Pose control method, device and system for floor grinding

Abstract

The invention discloses a position and posture control method, device and system for terrace grinding, and belongs to the technical field of intelligent control, and the method comprises the steps: carrying out the data extraction of the point cloud data of a ground region, obtaining ground point clouds, carrying out the screening of the ground point clouds according to the median of a vertical coordinate, and obtaining the selected ground point clouds; dividing a plurality of to-be-polished areas and center coordinates of the to-be-polished areas; distance measurement is conducted through a laser radar, and pose information of the floor grinding device relative to obstacles in the ground area is obtained; according to the collected current pose information of the terrace grinding device, the moving direction of the terrace grinding device is controlled through the center coordinates, and grinding is started until the terrace grinding device reaches the to-be-ground area; and when the floor polishing device moves, the moving speed and the moving angle of the floor polishing device are controlled for obstacle avoidance. Therefore, by implementing the method and the device, the problems that accurate obstacle avoidance is difficult to realize and the to-be-polished area cannot be accurately identified in the complex terrain in the prior art can be solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent control, and particularly relates to a pose control method, device and system for floor polishing. BACKGROUND

[0002] In the field of modern construction and industry, the flatness and smoothness of the floor are crucial to the overall engineering quality. Traditional floor polishing work mainly relies on manual operation of polishing equipment, which not only has high labor intensity and low work efficiency, but also is difficult to ensure the consistency of polishing quality, and is prone to problems such as uneven polishing and missed polishing. With the rapid development of automation technology, robot technology and intelligent control technology, various industries are actively exploring automation solutions to improve production efficiency and product quality. In the field of floor polishing, it is inevitable to introduce an automatic polishing device. Through the automatic equipment, the polishing process can be accurately controlled.

[0003] However, how to accurately control the pose movement of the automatic floor polishing device to ensure that the polishing tool can accurately reach the area to be polished and accurately avoid obstacles is a major problem that needs to be solved at present. Moreover, in the face of different shapes of the floor terrain, the automatic floor polishing device also needs to quickly and accurately identify the local high area to be polished, provide a good base flatness for floor paving, and improve the applicability and flexibility of the device. SUMMARY

[0004] The application provides a pose control method, device and system for floor polishing, which can solve the problem that it is difficult to accurately avoid obstacles and accurately identify the area to be polished in complex terrain in the prior art.

[0005] The first aspect of the application provides a pose control method for floor polishing, which comprises:

[0006] extracting point cloud data of the collected ground area to obtain ground point cloud, and screening the ground point cloud according to the median value of the longitudinal coordinate to divide a plurality of polishing areas and the center coordinates of the polishing areas;

[0007] measuring the distance in the ground area by using a laser radar to obtain the pose information of the floor polishing device relative to the obstacles in the ground area;

[0008] controlling the moving direction of the floor polishing device according to the current pose information of the floor polishing device in the ground point cloud through the center coordinates, until the floor polishing device reaches the polishing area and starts polishing;

[0009] controlling the moving speed and moving angle of the floor polishing device to avoid the obstacles according to the pose information when the floor polishing device moves.

[0010] The above scheme uses a three-dimensional laser range finder to collect point cloud data of the area that needs to be polished, and then filters and extracts the ground point cloud from the point cloud data to reduce the point cloud storage amount, which can speed up the subsequent point cloud processing speed. Then, according to the longitudinal coordinate standard, the locally high area is identified and polished to accurately identify the area to be polished in the complex terrain, and then the corresponding center coordinates are found to facilitate the subsequent movement of the floor polishing device to the area to be polished for path planning. Moreover, the laser radar is used to detect the distance between obstacles during the movement of the floor polishing device to ensure that the floor polishing device can avoid obstacles perfectly during polishing. In addition, the current pose information of the floor polishing device is accurately measured by the inertial measurement unit to control the moving direction of the device, and accurate path planning during polishing is realized.

