A car digital key calibration data collection method based on an autonomous navigation robot

Abstract

The application relates to a kind of automobile digital key calibration data acquisition methods based on autonomous navigation robot, comprising: three-dimensional point cloud map is constructed by robot acquisition, the robot is configured real-time positioning and mapping SLAM algorithm;The three-dimensional point cloud map is segmented, and vehicle point cloud is obtained;The vehicle point cloud is fitted as a cuboid, and the coordinate conversion matrix of point cloud coordinate system to vehicle body coordinate system is obtained;Robot calibration route navigation and calibration data acquisition are carried out.The beneficial effects of the application are: the application uses robot mapping and positioning function, realizes automatic calibration for vehicle positioning system, can help developer to quickly and accurately calibrate positioning system, reduces artificial cost.

Description

Technical Field

[0001] This invention relates to the field of calibration data acquisition technology, and more specifically, to a method for acquiring calibration data for automotive digital keys based on an autonomous navigation robot. Background Technology

[0002] The user experience of a vehicle keyless entry system depends on the positioning accuracy of the digital key by the vehicle positioning system. Therefore, the development of a vehicle keyless entry system involves a large amount of digital key calibration work, which is highly repetitive and mechanical, time-consuming and labor-intensive, and greatly affects development efficiency.

[0003] With the continuous development of robotics technology, robot mapping and localization technologies have gradually matured. Utilizing robot mapping and localization functions to automatically calibrate the positioning system of car digital keys has become a reality. This helps developers quickly and accurately calibrate the positioning system, reducing labor costs. Therefore, developing an autonomous navigation robot to collect data for car digital key calibration is particularly important. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for collecting calibration data for automotive digital keys based on an autonomous navigation robot.

[0005] Firstly, a method for collecting calibration data for automotive digital keys based on an autonomous navigation robot is provided, including:

[0006] Step 1: Construct a 3D point cloud map by collecting data using a robot equipped with a real-time localization and mapping (SLAM) algorithm;

[0007] Step 2: Segment the 3D point cloud map to obtain vehicle point clouds;

[0008] Step 3: Fit the vehicle point cloud into a cuboid and obtain the coordinate transformation matrix from the point cloud coordinate system to the vehicle coordinate system;

[0009] Step 4: Perform calibration route navigation and calibration data collection for the robot.

[0010] Preferably, step 1 includes:

[0011] Step 1.1: Install the robot perception module facing to the side so that the robot can directly observe the vehicle's feature points when moving around the vehicle; the robot perception module includes an RGBD camera and a LiDAR.

[0012] Step 1.2: Park the target vehicle in an open area and place the robot at a certain distance to the right of the vehicle. Roughly estimate the coordinates of the vehicle's center in the robot's perception module coordinate system, and denote it as point q.car ;

[0013] Step 1.3: Start the robot's SLAM mapping function. Manually control the robot to walk around the vehicle for a full circle to build a 3D map. When the map building is finished, temporarily save the 3D map as 3D map point cloud data.

[0014] Preferably, step 2 includes:

[0015] Step 2.1: Record the 3D map point cloud data as a point set. And take S map Points with z-coordinates greater than the threshold height are selected as a subset S to filter out feature points on the ground.

[0016] Step 2.2: Initialize the sphere center node set Traverse the point cloud S, if C is Or click p i If the distance to any point in set C is greater than ρ, then point p... i Add point p to the set of the center nodes C; otherwise, if point p i With point c in C j The distance is less than ρ, i.e., p i In the position of c j Within a sphere with center ρ and radius ρ, we consider p... i Belongs to node c j ;

[0017] Step 2.3: Traverse the set of sphere center nodes C. If the distance between two points c1 and c2 in C is less than 2ρ, then connect p1 and p2 with an undirected and unweighted edge to construct an undirected graph G.

[0018] Step 2.4: Traverse the set of sphere center nodes C and find the relationship between them and q. car The nearest sphere center node c car ;

[0019] Step 2.5, from c car Starting from graph G, traverse graph G and find the largest connected subgraph G′ on G;

[0020] Step 2.6: Traverse the point cloud S and select all points whose distance to any node in G′ is less than ρ, and denote their set as S. c S c This is the desired vehicle point cloud.

