A vehicle positioning method in a tunnel group scenario

Abstract

The application relates to a vehicle positioning method in a tunnel group scene and belongs to the technical field of vehicle positioning. The method is divided into the following three cases: a vehicle outside a tunnel: when a GPS signal exists, vehicle positioning is carried out by using GPS data, IMU data and visual odometer data, which is referred to as a G-I-V positioning method; when the GPS signal is weak, that is, when the GPS signal loss time is less than a time interval, vehicle positioning is carried out by using IMU data and visual odometer data, which is referred to as a V-I positioning method; and a vehicle in a tunnel: when the GPS signal loss time is greater than the time interval, that is, when it is judged that the vehicle enters the tunnel, vehicle positioning is carried out by using pseudolite data, IMU data and visual odometer data, which is referred to as a PL-V-I positioning method. The application can improve the accuracy and robustness of a vehicle positioning system in a tunnel scene.

Description

Technical Field

[0001] This invention belongs to the field of vehicle positioning technology and relates to a vehicle positioning method in a tunnel group scenario. Background Technology

[0002] Real-time high-precision positioning technology has wide applications in daily life, such as autonomous driving and intelligent transportation. Currently, the main positioning technologies include GPS positioning and inertial navigation positioning. However, each individual navigation and positioning method has its own set of problems.

[0003] GPS has become the most widely used navigation system due to its advantages such as all-weather operation, global coverage, low cost, high accuracy, and the fact that errors do not accumulate over time. The GPS system consists of three main parts: space segment, ground segment, and user segment. By processing data such as pseudorange between the receiver and satellites and satellite orbital parameters, it can obtain information such as the three-dimensional coordinates, velocity, and time of the user's current geographical location. However, its data update frequency is relatively low, making continuous positioning difficult, and the signal may be lost or damaged due to interference factors such as obstruction, multipath propagation, and severe weather.

[0004] Inertial Navigation Systems (INS) operate based on Newton's laws of motion. They rely on accelerometers in an Inertial Measurement Unit (IMU) measuring the vehicle's acceleration and gyroscopes measuring its angular velocity. A computer processes these measurements to determine the vehicle's attitude, velocity, and position. The IMU is an independent system used to calculate a vehicle's position, velocity, and attitude. This system is highly autonomous and stealthy, and boasts advantages such as high short-term navigation accuracy, good stability, and fast data update rates. However, errors in inertial devices accumulate over time, and initial alignment is required before each use.

[0005] Image information can be used to correct drift problems in inertial navigation. Therefore, effectively utilizing the complementarity of visual and inertial information to achieve an accurate and robust positioning and navigation system has become a hot topic and a challenge in the field of intelligent vehicle research.

[0006] Therefore, there is an urgent need for a positioning method that can effectively integrate GPS, IMU, and visual odometry data. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide a vehicle positioning method in tunnel scenarios. Based on Kalman filtering, it integrates GPS, IMU, and visual odometry data for vehicle positioning outside the tunnel; and incorporates pseudo-satellite data for vehicle positioning inside the tunnel, thereby improving the accuracy and robustness of the vehicle positioning system in tunnel scenarios.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A vehicle localization method in a tunnel complex scenario is divided into the following three cases:

[0010] Vehicles outside the tunnel:

[0011] (1) When GPS signal is present, vehicle positioning is performed using GPS data, IMU data and visual odometry data, referred to as GIV positioning method;

[0012] (2) When the GPS signal loss time is less than the time interval Δt, vehicle positioning is performed using IMU data and visual odometry data, referred to as the VI positioning method.

[0013] Vehicle inside the tunnel:

[0014] (3) When the GPS signal loss time is greater than the time interval Δt, it is determined that the vehicle has entered the tunnel. The vehicle is located using pseudo-satellite data, IMU data and visual odometry data, referred to as the PL-VI positioning method.

[0015] Furthermore, the GIV positioning method specifically involves: first, combining the position and velocity information output by the inertial navigation system and the visual odometry system, and then calculating the position and velocity using Kalman filtering by the inertial navigation platform and the visual odometry; next, combining the position and velocity output by the inertial navigation system with the ephemeris information of the satellite system, and then calculating the pseudorange and pseudorange rate by the inertial navigation platform; then, subtracting the pseudorange and pseudorange rate measured by the satellite navigation system from these values, and using the difference as the observation of the filtering model; then, feeding back the error estimated by the error model to the inertial navigation system; and finally, using the corrected inertial system calculation result as the output of the integrated navigation system.

[0016] Furthermore, the VI positioning method specifically involves: fusing the vehicle speed v1 calculated by the IMU and the vehicle speed v2 calculated by the visual odometry using Kalman filtering to obtain the vehicle speed v; using the previous known point as the reference feature and historical data to train the position information prediction network; then using the fused vehicle speed v, the reference feature, the accelerometer feature of the IMU, and the gyroscope feature as the input to the position information prediction network; finally, the network outputs the position information of the next point.

