Calibration method for a robot and robot applying it

Abstract

The application provides a kind of calibration method for robot and robot applied to it, the robot includes laser radar and wheeled odometry, robot is placed in calibration chamber, the calibration method includes: control robot rotates to a plurality of measurement angles in turn, obtains first point cloud by laser radar;First point cloud is matched with the point cloud map of calibration chamber;The distance deviation value of each point in first point cloud is calculated, and laser radar is calibrated according to distance deviation value;Control robot walks along preset trajectory, and obtains second point cloud by calibrated laser radar;Second point cloud is matched with the point cloud map of calibration chamber, and the first pose of robot is obtained;Second pose is obtained by wheeled odometry;According to first pose and second pose, wheeled odometry is calibrated.The application can improve the measurement accuracy of wheeled odometry and laser radar, and also can detect the wheeled odometry or laser radar with fault.

Description

Technical Field

[0001] This disclosure relates to the field of robotics, and more particularly to a calibration method for a robot and a robot using the same. Background Technology

[0002] The main modules used for localization and navigation in mobile robots include wheeled odometry, laser odometry, visual odometry, and inertial navigation modules. Especially in applications such as autonomous vehicles, unmanned vehicles, and service robots, the rapid development of mobile robots and the emergence of various automated tasks, such as object grasping and spatial exploration, have placed higher demands on their localization capabilities. Odometry technology plays a crucial role in these tasks. Among these, laser odometry and wheeled odometry are the most commonly used odometry methods to enhance a robot's environmental perception and self-localization.

[0003] However, laser odometers often exhibit individual differences, leading to significant individual variations in measurement accuracy. Furthermore, laser odometers may drift when stationary, resulting in cumulative errors.

[0004] Wheeled odometers use the rotational speeds of the left and right wheels to obtain velocity and angular velocity, which are then input into the kinematic model of a mobile robot to deduce the robot's current position and heading angle. However, due to measurement errors, actuator errors, model errors, and potential sideslip, wheeled odometer information also exhibits significant errors. Since wheeled odometers are derived through continuous calculations using geometric models, these errors accumulate in the results, eventually rendering the data unusable. However, wheeled odometers do not drift when stationary and have the advantages of low cost.

[0005] It is evident that different types of odometers have their own advantages and disadvantages. Therefore, using multiple sensors and performing joint calibration is an important measure to improve the accuracy of the corresponding sensors.

[0006] Calibrating wheeled odometers often requires manually measuring the actual odometer value and comparing it with the measured value, then compensating accordingly. This method introduces human error and is very inefficient.

[0007] The content of the background section only discloses the technology known to the inventors and does not necessarily represent the prior art in this field. Summary of the Invention

[0008] In view of one or more existing deficiencies, the present invention provides a calibration method for a robot, the robot including a lidar and a wheeled odometer, wherein the robot is placed in a calibration chamber, and the calibration method includes:

[0009] The robot is controlled to rotate sequentially to multiple measurement angles, and the first point cloud is acquired through the lidar.

[0010] Match the first point cloud with the point cloud map of the calibration room;

[0011] Calculate the distance deviation value for each point in the first point cloud, and calibrate the lidar based on the distance deviation value;

[0012] The robot is controlled to walk along a preset trajectory, and a second point cloud is acquired using a calibrated lidar.

[0013] The second point cloud is matched with the point cloud map of the calibration room to obtain the first pose of the robot;

[0014] The second pose is obtained through the wheel odometer;

[0015] The wheel odometer is calibrated based on the first pose and the second pose.

[0016] According to one aspect of the invention, the calibration chamber is rectangular, and the robot is placed in the middle of the calibration chamber near the long side.

[0017] According to one aspect of the invention, the dimensional error of the calibration chamber is on the order of millimeters.

[0018] According to one aspect of the invention, the point cloud map of the calibration chamber is a point cloud map that perfectly matches the size of the calibration chamber, and the robot pre-stores the point cloud map.

[0019] According to one aspect of the present invention, the step of matching the first point cloud with the point cloud map of the calibration room specifically includes: matching the first point cloud with the point cloud map of the calibration room using the ICP algorithm.

[0020] According to one aspect of the present invention, the step of calculating the distance deviation value of each point in the first point cloud and calibrating the lidar based on the distance deviation value specifically includes: fitting the four sides of the point cloud map with straight lines respectively, calculating the distance deviation value of the intersection point of each point in the first point cloud with the corresponding straight line, and compensating the measurement value of the lidar based on the distance deviation value.