[0021] Preferably, step 3 includes:

[0022] Step 3.1: Use a loop to enumerate the rotation angle θ within a preset value range and with a preset step size;

[0023] Step 3.2: Transfer the vehicle point cloud S cRotate the coordinates of all points in S' around the positive z-axis by θ, and denote the result as S'. c ;

[0024] Step 3.3, take x t ,x s S′ c The upper and lower α quantiles of the x-coordinate of the midpoint, y t ,x s S′ c The upper and lower α quantiles of the y-coordinate of the midpoint;

[0025] Step 3.4: Calculate the fitting evaluation index v = (x) for the rotation angle θ. t -x s ) 2 +(y t -y s ) 2 ;

[0026] Step 3.5, using variable v best To store the maximum value of v, and update v best While storing the value, save the corresponding θ. best and R c =[x s ,x t ,y s ,y t ],as well as

[0027] Step 3.6, at the end of the loop, θ best The x-axis stores the rotation angle of the vehicle body coordinate system relative to the point cloud coordinate system. c ,y c It stores the coordinates of the vehicle body coordinate system in the point cloud coordinate system, that is, the translation vector in the coordinate transformation, R. c The system stores the vehicle body outline information; suppose there is a point p = (x...) in the point cloud coordinate system. p ,y p ,z p Then, the combined coordinate transformation from the point cloud coordinate system to the vehicle body coordinate system is:

[0028]

[0029] Where p′ is the coordinate of point p in the vehicle body coordinate system.

[0030] Preferably, step 4 includes:

[0031] Step 4.1: Plan the automatically calibrated path;

[0032] Step 4.2: Use the PID algorithm to control the robot platform to move along the trajectory, perform trajectory navigation and collect calibration data.

[0033] Preferably, in step 4.1, the robot's motion path is a rounded rectangle. Assuming the car's width is 2w and its length is 2h, the robot's motion path is a rounded rectangle at a distance λ from the car's circumscribed rectangle, generated by the following parametric equation:

[0034]

[0035]

[0036] Where t is the system running time, at time 0, the robot starts from A8 and runs at speed v0, and t1, t2, ..., t8 are the times when the robot arrives at A1, A2, ..., A8 respectively.

[0037] Preferably, step 4.2 includes:

[0038] Step 4.2.1: Activate the SLAM automatic navigation function and convert the SLAM positioning result into P based on the coordinate transformation matrix. real (x real ,y real ,θ real ), where x real ,y real ,θ real These represent the robot's coordinates and heading angle in the vehicle's coordinate system; P is calculated in real time. real Parametric equations for the robot's motion path The shortest vector is denoted as the trajectory deviation factor E. rt In the input trajectory follower PID controller, the output heading adjustment angle θ adj , and feedforward heading angle θ rt Add them together to get θ tar ;

[0039] Step 4.2.2: Calculate the heading angle error E θ Input the heading angle to the PID controller to obtain ω bwd , and the feedforward heading angle rotation speed ω fwd Add them together to get the heading angle and rotation angle ω;

[0040] Step 4.2.3: Use constant linear velocity v and rotational angular velocity ω, and let the chassis control function execute it;

[0041] Step 4.2.4: Activate the keyless entry system. During the journey, collect real-time distance information from the digital key used for keyless entry, and P... real Stored in pairs for data calibration.

[0042] Preferably, in step 4.2.4, the digital key ranging distance information includes: the TOF distance to each base station measured by the tag held by the robot.

[0043] In a second aspect, a vehicle digital key calibration data acquisition system based on an autonomous navigation robot is provided, characterized in that it is used to perform vehicle digital key calibration data acquisition based on an autonomous navigation robot as described in any of the first aspects, comprising: an autonomous navigation robot and a vehicle body; the autonomous navigation robot is equipped with a sensing module, the sensing module being mounted facing the side, so that the autonomous navigation robot can directly observe vehicle feature points when moving around the vehicle.

