statistical values ​​in (4) may be, for example, the average value and variance of the vehicle speed within 1 s, the average value and variance of the acceleration values ​​in the x-, y-, and z-axis directions of the accelerometer within 10 s, and the average value and variance of the angular velocity values ​​in the x-, y-, and z-axis directions of the gyro.

[0135] In the above scenario 3, the GNSS sensor is effective, but since there is no map information, real-time positioning cannot be performed by combining the vehicle driving information collected by the sensor with map information.

[0136] In the scenario 3, the accuracy estimation model may specifically be a third accuracy estimation model. The third position information and the second intermediate variables in the training sample corresponding to the third accuracy estimation model are determined based on the second travel information. The second travel information includes travel information of the sample vehicle collected by the GNSS sensor.

[0137] As a specific example, in the training sample of the third accuracy estimation model, the second intermediate variables may include four types of intermediate variables, such as (3) the covariance of the optimization algorithm in step 220 above, and (4) statistical values ​​of parameters collected within a certain period of time by a statistical sensor.

[0138] After preparing a training sample set for each of the above scenarios, a model may be constructed for each of the above scenarios, that is, an accuracy estimation model for each scenario may be trained based on the training sample set corresponding to each scene. Here, the accuracy estimation model may be a machine learning model, such as, but not limited to, a random forest model, an xgboost model, a deep neural network model, etc.

[0139] Below is an example of an algorithm configuration group when the xgboost model is adopted as the accuracy estimation model.

[0140] Sample size: 300,000 'booster': 'gbtree' 'objective': 'reg:gamma' or 'reg:squarederror' 'gamma': 0.1, 'max_depth': 6, 'lambda': 3, 'subsample': 0.7, 'colsample_bytree': 0.7, 'min_child_weight': 3, 'silent': 1, 'eta': 0.1 As an example, the third location information, the fourth location information, and the second intermediate variable in the training sample in each scenario may be input into the accuracy estimation model of the corresponding scenario. The accuracy estimation model may obtain the fifth location information of the sample vehicle and the accuracy error of the third location information based on the third location information and the second intermediate variable. Here, the accuracy error may be an error of the third location information relative to the fifth location information, and the present invention is not limited thereto.

[0141] Preferably, in the process of model training, a first loss of the model may be determined based on the fifth location information and the fourth location information. Preferably, in the process of model training, a second loss of the model may be determined based on an accuracy error of the third location information relative to the fifth location information and an accuracy error of the third location information relative to the fourth location information. Then, a parameter of the accuracy estimation model may be updated based on at least one of the first loss and the second loss.

[0142] In some aspects, the accuracy error includes at least one of a lateral distance error, a longitudinal distance error, and a direction angle error.

[0143] Taking the error of the third position information with respect to the fourth position information as an example, the distance between the third position information and the fourth position information decomposed into components perpendicular to the road is called a horizontal distance error, and the distance between the third position information and the fourth position information decomposed into components in the road direction is called a vertical distance error. As an example, the direction of the positioning device (e.g., a true value collecting device) may be the road direction. The road direction may be called a driving direction, but is not limited thereto.

[0144] If the fourth position information includes longitude lon0, latitude lat0, and vehicle direction angle heading0, and the third position information includes longitude lon, latitude lat, and vehicle direction angle heading, it may be determined that the accuracy error of the third position information relative to the fourth position information includes a lateral distance error disthorizontal, a vertical distance error distvetical, and a direction angle error.

[0145] Here, the direction angle error may be the difference between the vehicle direction angle heading0 and the vehicle direction angle heading, that is, heading0-heading.

[0146] The calculation process of the horizontal distance error disthorizontal and the vertical distance error distvetical is as follows.

[0147] First, calculate the distance dist between the two positions corresponding to the fourth position information and the third position information. Specifically, the following formula may be obtained based on the Haversine formula:

[0148]

number

[0149]

number

[0150]

number

[0151] Then, a direction angle point_degree between two positions corresponding to the fourth position information and the third position information may be calculated. The direction angle point_degree is an angle by which the direction formed by the two positions corresponding to the fourth position information and the third position information is shifted from the direction of the positioning device, and its value ranges from 0 to 360°. The calculation process of the direction angle point_degree is specifically as follows.

[0152]

number

[0153]

number

[0154]

number

[0155]

number

[0156] Finally, the horizontal distance error disthorizontal and the vertical distance error distvetical may be calculated based on the direction angle point_degree. The specific calculation process is as follows:

[0157]

number

[0158]

number

[0159]

number

[0160]

number

[0161]

number

[0162]

number

[0163]

number

[0164]

number

[0165]

number

[0166] In addition, there is another kind of scenario, namely, tunnel scenario + no map scenario, in which even a high-precision sensor, such as an IMU sensor, cannot guarantee that the positioning result has a sufficiently high accuracy, so in this scenario, it is set that the positioning is unavailable and there is no need to estimate the positioning accuracy.