[0011] In a possible implementation method of the first aspect, the point cloud data of the collected ground area is data extracted to obtain the ground point cloud, specifically:

[0012] When the floor polishing device is in a stationary state, the ground area is scanned by a three-dimensional laser scanner to obtain point cloud data of the ground area;

[0013] The point cloud data below the preset floor elevation is deleted from the point cloud data to obtain first point cloud data;

[0014] The first point cloud data is sequentially subjected to bilateral filtering and down-sampling processing, and then the normal vector dot product of each point cloud in the down-sampled first point cloud data is calculated; wherein the local plane is a plane obtained by fitting the point cloud and other point clouds within a preset adjacent range;

[0015] After obtaining the normal vector dot product results of all the local planes, the ground point cloud is filtered from the down-sampled first point cloud data by comparing the size of the normal vector dot product results.

[0016] The above scheme first scans the ground area to obtain corresponding point cloud data, and then deletes the point cloud below the preset floor elevation from the point cloud data to complete the vertical positioning of the point cloud. Then, the first point cloud data is denoised by bilateral filtering, the point cloud storage amount is reduced by down-sampling processing, and the subsequent processing speed is accelerated by reducing the data amount. Then, the feature recognition is completed by calculating the normal vector of each point cloud and the corresponding local plane to find the locally high area to be polished, and the area to be polished is accurately identified from the complex environment.

[0017] In a possible implementation method of the first aspect, the ground point cloud is filtered according to the median value of the longitudinal coordinate, and a plurality of areas to be polished and the center coordinates of the areas to be polished are divided, specifically:

[0018] The median of the ordinate is calculated from the ordinate of the ground point cloud, and a horizontal plane equal to the median of the ordinate is constructed.

[0019] If there are three consecutive points in the ground point cloud with ordinates greater than the median ordinate within a preset size range, then the corresponding area to be polished is determined based on the three consecutive points with ordinates greater than the median ordinate, and the corresponding center coordinate is determined based on the horizontal coordinate range of the area to be polished; wherein the horizontal coordinate range is composed of the horizontal coordinate range and the vertical coordinate range of the area to be polished.

[0020] The above method identifies the range of each area to be polished and determines its corresponding center coordinates by using the median of the vertical coordinates. Subsequently, the forward path of the floor polishing device can be set based on these center coordinates to achieve path planning.

[0021] In one possible implementation of the first aspect, a lidar is used to measure distance within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area, specifically:

[0022] Using lidar located at both ends of the floor grinding device, the minimum lateral distance between the front and rear sides of the floor grinding device and obstacles in the ground area, as well as the directional angle between the floor grinding device and obstacles in the ground area are measured respectively.

[0023] The above solution uses lidar to collect the minimum lateral distance and azimuth angle of the floor grinding device to determine the distance between the device and obstacles, thereby enabling precise and timely obstacle avoidance during the device's movement.

[0024] In one possible implementation of the first aspect, based on the current pose information of the floor grinding device in the ground point cloud, the movement direction of the floor grinding device is controlled by the center coordinates until the floor grinding device reaches the area to be ground and grinding begins, specifically as follows:

[0025] After synchronizing the timestamps and coordinate systems of the inertial measurement unit, the 3D laser scanner, and the lidar, the position of the inertial measurement unit is projected onto the coordinate system of the 3D laser scanner to obtain the current pose information;

[0026] Based on the center coordinates and the current pose information, the angle between the floor grinding device and the area to be ground, as well as the direction of movement, are obtained.

[0027] Based on the included angle, the floor grinding device is controlled to move toward the center coordinate and the direction of movement is continuously adjusted until the included angle is zero.

[0028] When the included angle is zero, the grinding disc of the floor grinding device is driven by the differential drive wheel to start grinding.

