Sensor parameter adjustment method, robot and computer-readable storage medium

By performing multiple scans and point cloud data processing within the robot's range of activity, calculating sensor deviations, and using virtual plane to correct sensor parameters, the problem of sensor deviation from standard position during robot operation is solved, real-time correction and cost reduction are achieved.

CN114625095BActive Publication Date: 2024-11-29KEENON ROBOTICS CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202210337810.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-31
Publication Date
2024-11-29
Estimated Expiration
2042-03-31

AI Technical Summary

Technical Problem

During the operation of the robot, the sensor deviates from the standard position due to heat generation, vibration, etc., which makes the sensor calibration process cumbersome and expensive, especially in the absence of a specific calibration space, it is difficult to achieve real-time parameter adjustment.

Method used

By controlling the movement of the robot within the range of activity, the sensor performs multiple scans, segments the three-dimensional grid with a continuous field of view, marks and filters point cloud data, calculates the deviation between the sensor and the standard position, performs parameter correction, and uses virtual plane to correct the sensor to correct the sensor Deviation, realize sensor correction during motion.

Benefits of technology

The sensor calibration conditions are simplified, the time and economic cost of sensor calibration are reduced, and the specific calibration space is required, and the accuracy of sensor measurement data is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114625095B_ABST
    Figure CN114625095B_ABST
Patent Text Reader

Abstract

The present invention provides a sensor parameter adjustment method, a robot and a computer-readable storage medium, wherein the sensor parameter adjustment method is used to calibrate a sensor installed on a robot, and the field of view of the sensor can be divided into a plurality of continuous subspaces for processing during data processing, and the sensor parameter adjustment method comprises: controlling the robot to move within the range of activity, controlling the sensor to perform multiple scans, and statistically analyzing the point cloud data of each subspace at different times according to the scanning results; obtaining a point cloud data set according to the statistical results and a preset screening method; and calibrating the sensor parameters according to the point cloud data set. The embodiment of the present invention does not require a specific calibration space, so that the robot can perform sensor calibration during movement, simplifies the calibration conditions of the sensor, and reduces the calibration cost of the sensor. The robot and the computer-readable storage medium can be used to execute the aforementioned sensor parameter adjustment method.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention generally relates to the technical field of intelligent devices, and in particular to a sensor parameter adjustment method, a robot, and a computer-readable storage medium. Background Art

[0002] Robots rely on intelligent control systems and mechanical parts to move, and can replace manual labor in some areas. They are widely used in areas such as transportation and warehousing, precision machining, exhibition guidance, and even firefighting in special scenarios. They are an important direction for the development of intelligent devices. Sensors are important components used in robots to obtain information about the surrounding environment, such as visual sensors and laser radars. Sensors are fixedly installed on the main structure of the robot and have a certain scanning range. The sensors can obtain the positional relationship between obstacles and identification marks within the scanning range and the sensor based on the scanning results.

[0003] Normally, the positional relationship between the sensor and the robot is fixed, and the relative positional relationship between obstacles and identification marks and the robot can be calculated. However, due to the limitations of processing and installation processes, there is a certain deviation between the actual position of the sensor and the theoretical installation position. In addition, the robot will deviate from the standard position due to sensor heating, vibration offset, etc. during operation, and long-term operation will cause the deviation to gradually accumulate. Therefore, there is a need for real-time adjustment of the installation parameters of the robot sensor to ensure the stability provided. Traditional calibration methods need to be performed in a calibration space, but the normal working scene of the robot usually does not have a qualified calibration space, resulting in a cumbersome sensor calibration process with high time and economic costs.

[0004] The contents of the background technology section are merely the technologies known to the inventors and do not necessarily represent the prior art in the field. Summary of the invention

[0005] In view of one or more defects in the prior art, the present invention provides a sensor parameter adjustment method for calibrating the parameters of a sensor installed on a robot, wherein the field of view of the sensor can be divided into a plurality of continuous subspaces for processing during data processing, and the sensor parameter adjustment method comprises:

[0006] Control the robot to move within the range of activity, control the sensor to scan multiple times, and count the point cloud data of each subspace at different times based on the scanning results;

[0007] Obtain a point cloud data set based on statistical results and preset screening methods;

[0008] The sensor parameters are calibrated based on the point cloud data set.

[0009] According to one aspect of the present invention, wherein the subspace in the field of view of the sensor is a continuous three-dimensional grid, the step of acquiring a point cloud data set comprises:

[0010] Marking the three-dimensional grid according to the scanning result of the sensor;

[0011] The three-dimensional grids whose marking times are less than the first threshold are removed to obtain a three-dimensional grid set.