[0167] Therefore, in an embodiment of the present invention, based on the third position information of the sample vehicle determined using the driving information of the sample vehicle collected by the sensor, the second intermediate variable used in the process of determining the third position information and the fourth position information of the sample vehicle collected by the positioning device are determined, thereby realizing training of the accuracy estimation model and obtaining a trained accuracy estimation model. The accuracy estimation model according to the embodiment of the present invention can realize estimation of the accuracy error of the position information by fusing the vehicle position information determined based on the vehicle driving information collected by the sensor and the intermediate variable used in the process of determining the position information.

[0168] After obtaining an accurate estimation model through training, we may proceed to the online estimation stage, which is described below.

[0169] 5 is a schematic flow chart of a method 500 for estimating positioning accuracy according to an embodiment of the present invention. The method 500 may be performed by any electronic device having data processing capability. For example, the electronic device may be implemented as a server or a terminal device, and for example, the electronic device may be implemented as the calculation module 109 in FIG. 1, but the present invention is not limited thereto.

[0170] In some aspects, the machine learning model may be included (e.g., configured) in the electronic device, and may be the accuracy estimation model described above. Still referring to FIG. 3, the vehicle information collection module 301, the real-time positioning module 302, the information statistics module 303, the information determination module 306, the accuracy estimation model 307, and the accuracy estimation module 308 may be used to estimate the vehicle positioning accuracy. The following describes a method 500 for estimating the positioning accuracy with reference to FIG. 3.

[0171] As shown in FIG. 5, the method 500 includes steps 510-540.

[0172] Step 510: Obtain first driving information of the vehicle collected by a sensor.

[0173] As an example, the vehicle information collection module 301 in Fig. 3 may obtain the first driving information of the vehicle collected by a sensor. The sensor may collect the first driving information of the vehicle in the current and previous periods, but is not limited thereto. After obtaining the first driving information, the vehicle information collection module 301 may send the first driving information to the real-time positioning module 302 and the information determination module 306. Preferably, the first driving information may be sent to the information statistics module 303.

[0174] In some aspects, the first driving information includes vehicle driving information obtained by at least one of an inertial measurement unit (IMU) sensor, a vehicle speed sensor, and a visual sensor.

[0175] Specifically, the first driving information collected by the sensor is similar to the second driving information collected by the sensor in step 210 of FIG. 2, and the above description may be referred to, and the description will be omitted here.

[0176] Step 520: Obtain first position information of the vehicle and first intermediate variables used in the process of determining the first position information based on the first driving information.

[0177] 3 may obtain first position information of the vehicle and first intermediate variables used in the process of determining the first position information based on the first driving information. After obtaining the first position information and the first intermediate variables, the real-time positioning module 302 may transmit the first position information and the first intermediate variables to the accuracy estimation model 307.

[0178] In some aspects, the first location information includes a longitude, a latitude, and a heading angle of the vehicle.

[0179] In some aspects, first position information and first intermediate variables may be obtained based on the first travel information and the map information, where the first intermediate variables include error information between lane information determined based on the first travel information and lane information of the vehicle in the map information.

[0180] In some aspects, the first intermediate variables may include at least one of a covariance of an optimization algorithm used to determine the first location information and a statistical value of the first travel information within a first period of time.

[0181] 3 may collect statistical values ​​of parameters provided by the vehicle information collection module 301 within a predetermined period to obtain statistical values ​​of a first driving information within a first period. After obtaining the statistical values, the information statistics module 303 may send the statistical values ​​to the real-time positioning module 302, or send the statistical variables to the accuracy estimation model 307.

[0182] Specifically, the first position information is similar to the third position information in step 220 in FIG. 2, and the first intermediate variable is similar to the second intermediate variable in step 220 in FIG. 2, and the description thereof will be omitted here.

[0183] Step 530: Determine a target accuracy estimation model in a pre-trained accuracy estimation model according to the first driving information and the first intermediate variable. The accuracy estimation model is obtained by training a machine learning model based on a training sample set, and a training sample in the training sample set includes location information of a sample vehicle and intermediate variables used in the process of determining the location information.

[0184] Specifically, the training sample set may include third location information, fourth location information, and a second intermediate variable used in the process of determining the third location information, where the third location information is determined based on second driving information of the sample vehicle collected by a sensor at a first time, and the fourth location information includes location information of the sample vehicle collected by a positioning device at the first time.

[0185] Specifically, the training process of the accuracy estimation model may refer to the above description in FIG. 2, and the description is omitted here.