[0029] Before controlling the movement of the floor grinding device, the above scheme first performs spatiotemporal synchronization on the inertial measurement unit, 3D laser scanner, and lidar. This is because these three measuring instruments use different coordinate systems, and to ensure the consistency of the collected data, it is necessary to unify the timestamps and coordinate systems. After spatiotemporal synchronization, the floor grinding device is driven to move towards each central coordinate, and the direction of movement is continuously adjusted according to the data collected by the lidar during the movement, until the device reaches the central coordinate and begins grinding.

[0030] In one possible implementation of the first aspect, when the floor grinding device moves, the moving speed and moving angle of the floor grinding device are controlled according to the pose information to avoid the obstacle, specifically:

[0031] Based on the pose information and the planar dimensions of the floor grinding device, the angular velocity and linear velocity of the floor grinding device are obtained;

[0032] Adjust the angular and linear velocities of the differential drive wheel based on the angular and linear velocities of the floor grinding device and the directional angle between the floor grinding device and obstacles in the ground area.

[0033] One possible implementation of the first aspect also includes:

[0034] After all the areas to be polished have been polished, the ground area is scanned again using a 3D laser scanner to obtain the maximum vertical coordinates of the point cloud in this scan.

[0035] The maximum vertical coordinate of the scanned point cloud is compared with the median vertical coordinate of the point cloud data. The polishing process ends when the maximum vertical coordinate is equal to the median vertical coordinate.

[0036] After all areas to be sanded have been sanded, the maximum vertical coordinate is obtained and then compared with the median vertical coordinate during the previous scan. When the median vertical coordinate before sanding is equal to the maximum vertical coordinate after sanding, it indicates that there are no excessively high areas and the floor height is uniform, so the sanding work is complete.

[0037] The second aspect of this application provides a posture control device for floor grinding, the device comprising: a grinding area division module, a posture information acquisition module, an area grinding module, and a movement obstacle avoidance module;

[0038] The grinding area division module is used to extract ground point cloud data from the collected ground area, and to filter the ground point cloud according to the median of the vertical coordinate to divide it into several grinding areas and the center coordinates of the grinding areas.

[0039] The pose information acquisition module is used to measure distances using lidar within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area;

[0040] The area grinding module is used to control the movement direction of the floor grinding device through the center coordinates based on the current pose information of the floor grinding device on the ground point cloud, until the floor grinding device reaches the area to be ground and then starts grinding.

[0041] The mobile obstacle avoidance module is used to control the moving speed and moving angle of the floor grinding device to avoid obstacles when the floor grinding device moves, based on the posture information.

[0042] A third aspect of this application provides a posture control system for floor grinding, the system comprising: a posture control device for floor grinding and a floor grinding device.

[0043] The posture control device for floor grinding is used to implement the posture control method for floor grinding as described in any one of the embodiments of this application.

[0044] The floor grinding device is used to grind up excessively high areas on the floor.

[0045] In one possible implementation of the third aspect, the floor grinding device specifically comprises:

[0046] The floor grinding device includes a three-dimensional laser scanner, an inertial measurement unit, several lidar sensors, a grinding disc, and several differential drive wheels;

[0047] The three-dimensional laser scanner is used to scan the ground area and generate point cloud data of the area to be polished when the floor polishing device is stationary.

[0048] The inertial measurement unit is used to determine the current pose information of the floor grinding device;

[0049] The lidar is used to collect real-time positional information of the floor grinding device relative to obstacles in the ground area;

[0050] The grinding disc is used to grind up localized areas that are too high within the area to be ground.

