WO2022033867A1

WO2022033867A1 - Method for positioning with lane-level precision using road side unit

Info

Publication number: WO2022033867A1
Application number: PCT/EP2021/071020
Authority: WO
Inventors: Feng Wen; Yichen YAN
Original assignee: Continental Automotive Gmbh
Priority date: 2020-08-11
Filing date: 2021-07-27
Publication date: 2022-02-17
Also published as: CN111915921A

Abstract

A method for positioning with lane-level precision using a road side unit comprises: a vehicle acquiring a road safety message corresponding to a host vehicle from a broadcast of the road side unit; the vehicle extracting host vehicle state information detected by the road side unit from the road safety message; using Kalman filtering to perform filtering according to a host vehicle state estimated value at the previous moment in combination with a positioning value acquired by an on- board GNSS at the present moment and host vehicle state information detected by the road side unit at the present moment, so as to obtain a host vehicle state estimated value at the present moment, and setting a host vehicle state estimated value at an initial moment as a positioning value acquired by the on-board GNSS at that moment, the host vehicle state estimated value and host vehicle state information each comprising vehicle position and speed; continuing to use Kalman filtering to obtain a host vehicle state estimated value at each moment, and taking a position in the host vehicle state estimated value at each moment to be the present position of the host vehicle at that moment. The positioning method can meet the requirements of positioning with lane-level precision and has low costs.

Description

Method for positioning with lane-level precision using road side unit

Technical Field

[0001] The present invention relates to a vehicle positioning method, in particular to a method for positioning with lanelevel precision using a road side unit.

Background Art

[0002] In autonomous driving and V2X application scenarios, it is currently necessary to obtain positioning precision at the lane level in order to meet scenario requirements. The method by which a vehicle acquires positioning with high precision at the lane level is generally to perform augmentation by two types of method, based on GNSS: one type is satellite-based or road-based augmentation of the GNSS system, i.e. the use of RTK differential positioning; the other type is to use various types of on-board sensor for assistance to perform inertial measurement (Inertial Measurement Unit) , e.g. a gyroscope, accelerometer, wheel speed sensor, etc.

[0003] With RTK augmentation of the GNSS system, it is necessary to use widely deployed RTK base stations, and use cellular communication links to transmit RTK augmentation information, and perform correction calculation at the terminal or server end; the cost of the entire service is high, and in the case of urban roads, there is also the problem of error enlargement caused by the multi-path effect. IMU-based high-precision positioning provides limited improvement in absolute accuracy. Furthermore, if multiple sensor inputs are being relied on, it is still difficult to achieve a positioning standard with lane-level precision. Summary of the Invention

[ 0004 ] The problem solved by the present invention is the provision of a positioning method that can achieve lane-level precision at a low cost .

[ 0005 ] The method for pos itioning with lane-level precision using a road side unit as provided in the present invention comprises : a vehicle acquiring a road safety message corresponding to a host vehicle from a broadcast of the road side unit ; the vehicle extracting host vehicle state information detected by the road side unit from the road safety message ; using Kalman filtering to perform filtering according to a host vehicle state estimated value at the previous moment in combination with a positioning value acquired by an onboard GNSS at the present moment and host vehicle state information detected by the road side unit at the present moment , so as to obtain a host vehicle state estimated value at the present moment , and setting a host vehicle state estimated value at an initial moment as a positioning value acquired by the on-board GNSS at that moment , the host vehicle state estimated value and host vehicle state information each comprising vehicle position and speed; continuing to use Kalman filtering to obtain a host vehicle state estimated value at each moment , and taking a position in the host vehicle state estimated value at each moment to be the present position of the host vehicle at that moment .

[ 0006 ] Compared with the prior art , the above-mentioned solution has the following advantages : using an existing RSU broadcast mechanism in the V2X system, position information meeting lane-level precision is obtained with the aid of measurement of vehicles which are traf fic participants by the

RSU, and measurement values are fused with GNSS positioning values to estimate positioning information with lane-level precision . Furthermore , in the process of data fusion, Kalman filtering is used to further increase the confidence of data estimation . Since an existing system is used and there is no need for RTK service registration, the solution described above also has low costs .

Brief Description of the Drawings

[ 0007 ] Fig . 1 is a schematic implementation diagram of an embodiment of the positioning method of the present invention .

[ 0008 ] Fig . 2 is a schematic diagram of a vehicle communication system architecture suitable for the present invention .

Detailed Description of the Invention

[ 0009 ] Details are given in the following description so that those skilled in the art can understand the invention more comprehensively . However, it is obvious to those skilled in the art that the invention can be reali zed without some of these details . In addition, it should be understood that the invention is not limited to the specific embodiments described . On the contrary, the invention can be implemented by considering using any combination of the following characteristics and elements , no matter whether they relate to di f ferent embodiments . Furthermore , the aspects , features , embodiments and advantages below are merely explanatory, and should not be regarded as key elements or definitions of the claims , unless clearly stated in the claims .