[0186] In some aspects, the pre-trained accuracy estimation model includes a plurality of models each trained for different scenarios. In this case, a target accuracy estimation model may be determined in the pre-trained accuracy estimation model based on a first intermediate variable in the process of determining the first driving information and the first position information of the vehicle collected by the sensor. That is, a target accuracy estimation model applicable to the call may be determined according to a specific application scenario.

[0187] As an example, the information determination module 306 in FIG. 3 may determine an applicable accuracy estimation model for the call based on the first driving information and the first intermediate variable, i.e., based on the first driving information and whether there is currently available map information, such as a high-precision map.

[0188] As an example, when the first driving information includes vehicle driving information collected by a GNSS sensor and there is currently available map information (e.g., a high-precision map), the first accuracy estimation model may be determined as the target accuracy estimation model, i.e., it may be determined to call the first accuracy estimation model in scenario 1 above.

[0189] As another example, if the first driving information does not include vehicle driving information collected by a GNSS sensor and there is currently available map information (e.g., a high-precision map), the second accuracy estimation model may be determined as the target accuracy estimation model, i.e., it may be decided to call the second accuracy estimation model in scenario 2 above.

[0190] As another example, when the first driving information includes vehicle driving information collected by a GNSS sensor and there is no currently available map information (e.g., a high-precision map), the third accuracy estimation model may be determined as the target accuracy estimation model, i.e., it may be determined to call the third accuracy estimation model in scenario 3 above.

[0191] Specifically, for the first accuracy estimation model, the second accuracy estimation model, and the third accuracy estimation model, reference may be made to the description of step 250 in FIG. 2, and the description thereof will be omitted here.

[0192] Step 540: Input the first position information and a first intermediate variable into the target accuracy estimation model to obtain second position information of a vehicle and an accuracy error of the first position information relative to the second position information.

[0193] As an example, still referring to FIG. 3, after the information determination module 306 determines an applicable target accuracy estimation model for the call based on the first driving information of the vehicle acquired by the sensor in the vehicle information collection module 301, while referring to whether there is currently available map information, such as a high-precision map, the real-time positioning module 302 and the information statistics module 303 may send input-related features required for the accuracy estimation model to the target accuracy estimation model.

[0194] For example, when the first accuracy estimation model is invoked, the first intermediate variables may include (1) intermediate variables obtained based on information collected by a visual sensor and a high-precision map, (2) intermediate variables obtained based on a GNSS sensor, information collected by a visual sensor, and a high-precision map, (3) a covariance of the optimization algorithm, and (4) statistical values ​​of parameters collected by a statistical sensor within a certain period of time, which are used in the process of determining the first position information.

[0195] For example, when the second accuracy estimation model is invoked, the first intermediate variables may include (1) intermediate variables obtained based on information collected by a visual sensor and a high-precision map, (3) a covariance of an optimization algorithm, and (4) statistical values ​​of parameters collected by a statistical sensor within a certain period of time (not including parameters collected by a GNSS sensor) used in the process of determining the first position information.

[0196] Also, for example, when the third accuracy estimation model is invoked, the first intermediate variables may include (3) a covariance of the optimization algorithm used in the process of determining the first position information, and (4) statistical values ​​of parameters collected by a statistical sensor within a certain period of time.

[0197] The first accuracy estimation model, the second accuracy estimation model or the third accuracy estimation model may output predicted second position information of the vehicle and an accuracy error of the first position information relative to the second position when relevant features are input.

[0198] Here, the second position information may be the vehicle's position information collected by the positioning device predicted by the accuracy estimation model, or may be the vehicle's actual position information predicted by the accuracy estimation model, but is not limited thereto. As an example, the second position information includes the vehicle's longitude, latitude, and direction angle.

[0199] As an example, the target accuracy estimation model can realize prediction of actual position information of the vehicle collected by a positioning device and estimation of the accuracy error of the positioning information relative to the actual position information, using vehicle positioning information determined based on vehicle driving information collected by a sensor and intermediate variables used in the process of determining the positioning information.

[0200] Therefore, an embodiment of the present invention can effectively evaluate the accuracy error of high-precision positioning by obtaining first position information of a vehicle and a first intermediate variable used in the process of determining the first position information based on the first driving information, determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, inputting the first position information and the first intermediate variable into the target accuracy estimation model, and obtaining second position information of a vehicle and an accuracy error of the first position information relative to the second position information.

[0201] The method for estimating positioning accuracy according to the embodiment of the present invention has significantly improved both the accuracy error CEP90 and the average value compared to the conventional method using optimized covariance. Here, CEP90 means the value at the 90% position when all accuracy errors are sorted in ascending order. For example, in a data set of different road sections for about 2-3 hours each day for 20 days, the average value of the accuracy error determined by the embodiment of the present invention is improved by about 0.5 m on average compared to the covariance.