[0051] The differential drive wheel is used to drive the floor grinding device to move. Attached Figure Description

[0052] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0053] Figure 1 This is a schematic flowchart of a posture control method for floor grinding provided in a certain embodiment of this application;

[0054] Figure 2 This is a structural diagram of a floor grinding device according to a posture control method for floor grinding provided in a certain embodiment of this application;

[0055] Figure 3 This is a side view of a three-dimensional laser scanner of a pose control method for floor grinding provided in a certain embodiment of this application;

[0056] Figure 4 This is a lidar ranging image of a pose control method for floor grinding provided in a certain embodiment of this application;

[0057] Figure 5 This is a coordinate system top view of a posture control method for floor grinding provided in a certain embodiment of this application;

[0058] Figure 6 This is a moving diagram of a floor grinding device according to a posture control method for floor grinding provided in a certain embodiment of this application;

[0059] Figure 7 This is a structural diagram of a posture control device for floor grinding provided in a certain embodiment of this application;

[0060] Figure 8 This is a structural diagram of a posture control system for floor grinding provided in a certain embodiment of this application. Detailed Implementation

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0062] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0063] First Embodiment

[0064] To improve production efficiency and grinding quality, the introduction of automated grinding equipment in the floor grinding industry has become an inevitable trend. Although automated grinding equipment can reduce labor costs, there are still certain technical barriers in complex terrain environments, such as accurately identifying excessively high areas and planning paths to complete the grinding. Therefore, to improve grinding quality, precise positional control is also needed during automated grinding to ensure that the floor grinding equipment can accurately avoid obstacles and completely grind all excessively high areas during movement.

[0065] like Figure 1 As shown, to address the problems in existing technologies such as difficulty in achieving precise obstacle avoidance and inability to accurately identify the area to be polished in complex terrain, the first embodiment of this application provides a detailed flowchart of a pose control method for floor polishing. This embodiment of the pose control method for floor polishing includes steps S1 to S4, detailed below:

[0066] Step S1: Extract the ground point cloud data from the collected ground area to obtain the ground point cloud, and filter the ground point cloud according to the median of the vertical coordinate to divide it into several areas to be polished and the center coordinates of the areas to be polished.

[0067] This application embodiment designs a floor grinding device, including a 3D laser scanner, an inertial measurement unit, several lidar sensors, a grinding disc, and several differential drive wheels. The 3D laser scanner is used to scan the ground point cloud coordinates and generate point cloud data of the area to be ground. The inertial measurement unit can determine the orientation and motion information of the floor grinding device and determine the current pose information of the device; the lidar sensors are used to collect the pose information of the floor grinding device relative to obstacles in the ground area in real time, assisting the floor grinding device in accurately avoiding obstacles during movement.

[0068] To better demonstrate the specific structure of the floor grinding device Figure 2 A structural diagram of the floor grinding device is provided. In the diagram, 1 is the 3D laser scanner, 2 is the pan-tilt head of the 3D laser scanner, 3 is the inertial measurement unit, 4 is the drive motor of the grinding disc, 5 is the suspended up-and-down control system, 6 is the main structural frame, 7 is the lidar, 8 is the caster wheels, 9 is the grinding disc and its protective sleeve, 10 is the handrail, 11 is the control and analysis system, and 12 is the differential drive wheel. The floor grinding device has a total of four lidars, distributed at the front, back, left, and right sides of the device.

[0069] When the floor grinding device is stationary, a 3D laser scanner is used to scan the ground. The scanned data is then transmitted to a control and analysis system to obtain point cloud data of the ground area. In this embodiment, the forward direction of the floor grinding device is used as the longitudinal axis (y-axis) of the 3D laser scanner, the right side of the forward direction is the transverse axis (x-axis), and the vertically upward direction is the vertical axis (z-axis). Figure 3 A side view and coordinate system of a 3D laser scanner are provided. The center point of the laser emitter of the 3D laser scanner is used as the origin to show the y-axis and z-axis directions. The height h is the height of the set origin from the bottom surface of the floor grinding device.

[0070] Because the set h is located below the origin of the 3D laser scanner, data with z > -(h - 100mm) in the ground area point cloud data are first deleted, retaining only point cloud data above 100mm from the ground level, resulting in the first point cloud data. This first point cloud data is then preprocessed to improve data quality and reduce data volume. Specifically, bilateral filtering is used to denoise the first point cloud data, keeping it smooth and removing noise and outliers to improve data quality. Next, the denoised first point cloud data is downsampled using a uniform sampling method, maintaining an average point spacing of 10mm to filter the data, thereby reducing the amount of point cloud data that needs to be stored and improving the speed of subsequent point cloud processing.