[0012] According to one aspect of the present invention, the step of marking the three-dimensional grid includes: when the sensor obtains point cloud data in a certain three-dimensional grid, the three-dimensional grid is marked once.

[0013] According to one aspect of the present invention, the step of performing parameter calibration on the sensor includes:

[0014] Randomly selecting at least two sub-grid sets that are independent of each other and not on a preset straight line from the three-dimensional grid set, wherein the three-dimensional grids in the sub-grid sets are adjacent to each other;

[0015] According to the point cloud data of the sub-grid set, the deviation value between the sensor and the standard position is calculated;

[0016] Compensate for the scanning error caused by the sensor deviating from the standard position and adjust the sensor parameters.

[0017] According to one aspect of the present invention, the step of obtaining a stereo grid set further comprises: performing weighted averaging on the coordinates with the smallest standard deviation in the stereo grid set to obtain a virtual plane, wherein the weight of the weighted averaging is the number of times the stereo grid is marked.

[0018] According to one aspect of the present invention, the sensor parameter adjustment method also includes converting the point cloud data of the sub-grid set into coordinate values ​​in a three-dimensional rectangular coordinate system; the three-dimensional rectangular coordinate system takes the foot of the perpendicular of the standard position of the sensor in the virtual plane as the origin, takes the perpendicular line between the standard position of the sensor and the virtual plane as the Z axis, takes the projection of the midline of the sensor's field of view in the virtual plane as the X axis, and takes the perpendicular direction of the X axis in the virtual plane as the Y axis; wherein the preset straight lines are the X axis and its parallel line, and the Y axis and its parallel line.

[0019] According to one aspect of the present invention, the sensor parameter adjustment method further comprises:

[0020] Performing weighted averaging on the Z coordinates of the sub-grid set;

[0021] It is determined whether the weighted average value of the Z coordinates of the sub-grid set is less than a second threshold value. If it is not less than the second threshold value, the sub-grid set is deleted and a new sub-grid set is selected until at least two independent sub-grid sets that are not on a preset straight line are obtained.

[0022] According to one aspect of the present invention, the step of calculating the deviation value of the sensor from the standard position comprises:

[0023] Calculate the deviation of the sensor from the standard position in a direction perpendicular to the virtual plane according to the weighted average value of the Z coordinates of the sub-grid set;

[0024] Calculating the pitch angle deviation of the sensor from the standard position according to the sub-grid sets with different weighted average values ​​of two X coordinates;

[0025] The roll angle deviation of the sensor from the standard position is calculated according to two sub-grid sets with different weighted average values ​​of Y coordinates.

[0026] According to one aspect of the present invention, the present invention also includes a robot, comprising:

[0027] main body;

[0028] A sensor, the sensor is disposed on the main body and configured to scan the surrounding environment of the robot to obtain point cloud data;

[0029] a motion device, the motion device being disposed on the main body and capable of being driven to drive the robot to move; and

[0030] A control system is disposed on the main body and communicates with the sensor and the motion device. The control system is configured to execute the sensor parameter adjustment method as described above.

[0031] According to one aspect of the present invention, the present invention also includes a computer-readable storage medium, including computer executable commands stored thereon, and when the executable commands are executed by a processor, the sensor parameter adjustment method as described above is implemented.

[0032] Compared with the prior art, the embodiments of the present invention provide a sensor parameter adjustment method, which controls the robot to move within the range of activity, uses the sensor installed on the robot to scan multiple times within the scanning range, obtains a virtual plane that can be used for calibration and correction, and uses the virtual plane to correct the deviation of the sensor from the standard position, without the need for a specific calibration space, so that the robot can perform sensor calibration during movement, simplifies the sensor calibration conditions, and reduces the sensor calibration cost. The embodiments of the present invention also include a robot and a computer-readable storage medium for executing the aforementioned sensor parameter adjustment method. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0034] Figure 1 is a flow chart of a sensor parameter adjustment method in one embodiment of the present invention;

[0035] Figure 2 is a flow chart of a sensor parameter adjustment method including a three-dimensional grid marking process in one embodiment of the present invention;

[0036] Figure 3 is a flow chart of a sensor parameter adjustment method including a process of determining a virtual plane in one embodiment of the present invention;

[0037] Figure 4 is a flow chart of a sensor parameter adjustment method including a coordinate conversion process in one embodiment of the present invention;

