Lane line based multi-source fusion positioning method and system

Abstract

This invention relates to a multi-source fusion localization method and system based on lane lines, belonging to the field of autonomous driving technology. It acquires laser SLAM localization data and lane line data, calculates the difference between the current frame and the previous frame for both, determines whether laser SLAM jumps exceed a set threshold based on the difference, and determines the vehicle's deviation direction based on the difference in lane line data. If the difference exceeds the set threshold, the laser SLAM jump direction is further determined, and its consistency with the vehicle's deviation direction is checked. If they are inconsistent, the fusion weight of the laser SLAM localization data is reduced. When a laser SLAM jump is detected, its jump direction is checked for consistency with the vehicle deviation direction determined by the lane line data, and the fusion weight of the laser SLAM is reduced to minimize the impact of laser SLAM jumps on the vehicle and ensure normal vehicle operation.

Description

Technical Field

[0001] This invention relates to a multi-source fusion localization method and system based on lane lines, belonging to the field of autonomous driving technology. Background Technology

[0002] High-precision positioning is one of the essential conditions for ensuring the safe operation of autonomous vehicles. Currently, the high-precision positioning results used by autonomous vehicles are typically fused from multiple sources. Multi-source fusion positioning refers to improving the accuracy and robustness of the positioning system by fusing data from different sensors and utilizing the complementarity of these data. In autonomous driving technology, the fused positioning data sources usually include laser SLAM (Simultaneous Localization and Mapping) positioning results, integrated navigation positioning results, and odometer results. However, in locations with significant environmental changes, laser SLAM positioning results may exhibit abrupt changes, thus affecting vehicle operation; satellite positioning results are unusable in occluded environments; and over time, odometer-based estimations show a gradual drift in positioning results.

[0003] Chinese invention patent application CN114199259A discloses a multi-source fusion navigation and positioning method based on motion state and environmental perception. This method acquires the positions of feature points in an image using a visual camera and directly provides the visual positioning result. While this method improves positioning accuracy, positioning anomalies still occur in low-light or poor lighting conditions.

[0004] Chinese invention patent application CN116660961A discloses a high-precision positioning method based on multi-source fusion technology. This method combines laser SLAM and integrated navigation, solving the problem of abnormal positioning results under poor lighting conditions. However, in occluded environments or environments with unclear features, laser SLAM is prone to jumps, which can lead to positioning anomalies and fail to support the requirements of autonomous driving. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-source fusion positioning method and system based on lane lines to solve the problem that laser SLAM jumps during fusion positioning affect positioning accuracy.

[0006] To achieve the above objectives, the present invention includes:

[0007] The present invention provides a multi-source fusion localization method based on lane lines, comprising: acquiring laser SLAM localization data and lane line data; calculating the difference components between the current frame and the previous frame of the laser SLAM localization data and lane line data respectively; determining whether a laser SLAM jump has occurred based on the difference components of the laser SLAM; determining the jump direction of the laser SLAM when a laser SLAM jump occurs; further determining the vehicle offset direction based on the difference components of the lane line data; and performing a consistency judgment between the laser SLAM jump direction and the vehicle offset direction. If they are inconsistent, reducing the fusion weight of the laser SLAM localization data.

[0008] Furthermore, determining whether a jump has occurred in laser SLAM by using the difference component between the current frame and the previous frame includes the following steps: calculating the difference component between the current frame and the previous frame of laser SLAM positioning data; converting the difference component to the vehicle coordinate system; and determining whether the lateral difference component between the current frame and the previous frame of laser SLAM in the vehicle coordinate system is greater than a set threshold.