[0071] After completing the bilateral filtering and downsampling processing of the first point cloud data, feature recognition is performed on the optimized first point cloud data, and the area to be polished is determined by identifying the locally excessively high areas.

[0072] Specifically, for each point in the optimized first point cloud data, the normal vector of the corresponding local plane is determined. The dot product of the normal vectors of any two local planes is calculated, and identifiers 'a' and 'b' are created for the point clouds contained in the two local planes respectively. If the dot product of the normal vectors of any two local planes is 0, it indicates that the two local planes are perpendicular. After traversing all the local planes, the dot product result is judged. If identifier 'a' > identifier 'b', the point cloud corresponding to identifier 'b' is deleted; if identifier 'a' < identifier 'b', the point cloud corresponding to identifier 'a' is deleted. The last remaining local planes are used as the ground point cloud, which can identify areas that are excessively high.

[0073] The above process determines whether two local planes belong to the ground or upright objects such as walls and columns by judging whether they are on the same plane. This helps to find the ground point cloud, providing data support for the subsequent division of the area to be polished. The dot product of the normal vectors is used to determine whether the relationship between the two local planes is perpendicular, coplanar, or slightly inclined, and thus determine whether the point cloud of the fitted small plane belongs to the ground or the wall or column.

[0074] The local plane is a plane obtained by fitting a point cloud with other point clouds within a preset adjacent range.

[0075] Optionally, in this embodiment of the application, the adjacent range is set to a range of 100mm × 100mm centered on the location of a point cloud.

[0076] Then, the median ordinate of the ground point cloud is calculated using its ordinate, and a horizontal plane with z = median ordinate is constructed. The ordinates of the points in the ground point cloud are then evaluated using this horizontal plane. If a point cloud's ordinate is greater than its median ordinate, that point cloud is retained. If there are n consecutive point clouds with ordinates greater than their median ordinates within a preset size range, the corresponding horizontal coordinate range and center coordinates are determined based on these n point clouds. A corresponding point cloud index is then constructed for this horizontal coordinate range and center coordinates to facilitate subsequent path planning for the floor grinding device to move towards the area to be ground.

[0077] By retaining the range of horizontal coordinates where the ordinate of the point cloud is greater than the median ordinate, and the corresponding center coordinates, the corresponding areas to be polished with local protrusions are obtained. Using the center coordinates, the subsequent floor polishing device can automatically set its path based on the point cloud index, and the device will move towards each area to be polished and perform polishing. The range of horizontal coordinates is composed of the range of horizontal and vertical coordinates of the areas to be polished, that is, determined by the maximum / minimum horizontal coordinates and maximum / minimum vertical coordinates of the n point clouds whose ordinates are greater than the median ordinate.

[0078] For example, assuming the median value of the ordinate is mean, then constructing a horizontal plane with z = mean, parallel to the y-axis and x-axis, yields a horizontal coordinate range {x...}. min ,,x max} and {y min ,,y max The center coordinates corresponding to the horizontal coordinate range are [x mean y mean ].

[0079] Step S2: Use lidar to measure distance within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area.

[0080] In this embodiment of the application, four lidars located in front of, behind, to the left and right of the floor grinding device are used to measure in real time the minimum lateral distance between the front and rear sides of the floor grinding device and obstacles in the ground area, as well as the directional angle between the floor grinding device and obstacles in the ground area.

[0081] Specifically, Figure 4 A lidar ranging diagram is provided, illustrating how the lidar measures the pose information between the floor grinding device and the wall. As shown, the lidar units are positioned at the front, rear, left, and right ends of the floor grinding device. The lidar unit closest to the wall collects the minimum lateral distances d1 and d2 between the floor grinding device and the wall, and simultaneously collects the azimuth angle θ between the moving direction of the floor grinding device and the wall. Based on the collected minimum lateral distances and azimuth angles, the next movement direction and speed of the floor grinding device are controlled to achieve precise obstacle avoidance and prevent the device from colliding with the wall during its movement.