[0202] 6 is a diagram illustrating a specific example of a method 600 for estimating positioning accuracy according to an embodiment of the present invention. The method 600 may be performed by any electronic device having data processing capabilities. For example, the electronic device may be implemented as a server or a terminal device, and for example, the electronic device may be implemented as the computing module 109 in FIG. 1, but the present invention is not limited thereto.

[0203] It should be noted that, although Figure 6 illustrates steps or operations of a method for estimating positioning accuracy, these steps or operations are merely examples, and embodiments of the present invention may perform other operations or variations of each operation in the figure. Also, each step in Figure 6 may be performed in an order different from that shown in the figure, and not all of the operations in Figure 6 may be performed.

[0204] As shown in FIG. 6, the method 600 includes steps 601-610.

[0205] Step 601: Obtain first driving information of a vehicle.

[0206] Specifically, step 601 may refer to the description of step 510, and the description thereof will be omitted here.

[0207] Step 602: It is determined whether it is a tunnel.

[0208] Specifically, it may be determined whether the vehicle is in a tunnel section based on the first travel information in step 601. For example, if the first travel information includes vehicle travel information collected by a GNSS sensor, it may be determined that the vehicle is not in a tunnel section, and if the first travel information does not include vehicle travel information collected by a GNSS sensor, it may be determined that the vehicle is in a tunnel section.

[0209] For example, if the vehicle is in a tunnel section, the next step 603 is executed. If the vehicle is not in a tunnel section, the next step 605 is executed.

[0210] Step 603: It is determined whether or not there is a high-precision map.

[0211] That is, in the tunnel scenario, it is further determined whether or not there is a high-precision map. If there is a high-precision map, the next step 604 is executed, and if there is no high-precision map, the next step 610 is executed.

[0212] Step 604: Invoke a second accuracy estimation model.

[0213] Step 605: It is determined whether or not there is a high-precision map.

[0214] That is, in a non-tunnel scenario, it is further determined whether there is a high-precision map. If there is a high-precision map, the next step 606 is executed, and if there is no high-precision map, the next step 607 is executed.

[0215] Step 606: Invoke the first accuracy estimation model.

[0216] Step 607: Invoke the third accuracy estimation model.

[0217] Specifically, for the first accuracy estimation model, the second accuracy estimation model, and the third accuracy estimation model, reference may be made to the explanations in FIG. 2 and FIG. 5, and the explanations thereof will be omitted here.

[0218] Step 608: Perform parameter statistics.

[0219] Specifically, statistics may be performed on the driving parameters in the first driving information in step 601 to obtain statistical values ​​of each driving parameter.

[0220] Step 609: Perform real-time positioning.

[0221] Specifically, the first location information of the vehicle may be determined based on the first traveling information in step 601 and the statistical value obtained in step 608. The real-time positioning result (i.e., the first location information) and intermediate variables used in the process of determining the first location information may be input into corresponding accuracy estimation models. Specifically, the parameters input into different accuracy estimation models may refer to the description in FIG. 5, and the description thereof will be omitted here.

[0222] Step 610: Determine that a position fix is ​​unavailable.

[0223] That is, in the tunnel scenario and without high precision map, it is determined that positioning is unavailable.

[0224] Step 611: Determine the accuracy error.

[0225] Specifically, the accuracy error of the real-time positioning result in different scenarios may be determined based on the results of invoking the accuracy estimation model in different scenarios.

[0226] Therefore, an embodiment of the present invention can effectively evaluate the accuracy error of high-precision positioning by obtaining first position information of a vehicle and a first intermediate variable used in the process of determining the first position information based on the first driving information, determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, inputting the first position information and the first intermediate variable into the target accuracy estimation model, and obtaining second position information of a vehicle and an accuracy error of the first position information relative to the second position information.

[0227] The above describes in detail specific embodiments of the present invention with reference to the accompanying drawings, but the present invention is not limited to the specific details of the above-mentioned embodiments, and many simple modifications are possible to the technical solutions of the present invention within the scope of the technical idea of ​​the present invention, and these simple modifications belong to the technical scope of the present invention. For example, the individual specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction, and in order to avoid unnecessary repetition, further description of various possible combinations will be omitted in this specification. For example, various embodiments of the present invention can be combined in any manner, and as long as it is not contrary to the idea of ​​the present invention, it will be considered as being disclosed in the present specification as well.

[0228] In addition, in various method embodiments of the present invention, the magnitude of the sequence numbers of the above-mentioned processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and inherent logic, and should not impose any limitations on the execution process of the embodiments of the present invention. In addition, these sequence numbers can be appropriately exchanged so that the described embodiments of the present invention can be implemented in an order other than that shown or described.