[0009] Furthermore, the difference between the current frame and the previous frame of laser SLAM positioning data is calculated using the following formula:

[0010]

[0011] Where Δxslam is the lateral difference component of laser SLAM in the global coordinate system; Δyslam is the longitudinal difference component of laser SLAM in the global coordinate system; xslam current The lateral component of the current frame's laser SLAM in the global coordinate system; yslam currten xslam represents the longitudinal component of the current frame's laser SLAM in global coordinates. pre The lateral component of the previous frame's laser SLAM in the global coordinate system; yslam pre This represents the longitudinal component of the previous frame of laser SLAM in the global coordinate system.

[0012] Furthermore, the laser SLAM differential components are transformed to the vehicle coordinate system using the following formula:

[0013]

[0014] Where Δxslamb is the lateral difference component of laser SLAM in the vehicle coordinate system; Δyslamb is the longitudinal difference component of laser SLAM in the vehicle coordinate system. This is the vehicle's heading angle.

[0015] Furthermore, when entering an area where neither laser SLAM positioning data nor integrated navigation data is available, the fusion weight of integrated navigation data and laser SLAM positioning data is reset to 0, and the laser SLAM positioning data at this moment is used as the initial value of speed odometer to initialize the speed odometer. The speed odometer is used to recursively derive the positioning result. When lane line data is available, the speed odometer recursive positioning result is fused with the lane line data to correct the speed odometer recursive positioning result.

[0016] A multi-source fusion positioning system based on lane lines includes a processor, characterized in that the processor executes a computer program to perform the following steps:

[0017] Acquire laser SLAM positioning data and lane line data, calculate the difference components between the current frame and the previous frame of laser SLAM positioning data and lane line data respectively, and determine whether laser SLAM has a jump based on the difference components; when laser SLAM has a jump, determine the jump direction of laser SLAM, and also determine the vehicle deflection direction based on the difference components of lane line data, and check the consistency between the laser SLAM jump direction and the vehicle deflection direction. If they are inconsistent, reduce the fusion weight of laser SLAM positioning data.

[0018] Furthermore, determining whether a jump has occurred in laser SLAM by using the difference component between the current frame and the previous frame includes the following steps: calculating the difference component between the current frame and the previous frame of laser SLAM positioning data; converting the difference component to the vehicle coordinate system; and determining whether the lateral difference component between the current frame and the previous frame of laser SLAM in the vehicle coordinate system is greater than a set threshold.

[0019] Furthermore, the difference between the current frame and the previous frame of laser SLAM positioning data is calculated using the following formula:

[0020]

[0021] Where Δxslam is the lateral difference component of laser SLAM in the global coordinate system; Δyslam is the longitudinal difference component of laser SLAM in the global coordinate system; xslam current The lateral component of the current frame's laser SLAM in the global coordinate system; yslam currten xslam represents the longitudinal component of the current frame's laser SLAM in global coordinates. pre The lateral component of the previous frame's laser SLAM in the global coordinate system; yslam pre This represents the longitudinal component of the previous frame of laser SLAM in the global coordinate system.

[0022] Furthermore, the laser SLAM differential components are transformed to the vehicle coordinate system using the following formula:

[0023]

[0024] Where Δxslamb is the lateral difference component of laser SLAM in the vehicle coordinate system; Δyslamb is the longitudinal difference component of laser SLAM in the vehicle coordinate system. This is the vehicle's heading angle.

[0025] Furthermore, when entering an area where neither laser SLAM positioning data nor integrated navigation data is available, the fusion weight of integrated navigation data and laser SLAM positioning data is reset to 0, and the laser SLAM positioning data at this moment is used as the initial value of speed odometer to initialize the speed odometer. The speed odometer is used to recursively derive the positioning result. When lane line data is available, the speed odometer recursive positioning result is fused with the lane line data to correct the speed odometer recursive positioning result.