[0082] Step S3: Based on the current pose information of the floor grinding device in the ground point cloud, control the movement direction of the floor grinding device through the center coordinates until the floor grinding device reaches the area to be ground and starts grinding.

[0083] Before path planning, the timestamps and coordinate systems of the inertial measurement unit (IMU), 3D laser scanner, and LiDAR are synchronized. Because the IMU, 3D laser scanner, and LiDAR use different coordinate systems during data acquisition, the acquired data needs to be unified in terms of coordinate system and timestamps to ensure real-time updates of the floor grinding device's pose information, providing data support for precise pose control.

[0084] In this embodiment, the coordinate systems of the 3D laser scanner and the inertial measurement unit (IMU) are provided. For the 3D laser scanner's coordinate system, the forward direction of the floor grinding device is taken as the y-axis, the right side of the forward direction is the x-axis, and the vertically upward direction is the z-axis. For the IMU's coordinate system, the forward direction of the floor grinding device is the x-axis, the right side of the forward direction is the y-axis, the vertically upward direction relative to the ground is the z-axis, and the center of the IMU is the origin on the left. Because the positions of the origins are different, the coordinate systems of the 3D laser scanner and the IMU are also different.

[0085] Because the inertial measurement unit (IMU) and the 3D laser scanner have different measurement frequencies and spatial coordinate systems, spatiotemporal calibration is required so that the IMU and the 3D laser scanner can describe the same object at the same time and place.

[0086] Specifically, hardware synchronization is achieved by providing UTC timestamps to the 3D laser scanner through the connection of the inertial measurement unit (IMU) and the NMEA protocol. The NMEA protocol, short for National Marine Electronics Association, is the unified standard protocol for current GPS navigation devices. After timestamp synchronization, the pose calibration of the IMU and the 3D laser scanner in space needs to be completed. For the same object, its coordinates differ in different coordinate systems; therefore, it is necessary to describe the transformation process of the object in different coordinate systems and determine the external relationships between the coordinate systems, i.e., to complete the calculation of external parameters.

[0087] For the spatial calibration between the coordinates of the inertial measurement unit (IMU) and the 3D laser scanner, this embodiment employs a hand-eye calibration method. The hand (represented here as the IMU) and the eye (represented here as the 3D laser scanner) are fixed on the same carrier. When the carrier moves, the pose changes of the "hand" and "eye" satisfy certain constraints. The coordinate transformation relationship between the "hand" and "eye" can be solved using statically indeterminate equations. According to the hand-eye calibration principle, although the coordinates of the same object differ in different coordinate systems, its position in world coordinates is fixed. Therefore, solving the extrinsic parameter equation between the two coordinate systems in the form AX = XB completes the coordinate system unification.

[0088] The same process can be used to unify the coordinate system of the lidar, thereby achieving spatiotemporal synchronization of the inertial measurement unit, the 3D laser scanner, and the lidar.

[0089] in, Figure 5 A top view of the coordinate system of the 3D laser scanner and the inertial measurement unit is provided. Figure (a) is a top view of the coordinate system of the 3D laser scanner, and Figure (b) is a top view of the coordinate system of the inertial measurement unit.

[0090] After completing spatiotemporal synchronization, the position of the inertial measurement unit is projected onto the coordinate system of the 3D laser scanner. Based on the gyroscope and accelerometer built into the inertial measurement unit, the pose of the floor grinding device in the point cloud can be determined, and the current pose information of the floor grinding device in the ground point cloud can be obtained.

[0091] Another reason for unifying the coordinate system of the inertial measurement unit and the 3D laser scanner is that sometimes the position information determined by a single instrument is not accurate enough. In order to ensure that the position and orientation of the instrument can be updated in real time, it is necessary to combine the point cloud data collected by the inertial measurement unit and the 3D laser scanner to determine the position.