[0038] Figure 5 is a flow chart of a sensor parameter adjustment method including a sub-grid set determination process in one embodiment of the present invention;

[0039] Figure 6 is a flow chart of a sensor parameter adjustment method including a process of determining an external point ratio in one embodiment of the present invention;

[0040] Figure 7 is a flow chart of a sensor parameter adjustment method including a deviation calculation process in one embodiment of the present invention;

[0041] Figure 8 is a schematic diagram of a sensor scanning range and a stereo grid set in one embodiment of the present invention;

[0042] Fig. 9 is a schematic diagram of a virtual plane and a sub-grid set in one embodiment of the present invention;

[0043] Fig.10 1 is a block diagram of a robot in one embodiment of the present invention. DETAILED DESCRIPTION

[0044] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0045] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0046] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, or can communicate with each other; it can be directly connected, or indirectly connected through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0047] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply means that the first feature is lower in level than the second feature.

[0048] The disclosure below provides many different embodiments or examples to realize different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or reference letters in different examples, and this repetition is for the purpose of simplicity and clarity, which itself does not indicate the relationship between the various embodiments and / or settings discussed. In addition, the present invention provides various specific examples of processes and materials, but those of ordinary skill in the art can be aware of the application of other processes and / or the use of other materials.

[0049] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0050] Figure 1 The specific process of the sensor parameter adjustment method 1 in the preferred embodiment of the present invention is shown below. Figure 1 Detailed description.

[0051] like Figure 1 As shown, in step S1, the robot is controlled to move within the range of activity, the sensor is controlled to scan multiple times, and the point cloud data of each subspace at different times are counted according to the scanning results. The robot in this embodiment has a specific range of activity, for example, a storage robot runs in a warehouse. During the movement of the robot, the position scanned by the sensor also changes. In this step, the point cloud data of each subspace within the field of view at different times are counted according to the scanning results of the sensor. According to a preferred embodiment of the present invention, the number of times the sensor obtains point cloud data in each subspace during the movement of the robot is counted. For example, the sensor is a laser radar. During the movement of the robot, the laser radar obtains point cloud data of the surrounding environment and counts the number of times point cloud data appears in different subspaces. In step S2, a point cloud data set is obtained according to the statistical results and a preset screening method. In step S3, the sensor is parameterized according to the point cloud data set. The specific method of obtaining the point cloud data set and parameter correction is described in detail in the subsequent embodiments.

[0052] Figure 2 The specific process of the sensor parameter adjustment method 100 according to the preferred embodiment of the present invention is shown. Figure 8 The schematic diagram of the sensor scanning range according to the preferred embodiment of the present invention is shown below. Figure 2 and Figure 8 Detailed description.

[0053] The sensor parameter adjustment method 100 in this embodiment is used to calibrate the sensors installed on the robot, such as Figure 8 As shown, the scanning range of the sensor in this embodiment is divided into continuous three-dimensional grids. In actual application scenarios, the scanning range of the sensor is usually a hemispherical or spherical range with the sensor as the center. However, at the edge of the hemispherical range, the point cloud data obtained by the sensor is severely distorted, which is generally considered to exceed the effective scanning range of the sensor. According to a specific embodiment of the present invention, the approximate center position of the actual scanning range of the sensor can be selected as the scanning range in this embodiment, for example Figure 8 As shown, the scanning range of the sensor is a cubic space with an edge length of 3 meters. The scanning range of the sensor can be divided into a plurality of continuous three-dimensional grids. Specifically, each individual three-dimensional grid can be set as a cube with an edge length of 2 centimeters. In different embodiments of the present invention, the size and shape of the individual three-dimensional grids can be different. The smaller the size of the three-dimensional grid, the higher the accuracy of calibrating the sensor using the sensor parameter adjustment method 100, and accordingly, the more computing resources are occupied.

[0054] In step S101, the robot is controlled to move within the range of movement, the sensor is controlled to scan multiple times, and the three-dimensional grid is marked according to the scanning results. During the movement of the robot, the scanning range of the sensor will also change with the movement of the robot's position, but the relative position of the three-dimensional grid and the robot has not changed. During the movement, the sensor is controlled to scan multiple times to obtain multiple scanning results, and the three-dimensional grid is marked according to the scanning results. According to a preferred embodiment of the present invention, the sensor obtains point cloud data at a certain three-dimensional grid position, which is specifically determined according to the working principle of the sensor, such as the laser radar receiving the reflected echo of an obstacle. In this embodiment, point cloud data is obtained at a specific three-dimensional grid position, that is, a mark is made on the corresponding three-dimensional grid. After multiple scans by the sensor, the number of times different three-dimensional grids are marked will also vary.