[0026] The beneficial effects of this invention are as follows: when a laser SLAM jump is detected, the jump direction is consistent with the vehicle offset direction determined by the lane line data, the laser SLAM fusion weight is reduced, the impact of laser SLAM jump on the vehicle is reduced, and the positioning is more accurate. Attached Figure Description

[0027] Figure 1 This is a flowchart of the multi-source fusion localization method based on lane lines provided in an embodiment of the present invention;

[0028] Figure 2 This is a flowchart of a scheme for recursively locating using an odometer, provided by an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0030] The inventive concept of this invention is as follows: when fusing laser SLAM positioning results with other positioning data sources, the laser SLAM positioning results are detected. If the laser SLAM positioning results change abruptly, their consistency with the direction of lane line position change is determined, thereby adjusting the fusion weight of the laser SLAM positioning results in real time.

[0031] Method Example 1:

[0032] like Figure 1 The diagram shown is a flowchart of a multi-source fusion positioning method based on lane lines provided in this embodiment, including combined navigation positioning data, laser SLAM positioning data, and odometer positioning data.

[0033] Step 1: Receive laser SLAM positioning data and determine the availability of the laser SLAM positioning status. If the laser SLAM positioning status is available, proceed to the next step; if the laser SLAM positioning status is unavailable, return to continue receiving laser SLAM positioning data. Receive lane line data and determine the availability of the lane line data. If the lane line data is available, proceed to the next step; if it is unavailable, return to continue receiving lane line data.

[0034] Step 2: Calculate the difference between the current frame and the previous frame in the laser SLAM localization result:

[0035]

[0036] Δxslam: Lateral difference component of laser SLAM in global coordinate system; Δyslam: Longitudinal difference component of laser SLAM in global coordinate system; xslam current : The lateral component of the current frame's laser SLAM in the global coordinate system; yslam currten : The longitudinal component of the current frame's laser SLAM in global coordinates; xslam pre : The lateral component of the previous frame of laser SLAM in the global coordinate system; yslam pre : The longitudinal component of the previous frame of laser SLAM in the global coordinate system.

[0037] Calculate the difference between the current frame and the previous frame of lane line data, and determine the vehicle's deviation direction based on the difference between the lane lines, using the lane lines as a reference to determine whether the vehicle deviates to the left or right.

[0038] Step 3: Transform the frame difference components of laser SLAM to the vehicle coordinate system;

[0039]

[0040] Δxslamb: Lateral difference component of laser SLAM in vehicle coordinate system; Δyslamb: Longitudinal difference component of laser SLAM in vehicle coordinate system;

[0041] Since both the pitch angle θ and roll angle γ can be approximated as 0° during vehicle operation, the rotation matrix in the above formula can be directly used as the heading angle. It can be simplified, or a complete rotation matrix can be used for transformation.

[0042] Step 4: Determine whether the lateral distance of the front and rear frame difference components of the laser SLAM under the vehicle body system is greater than the set threshold. If it is greater than the set threshold, it is determined that there is a jump in the laser SLAM positioning result and proceeds to step 5 to adjust the fusion weight of the laser SLAM. If it is less than the set threshold, it proceeds normally to the fusion module.

[0043] Step 5: Determine if the direction of the laser SLAM jump is consistent with the vehicle offset direction determined by the lane lines. If they are consistent, maintain the laser SLAM fusion weights unchanged; otherwise, reduce the laser SLAM fusion weights. Specifically, when a vehicle is traveling straight along the center of the lane, if the laser SLAM positioning data differential component jumps to the right, while the lane line data differential component indicates the vehicle should still be traveling along the center of the lane, the laser SLAM jump direction is inconsistent with the vehicle offset direction determined by the lane line data. Maintaining the laser SLAM fusion weights unchanged will cause the vehicle to offset to the right. To reduce the impact of the laser SLAM jump on the vehicle, the laser SLAM fusion weights need to be reduced to ensure the vehicle continues to travel as straight as possible.

[0044] Step 6: Perform localization data fusion based on the received other data sources and laser SLAM data, along with the assigned weights, and output the fused localization result.

[0045] When the laser SLAM data positioning result does not change, one feasible approach is to fuse the laser SLAM data positioning result with other available data sources according to the system's predetermined weights, and output the positioning result to achieve vehicle positioning.