[0092] For areas with localized protrusions to be polished, the angle between the floor polishing device and the area to be polished, as well as the direction of movement, are obtained based on the center coordinates of the area and the current pose information. Then, based on the angle, the floor polishing device is controlled to move towards the center coordinates, and the direction of movement is continuously adjusted to avoid obstacles, until the angle is zero. This indicates that the floor polishing device has reached its center position and polishing can begin.

[0093] Specifically, Figure 6 A diagram showing the movement of a floor grinding device is provided. In the diagram, the angle between the floor grinding device and the area to be ground is θ, and the center coordinates of the area to be ground are [x...]. mean y mean Based on the included angle and direction of movement, steering is achieved through differential drive wheels, and the value of the included angle is continuously fed back during forward movement until the included angle reaches zero. The total moving length of the floor grinding device is the distance from the center coordinate to the center point of the 3D laser scanner.

[0094] When the floor grinding device moves to the central coordinate, the grinding disc is driven to begin grinding, and the differential drive wheel moves back and forth and left and right to complete the grinding. During the grinding process, the moving distance of the floor grinding device should be at least equal to the following formula to achieve full coverage of the area to be ground:

[0095] max{max{x mean -x min x max -x mean}, max{y mean -y min y max -y mean}};

[0096] After completing one grinding cycle, the floor grinding device is driven to the next grinding area by using the stored horizontal coordinate range and center coordinates of the area to be ground.

[0097] After all the areas to be polished have been polished, the ground area is scanned again using a 3D laser scanner to obtain the maximum vertical coordinates of the scanned point cloud. Then, the maximum vertical coordinates of the scanned point cloud are compared with the median vertical coordinates of the point cloud data. When the maximum vertical coordinates are equal to the median vertical coordinates, it indicates that there are no locally protruding areas, and therefore the polishing is complete.

[0098] Step S4: When the floor grinding device moves, the moving speed and moving angle of the floor grinding device are controlled according to the position information to avoid the obstacle.

[0099] During the movement of the floor grinding device, the angular velocity and linear velocity of the device need to be deduced based on the aforementioned posture information and the planar dimensions of the device. The specific formulas are as follows:

[0100] v1 = ω1·r;

[0101] v2 = ω2·r;

[0102]

[0103] v = ω·R + ω·(Bb) / 2;

[0104]

[0105] In the formula, v1 and v2 are the linear velocities of the left and right wheels of the floor grinding device in the forward direction, respectively; ω1 and ω2 are the angular velocities of the left and right wheels of the floor grinding device in the forward direction, respectively; r is the radius of the differential drive wheel; ω is the angular velocity of the floor grinding device; B is the width of the floor grinding device; b is the minimum measurement distance between the side of the floor grinding device and the obstacle; R is the radius from the right wheel in the forward direction to the turning center radius; v is the linear velocity of the floor grinding device; θ is the direction angle between the forward direction of the floor grinding device and the obstacle; d1 and d2 are the minimum lateral distances between the front and rear sides of the floor grinding device and the obstacle in the ground area, respectively; and L is the length of the floor grinding device.

[0106] The specific calculation of the above formula can be found here. Figure 4 .

[0107] Based on the angular and linear velocities of the floor grinding device and the directional angle between the floor grinding device and obstacles in the ground area, the angular and linear velocities of the differential drive wheel are adjusted to change the direction of movement of the floor grinding device, so that the floor grinding device can accurately reach the area to be ground while avoiding obstacles.