[ 0010 ] In the "Cooperative Intelligent Transportation System/Vehicle Communication System Application Layer and Application Layer Data Exchange Standard T/CSAE 53-2017" published on 18 September 2017 by the China Society o f Automotive Engineers (hereinafter abbreviated as the "V2X standard" , a type of Road Safety Message (RSM) was mentioned in relation to communication between an on-board unit and a road side unit (V2 I ) . The road safety message is defined as follows : The road side unit obtains real-time state information of surrounding traf fic participants ( e . g . surrounding vehicles , non-motor vehicles , pedestrians , etc . ) by means o f a corresponding detection method possessed by the road side itsel f , and organi zes this information into a format defined by the message body in question, to serve as basic safety state information of these traffic participants , for broadcast to the surrounding vehicles , to support associated applications of these vehicles . On this basis , the inventor of the present invention proposes that the abovementioned road safety message mechanism to which V2 I communication in the V2X standard relates can be utili zed to assist a vehicle in achieving precise positioning at the lane level , by inserting a vehicle position measured by a road side unit into a road safety message .

[ 0011 ] On this basis , the method for positioning with lanelevel precision proposed in the present invention compri ses : a vehicle acquiring an RSM message corresponding to a host vehicle from a broadcast of a road side unit ; the vehicle extracting host vehicle state information detected by the road side unit from the RSM message ; using Kalman filtering to perform filtering according to a host vehicle state estimated value at the previous moment in combination with a positioning value acquired by an on-board GNSS at the present moment and host vehicle state information detected by the road side unit at the present moment , so as to obtain a host vehicle state estimated value at the present moment , setting an estimated value at an initial moment as a positioning value acquired by the on-board GNSS , the host vehicle state estimated value and host vehicle state information each comprising vehicle position and speed; continuing to use Kalman filtering to obtain a host vehicle state estimated value at each moment , and taking the host vehicle state estimated value at each moment to be the present position of the host vehicle at that moment .

[ 0012 ] The process of implementation of the positioning method of the present invention is explained further below by means of particular embodiments . Referring to Fig . 1 , multiple road side units are disposed along one or two sides of a road, each road side unit being able to cover a portion of a lane range , detect vehicles within the lane range and broadcast road safety messages . Each vehicle must be equipped with an on-board unit ( OBU) in order to communicate with the road side units . Referring to Fig . 2 , in a vehicle communication architecture suitable for the present invention, an on-board unit 200 comprises : a radio communication sub-system 203 , which receives and transmits air signals ; a global navigation satellite system ( GNSS ) 201 , for providing information such as vehicle position, direction, speed and time ; an on-board device processing unit 202 , which runs a program to generate air signals to be transmitted and processes received air signals ; and an antenna, for receiving and transmitting RF signals . The on-board unit 200 is connected via an interface to an application electronic control unit 300 , in which a program is run to reali ze an application of a vehicle communication system, and reminders are issued to a driver by images , sound or vibration, etc . via a human-machine interface 500 . In some embodiments , the application electronic control unit and the on-board device processing unit may be implemented in a single physical device . [0013] Continuing to refer to Fig. 1, the road side units along the road, e.g. a road side unit 100, will detect vehicles within the range of coverage of each sensor thereof (e.g. camera, laser radar, millimetre-wave radar, etc.) in order to acquire vehicle state information. The vehicle state information detected by the road side unit may comprise vehicle position and speed, and may also comprise information such as vehicle direction. The vehicle state information detected by the road side unit will be bound to a vehicle identifier, and then broadcasted in a manner complying with the road safety message type in the V2X standard.

[0014] Veh icles travelling on the road, e.g. a vehicle 1 and a vehicle 2, acquire GNSS positioning values via their respective on-board units. When vehicle 1 and vehicle 2 enter the broadcast communication range of the road side unit 100, they receive the abovementioned road safety message via their respective on-board units. Next, vehicle 1 and vehicle 2 extract host vehicle state information from the road safety message via their respective on-board units. For example, when the vehicle identifier is a license plate, vehicle 1 and vehicle 2 each extract state information corresponding to their license plates. The extracted host vehicle state information will undergo data fusion with the host vehicle GNSS positioning value, in order to estimate positioning information with lanelevel precision at the present moment.

[0015] Specifically, Kalman filtering is used to perform data fusion. As is well known, Kalman filtering is used in any dynamic system containing indeterminate information, to make a well-founded prediction for the next course of the system (positioning information conforming to lane-level precision) , and even when accompanied by various types of interference, Kalman filtering can always indicate the situation that is really happening . The use of Kalman filtering in a continuous ly varying system (with continuous variation in position as the vehicle is travelling) is ideal ; it has the advantage of occupying a small amount of internal memory ( apart from the previous state quantity, other historical data need not be retained) , and is also very fast , so is well-suited to use in real-time problems and embedded systems .