[0229] An embodiment of the method of the present invention has been described in detail above. In the following, an embodiment of the apparatus of the present invention will be described in detail with reference to FIGS.

[0230] 7 is a schematic block diagram of a positioning accuracy estimation apparatus 700 according to an embodiment of the present invention. As shown in FIG. 7, the positioning accuracy estimation apparatus 700 may include an acquisition unit 710, a processing unit 720, a determination unit 730 and a target accuracy estimation model 740.

[0231] The acquisition unit 710 acquires first driving information of the vehicle from a sensor.

[0232] The processing unit 720 obtains, based on the first driving information, first position information of the vehicle and first intermediate variables used in the process of determining the first position information.

[0233] The determining unit 730 determines a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable. The accuracy estimation model is obtained by training a machine learning model based on a training sample set, and a training sample in the training sample set includes location information of a sample vehicle and intermediate variables used in the process of determining the location information.

[0234] The first position information and the first intermediate variable are input to the target accuracy estimation model 740, and the target accuracy estimation model 740 obtains second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

[0235] In some aspects, the processing unit 720 obtains the first position information and the first intermediate variable based on the first driving information and map information, where the first intermediate variable includes error information between lane information determined based on the first driving information and lane information of the vehicle in the map information.

[0236] In some aspects, the first travel information includes travel information of the vehicle collected by a Global Navigation Satellite System (GNSS) sensor.

[0237] Here, the determination unit 730 determines a first accuracy estimation model as the target accuracy estimation model based on the first driving information and the first intermediate variable, where the position information and the intermediate variable in the training sample set of the first accuracy estimation model are determined based on the driving information and the map information of the sample vehicle collected by a GNSS sensor.

[0238] In some aspects, the determiner 730 determines a second accuracy estimation model as the target accuracy estimation model based on the first driving information and the first intermediate variable, where the position information and the intermediate variable in the training sample set corresponding to the second accuracy estimation model are determined based on a valid part of the driving information of the sample vehicle collected by a GNSS sensor and the map information, and the driving information of the sample vehicle collected by the GNSS sensor is set to be partially invalid.

[0239] In some aspects, the sample vehicle driving information collected by the GNSS sensor is periodically set to invalid and valid in a time sequence.

[0240] In some aspects, the first travel information includes travel information of the vehicle collected by a GNSS sensor.

[0241] Here, the determination unit 730 determines a third accuracy estimation model as the accuracy estimation model based on the first driving information and the first intermediate variable, where the position information and the intermediate variable in the training sample set corresponding to the third accuracy estimation model are determined based on the driving information of the sample vehicle collected by a GNSS sensor.

[0242] In some aspects, the first driving information includes driving information of the vehicle collected by at least one of an inertial measurement unit (IMU) sensor, a vehicle speed sensor, and a visual sensor.

[0243] In some aspects, the first intermediate variable includes at least one of a covariance of an optimization algorithm used to determine the first position information and a statistical value of the first driving information during a first period of time.

[0244] In some aspects, the first location information includes a first longitude, a first latitude, and a first vehicle direction angle of the vehicle, and the second location information includes a second longitude, a second latitude, and a second vehicle direction angle of the vehicle.

[0245] In some aspects, the accuracy error includes at least one of a lateral distance error, a longitudinal distance error, and a direction angle error.

[0246] It should be noted that the embodiment of the device corresponds to the embodiment of the method, and the description of the embodiment of the method may be referred to. In order to avoid repetition, the description is omitted here. Specifically, the estimation device 700 of the positioning accuracy according to the embodiment can correspond to the subject that executes the method 500 or 600 according to the embodiment of the present invention, and the above operations and / or functions of each module in the estimation device 700 of the positioning accuracy are for realizing the corresponding flow of the above method 500 or 600, respectively, and the description is omitted here for brevity.

[0247] 8 is a schematic block diagram of an apparatus 800 for training an accuracy estimation model according to an embodiment of the present invention. As shown in FIG. 8, the apparatus 800 for training an accuracy estimation model may include a first acquisition unit 810, a processing unit 820, a second acquisition unit 830, a determination unit 840 and a training unit 850.

[0248] The first acquisition unit 810 acquires second driving information of the sample vehicle at a first time from a sensor.

[0249] The processing unit 820 obtains, based on the second driving information, third location information of the sample vehicle and second intermediate variables used in the process of determining the third location information.

[0250] The second acquisition unit 830 acquires fourth position information of the sample vehicle at the first time from a positioning device.

[0251] A determination unit 840 determines a training sample set including the third position information, the fourth position information, and the second intermediate variable.

[0252] A training unit 850 trains the accuracy estimation model based on the training sample set.

[0253] In some aspects, the accuracy estimation model includes a first accuracy estimation model and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor.