[0044] Thirdly, a computer storage medium is provided, wherein a computer program is stored in the computer storage medium; when the computer program is run on a computer, the computer performs the acquisition of digital key calibration data for a car based on an autonomous navigation robot as described in any of the first aspects.

[0045] The beneficial effects of this invention are: by utilizing the robot's mapping and positioning functions, this invention can automatically calibrate the vehicle positioning system, which can help developers quickly and accurately calibrate the positioning system and reduce labor costs. Attached Figure Description

[0046] Figure 1 A flowchart of a method for collecting calibration data for a car digital key;

[0047] Figure 2 A top view schematic diagram of a car digital key calibration data acquisition system;

[0048] Figure 3 This is a schematic diagram of the calibration route;

[0049] Figure 4 This is a block diagram of the navigation algorithm. Detailed Implementation

[0050] The present invention will be further described below with reference to embodiments. The description of the embodiments below is only for the purpose of helping to understand the present invention. It should be noted that those skilled in the art can make several modifications to the present invention without departing from the principle of the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

[0051] Example 1:

[0052] like Figure 1 As shown in the figure, this application provides a method for collecting calibration data for a car digital key, including:

[0053] Step 1: Construct a 3D point cloud map by collecting data using a robot equipped with a Simultaneous Localization and Mapping (SLAM) algorithm.

[0054] Step 1 includes:

[0055] Step 1.1: Install the robot perception module facing to the side, so that the robot can directly observe the vehicle's feature points when moving around the vehicle, thereby enabling the robot to obtain its relative coordinates with the vehicle more accurately during movement; the robot perception module includes an RGBD camera and a LiDAR.

[0056] Step 1.2: Park the target vehicle in an open area, ensuring there are no other obvious objects within a 4m radius. Place the robot approximately 2m to the right of the vehicle. Roughly estimate the coordinates of the vehicle's center in the robot's perception module coordinate system, denoted as point q. car ;

[0057] Step 1.3: Start the robot's SLAM mapping function. Manually control the robot to walk around the vehicle for a full circle to build a 3D map. When the map building is finished, temporarily save the 3D map as 3D map point cloud data.

[0058] In a 3D map, the point cloud not only contains vehicle point clouds but may also include information about other vehicles or objects in the distance. Therefore, it is necessary to segment the vehicle point cloud.

[0059] Step 2: Segment the 3D point cloud map to obtain vehicle point clouds.

[0060] Step 2 includes:

[0061] Step 2.1: Record the 3D map point cloud data as a point set. And take S map Points with z-coordinates greater than a threshold height (e.g., 0.3m) are used as a subset S to filter out feature points on the ground.

[0062] Step 2.2: Initialize the sphere center node set Traverse the point cloud S, if C is Or click p i If the distance to any point in set C is greater than ρ, then point p... i Add point p to the set of the center nodes C; otherwise, if point p i With point c in C j The distance is less than ρ, i.e., p i In the position of c j Within a sphere with center ρ and radius ρ, we consider p... i Belongs to node c j ;

[0063] Step 2.3: Traverse the set of sphere center nodes C. If the distance between two points c1 and c2 in C is less than 2ρ, then connect p1 and p2 with an undirected and unweighted edge to construct an undirected graph G.

[0064] Step 2.4: Traverse the set of sphere center nodes C and find the relationship between them and q. car The nearest sphere center node c car ;

[0065] Step 2.5, from c car Starting from graph G, traverse graph G and find the largest connected subgraph G′ on G;

[0066] Step 2.6: Traverse the point cloud S and select all points whose distance to any node in G′ is less than ρ, and denote their set as S. c S c This is the desired vehicle point cloud.

[0067] Step 3: Fit the vehicle point cloud into a cuboid and obtain the coordinate transformation matrix from the point cloud coordinate system to the vehicle coordinate system.

[0068] Step 3 includes:

[0069] Step 3.1: Use a loop to enumerate the rotation angle θ within a preset value range with a preset step size. The step size can be 0.001 or other values. The preset value range can be from 0 to π, or other value ranges can be set according to the scenario.