[0046] Method Example 2:

[0047] Based on Example 1, if extreme conditions are encountered that render both laser SLAM and integrated navigation data unavailable, such as tunnels with continuously similar internal scenes or changes in the surrounding environment compared to the pre-obtained point cloud data of that scene, and satellite data is lost, the following scheme of using odometry to recursively deduce positioning results is further included.

[0048] like Figure 2 The diagram shown is a flowchart of a scheme for recursively locating using an odometer, as provided in this embodiment.

[0049] Specifically, in step one, after receiving the laser SLAM positioning data, it is determined whether the vehicle has entered a specific area based on the laser SLAM positioning results. If the vehicle has not entered the specific area, the laser SLAM positioning results are output and fused normally. If the vehicle has entered the specific area, the next step is performed.

[0050] Step 2: After entering a specific area, reset the fusion weights of both the laser SLAM data and the integrated navigation data to 0, so that the total weight ratio of the data participating in the fusion is 1.

[0051] Step 3: Use the laser SLAM positioning result at the moment of entering the specific area as the initial value of the velocity odometer to initialize the velocity odometer.

[0052] Step 4: Use the speedometer to recursively calculate the current positioning result, and when lane line data is available, fuse the lane line data and the positioning result calculated by the speedometer to output the fused positioning result.

[0053] Specifically, during vehicle operation, the odometer simultaneously calculates the vehicle's longitudinal (the direction of travel along the road) and lateral (the direction perpendicular to the direction of travel) travel. In the lateral calculation, the odometer uses calibrated unit pulse distances to obtain the distance traveled by the left and right wheels per unit time. Then, based on the principle of integration, the difference between the distances traveled by the left and right wheels within a certain time is used to calculate the change in the heading angle, thus obtaining the vehicle's current heading angle and coordinates (X, Y), completing the lateral positioning calculation. However, the lateral calculation based on integration can drift as the calculation time increases. To correct this, when lane line data is detected and available, it is fused with the lane line data, making the lateral calculation more accurate and the positioning result more precise.

[0054] System Implementation Example:

[0055] A lane-line-based multi-source fusion positioning system includes a processor. To address the issue of laser SLAM jumps affecting positioning accuracy during fusion positioning, the lane-line-based multi-source fusion positioning system of this embodiment executes the lane-line-based multi-source fusion positioning method as described in Method Embodiments 1 and 2. This lane-line-based multi-source fusion positioning method has been sufficiently clearly described in Method Embodiments 1 and 2, and will not be repeated here.

Claims

1. A multi-source fusion localization method based on lane lines, characterized in that, include: Acquire laser SLAM positioning data and lane line data, calculate the difference components between the current frame and the previous frame of laser SLAM positioning data and lane line data respectively, and determine whether laser SLAM has a jump based on the difference components; when laser SLAM has a jump, determine the jump direction of laser SLAM, and also determine the vehicle deflection direction based on the difference components of lane line data, and check the consistency between the laser SLAM jump direction and the vehicle deflection direction. If they are inconsistent, reduce the fusion weight of laser SLAM positioning data.

2. The multi-source fusion localization method based on lane lines according to claim 1, characterized in that, Determining whether a jump has occurred in laser SLAM by using the difference between the current frame and the previous frame includes the following steps: calculating the difference between the current frame and the previous frame of laser SLAM positioning data; converting the difference to the vehicle coordinate system; and determining whether the lateral difference between the current frame and the previous frame of laser SLAM in the vehicle coordinate system is greater than a set threshold.