[0055] In step S102, the three-dimensional grids whose marking times are less than the first threshold are removed to obtain a three-dimensional grid set. As mentioned above, in the multiple scanning processes of the sensor, different three-dimensional grids are marked at different times. The three-dimensional grids with more marking times indicate that the number of obstacles scanned at this three-dimensional grid (a certain position relative to the robot) is large. Correspondingly, the three-dimensional grids with fewer marking times indicate that the number of obstacles scanned at this three-dimensional grid (a certain position relative to the robot) is small. During the movement of the robot, the frequency of obstacles at this position is low. Step S102 can also remove dynamic obstacles. Within the scanning range of the sensor, dynamic obstacles appear or affect the scanning results of the sensor. At the same time, the number of times the point cloud data of the dynamic obstacles is obtained is less than the number of times the static plane point cloud data is obtained. Removing the three-dimensional grids whose marking times are less than the first threshold can reduce the interference of dynamic obstacles on sensor calibration.

[0056] The first threshold can be a specific value. For example, during the movement of the robot, the sensor is controlled to scan 100 times, and the first threshold can be set to 70. The first threshold can also be a proportional value. For example, during the movement of the robot, the sensor is scanned multiple times, and the first threshold is set to 50%. When the number of times the three-dimensional grid is marked is less than 50% of the number of times the sensor scans, the three-dimensional grid is removed. Take the visual sensor set in front of the robot as an example. During the movement of the robot, the sensor can scan the range in front of the robot. For the scanning result of the sensor, such as Figure 8 As shown in the figure, the ground, obstacles, moving human bodies, etc. are scanned and marked at the corresponding three-dimensional grids. After the selection in this step, a set of three-dimensional grids that frequently appear in the sensor scanning range is obtained, and the actual position corresponding to the three-dimensional grid set is considered to be the ground within the robot's activity range. In this embodiment, the size of the three-dimensional grid is set to a cube with an edge length of 1 cm. Considering the scanning accuracy of the sensor and the flatness of the ground, the actual three-dimensional grid set obtained is not necessarily as Figure 8 The continuous layer shown may also be up and down or missing. According to other embodiments of the present invention, the three-dimensional grid set obtained according to this step may also represent other physical planes, such as a ceiling or a straight wall.

[0057] In step S103, at least two independent sub-grid sets that are not on a preset straight line are randomly selected from the stereo grid set, wherein the stereo grids in the sub-grid sets are adjacent. Mutually independent means that the ranges of the sub-grid sets do not overlap, and not on a preset straight line means that there is an angle between the line connecting the two sub-grid sets and the preset straight line, but not all lines connecting any two sub-grid sets have an angle with the preset straight line. The preset straight line is combined with Fig. 9 Detailed description. The number of sub-grid sets is at least two, and more than one may be selected. For example, according to a preferred embodiment of the present invention, four sub-grid sets are selected from the point cloud data of the stereo grid set, and there is an angle between the connecting line of two of the four sub-grid sets and the preset straight line. According to a preferred embodiment of the present invention, the spacing between adjacent sub-grid sets in the actual space is not less than a second preset distance, such as 20 centimeters. The larger the spacing between the sub-grid sets, the more accurately it can reflect the deviation value of the sensor from the standard position.

[0058] In step S104, the deviation value between the sensor and the standard position is calculated based on the point cloud data of the sub-grid set. As mentioned above, the relative position between the three-dimensional grid and the robot is fixed. After obtaining the three-dimensional grid set, the standard position of the sensor is known, and the angle and distance between the standard position of the sensor and the ground can be obtained by the height at which the sensor is installed on the robot. The vector composed of the angle and distance is the point cloud data of the sub-grid set. According to the point cloud data of the sensor and the sub-grid set, the deviation value between the sensor and the standard position is calculated. In step S105, the scanning error caused by the deviation of the sensor from the standard position is compensated to complete the calibration of the sensor. The standard position refers to the theoretical installation position of the sensor on the robot, but in the actual processing, installation and subsequent use, the sensor may deviate from the standard position, resulting in errors in the scanning data of the sensor. After obtaining the point cloud data of the sub-grid set, the scanning error of the sensor can be obtained by calculation according to the characteristics of the point cloud data, and then the measured value is compensated according to the error value to be closer to the true value, completing the calibration of the sensor.