[0070] Step 3.2: Transfer the vehicle point cloud S c Rotate the coordinates of all points in S' around the positive z-axis by θ, and denote the result as S'. c ;

[0071] Step 3.3, take x t ,x s S′ c The upper and lower α quantiles of the x-coordinate of the midpoint, y t ,x s S′ c The upper and lower α quantiles of the y-coordinate of the midpoint;

[0072] Step 3.4: Calculate the fitting evaluation index v = (x) for the rotation angle θ. t -x s ) 2 +(y t -y s ) 2 ;

[0073] Step 3.5, using variable v best To store the maximum value of v, and update v best While storing the value, save the corresponding θ.best and R c =[x s ,x t ,y s ,y t ],as well as

[0074] Step 3.6, at the end of the loop, θ best The x-axis stores the rotation angle of the vehicle body coordinate system relative to the point cloud coordinate system. c ,y c It stores the coordinates of the vehicle body coordinate system in the point cloud coordinate system, that is, the translation vector in the coordinate transformation, R. c The system stores the vehicle body outline information; suppose there is a point p = (x...) in the point cloud coordinate system. p ,y p ,z p Then, the combined coordinate transformation from the point cloud coordinate system to the vehicle body coordinate system is:

[0075]

[0076] Where p′ is the coordinate of point p in the vehicle body coordinate system.

[0077] Step 4: Perform calibration route navigation and calibration data collection for the robot.

[0078] Example 2:

[0079] Based on Embodiment 1, Embodiment 2 of this application provides a more specific method for collecting calibration data for automotive digital keys, including:

[0080] Step 1: Construct a 3D point cloud map by collecting data using a robot equipped with a real-time localization and mapping (SLAM) algorithm.

[0081] Step 2: Segment the 3D point cloud map to obtain vehicle point clouds.

[0082] Specifically, based on the point set Construct an undirected graph, initialize graph nodes using several envelope spheres of radius ρ, initialize edges based on the distance relationships between the envelope spheres, and segment the vehicle point cloud S using a maximum connected subgraph method for undirected graphs. c .

[0083] Step 3: Fit the vehicle point cloud into a cuboid and obtain the coordinate transformation matrix from the point cloud coordinate system to the vehicle coordinate system.

[0084] Specifically, in order to obtain the vehicle center coordinates and the orientation of the vehicle body coordinate system, point set S is used. cBy fitting it to a cuboid, the coordinate transformation matrix of the vehicle body coordinate system can be obtained. In the range [0, π), an optimal θ is found using enumeration, such that after rotating the vehicle body coordinate system relative to the point cloud coordinate system, v = (x...). t -x s ) 2 +(y t -y s ) 2 Lowest score, O c It stores the coordinates of the vehicle body coordinate system in the point cloud coordinate system, that is, the translation vector in the coordinate transformation, R. c The system stores information about the vehicle's outline.

[0085] Step 4, as follows Figure 4 As shown, the robot performs calibration route navigation and calibration data collection.

[0086] Step 4 includes:

[0087] Step 4.1: Plan the automatically calibrated path.

[0088] In a keyless entry system for automobiles, the business logic is that when the digital key is a certain distance from the car, the keyless entry system executes the action corresponding to that distance. The boundary of the area is a rounded rectangle. The robot's movement path is designed as follows: Figure 3 As shown. Assume the car's width is 2w and its length is 2h. Use the following parametric equations to generate the equation of the rounded rectangle at a distance λm from the car.

[0089]

[0090]

[0091] Where t is the system running time, at time 0, the robot starts from A8 and runs at speed v0, and t1, t2, ..., t8 are the times when the robot arrives at A1, A2, ..., A8, respectively, as shown in the attached figure. Figure 3 As shown.

[0092] Step 4.2: Use the PID algorithm to control the robot platform to move along the trajectory, perform trajectory navigation and collect calibration data.