[0017] Visual and inertial navigation information have inherent complementary characteristics in localization. The IMU can provide its own three-axis acceleration and angular velocity, and use pre-integration technology to estimate its own attitude, effectively solving the scale uncertainty of the system when only a monocular camera is available. At the same time, it can improve the accuracy and robustness of the system by using pre-integration to increase the system's constraint terms in the case of visual tracking errors or loss caused by dynamic objects, sparse features, and blurred images.

[0018] Furthermore, the PL-VI positioning method specifically involves: firstly, using the relationship between pseudorange measurements and user coordinates, the user's spatial coordinate position is obtained; then, the calculated user coordinates are combined with the data obtained by the VI method, and fused using Kalman filtering and a loose combination method.

[0019] The user's spatial coordinate position expression is as follows:

[0020]

[0021] Where, ρ i Let x be the pseudorange value of pseudosatellite i, (x i ,y i ,z i Let (x) be the coordinates of pseudosatellite i, and (x) be the coordinates of pseudosatellite i. u ,y u ,z u ) represents the user receiver coordinates, Δt u Let c be the clock delay error of the user receiver, and c be the speed of light.

[0022] The beneficial effects of this invention are as follows: This invention considers three situations: strong, weak, and completely lost GPS signals inside and outside the tunnel. It uses the fusion of different data to achieve vehicle positioning, thereby improving the accuracy and robustness of the vehicle positioning system inside and outside the tunnel.

[0023] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0024] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0025] Figure 1 Flowchart of GIV localization method;

[0026] Figure 2 Flowchart for VI positioning method;

[0027] Figure 3 Flowchart of PL-VI positioning method;

[0028] Figure 4 A schematic diagram of the tunnel's overall positioning;

[0029] Figure 5 This is a flowchart of the overall positioning method of the present invention. Detailed Implementation

[0030] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0031] Please see Figures 1-5 This invention proposes a vehicle positioning method in tunnel cluster scenarios. When GPS signals are present, Kalman filtering is used to fuse data from GPS, IMU, and visual odometry. The main process is as follows: outside the tunnel, a compact combination fusion method is used to fuse the three sets of data.

[0032] This method (i.e., the GIV positioning method) first combines the position and velocity information output by the inertial navigation system and the visual odometry system. The inertial navigation platform and the visual odometry system then calculate the position and velocity using Kalman filtering. Next, the position and velocity output by the inertial navigation system are combined with the ephemeris information from the satellite system. The inertial navigation platform calculates the pseudorange and pseudorange rate, and then subtracts them from the pseudorange and pseudorange rate measured by the satellite navigation system. This difference is used as the observation of the filtering model. Then, the error estimated by the error model is fed back to the inertial navigation system. Finally, the corrected inertial system calculation result is used as the output of the integrated navigation system.

[0033] When GPS signal acquisition frequency is low, if two adjacent GPS data points have times T0 and T1 respectively, with a time interval of Δt, only IMU data and visual odometry data exist during this period. In this case, this invention utilizes the VI method for vehicle localization. Specifically, the vehicle speed v1 calculated by the IMU and the vehicle speed v2 calculated by the visual odometry are fused using Kalman filtering to obtain the vehicle speed v. Using a known point as a reference feature, a location information prediction network is trained using historical data. The fused vehicle speed v, along with the reference feature, IMU accelerometer features, and gyroscope features, are then used as inputs to the location information prediction network. Finally, the network outputs the location information for the next point, thereby improving the localization frequency.

[0034] When the GPS signal loss time interval is greater than Δt, it is determined that the vehicle has entered a tunnel. At this point, the PL-VI method is used for vehicle positioning. This involves using pseudo-satellites to locate the vehicle within the tunnel. Pseudo-satellites, as an enhancement and supplement to the GNSS system, can effectively overcome the inherent defects of the GNSS system itself, significantly improve the integrity and reliability of existing GNSS navigation systems for train positioning, expand the application scenarios of GNSS navigation systems, and are conducive to the further development of integrated navigation systems, improving the safety performance of train control systems. GNSS / PL integrated navigation technology comprehensively utilizes the resources of both navigation systems, complementing each other's advantages and meeting current positioning requirements for high precision, high availability, and strong anti-interference capabilities, facilitating seamless vehicle positioning. The main process is as follows: pseudo-satellites are installed inside the tunnel to ensure that the vehicle can receive data from four pseudo-satellites at any position within the tunnel, and the pseudo-satellite positions are accurately calibrated. The distance is obtained by multiplying the time delay of pseudo-satellite signal transmission and reception by the speed of light. However, in actual pseudo-satellite systems, there is a certain deviation between the time systems of the pseudo-satellites and the user receiver and the standard time, resulting in a certain error between the pseudorange and the actual distance. The time when a pseudosatellite transmits a signal is measured by the pseudosatellite clock and is called the signal transmission time. The time when the user receives the signal is obtained from the receiver and is called the signal reception time. Neither the pseudosatellite nor the user receiver's clocks are high-precision atomic clocks, causing them to be out of sync and have some error compared to system time. The standard clock for system time can generally be understood as a high-precision atomic clock, but it is expensive and generally unsuitable for use in real pseudosatellite systems due to cost issues. In this invention, it is only used as a reference time for the theoretical analysis of pseudorange calculations. Assume the time when the user receiver receives the signal is t. r The corresponding system time is t, and the receiver clock leads the system time by Δt. r , representing the clock difference of the receiver clock, is an unknown quantity. Therefore, the signal reception time t... r The relationship between the system time t and the system time t can be expressed by formula (1):