[0059] Figure 3 The specific process of the sensor parameter adjustment method 200 according to the preferred embodiment of the present invention is shown below. Figure 3 and Figure 8 Detailed description.

[0060] Steps S201, S202, S204, S205 and S206 in the sensor parameter adjustment method 200 are substantially the same as steps S101, S102, S103, S104 and S105 in the sensor parameter adjustment method 100, and are not described in detail here. In step S203, after obtaining the stereo grid set, the coordinates with the smallest standard deviation in the stereo grid set are weighted averaged to obtain a virtual plane, wherein the weight of the weighted average is the number of times the stereo grid is marked. As described in the above embodiments, the stereo grids in the stereo grid set do not completely belong to a continuous plane, wherein there are differences due to scanning errors and physical plane fluctuations. In this step, the coordinates with the smallest standard deviation are weighted averaged to obtain a virtual plane with unified coordinates. The coordinates with the smallest standard deviation represent the direction with the smallest coordinate difference, thereby eliminating the stereo grids that are obviously not in the virtual plane. In specific applications, the coordinate with the smallest standard deviation only represents a certain coordinate direction, not the actual calculation of the standard deviation. Still taking the ground as an example, in the aforementioned steps, the robot movement is controlled, and the direction represented by the ground is known. According to a specific embodiment of the present invention, the direction of the virtual plane can be directly determined. In the multiple scanning results of the sensor, the three-dimensional grids in the three-dimensional grid set are marked multiple times. The unified coordinates obtained by weighted averaging with this as the weight can more accurately determine the virtual plane. The coordinates in this step can be the coordinates in the three-dimensional rectangular coordinate system specifically described in the subsequent embodiments, or they can be the coordinates established by the specific division method of the sensor's scanning range into three-dimensional grids. For example, one of the three-dimensional grids is taken as the origin, recorded as (0,0,0), and the positive direction is arbitrarily determined to determine the coordinates of all three-dimensional grids relative to the origin, and expressed in the same way.

[0061] After determining the virtual plane, according to some preferred embodiments of the present invention, the deviation values ​​of the sensor from the standard position are decomposed into the distance deviation in the direction perpendicular to the virtual plane, the pitch angle deviation in the direction perpendicular to the virtual plane, and the roll angle deviation in the direction parallel to the virtual plane.

[0062] Figure 4 The specific process of the sensor parameter adjustment method 300 including the coordinate conversion process in the preferred embodiment of the present invention is shown below. Figure 4 Steps S301, S302, S303, S304, S306 and S307 in the sensor parameter adjustment method 300 are substantially the same as steps S201, S202, S203, S204, S205 and S206 in the sensor parameter adjustment method 200, and are not described in detail.

[0063] In step S305 of this embodiment, the point cloud data of the sub-grid set is converted into coordinate values ​​in a three-dimensional rectangular coordinate system. The three-dimensional rectangular coordinate system in this embodiment is established according to the standard position of the sensor and the virtual plane, with the foot of the perpendicular of the standard position of the sensor in the virtual plane as the origin, the perpendicular line between the standard position of the sensor and the virtual plane as the Z axis, the projection of the midline of the sensor's field of view in the virtual plane as the X axis, and the vertical direction of the X axis in the virtual plane as the Y axis, as shown in FIG. Fig. 9 As shown, the virtual plane of the rectangular shaded part has the X axis at the center line, the virtual plane is represented by the XY plane, and the Z axis is the vertical line of the virtual plane. Further, the preset straight line described in step S304 is the X axis and its parallel line, and the Y axis and its parallel line, that is, the X coordinates of two or more sub-grid sets are not completely the same, and the Y coordinates are not completely the same. Based on the (X, Y, Z) coordinates of the sub-grid set, the distance deviation of the sensor in the vertical direction of the virtual plane, the pitch angle deviation in the vertical direction of the virtual plane, and the roll angle deviation in the parallel direction of the virtual plane can be calculated, and compensation is performed according to the deviation value to complete the calibration of the sensor.

[0064] Figure 5 The specific flow of the sensor parameter adjustment method 400 including the process of determining the sub-grid set according to the preferred embodiment of the present invention is shown below. Figure 5 In this embodiment, steps S401, S402, S403, S404, S405, S408 and S409 are substantially the same as steps S301, S302, S303, S304, S305, S306 and S307 in the sensor parameter adjustment method 300, and are not described in detail.