[0021] According to one aspect of the present invention, the step of calculating the distance deviation value of each point in the first point cloud and calibrating the lidar based on the distance deviation value further includes: performing multiple samplings at each measurement angle, calculating the average distance deviation value between the multiple sampling values ​​of each point and the intersection point of the corresponding straight line, and compensating the measurement value of the lidar based on the average distance deviation value.

[0022] According to one aspect of the invention, the preset trajectory is rectangular and has approximately the same aspect ratio as the calibration chamber.

[0023] According to one aspect of the present invention, the step of matching the second point cloud with the map of the calibration room to obtain the first pose of the robot specifically includes: matching the second point cloud with the map of the calibration room using the ICP algorithm to obtain the first pose of the robot.

[0024] According to one aspect of the present invention, the step of calibrating the wheel odometer based on a first pose and a second pose includes: comparing the second pose with the first pose, and correcting the parameters of the wheel odometer until the deviation between the second pose and the first pose is less than a threshold value.

[0025] The present invention also relates to a robot, comprising:

[0026] LiDAR is used to acquire point clouds;

[0027] A wheeled odometer is used to acquire the linear velocity and angular velocity of the robot and to calculate its pose based on a kinematic model.

[0028] Memory, used to store maps of the calibration room; and

[0029] The processor is used to execute the above calibration method to jointly calibrate the lidar and the wheel odometer.

[0030] This invention provides a robot that simultaneously utilizes LiDAR and a wheeled odometer, and also provides a calibration scheme based on a high-precision calibration chamber and LiDAR. Before calibrating the wheeled odometer, the LiDAR is calibrated first, ensuring the reliability of the wheeled odometer. This invention can simultaneously improve the measurement accuracy of both the wheeled odometer and the LiDAR, and can also detect faulty wheeled odometers or LiDARs. Attached Figure Description

[0031] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure. In the drawings:

[0032] Figure 1a A schematic diagram of a robot according to an embodiment of the present invention is shown;

[0033] Figure 1b It shows Figure 1a Internal module diagram;

[0034] Figure 2 A flowchart of a calibration method for a robot according to an embodiment of the present invention is shown;

[0035] Figure 3aA schematic diagram of a robot acquiring a first point cloud in a calibration chamber according to an embodiment of the present invention is shown;

[0036] Figure 3b It shows Figure 3a A magnified view of the point cloud in the image;

[0037] Figure 4a A schematic diagram of a robot walking along a preset trajectory according to an embodiment of the present invention is shown;

[0038] Figure 4b A schematic diagram of a robot walking along a self-calibrated trajectory according to an embodiment of the present invention is shown;

[0039] Figure 5 A schematic diagram of the calibration process of a wheeled odometer according to an embodiment of the present invention is shown. Detailed Implementation

[0040] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0041] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0042] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection, an electrical connection, or a connection that allows for communication; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0043] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0044] The following disclosure provides many different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. In addition, examples of various specific processes and materials are provided in this invention, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0045] This invention provides a calibration method for a robot and a robot using the same. The robot includes a lidar and a wheeled odometer. The robot is placed in a calibration chamber. The calibration method includes: controlling the robot to rotate sequentially to multiple measurement angles to acquire a first point cloud using the lidar; matching the first point cloud with a point cloud map in the calibration chamber; calculating the distance deviation value of each point in the first point cloud and calibrating the lidar based on the distance deviation value; controlling the robot to walk along a preset trajectory and acquiring a second point cloud using the calibrated lidar; matching the second point cloud with the point cloud map in the calibration chamber to obtain a first pose of the robot; acquiring a second pose using the wheeled odometer; and calibrating the wheeled odometer based on the first and second poses. This invention can improve the measurement accuracy of the wheeled odometer and lidar, and can also detect faulty wheeled odometers or lidars.