3. The multi-source fusion localization method based on lane lines according to claim 2, characterized in that, The difference between the current frame and the previous frame of laser SLAM positioning data is calculated using the following formula: Where Δxslam is the lateral difference component of laser SLAM in the global coordinate system; Δyslam is the longitudinal difference component of laser SLAM in the global coordinate system; xslam current The lateral component of the current frame's laser SLAM in the global coordinate system; yslam currten xslam represents the longitudinal component of the current frame's laser SLAM in global coordinates. pre The lateral component of the previous frame's laser SLAM in the global coordinate system; yslam pre This represents the longitudinal component of the previous frame of laser SLAM in the global coordinate system.

4. The multi-source fusion localization method based on lane lines according to claim 2, characterized in that, The laser SLAM differential component transformation to the vehicle coordinate system is performed using the following formula: Where Δxslamb is the lateral difference component of laser SLAM in the vehicle coordinate system; Δyslamb is the longitudinal difference component of laser SLAM in the vehicle coordinate system. This is the vehicle's heading angle.

5. The multi-source fusion localization method based on lane lines according to claim 1, characterized in that, When entering an area where neither laser SLAM positioning data nor integrated navigation data is available, the fusion weight of the integrated navigation data and laser SLAM positioning data is reset to 0, and the laser SLAM positioning data at this moment is used as the initial value of the speed odometer to initialize the speed odometer. The speed odometer is used to recursively derive the positioning result. When lane line data is available, the speed odometer recursive positioning result is fused with the lane line data to correct the speed odometer recursive positioning result.

6. A multi-source fusion positioning system based on lane lines, comprising a processor, characterized in that, The processor executes a computer program to perform the following steps: Acquire laser SLAM positioning data and lane line data, calculate the difference between the current frame and the previous frame of laser SLAM positioning data and lane line data respectively, determine whether the laser SLAM jump exceeds a set threshold based on the difference of laser SLAM, determine the vehicle deflection direction based on the difference of lane line data, and further determine the laser SLAM jump direction when the difference of laser SLAM exceeds the set threshold, and check the consistency between the laser SLAM jump direction and the vehicle deflection direction. If they are inconsistent, reduce the fusion weight of laser SLAM positioning data.

7. The multi-source fusion positioning system based on lane lines according to claim 6, characterized in that, Determining whether a jump has occurred in laser SLAM by using the difference between the current frame and the previous frame includes the following steps: calculating the difference between the current frame and the previous frame of laser SLAM positioning data; converting the difference to the vehicle coordinate system; and determining whether the lateral difference between the current frame and the previous frame of laser SLAM in the vehicle coordinate system is greater than a set threshold.

8. The multi-source fusion positioning system based on lane lines according to claim 7, characterized in that, The difference between the current frame and the previous frame of laser SLAM positioning data is calculated using the following formula: Where Δxslam is the lateral difference component of laser SLAM in the global coordinate system; Δyslam is the longitudinal difference component of laser SLAM in the global coordinate system; xslam current The lateral component of the current frame's laser SLAM in the global coordinate system; yslam currten xslam represents the longitudinal component of the current frame's laser SLAM in global coordinates. pre The lateral component of the previous frame's laser SLAM in the global coordinate system; yslam pre This represents the longitudinal component of the previous frame of laser SLAM in the global coordinate system.

9. The multi-source fusion positioning system based on lane lines according to claim 7, characterized in that, The laser SLAM differential component transformation to the vehicle coordinate system is performed using the following formula: Where Δxslamb is the lateral difference component of laser SLAM in the vehicle coordinate system; Δyslamb is the longitudinal difference component of laser SLAM in the vehicle coordinate system. This is the vehicle's heading angle.

10. The multi-source fusion positioning system based on lane lines according to claim 6, characterized in that, When entering an area where neither laser SLAM positioning data nor integrated navigation data is available, the fusion weight of the integrated navigation data and laser SLAM positioning data is reset to 0, and the laser SLAM positioning data at this moment is used as the initial value of the speed odometer to initialize the speed odometer. The speed odometer is used to recursively derive the positioning result. When lane line data is available, the speed odometer recursive positioning result is fused with the lane line data to correct the speed odometer recursive positioning result.

Images