[0065] In step S406, the Z coordinates of the sub-grid set are weighted averaged. Further, according to a preferred embodiment of the present invention, the weight of the weighted average can be selected from the number of times the stereo grid is marked. In the three-dimensional rectangular coordinate system established in the above steps, the Z coordinate represents the distance from the virtual plane. The stereo grids in the sub-grid set are theoretically located in the virtual plane, but there are still factors such as scanning errors. Therefore, in this step, the Z coordinates of the sub-grid set are weighted averaged to obtain a unified Z coordinate corresponding to a sub-grid set. In step S407, it is determined whether the weighted average value of the Z coordinates of the sub-grid set is less than the second threshold value. The specific second threshold value can be set to 5 cm. In theory, the sub-grid set should be located in the virtual plane, but due to sensor scanning errors and sensor deviations from the standard position, the Z coordinates of the sub-grid set should vary within the range allowed by the deviation and error. If the weighted average value of the Z coordinates of the sub-grid set is not less than the second threshold value, in step S410, the sub-grid set is deleted, and a sub-grid set that meets the above conditions is selected again in the stereo grid set. If the weighted average value of the Z coordinates of the sub-grid set is not less than the second threshold value, it means that there is a large deviation in the scanning result of the sensor at the position corresponding to the sub-grid set. For example, there are continuous and uniform obstacles within the range of the robot's activity, such as the skirting on the side of the wall. The sub-grid set at this position needs to be deleted and re-selected. According to a preferred embodiment of the present invention, when re-selecting the sub-grid set, the distance between the re-selected sub-grid set and the deleted sub-grid set is not less than the first preset distance, such as 10 cm, to eliminate possible interference from obstacles. If the weighted average value of the Z coordinates of the sub-grid set is less than the second threshold value, in step S408, the deviation value of the sensor from the standard position is calculated based on the coordinate value of the sub-grid set.

[0066] Figure 6 The specific flow of the sensor parameter adjustment method including the process of determining the external point ratio in a preferred embodiment of the present invention is shown below. Figure 6 Steps S501, S502, S503, S504, S505, S506, S507, S510 and S511 in this embodiment are substantially the same as steps S401, S402, S403, S404, S405, S406, S407, S408 and S409 in the sensor parameter adjustment method 400, and are not described in detail.

[0067] In step S508, the outliers in the sub-grid set are counted and the outlier ratio is calculated. In this embodiment, the outliers refer to the stereo grids in the sub-grid set whose Z coordinates are not less than a certain threshold value, and isolated stereo grids. The certain threshold value can be selected according to the scanning accuracy of the sensor, for example, 7 cm. Due to inevitable scanning errors and obstacles, there may be some outliers in the sub-grid set, namely error points. The outlier ratio refers to the ratio of stereo grids belonging to outliers to the total number of stereo grids in the sub-grid set.

[0068] In step S509, it is determined whether the ratio of the external points conforms to the virtual plane. If the ratio of the external points is small and conforms to the virtual plane, then steps S510 and S511 are performed to calibrate the sensor using the selected sub-grid set. If the ratio of the external points is too large, it means that there are large fluctuations in the sub-grid set, or there are large errors in the sensor measurement data, which do not conform to the virtual plane. Therefore, in step S512, it is necessary to delete the sub-grid set and reselect the sub-grid set. Furthermore, the distance between the reselected sub-grid set and the deleted sub-grid set is not less than the first preset distance.

[0069] like Figure 6 As shown, in step S513, it is determined whether the number of times the sub-grid set is reselected is greater than a third threshold, for example, 5 times. If the number of times the sub-grid set is reselected is not greater than the third threshold, step S505 is repeated to convert the point cloud data of the newly selected sub-grid set into coordinate values ​​in a three-dimensional rectangular coordinate system. If the number of times the sub-grid set is reselected is greater than the third threshold, in step S514, the robot is controlled to issue an alarm. The virtual plane in this embodiment is not the physical plane actually scanned by the sensor, but a virtual plane calculated based on the results of multiple scans of the sensor. When multiple sub-grid sets are selected and none of them meet the virtual plane, it means that the physical plane in the robot motion path has large fluctuations, and a virtual plane that meets the requirements cannot be obtained, or the deviation value between the sensor and the standard position is large, which exceeds the range of parameter adjustment.