[0254] Here, the processing unit 820 acquires the third position information and the second intermediate variable based on the second driving information and the map information, where the second intermediate variable includes error information between the lane information determined based on the second driving information and the lane information of the sample vehicle in the map information.

[0255] Here, a training unit 850 trains the first accuracy estimation model based on the training sample set.

[0256] In some aspects, the accuracy estimation model includes a second accuracy estimation model, and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor.

[0257] The processing unit 820 further sets the travel information of the sample vehicle collected by the GNSS sensor to be partially invalid, and acquires the third position information and the second intermediate variable based on the valid part of the second travel information and the map information, where the second intermediate variable includes error information between lane information determined based on the second travel information and the lane information of the sample vehicle in the map information.

[0258] Here, a training unit trains the second accuracy estimation model based on the training sample set.

[0259] In some aspects, the accuracy estimation model includes a third accuracy estimation model, and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor.

[0260] Here, a training unit 850 trains the third accuracy estimation model based on the training sample set.

[0261] In some aspects, the second driving information includes driving information of the sample vehicle collected by at least one of an inertial measurement unit (IMU) sensor, a vehicle speed sensor, and a visual sensor.

[0262] In some aspects, the second intermediate variables include at least one of a covariance of an optimization algorithm used to determine the third position information and a statistical value of the second driving information during a second period of time.

[0263] It should be noted that the embodiment of the device corresponds to the embodiment of the method, and the description of the embodiment of the method may be referred to. In order to avoid repetition, the description is omitted here. Specifically, the training device 800 of the accuracy estimation model according to the embodiment can correspond to the subject that executes the method 200 according to the embodiment of the present invention, and the above operations and / or functions of each module in the training device 800 of the accuracy estimation model are for realizing the corresponding flow of the above method 200, respectively, and the description is omitted here for brevity.

[0264] The above describes the apparatus and system of the embodiment of the present invention in terms of functional modules with reference to the accompanying drawings. The functional modules may be implemented by instructions in hardware, software, or a combination of hardware and software modules. Specifically, the steps of the method aspects of the embodiment of the present invention may be performed by instructions in the form of integrated logic circuits and / or software in hardware in a processor, and the steps of the method disclosed in relation to the embodiment of the present invention may be performed directly by a hardware decoding processor or may be performed using a combination of hardware and software modules in a decoding processor. Preferably, the software modules may be located in a storage medium established in the art, such as a random memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically rewritable programmable memory, a register, etc. Preferably, the storage medium is in a memory, and the processor reads information in the memory and performs the steps in the above embodiment of the method in combination with the hardware.

[0265] FIG. 9 is a schematic block diagram of an electronic device 1100 according to an embodiment of the present invention.

[0266] As shown in FIG. 9, the electronic device 1100 may include a memory 1110 and a processor 1120.

[0267] The memory 1110 stores a computer program and transmits the program code to the processor 1120. In other words, the processor 1120 can call and execute the computer program from the memory 1110 to implement the method according to the embodiment of the present invention.

[0268] For example, the processor 1120 may perform the steps of the method 200 described above according to instructions in a computer program.

[0269] In some embodiments of the invention, processor 1120 may include, but is not limited to, a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc.

[0270] In some embodiments of the invention, memory 1110 may include, but is not limited to, volatile and / or non-volatile memory, where non-volatile memory may be Read-Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM) or flash memory, and volatile memory may be Random Access Memory (RAM) acting as external cache memory. Various forms of RAM may be used, including, by way of example and not limitation, static random access memory (Static RAM (SRAM)), dynamic random access memory (DRAM)), synchronous dynamic random access memory (Synchronous DRAM (SDRAM)), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM (DDR SDRAM)), enhanced synchronous dynamic random access memory (Enhanced SDRAM (ESDRAM)), synchronous link dynamic random access memory (synch link DRAM (SLDRAM)), and direct memory bus random access memory (Direct Rambus RAM (DR RAM)).

[0271] In some embodiments of the present invention, a computer program may be divided into one or more modules that are stored in memory 1110 and executed by processor 1120 to perform a method according to the present invention. The one or more modules may be a sequence of computer program instruction segments that describe the execution of the computer program in electronic device 1100 and may perform a particular function.

[0272] 9, the electronic device 1100 may further include a transceiver 1130 that can be connected to the processor 1120 or the memory 1110. Here, the processor 1120 may control the transceiver 1130 to communicate with other devices, specifically, to transmit information or data to other devices, or to receive information or data transmitted by other devices. The transceiver 1130 may include a transmitter and a receiver. The transceiver 1130 may further include an antenna, and the number of antennas may be one or more.

[0273] The components in electronic device 1100 are connected by a bus system. Here, the bus system includes a power bus, a control bus, and a status signal bus in addition to a data bus.