[ 0016 ] In this embodiment , standard Kalman filtering is used to perform filtering according to a host vehicle state estimated value at the previous moment in combination with the GNSS positioning value at the present moment and host vehicle state information detected by the road side unit at the present moment , so as to obtain a host vehicle state estimated value at the present moment . Since the on-board unit will continuously receive GNSS positioning values and host vehicle state information detected by the road side unit , the on-board unit also continuously uses Kalman filtering to obtain a host vehicle state estimated value at each moment , and takes the position in the host vehicle state estimated value at each moment to be the present position of the host vehicle at that moment . Since Kalman filtering has very high confidence , when the position in the host vehicle state estimated value at each moment is taken to be the present position of the host vehicle at that moment , the requirements of positioning with lanelevel precision can also be met .

[ 0017 ] To facilitate understanding, let the host vehicle state estimated value at each moment ( including position and vehicle speed) be Xo, Xi, X2 X_n-i, X_n, let the GNSS positioning value at each moment ( longitude and latitude coordinates ) be Po, Pi, P2 Pn-iz Pnz and let the host vehicle state information detected by the road side unit at each moment ( including position and vehicle speed) be Xi" , X₂ ' > X_n-i" , X_n" . Let X_o= Po, i.e. the host vehicle state estimated value at an initial moment is the GNSS positioning value at that moment. Thus, the host vehicle state estimated value at moment n can be expressed by the following formula:

X_n = Karlmann(X_n-1, P_n,X^), n=l,2,3. N

[0018] The vehicle speed at each moment can be calculated on the basis of the GNSS positioning values (the difference between two consecutive positioning values) , and may also come from an inertial measurement unit. For example, it is obtained from a wheel speed sensor, or/and a preliminary input is obtained from the wheel speed sensor, and an acceleration measured by an accelerometer is then used to perform a correction in order to obtain a more accurate vehicle speed.

X_n = Karlmann(X_n-1, P_n,X^, M_n), n=l,2,3. N where M_n denotes data coming from the inertial measurement unit (IMU) at moment n.

[0019] Confidence may also be used when perfoming Kalman filtering. Specifically, confidence is attached to both the GNSS positioning value at each moment and the host vehicle state information detected by the road side unit at each moment. The influence of information with higher confidence on the Kalman filtering result has a greater weighting. That is to say, the Kalman filtering result is closer to information with higher confidence. Since the position information detected by the road side unit has higher confidence than the GNSS positioning value, the position obtained by the Kalman filtering result at each moment is closer to the position information detected by the road side unit. [ 0020 ] Since positioning information conforming to lanelevel precision at each moment is obtained by Kalman filtering, other applications in the V2X system can be further supported, e . g . FCW ( Forward Collision Warning) mentioned in the V2X standard . FCW means that when the host vehicle is travelling in a lane and is in danger of having a rear-end collision with a distant vehicle in the same lane directly in front , the FCW application will issue a warning to the driver of the host vehicle . Since the j udgement logic of FCW is based on whether the host vehicle and the distant vehicle are in the same lane , the positioning value conforming to lane-level precis ion that is obtained by the method described above can greatly increase the accuracy of the FCW application .

[ 0021 ] Although the invention is already disclosed above with preferred embodiments , the invention is not limited to the preferred embodiments . Any change or modi fication made by those skilled in the art without departing from the spirit and scope of the present invention should fall within the scope of protection of the invention . Therefore , the scope of protection of the invention should be subj ect to the scope defined by the claims .

Claims

1 . A method for positioning with lane-level precision using a road side unit , characteri zed by comprising : a vehicle acquiring a road safety message corresponding to a host vehicle from a broadcast of the road side unit ; the vehicle extracting host vehicle state information detected by the road side unit from the road safety message ; using Kalman filtering to perform filtering according to a host vehicle state estimated value at the previous moment in combination with a positioning value acquired by an on-board GNSS at the present moment and host vehicle state information detected by the road side unit at the present moment, so as to obtain a host vehicle state estimated value at the present moment , and setting a host vehicle state estimated value at an initial moment as a positioning value acquired by the on-board GNSS at that moment , the host vehicle state estimated value and host vehicle state information each comprising vehicle position and speed; continuing to use Kalman filtering to obtain a host vehicle state estimated value at each moment , and taking a position in the host vehicle state estimated value at each moment to be the present position of the host vehicle at that moment .

2 . The method for positioning with lane-level precision as claimed in claim 1 , characteri zed in that a vehicle speed is obtained in combination with one or more inertial measurement value .

3 . The method for positioning with lane-level precision as claimed in claim 2 , characteri zed in that the inertial measurement value comes from one or more of the fol lowing sensors : an accelerometer and a wheel speed sensor .

4. The method for positioning with lane-level precision as claimed in claim 1, characterized in that a standard Kalman filter is used.

5. The method for positioning with lane-level precision as claimed in claim 1, characterized in that the road safety message broadcasted by the road side unit at least comprises a vehicle identifier and corresponding position information; and the vehicle extracts a position message corresponding to the host vehicle on the basis of the vehicle identifier.

6. The method for positioning with lane-level precision as claimed in claim 5, characterized in that the vehicle identifier is a license plate.

7. The method for positioning with lane-level precision as claimed in claim 5, characterized in that the road safety message broadcasted by the road side unit further comprises a vehicle direction.