[0070] Figure 7 The specific process of the sensor parameter adjustment method including the calculation deviation process in the preferred embodiment of the present invention is shown below. Figure 7 In this embodiment, S601, S602, S603, S604, S605, S606, S607, S608, S609, S613, S614, S615 and S616S are substantially the same as S501, S502, S503, S504, S505, S506, S507, S508, S509, S511, S512, S513 and S514, respectively, and are not described in detail.

[0071] In this embodiment, the deviation between the sensor and the standard position is decomposed into the vertical deviation on the Z axis, the roll angle deviation in the ZY plane and the pitch angle deviation in the ZX plane. In step S610, the deviation between the sensor and the standard position in the direction perpendicular to the virtual plane, that is, the vertical deviation on the Z axis, is calculated according to the weighted average value of the Z coordinates of different sub-grid sets. Still taking the ground as an example, the distance between the standard position of the sensor and the ground can be obtained according to the height of the standard position of the sensor on the robot. Theoretically, the Z coordinate of the sub-grid set should be 0. The distance deviation of the sensor on the Z axis can be obtained according to the weighted average value of the measured data of a group of sub-grid sets, or the vertical deviation of the sensor on the Z axis can be obtained according to the average value of the weighted average values ​​of the Z coordinates of multiple sub-grid sets.

[0072] In step S611, the pitch angle error between the sensor and the standard position is calculated based on two sub-grid sets with different X coordinates, for example Fig. 9 In the sub-grid set A1 and the sub-grid set A3 or the sub-grid set A4, when the sensor is in the standard position, the angle between the midline of the sensor scanning range and the virtual plane is a known fixed value. According to the change of the X coordinates in the two sub-grid sets with different X coordinates, the pitch angle deviation of the sensor from the standard position can be calculated. In step S612, according to the two sub-grid sets with different Y coordinates, the roll angle deviation of the sensor from the standard position is calculated. For example, Fig. 9 In the sub-grid set A1 and the sub-grid set A2 or the sub-grid set A3, according to the change of the Y coordinate of the sub-grid set, when the sensor is in the standard position, the angle between the center line of the sensor scanning range and the virtual plane is a known fixed value, and according to the change of the Y coordinate in the two sub-grid sets with different Y coordinates, the roll angle deviation of the sensor from the standard position can be calculated. Of course, in different embodiments of the present invention, the distance deviation on the Z axis, the roll angle deviation and the pitch angle deviation can be calculated as Figure 7 The three deviation values ​​can be calculated in sequence, or simultaneously, or in any sequence. After the three deviation values ​​are calculated, the measurement data of the sensor is compensated to complete the calibration of the sensor.

[0073] According to a preferred embodiment of the present invention, when sensor calibration is required, a preset instruction can be input to enable the robot to execute the aforementioned sensor parameter adjustment method, or the robot can be controlled to execute the sensor parameter adjustment method after the robot operates normally for a preset time, for example, every 24 hours. In this embodiment, the sensor parameter adjustment method does not need to suspend the normal operation of the robot. When the robot moves to perform other tasks, the deviation between the sensor and the standard position can be corrected at the same time to ensure that the robot can move normally, and there is no need to provide a separate calibration space or arrange calibration time, which greatly reduces the sensor calibration cost and improves the sensor calibration efficiency.

[0074] The present invention also includes an embodiment of a robot, such as Fig.10 As shown, the robot 1 includes a main body, a sensor 10, a control system 20 and a motion device 30. The main body is the external frame of the robot 1, which is used to support and install the components inside it. The sensor 10 is installed on the main body and can scan the environment around the robot 1 to obtain point cloud data of the surrounding environment, such as a visual sensor, a laser radar, etc. The motion device 30 communicates with the control system 20 and can be controlled to drive the robot to move. The control system 20 is set on the main body and communicates with the sensor 10. It can execute the aforementioned sensor parameter adjustment method, calibrate the position of the sensor, and improve the accuracy of the sensor measurement data. The control system 20 can execute the aforementioned sensor calibration method after receiving a preset instruction or after a certain preset time interval.

[0075] The present invention also includes an embodiment of a computer-readable storage medium, including computer executable instructions stored thereon, and when the executable instructions are executed by a processor, the sensor calibration method as described above is implemented.