[0274] According to one aspect of the present invention, there is provided a communications device including a memory having a computer program stored therein, and a processor, the processor calling and executing the computer program stored in the memory to cause an encoder to perform a method of the above-described method embodiment.

[0275] According to one aspect of the present invention, there is provided a computer storage medium having a computer program stored thereon, the computer program being configured to, when executed by a computer, cause the computer to perform the method of the above method embodiment. In other words, an embodiment of the present invention further provides a computer program product comprising instructions, the instructions being configured to, when executed by a computer, cause the computer to perform the method of the above method embodiment.

[0276] According to another aspect of the invention there is provided a computer program product or computer program comprising computer instructions stored on a computer readable storage medium, the computer instructions being read by a processor of a computing device from the computer readable storage medium and executed by the processor to cause the computing device to perform a method according to any of the method embodiments described above.

[0277] In other words, when implemented using software, it can be implemented as a computer program product, in whole or in part. A computer program product includes one or more computer instructions. When the computer program instructions are loaded into a computer and executed, a flow or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, the computer instructions can be transmitted from a website site, computer, server, or data center to another website site, computer, server, or data center by wire (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (infrared, radio, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer, or a data storage device such as a server, data center, etc. that integrates one or more available media. The computer readable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), a semiconductor medium (e.g., a solid state disk (SSD)), or the like.

[0278] It should be noted that the computer-readable medium may be targeted at relevant data such as user information in certain embodiments of the present invention. When the computer-readable medium is applied to a specific product or technology in the above embodiment of the present invention, the user's permission or consent must be obtained, and the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0279] Those skilled in the art will understand that the exemplary modules and algorithm steps described in connection with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether the functions disclosed herein are implemented in hardware or software depends on the specific application and design constraints of the technical solution. The functions described herein may be implemented in different ways for each specific application by those skilled in the art, but such implementations should not be considered outside the scope of the present specification.

[0280] It should be noted that in some embodiments provided herein, the disclosed apparatus, device, and method may be realized in other ways. For example, the above-described apparatus embodiments are merely exemplary, and for example, the division of modules is merely a division of logical functions, and when actually implemented, there may be additional divisions, for example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not implemented. In another respect, the illustrated or described couplings or direct couplings or communication connections between each other may be indirect couplings or communication connections through some interfaces, devices, or modules, which may be electrical, mechanical, or other forms.

[0281] Modules illustrated as separate components may or may not be physically separated, and components illustrated as modules may or may not be physical modules, i.e., located in one location or distributed across multiple network elements. Separate or complete modules may be selected according to actual needs to achieve the objectives of the present embodiment. Separate physical presence may be possible, or two or more modules may be integrated into one module.

[0282] The above are only specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any modifications or replacements that can be easily thought of by those skilled in the art within the technical scope disclosed in the present invention are within the scope of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for estimating positioning accuracy, comprising: acquiring first driving information of a vehicle from a sensor; obtaining first position information of the vehicle based on the first driving information, and a first intermediate variable used in a process of determining the first position information; determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variables, the accuracy estimation model being obtained by training a machine learning model based on a training sample set, wherein a training sample in the training sample set includes location information of a sample vehicle and intermediate variables used in a process of determining the location information; and inputting the first position information and the first intermediate variable into the target accuracy estimation model to obtain second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

2. The step of acquiring first position information of the vehicle and a first intermediate variable used in a process of determining the first position information based on the first driving information includes:

2. The method according to claim 1, comprising: a step of acquiring the first position information and the first intermediate variable based on the first driving information and map information, the first intermediate variable including error information between lane information determined based on the first driving information and lane information of the vehicle in the map information.

3. The first driving information includes driving information of the vehicle collected by a Global Navigation Satellite System (GNSS) sensor; The step of determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable includes: The method of claim 2, comprising: a step of determining a first accuracy estimation model as the target accuracy estimation model based on the first driving information and the first intermediate variables, wherein the position information and the intermediate variables in a training sample set of the first accuracy estimation model are determined based on the driving information and the map information of the sample vehicle collected by a GNSS sensor.

4. The step of determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable includes: The method of claim 2, further comprising: a step of determining a second accuracy estimation model as the target accuracy estimation model based on the first driving information and the first intermediate variables, wherein the position information and the intermediate variables in a training sample set corresponding to the second accuracy estimation model are determined based on a valid portion of the driving information of the sample vehicle collected by a GNSS sensor and the map information, and the driving information of the sample vehicle collected by the GNSS sensor is set partially invalid.

5. The method according to claim 4 , wherein the driving information of the sample vehicle collected by the GNSS sensor is periodically set to invalid and valid according to a time sequence.