[0093] Step 4.2 includes:

[0094] Step 4.2.1: Activate the SLAM automatic navigation function and convert the SLAM positioning result into P based on the coordinate transformation matrix. real (x real ,y real ,θ real ), where x real ,y real ,θ realThese represent the robot's coordinates and heading angle in the vehicle's coordinate system; P is calculated in real time. real Parametric equations for the robot's motion path The shortest vector is denoted as the trajectory deviation factor E. rt In the input trajectory follower PID controller, the output heading adjustment angle θ adj , and feedforward heading angle θ rt Add them together to get θ tar ;

[0095] Step 4.2.2: Calculate the heading angle error E θ Input the heading angle to the PID controller to obtain ω bwd , and the feedforward heading angle rotation speed ω fwd Add them together to get the heading angle and rotation angle ω;

[0096] Step 4.2.3: Use a constant linear velocity v = 0.2 m / s and a rotational angular velocity ω, and let the chassis control function execute it;

[0097] Step 4.2.4: Activate the keyless entry system. During the journey, collect real-time distance information from the digital key used for keyless entry, and p. real Stored in pairs for data calibration.

[0098] In step 4.2.4, the distance information measured by the digital key includes: the TOF distance to each base station measured by the tag held by the robot.

[0099] It should be noted that the parts in this embodiment that are the same as or similar to those in Embodiment 1 can be referred to each other, and will not be repeated in this application.

[0100] Example 3:

[0101] Based on embodiments 1 and 2, embodiment 3 of this application provides a car digital key calibration data acquisition system, such as... Figure 2 As shown, the system includes an autonomous navigation robot and a car body. The autonomous navigation robot is equipped with a perception module, which is mounted to the side, allowing the robot to directly observe vehicle feature points as it moves around the vehicle. The robot already possesses SLAM mapping and chassis control capabilities.

[0102] The embodiments of this application obtain point cloud maps and real-time positioning coordinates from the SLAM algorithm, then run the relevant algorithms, and finally call the chassis control function to control the robot chassis, so as to realize autonomous calibration route navigation and calibration data acquisition.

[0103] Specifically, the system provided in this embodiment is the system corresponding to the method provided in Embodiments 1 and 2. Therefore, the parts in this embodiment that are the same as or similar to those in Embodiments 1 and 2 can be referred to each other, and will not be repeated in this application.

Claims

1. A method for collecting calibration data for automotive digital keys based on an autonomous navigation robot, characterized in that, include: Step 1: Construct a 3D point cloud map by collecting data using a robot equipped with a real-time localization and mapping (SLAM) algorithm; Step 2: Segment the 3D point cloud map to obtain vehicle point clouds; Step 3: Fit the vehicle point cloud into a cuboid and obtain the coordinate transformation matrix from the point cloud coordinate system to the vehicle coordinate system; Step 4: Perform robot calibration route navigation and calibration data collection; Step 4 includes: Step 4.1: Plan the automatically calibrated path; In step 4.1, the robot's movement path is a rounded rectangle. Let the width of the car be... , length is The robot's movement path is a distance from the circumscribed rectangular outline of the car. A rounded rectangle is generated by the following parametric equation: in, For system runtime, at time 0, the robot starts from... Start-up, at speed run, The robots arrived respectively The moment; Step 4.2: Use the PID algorithm to control the robot platform to move along the trajectory, perform trajectory navigation and collect calibration data; Step 4.2 includes: Step 4.2.1: Activate the SLAM automatic navigation function and convert the SLAM positioning result into [the desired coordinates] according to the coordinate transformation matrix. ,in These are the robot's coordinates and heading angle in the vehicle's coordinate system; real-time calculation. Parametric equations for the robot's motion path The shortest vector is denoted as the trajectory deviation factor. In a PID controller that follows the input trajectory, the output heading adjustment angle is... , and feedforward heading angle Add them together to get ; Step 4.2.2: Calculate the heading angle error Input the heading angle PID controller to obtain , and feedforward heading angle rotation speed Add them together to get the heading angle rotation angle. ; Step 4.2.3: Use constant linear velocity and rotational angular velocity The chassis control function will then take over. Step 4.2.4: Activate the keyless entry system. During the journey, collect real-time distance information from the digital key used in the keyless entry system. Stored in pairs for data calibration.