[0046] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0047] Figure 1a A schematic diagram of a robot according to an embodiment of the present invention is shown. Figure 1b It shows Figure 1a Internal module diagram, combined with Figure 1a and Figure 1b The robot includes a housing 20 for carrying objects and a mobile chassis 10. The mobile chassis 10 has at least two sets of drive wheels 121, each set located on one side of the mobile chassis 10. The robot also includes a function controller for user operation, a low-level controller for map generation and path planning, and a component controller for controlling the movement unit and the environment detection unit, wherein the component controller controls the travel speed of the drive wheels 121. At least one set of drive wheels 121 serves as the left drive wheel, and at least one set of drive wheels 121 serves as the right drive wheel, located on opposite sides of the chassis 10. Optionally, the robot may also include at least two sets of driven wheels, one set of drive wheels corresponding to one set of driven wheels, wherein at least one set of driven wheels serves as the left driven wheel, and at least one set of driven wheels serves as the right driven wheel. The left and right driven wheels assist the left and right drive wheels in moving the robot's housing 20 and mobile chassis 10, thereby reducing the load pressure on the drive wheels 120.

[0048] Continue to refer to Figures 1a-1b The robot also includes: a lidar 110, a wheeled odometer 120, a memory 130, and a processor 140.

[0049] The lidar 110 can be positioned at an opening in the robot's housing, facilitating the emission of laser signals to detect surrounding objects. In one specific embodiment, the lidar 110 includes a photoelectric receiving array and a laser emitting unit array. When the lidar 110 rotates along a set plane, the photoelectric receiving array forms a scanning cylinder, increasing the scanning area and facilitating the acquisition of detailed object shapes, thus preventing the robot from bumping into objects. In another specific embodiment, the lidar 110 contains only a single photoelectric receiving unit and a single laser emitting unit. After rotating along the set plane, the lidar 110 can measure the shape of an object around a circle, thereby reducing costs. Optionally, the set plane can be a horizontal plane, facilitating object detection during robot movement. Furthermore, other set planes, such as a vertical plane, can be selected according to user needs; this invention does not limit the selection of such planes.

[0050] The wheeled odometer 120 uses the rotational speed values ​​of at least two sets of drive wheels 121 to obtain the linear velocity and angular velocity of the robot, substitutes them into the kinematic model of the mobile robot, and deduce the robot's current pose, that is, position and heading angle information.

[0051] Memory 130 is used to store maps of the calibration room. In some specific embodiments, memory 130 includes: a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, mechanical encoding device, such as a punch card or recessed protrusion structure storing instructions thereon, and any suitable combination thereof.

[0052] Processor 140 is used to execute a calibration method to jointly calibrate lidar 110 and wheel odometer 120.

[0053] The system architecture and module composition of the robot have been introduced above. The calibration methods used for the robot will be introduced below.

[0054] Figure 2 A flowchart of a calibration method for a robot according to an embodiment of the present invention is shown. The robot includes a lidar 110 and a wheeled odometer 120. The robot is placed in a calibration chamber. The calibration method includes:

[0055] In step S11, the robot is controlled to rotate sequentially to multiple measurement angles, and the first point cloud is acquired through the lidar 110.

[0056] Figure 3a A schematic diagram of a robot acquiring a first point cloud in a calibration chamber, according to an embodiment of the present invention, is shown. The calibration chamber is, for example, represented by a rectangular frame, and the robot is, for example, represented by a solid large circle within it. The robot is controlled to rotate to multiple measurement angles, preferably, the last position rotated to is the initial position.

[0057] According to a preferred embodiment of the present invention, the calibration chamber is rectangular, and the robot is placed in the middle of the calibration chamber near the long side.

[0058] For example, the robot's initial position is facing the long side of the calibration chamber with a heading angle of 0°. The robot is controlled to rotate in place at 10° intervals until it rotates 360° and returns to the initial position. The interval angles can be set according to the scanning area of ​​the LiDAR 110 and the size of the calibration chamber, so that the first point cloud covers all the edges of the calibration chamber.

[0059] In step S12, the first point cloud is matched with the point cloud map of the calibration room.

[0060] Continue to refer to Figure 3a The point cloud map in the calibration chamber consists of regularly arranged black dots (solid squares), and its shape perfectly matches the size of the calibration chamber. The first point cloud image acquired after the robot rotates once within the calibration chamber is shown below. Figure 3a The image shows a series of dark gray dots (represented by hollow ellipses in a magnified view). The first point cloud includes the boundary information of the calibration chamber, and its shape is roughly the same as that of the calibration chamber. The first point cloud is matched with the point cloud map of the calibration chamber to obtain the robot's actual position.

[0061] According to a preferred embodiment of the invention, the dimensional error of the calibration chamber is on the order of millimeters.