6. The first travel information includes travel information of the vehicle collected by a GNSS sensor, The step of determining a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable includes:

2. The method of claim 1, comprising: a step of determining a third accuracy estimation model as the accuracy estimation model based on the first driving information and the first intermediate variables, wherein the position information and the intermediate variables in a training sample set corresponding to the third accuracy estimation model are determined based on driving information of the sample vehicle collected by a GNSS sensor.

7. The method of claim 1 , wherein the first driving information includes driving information of the vehicle collected by at least one of an inertial measurement unit (IMU) sensor, a vehicle speed sensor, and a visual sensor.

8. 2. The method of claim 1, wherein the first intermediate variables include at least one of a covariance of an optimization algorithm used to determine the first position information and a statistical value of the first driving information during a first time period.

9. 2. The method of claim 1, wherein the first location information includes a first longitude, a first latitude, and a first vehicle direction angle of the vehicle, and the second location information includes a second longitude, a second latitude, and a second vehicle direction angle of the vehicle.

10. The method of claim 1 , wherein the accuracy error includes at least one of a lateral distance error, a longitudinal distance error, and a direction angle error.

11. A method for training an accuracy estimation model, comprising the steps of: acquiring second driving information of the sample vehicle from a sensor at a first time; obtaining third position information of the sample vehicle based on the second driving information, and second intermediate variables used in a process of determining the third position information; obtaining fourth position information of the sample vehicle at the first time from a positioning device; determining a training sample set including the third location information, the fourth location information, and the second intermediate variable; training the accuracy estimation model based on the training sample set.

12. The accuracy estimation model includes a first accuracy estimation model, and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor; The step of acquiring third position information of the sample vehicle and a second intermediate variable used in the process of determining the third position information based on the second driving information includes: A step of acquiring the third position information and the second intermediate variable based on the second traveling information and map information, the second intermediate variable including error information between lane information determined based on the second traveling information and lane information of the sample vehicle in the map information; Training the accuracy estimation model based on the training sample set includes: The method of claim 11 , further comprising: training the first accuracy estimation model based on the training sample set.

13. The accuracy estimation model includes a second accuracy estimation model, and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor; The method further includes a step of partially invalidating the driving information of the sample vehicle collected by the GNSS sensor; The step of acquiring third position information of the sample vehicle and a second intermediate variable used in the process of determining the third position information based on the second driving information includes: A step of acquiring the third position information and the second intermediate variable based on a valid portion of the second travel information and map information, the second intermediate variable including error information between lane information determined based on the second travel information and lane information of the sample vehicle in the map information; Training the accuracy estimation model based on the training sample set includes: The method of claim 11 , further comprising: training the second accuracy estimation model based on the training sample set.

14. The accuracy estimation model includes a third accuracy estimation model, and the second driving information includes driving information of the sample vehicle collected by a Global Navigation Satellite System (GNSS) sensor; Training the accuracy estimation model based on the training sample set includes: The method of claim 11 , further comprising: training the third accuracy estimation model based on the training sample set.

15. A positioning accuracy estimation device, comprising: an acquisition unit that acquires first driving information of the vehicle from a sensor; A processing unit that acquires first position information of the vehicle and a first intermediate variable used in a process of determining the first position information based on the first driving information; A determination unit that determines a target accuracy estimation model in a pre-trained accuracy estimation model based on the first driving information and the first intermediate variable, the accuracy estimation model being obtained by training a machine learning model based on a training sample set, and a training sample in the training sample set includes location information of a sample vehicle and intermediate variables used in a process of determining the location information; The target accuracy estimation model receives the first position information and the first intermediate variable as input, and obtains second position information of the vehicle and an accuracy error of the first position information relative to the second position information.

16. A training device for an accuracy estimation model, comprising: a first acquisition unit that acquires second driving information of the sample vehicle at a first time from a sensor; a processing unit that acquires third position information of the sample vehicle and second intermediate variables used in a process of determining the third position information based on the second driving information; a second acquisition unit that acquires fourth position information of the sample vehicle at the first time from a positioning device; a determination unit that determines a training sample set including the third position information, the fourth position information, and the second intermediate variable; a training unit for training the accuracy estimation model based on the training sample set.

17. 15. An electronic device comprising a processor and a memory having instructions stored thereon, the processor, when executing the instructions, performing a method according to any one of claims 1 to 14.

18. A computer program which, when executed by an electronic device, causes the electronic device to carry out a method according to any one of claims 1 to 14.

Citation Information

Patent Citations

  • Data processing method, device and equipment, and storage medium

    CN112161633A

  • GNSS / INS integrated navigation method based on Elman neural network online learning assistance

    CN112505737A

  • Present position detection apparatus, map display device and present position detecting method

    JP2007232690A

  • Driving support device and driving support method

    JP2018040693A

  • Vehicle's self-position estimating device

    JP2020003463A