[0076] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A sensor parameter adjustment method for calibrating the parameters of a sensor installed on a robot, wherein the field of view of the sensor can be divided into a plurality of continuous subspaces for processing during data processing. include: Control the robot to move within the range of activity, control the sensor to scan multiple times, and count the point cloud data of each subspace at different times based on the scanning results; Obtain a point cloud data set based on statistical results and preset screening methods; According to the point cloud data set, the sensor parameters are calibrated; Wherein the subspace in the field of view of the sensor is a continuously arranged three-dimensional grid, and the step of acquiring a point cloud data set comprises: Marking the three-dimensional grid according to the scanning result of the sensor; Remove three-dimensional grids whose marking times are less than a first threshold, and obtain a three-dimensional grid set; The step of performing parameter calibration on the sensor comprises: Randomly selecting at least two sub-grid sets that are independent of each other and not on a preset straight line from the three-dimensional grid set, wherein the three-dimensional grids in the sub-grid sets are adjacent to each other; According to the point cloud data of the sub-grid set, the deviation value between the sensor and the standard position is calculated; Compensate for the scanning error caused by the sensor deviating from the standard position and adjust the sensor parameters.

2. The sensor parameter adjustment method according to claim 1, wherein the step of marking the three-dimensional grid include: When the sensor obtains point cloud data in a certain three-dimensional grid, the three-dimensional grid is marked.

3. The sensor parameter adjustment method according to claim 1, wherein the step of obtaining a stereo grid set further comprises: include: A weighted average is performed on the coordinates with the smallest standard deviation in the three-dimensional grid set to obtain a virtual plane, and the weight of the weighted average is the number of times the three-dimensional grid is marked.

4. The sensor parameter adjustment method according to claim 3 further includes converting the point cloud data of the sub-grid set into coordinate values ​​in a three-dimensional rectangular coordinate system; the three-dimensional rectangular coordinate system takes the foot of the perpendicular of the standard position of the sensor in the virtual plane as the origin, takes the perpendicular line between the standard position of the sensor and the virtual plane as the Z axis, takes the projection of the midline of the field of view of the sensor in the virtual plane as the X axis, and takes the perpendicular direction of the X axis in the virtual plane as the Y axis; wherein the preset straight lines are the X axis and its parallel line, and the Y axis and its parallel line.

5. The sensor parameter adjustment method according to claim 4, further comprising: include: Performing weighted averaging on the Z coordinates of the sub-grid set; It is determined whether the weighted average value of the Z coordinates of the sub-grid set is less than a second threshold value. If it is not less than the second threshold value, the sub-grid set is deleted and a new sub-grid set is selected until at least two independent sub-grid sets that are not on a preset straight line are obtained.

6. The sensor parameter adjustment method according to claim 5, further comprising: include: Count the outliers in the subgrid set and calculate the outlier ratio; Determine whether the ratio of the external point conforms to the virtual plane, and if so, calculate the deviation value between the sensor and the standard position according to the point cloud data of the sub-grid set; If it does not conform to the virtual plane, the subgrid set is deleted, and a new subgrid set is selected, and the distance between the new subgrid set and the deleted subgrid set is not less than a first preset distance.

7. The sensor parameter adjustment method according to claim 6, further comprising: include: Determining whether the number of times the subgrid set is reselected is greater than a third threshold; If the number of times the sub-grid set is reselected is not greater than the third threshold, repeatedly converting the point cloud data of the sub-grid set into coordinate values ​​in a three-dimensional rectangular coordinate system; If the number of times the sub-grid set is reselected is greater than a third threshold, the robot is controlled to issue an alarm.

8. The sensor parameter adjustment method according to claim 4, wherein the step of calculating the deviation value of the sensor from the standard position include: Calculate the deviation of the sensor from the standard position in a direction perpendicular to the virtual plane according to the weighted average value of the Z coordinates of the sub-grid set; Calculating the pitch angle deviation of the sensor from the standard position according to the sub-grid sets with different weighted average values ​​of two X coordinates; The roll angle deviation of the sensor from the standard position is calculated according to two sub-grid sets with different weighted average values ​​of Y coordinates.

9. A robot, include: main body; A sensor, the sensor is disposed on the main body and configured to scan the surrounding environment of the robot to obtain point cloud data; A motion device, which is disposed on the main body and can be driven to drive the robot to move; and A control system, wherein the control system is disposed on the main body and communicates with the sensor and the motion device, and the control system is configured to execute the sensor parameter adjustment method according to any one of claims 1-8. 10 . A computer-readable storage medium, comprising computer executable commands stored thereon, wherein the executable commands implement the sensor parameter adjustment method according to any one of claims 1 to 8 when executed by a processor.

Citation Information

Patent Citations

  • External parameter calibration method and device, computer equipment and storage medium

    CN112578356A