[0062] For example, the internal dimensions of the calibration chamber are 10m * 2m, with a dimensional error of 1mm. The higher the dimensional accuracy of the calibration chamber, the higher the accuracy of the calibration method in this embodiment. The calibration chamber can be designed according to the required calibration accuracy to obtain the required dimensional error.

[0063] According to a preferred embodiment of the present invention, the point cloud map of the calibration chamber is a point cloud map that perfectly matches the size of the calibration chamber, and the robot pre-stores the point cloud map.

[0064] Continue to refer to Figure 3a The point cloud map of the calibration chamber is composed of a regular arrangement of black dots. The more dots and the denser the point cloud map, the higher its accuracy. Preferably, the point cloud map of the calibration chamber can be drawn manually according to its dimensions, or it can be drawn using CAD or ordinary drawing software. As long as the dimensions of the calibration chamber can be accurately matched, this invention does not limit the specific implementation method. The point cloud map of the calibration chamber is pre-stored in the robot's memory 130 for direct access by the processor 140.

[0065] According to a preferred embodiment of the present invention, the step of matching the first point cloud with the point cloud map of the calibration room specifically includes: matching the first point cloud with the point cloud map of the calibration room using the ICP algorithm.

[0066] The Iterative Closest Point (ICP) algorithm is a point cloud matching algorithm. It can unify two sets of point cloud data from different coordinate systems into the same reference coordinate system through certain rotation and translation transformations.

[0067] For example, the first point cloud is the target point cloud P to be matched, and the point cloud map in the calibration room is the source point cloud Q. According to certain constraints, the nearest neighbor points (pi, qi) are found, and then the optimal matching parameters R and t are calculated to minimize the error function. The error function E(R, t) is:

[0068]

[0069] Where n is the number of nearest neighbor pairs, pi is a point in the target point cloud P, qi is the nearest neighbor point in the source point cloud Q corresponding to pi, R is the rotation matrix, and t is the translation vector.

[0070] In a specific embodiment, the steps for matching point clouds using the ICP algorithm are as follows:

[0071] (1) Take a point set pi∈P in the target point cloud P;

[0072] (2) Find the corresponding point set qi∈Q in the source point cloud Q such that ||qi-pi||=min;

[0073] (3) Calculate the rotation matrix R and translation vector t to minimize the error function E(R, t);

[0074] (4) Perform rotation and translation transformations on pi using rotation matrix R and translation vector t to obtain a new set of corresponding points pi' = {pi' = Rpi + t, pi ∈ P};

[0075] (5) Calculate the average distance between pi' and the corresponding point set qi using the following formula:

[0076]

[0077] (6) If the average distance d is less than the preset threshold or greater than the preset maximum number of iterations, stop the iterative calculation; otherwise, return to step (2) until the convergence condition is met, thereby completing the matching of the first point cloud with the point cloud map of the calibration room.

[0078] The above is merely an illustrative description and does not constitute a limitation on the specific implementation of matching the first point cloud with the point cloud map of the calibration room using the ICP algorithm.

[0079] In step S13, the distance deviation value of each point in the first point cloud is calculated, and the lidar 110 is calibrated based on the distance deviation value.

[0080] Continue to refer to Figure 3a The point cloud map in the calibration chamber is composed of regularly arranged black dots. The point cloud map perfectly matches the size of the calibration chamber. Because the calibration chamber has a size error at the millimeter level, the point cloud map also has very high accuracy. The first point cloud acquired by the robot is as follows: Figure 3a As shown by the individual dark gray dots, Figure 3b It shows Figure 3a A magnified view of the point cloud in the image, from... Figure 3b As can be seen, there is a distance deviation between the points on the first point cloud and the points on the point cloud map; that is, there is a distance deviation between the measured value and the ideal value. In a specific embodiment, after matching the first point cloud with the point cloud map, each point on the first point cloud corresponds one-to-one with a point on the point cloud map. At the same time, the actual position of the robot can be determined, thereby obtaining the distance value of each point. Then, the lidar 110 is calibrated based on the distance deviation value between the corresponding point pairs on the first point cloud and the point cloud map.

[0081] According to a preferred embodiment of the present invention, the step of calculating the distance deviation value of each point in the first point cloud and calibrating the lidar 110 based on the distance deviation value specifically includes: fitting the four sides of the point cloud map with straight lines respectively, calculating the distance deviation value of the intersection point of each point in the first point cloud with the corresponding straight line, and compensating the measurement value of the lidar 110 based on the distance deviation value.