[0035] t r =t+Δt r (1)

[0036] Although the clock inside the pseudosatellite system is synchronized, it still has an error compared to the system time. Let t be the time when the pseudosatellite transmits its signal. s The clock bias of the pseudosatellite clock is Δt. s , is a known quantity, and the actual time used in the navigation signal transmission process is τ. Then the relationship between the signal transmission time and the system time t can be expressed by equation (2).

[0037] t s =t-τ+Δt s (2)

[0038] When calculating the pseudorange, the difference between the signal reception time and the signal transmission time is multiplied by the speed of light to obtain the pseudorange value, as shown in formula (3):

[0039] ρ=c*(t r -t s (3) Substituting formula (1) and formula (2) into the equation, we get:

[0040] ρ=c*(Δt r -Δt s )+r (4)

[0041] Where c is the speed of light. Equation (4) shows that the pseudorange includes the true distance r between the receiver and the pseudosatellite, the pseudosatellite clock error, and the receiver clock error. However, the satellite pseudorange also includes other interference terms, whose errors are represented by ∈. Therefore, the pseudorange observation equation is:

[0042] ρ=c*(Δt r -Δt s )+r+∈ (5)

[0043] Assume the user receiver coordinates are (x u ,y u ,z u The pseudo-satellite coordinates are (x... i ,y i ,z i If the distance between the pseudo-satellite and the user receiver is:

[0044]

[0045] Substituting formula (6) into formula (5), and temporarily omitting the error ∈ for ease of calculation, we obtain the relationship between the pseudorange measurement value and the user's coordinates. By combining these equations into a nonlinear system, we can solve the system to obtain the user's spatial coordinate position, as shown in formula (7).

[0046]

[0047] It contains four unknowns, namely the user receiver coordinates (x... u ,y u ,z u ) and Δt u , where Δt u The user receiver clock delay error is considered. The calculated user coordinates are combined with data obtained through the VI method, and then fused using a Kalman filter and a loose combination method to achieve positioning within the tunnel.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A vehicle positioning method in a tunnel complex scenario, characterized in that, This method can be divided into the following three cases: Vehicles outside the tunnel: (1) When GPS signal is available, vehicle positioning is performed using GPS data, IMU data and visual odometry data, referred to as GIV positioning method; (2) When the GPS signal is weak, that is, the GPS signal loss time is less than the time interval. At that time, vehicle positioning is performed using IMU data and visual odometry data, referred to as the VI positioning method; Vehicle inside the tunnel: (3) When the GPS signal loss time is greater than the time interval When a vehicle enters a tunnel, it is determined that the vehicle has entered the tunnel. The vehicle is located using pseudosatellite data, IMU data, and visual odometry data, which is referred to as the PL-VI positioning method. The VI positioning method is as follows: the vehicle speed v1 calculated by the IMU and the vehicle speed v2 calculated by the visual odometry are fused by Kalman filtering to obtain the vehicle speed v. The previous known point is used as the reference feature, and historical data is used to train the position information prediction network. Then, the fused vehicle speed v, the reference feature, the accelerometer feature of the IMU, and the gyroscope feature are used as the input of the position information prediction network. Finally, the network outputs the position information of the next point.

2. The vehicle positioning method in a tunnel group scenario according to claim 1, characterized in that, The GIV positioning method is as follows: First, the position and velocity information output by the inertial navigation system and the visual odometry system are combined, and the position and velocity are calculated by the inertial navigation platform and the visual odometry through Kalman filtering; then, the position and velocity output by the inertial navigation system are combined with the ephemeris information of the satellite system, and the pseudorange and pseudorange rate are calculated by the inertial navigation platform; then, the difference is calculated with the pseudorange and pseudorange rate measured by the satellite navigation system, and the difference is used as the observation of the filtering model. Then, the error estimated by the error model is fed back to the inertial navigation system; finally, the corrected calculation results of the inertial system are used as the output of the integrated navigation system.

3. The vehicle positioning method in a tunnel group scenario according to claim 1, characterized in that, The PL-VI positioning method is as follows: First, the spatial coordinates of the user are obtained by using the relationship between pseudorange measurement values ​​and user coordinates; then, the calculated user coordinates are combined with the data obtained by the VI method, and fused using Kalman filtering and a loose combination method.

4. The vehicle positioning method in a tunnel group scenario according to claim 3, characterized in that, The expression for the user's spatial coordinates is: in, pseudo-satellite i The pseudo-range value, pseudo-satellite i coordinates For the user receiver coordinates, For user receiver clock delay error, It is the speed of light.

Images