2. The method for collecting calibration data for automotive digital keys based on an autonomous navigation robot according to claim 1, characterized in that, Step 1 includes: Step 1.1: Install the robot perception module facing to the side so that the robot can directly observe the vehicle's feature points when moving around the vehicle; the robot perception module includes an RGBD camera and a LiDAR. Step 1.2: Park the target vehicle in an open area and place the robot at a certain distance to the right of the vehicle. Roughly estimate the coordinates of the vehicle's center in the robot's perception module coordinate system, and denot it as point . ; Step 1.3: Start the robot's SLAM mapping function. Manually control the robot to walk around the vehicle for a full circle to build a 3D map. When the map building is finished, temporarily save the 3D map as 3D map point cloud data.

3. The method for collecting calibration data for automotive digital keys based on an autonomous navigation robot according to claim 2, characterized in that, Step 2 includes: Step 2.1: Record the 3D map point cloud data as a point set. and take Points with z-coordinates greater than the threshold height are considered as a subset. To filter out feature points on the ground; Step 2.2: Initialize the sphere center node set Traversing point clouds ,like for Or click and The distance to any point in the set is greater than Then the point Add to the set of the center nodes of the sphere; otherwise, if the point... and The distance to the middle point is less than ,Right now In the position of For the center of the ball, Within a sphere with radius , it is considered that... Belongs to node ; Step 2.3: Traverse the set of sphere center nodes ,like Two points , The distance is less than Then connect the two sides with an undirected, unweighted edge. This is how an undirected graph is constructed. ; Step 2.4: Traverse the set of sphere center nodes Find the result of the above. The nearest center node ; Step 2.5, from Starting point traversal graph ,beg Maximum connected subgraph on ; Step 2.6: Traverse the point cloud Filter out all that are related to The distance between any node in the middle is less than Let the set of points be denoted as . , This is the desired vehicle point cloud.

4. The method for collecting calibration data for automotive digital keys based on an autonomous navigation robot according to claim 3, characterized in that, Step 3 includes: Step 3.1: Using a loop, enumerate the rotation angles within a preset numerical range and with a preset step size. ; Step 3.2: Create vehicle point clouds Rotate the coordinates of all points around the positive z-axis Record the result as ; Step 3.3, Take They are respectively midpoint coordinates up and down Quantiles They are respectively midpoint coordinates up and down quantiles; Step 3.4: Calculate the rotation angle Fit evaluation index = ; Step 3.5, Using Variables To save The maximum value, in updating Save the corresponding value at the same time and = [ , , , ],as well as , ; Step 3.6, at the end of the loop, The system stores the rotation angle of the vehicle body coordinate system relative to the point cloud coordinate system. It stores the coordinates of the vehicle body coordinate system in the point cloud coordinate system, that is, the translation vector in the coordinate transformation. The system stores the vehicle body outline information; let there be a point in the point cloud coordinate system. The combined coordinate transformation from the point cloud coordinate system to the vehicle body coordinate system is: in, For point Coordinates in the vehicle body coordinate system.

5. The method for collecting calibration data for automotive digital keys based on an autonomous navigation robot according to claim 4, characterized in that, In step 4.2.4, the distance information measured by the digital key includes: the TOF distance to each base station measured by the tag held by the robot.

6. A car digital key calibration data acquisition system based on an autonomous navigation robot, characterized in that, The method for collecting calibration data for a car digital key based on an autonomous navigation robot as described in any one of claims 1 to 5 includes: an autonomous navigation robot and a car body; the autonomous navigation robot is equipped with a sensing module, which is mounted facing to the side, so that the autonomous navigation robot can directly observe vehicle feature points when it travels around the vehicle.

7. A computer storage medium, characterized in that, The computer storage medium stores a computer program; when the computer program is run on the computer, it causes the computer to execute the vehicle digital key calibration data acquisition method based on an autonomous navigation robot as described in any one of claims 1 to 5.

Images