[0082] Continue to refer to Figure 3b A magnified view of the area is used, and the ICP algorithm is employed to ensure that the first point cloud coincides with the point cloud map (due to measurement errors in the first point cloud, a perfect fit is impossible), thus determining the robot's actual position. Furthermore, straight-line fitting is performed on the four edges of the point cloud map, and the distance difference between each point in the first point cloud and the intersection point of the corresponding straight line is calculated. The intersection points are shown below. Figure 3b As shown by the hollow circles in the diagram, the intersection points are taken as ideal laser points. Finally, the lidar 110 is calibrated based on the distance deviation between the corresponding point pairs on the first point cloud and the point cloud map.

[0083] According to a preferred embodiment of the present invention, the step of calculating the distance deviation value of each point in the first point cloud and calibrating the lidar 110 based on the distance deviation value further includes: performing multiple samplings at each measurement angle, calculating the average distance deviation value between the multiple sampling values ​​of each point and the intersection point of the corresponding straight line, and compensating the measurement value of the lidar 110 based on the average distance deviation value.

[0084] In one specific embodiment, points in the first point cloud for each measurement angle are sampled 20 times. The distance difference between the actual laser point and the ideal laser point is calculated. That is, the average distance deviation between the points in the first point cloud and the intersection points is used as the compensation value. The correspondence between the laser measurement distance and the compensation value is recorded, as shown in the table below:

[0085]

[0086] The calibrated laser distance measurement value is the actual measurement value plus a compensation value. The compensation value is the same as the measurement distance value in the table above that is closest to the actual measurement value.

[0087] For example, the robot's initial position is facing the long side of the calibration chamber. It is controlled to rotate in place at 10° intervals, and the point cloud acquired at each measurement angle is matched with the point cloud map in the calibration chamber to calculate the distance deviation value, until it rotates 360° and returns to the initial position. Alternatively, the robot's initial position is facing the long side of the calibration chamber, and it is controlled to rotate in place at 10° intervals. The point cloud acquired at each measurement angle is stitched together to obtain the first point cloud. This first point cloud is matched with the point cloud map in the calibration chamber to calculate the distance deviation value for each point. The deviation value is used as a compensation value to obtain a table showing the correspondence between laser measurement distance and compensation value, or a fitting curve showing the laser measurement distance and compensation value, thus completing the calibration of the lidar 110. In practical applications, the compensation value corresponding to the measurement distance is determined by looking up the table or curve, thereby improving the measurement accuracy of the lidar 110.

[0088] The inventors devised a method for jointly calibrating the lidar 110 and the wheeled odometer 120. First, the lidar 110 is calibrated in a high-precision calibration chamber. Then, the calibrated lidar 110 is used to calibrate the wheeled odometer 120, ensuring the reliability of the wheeled odometer 120. Through actual testing, this scheme significantly improves the accuracy of both the wheeled odometer 120 and the lidar 110. In one specific embodiment, the accuracy of the wheeled odometer is improved from 5% to 1%; the lidar accuracy is improved from 3cm to 1cm; and faulty odometers or lidar sensors are effectively detected. The calibration of the lidar 110 is achieved in steps S11-S13. The calibration of the wheeled odometer 120 is further described below.

[0089] In step S14, the robot is controlled to walk along a preset trajectory, and the second point cloud is obtained by the calibrated lidar 110.

[0090] According to a preferred embodiment of the present invention, the preset trajectory is rectangular and has approximately the same aspect ratio as the calibration chamber.

[0091] Figure 4a A schematic diagram of a robot walking along a preset trajectory according to an embodiment of the present invention is shown. The preset trajectory is rectangular and has approximately the same aspect ratio as the calibration chamber, for example, based on a safe distance that the robot can walk along the edge. During the robot's movement along the rectangular trajectory, a second point cloud is acquired using a calibrated lidar.

[0092] In step S15, the second point cloud is matched with the point cloud map of the calibration room to obtain the first pose of the robot.

[0093] The operation of matching the second point cloud with the point cloud map of the calibration room is the same as the operation of matching the first point cloud with the point cloud map of the calibration room described above. That is, the second point cloud includes the edge information of the calibration room, and the second point cloud is matched with the point cloud map of the calibration room to obtain the robot's actual position.

[0094] According to a preferred embodiment of the present invention, the step of matching the second point cloud with the map of the calibration room to obtain the first pose of the robot specifically includes: matching the second point cloud with the map of the calibration room using the ICP algorithm to obtain the first pose of the robot.

[0095] The operation of matching the second point cloud with the point cloud map in the calibration room using the ICP algorithm is the same as the operation of matching the first point cloud with the point cloud map in the calibration room described above, and will not be repeated here.

[0096] By matching the calibrated laser information with the point cloud map using the ICP algorithm, the robot's current pose, i.e., the first pose, can be obtained in real time.

[0097] In step S16, the second pose is obtained through the wheel odometer.

[0098] The wheeled odometer 120 uses the rotational speed values ​​of at least two sets of drive wheels 121 to obtain the linear velocity and angular velocity of the robot, substitutes them into the kinematic model of the mobile robot, and deduces the robot's current pose, i.e., the second pose.

[0099] According to a preferred embodiment of the present invention, before the step of obtaining the second pose through the wheeled odometer 120, the calibration method further includes: the robot walking along a self-calibration trajectory in the calibration chamber to automatically calibrate the wheeled odometer.

[0100] Figure 4b A schematic diagram of a robot walking along a self-calibrated trajectory according to an embodiment of the present invention is shown. The self-calibrated trajectory is a route that involves translation and rotation relative to the edge of the calibration chamber. The self-calibrated trajectory has more rotation angles because the accuracy of the robot's mobile chassis 10 is an important indicator of the chassis's performance. The chassis accuracy is affected by the straight-line and angular errors of the wheeled odometer 120. Therefore, both the straight-line and angular errors of the wheeled odometer 120 are calibrated; that is, the self-calibrated trajectory must allow the robot to both translate and rotate to minimize errors.

[0101] Prolonged robot operation leads to tire wear and a reduction in wheel diameter, affecting the accuracy of the wheeled odometer. Theoretically, if the pose optimization process is correct, the deviations in the wheeled odometer's pose prediction process can be calculated and corrected. This is the wheeled odometer calibration process.

[0102] Assuming the robot's pose change predicted by the wheel odometry 120 over a period of time is u, and after pose optimization through the SLAM process, the final pose change is u * . u and u * The difference is caused by the error in the wheel odometer parameters, so it is necessary to calculate this error in order to compensate for the pose prediction of the subsequent odometer.

[0103] u and u * They roughly satisfy a linear relationship, that is:

[0104]

[0105] After unfolding:

[0106]

[0107] Since formula (3) is a linear equation, it can be solved using the linear least squares method. To facilitate the solution, the expression of formula (4) is adjusted:

[0108]

[0109] Formula (5) can be expressed in the familiar form AX = Y, thus yielding the final result for X:

[0110] X = (A T A) -1 A T Y#(6)

[0111] Where A and Y are combinations of data from multiple measurements:

[0112]

[0113] The robot according to Figure 4b The self-calibration trajectory travels in the calibration chamber, and the wheeled odometer 120 is automatically calibrated using the above algorithm.

[0114] Figure 5 This diagram illustrates a wheeled odometer calibration process according to an embodiment of the present invention. During robot movement, n odometer-predicted pose changes u and the corresponding optimized pose transformation values ​​u are recorded. *(As the true value), and as the sampled data. To ensure data diversity, the sampled data should include obvious translational and rotational motion information. The odometer information is calibrated using n sampled data. After successful calibration, the new odometer pose prediction information is updated to the calibrated result (theoretically, after calibration, the pose prediction information of the wheeled odometer 120 has a small error compared to the true value).

[0115] In step S17, the wheel odometer 120 is calibrated according to the first pose and the second pose.

[0116] According to a preferred embodiment of the present invention, the step of calibrating the wheel odometer 120 based on the first pose and the second pose includes: comparing the second pose with the first pose, correcting the parameters of the wheel odometer, until the deviation between the second pose and the first pose is less than a threshold value.

[0117] Since the lidar 110 has been calibrated and the point cloud map in the calibration room has high accuracy, the accuracy of the first pose obtained by the ICP algorithm will also be high. The first pose is compared with the second pose calculated by the wheel odometer 120, and the parameters of the wheel odometer 120 are continuously modified until the calculation result of the wheel odometer 120 is basically consistent with the calculation result matched by the ICP algorithm, that is, the deviation value is less than the threshold, and finally the calibration of the odometer parameters is achieved.

[0118] This invention proposes a high-precision calibration chamber and a calibration scheme for the lidar 110. Before calibrating the wheel odometer 120, the lidar 110 is calibrated first, ensuring the reliability of the wheel odometer 120. This invention can simultaneously improve the measurement accuracy of both the wheel odometer 120 and the lidar 110, and can also detect faulty wheel odometers 120 or lidar 110s.

[0119] This specification provides the operational steps of the methods described in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operational steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiments or flowcharts can be executed sequentially or in parallel.

[0120] The present invention also relates to a robot, as shown in the reference. Figures 1a-1b The robots include:

[0121] LiDAR 110 is used to acquire point clouds;

[0122] A wheeled odometer 120 is used to acquire the linear velocity and angular velocity of the robot and calculate its pose based on the kinematic model.

[0123] Memory 130 is used to store maps of the calibration room; and

[0124] The processor 140 is configured to perform the calibration method described above to jointly calibrate the lidar 110 and the wheel odometer 120.

[0125] This invention provides a robot that simultaneously utilizes a lidar 110 and a wheeled odometer 120, and also provides a calibration scheme based on a high-precision calibration chamber and lidar. Before calibrating the wheeled odometer 120, the lidar 110 is calibrated first, ensuring the reliability of the wheeled odometer 120. This invention can simultaneously improve the measurement accuracy of the wheeled odometer 120 and the lidar 110, and can also detect faulty wheeled odometers 120 or lidar 110s.

[0126] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A calibration method for a robot, the robot including a lidar and a wheeled odometer, wherein the robot is placed in a calibration chamber, the calibration method comprising: The robot is controlled to rotate sequentially to multiple measurement angles, and the first point cloud is acquired through the lidar. Matching the first point cloud with the point cloud map of the calibration room includes: matching the first point cloud with the point cloud map of the calibration room using the ICP algorithm; wherein the point cloud map of the calibration room is a point cloud map that perfectly matches the size of the calibration room, and the robot pre-stores the point cloud map; the point cloud map of the calibration room is drawn based on the size of the calibration room. Calculate the distance deviation value of each point in the first point cloud, and calibrate the lidar based on the distance deviation value, including: fitting the four sides of the point cloud map with straight lines respectively, calculating the distance deviation value of the intersection point of each point in the first point cloud with the corresponding straight line, and compensating the measurement value of the lidar based on the distance deviation value. The robot is controlled to walk along a preset trajectory and acquire a second point cloud through a calibrated lidar; wherein the preset trajectory is rectangular and has approximately the same aspect ratio as the calibration chamber; The second point cloud is matched with the point cloud map of the calibration room to obtain the first pose of the robot; The second pose is obtained through the wheel odometer; The wheel odometer is calibrated based on the first pose and the second pose.

2. The calibration method according to claim 1, wherein the calibration chamber is rectangular, and the robot is placed in the middle of the calibration chamber near the long side.

3. The calibration method according to claim 2, wherein the dimensional error of the calibration chamber is on the order of millimeters.

4. The calibration method according to claim 1, wherein the step of calculating the distance deviation value of each point in the first point cloud and calibrating the lidar based on the distance deviation value further includes: At each measurement angle, multiple samples are taken, and the average distance deviation between the multiple sampled values ​​at each point and the intersection point of the corresponding straight line is calculated. The measurement value of the lidar is then compensated based on the average distance deviation value.

5. The calibration method according to any one of claims 1-4, wherein the step of matching the second point cloud with the map of the calibration room to obtain the robot's first pose specifically includes: The second point cloud is matched with the map in the calibration room using the ICP algorithm to obtain the robot's first pose.

6. The calibration method according to claim 5, wherein the step of calibrating the wheel odometer based on the first pose and the second pose includes: The second pose is compared with the first pose, and the parameters of the wheel odometer are corrected until the deviation between the second pose and the first pose is less than a threshold.

7. A robot, comprising: LiDAR is used to acquire point clouds; A wheeled odometer is used to acquire the linear velocity and angular velocity of the robot and to calculate its pose based on a kinematic model. Memory, used to store maps of the calibration room; and A processor is configured to execute the calibration method according to any one of claims 1-6 to jointly calibrate the lidar and the wheel odometer.

Images