[0108] Implementing the embodiments of this application has the following beneficial effects:

[0109] This application embodiment uses a 3D laser rangefinder to collect point cloud data of the area to be polished, then filters and extracts the ground point cloud to reduce storage requirements and speed up subsequent point cloud processing. Based on the vertical coordinates, areas with excessively high elevations are accurately identified and polished, enabling precise identification of the polishing area in complex terrain. The corresponding center coordinates are also located to facilitate path planning for the movement of the floor polishing device towards the polishing area. Furthermore, a lidar sensor is used to detect the distance between obstacles during the movement of the floor polishing device, ensuring perfect obstacle avoidance. Additionally, an inertial measurement unit accurately measures the current pose information of the floor polishing device to control its movement direction, achieving precise path planning during the polishing process.

[0110] Second Embodiment

[0111] Furthermore, in order to execute the posture control device for floor grinding corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects, Figure 7 A structural diagram of a posture control device for floor grinding is provided. For ease of explanation, only the parts relevant to this embodiment are shown. The posture control device for floor grinding provided in this application embodiment includes:

[0112] The grinding area division module 201 is used to extract ground point cloud data from the collected ground area point cloud data, and to filter the ground point cloud according to the median of the vertical coordinate to divide it into several grinding areas and the center coordinates of the grinding areas.

[0113] In this embodiment of the application, when the floor grinding device is in a stationary state, the ground area is scanned by a three-dimensional laser scanner to obtain point cloud data of the ground area;

[0114] Delete the point cloud data that is lower than the preset ground elevation to obtain the first point cloud data;

[0115] The first point cloud data is sequentially subjected to bilateral filtering and downsampling. Then, the dot product of the normal vectors of the local plane corresponding to each point cloud in the downsampled first point cloud data is calculated. The local plane is a plane obtained by fitting a point cloud with other point clouds within a preset adjacent range.

[0116] After obtaining the dot product results of the normal vectors of all the local planes, the ground point cloud is selected from the first point cloud data after downsampling by comparing the magnitude of the dot product results.

[0117] The median of the ordinate is calculated from the ordinate of the ground point cloud, and a horizontal plane equal to the median of the ordinate is constructed.

[0118] If there are three consecutive points in the ground point cloud with ordinates greater than the median ordinate within a preset size range, then the corresponding area to be polished is determined based on the three consecutive points with ordinates greater than the median ordinate, and the corresponding center coordinate is determined based on the horizontal coordinate range of the area to be polished; wherein the horizontal coordinate range is composed of the horizontal coordinate range and the vertical coordinate range of the area to be polished.

[0119] The pose information acquisition module 202 is used to measure distances using lidar within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area.

[0120] In this embodiment of the application, four lidars located in front of, behind, to the left and right of the floor grinding device are used to measure in real time the minimum lateral distance between the front and rear sides of the floor grinding device and obstacles in the ground area, as well as the directional angle between the floor grinding device and obstacles in the ground area.

[0121] For example, in this embodiment, lidar is located at the front, rear, left, and right ends of the floor grinding device, respectively. The lidar collects the pose information between the floor grinding device and the wall. The lidar on the side closest to the wall collects the minimum lateral distances d1 and d2 between the floor grinding device and the wall, and simultaneously collects the azimuth angle θ between the forward direction of the floor grinding device and the wall. Then, based on the collected minimum lateral distances and azimuth angles, the next movement direction and speed of the floor grinding device are controlled to achieve precise obstacle avoidance and prevent the device from colliding with the wall during its forward movement.

[0122] The area grinding module 203 is used to control the movement direction of the floor grinding device through the center coordinates based on the current pose information of the floor grinding device on the ground point cloud, until the floor grinding device reaches the area to be ground and then starts grinding.

[0123] In this embodiment of the application, after synchronizing the timestamps and coordinate systems of the inertial measurement unit, the 3D laser scanner, and the lidar, the position of the inertial measurement unit is projected onto the coordinate system of the 3D laser scanner to obtain the current pose information;

[0124] Based on the center coordinates and the current pose information, the angle between the floor grinding device and the area to be ground, as well as the direction of movement, are obtained.

[0125] Based on the included angle, the floor grinding device is controlled to move toward the center coordinate and the direction of movement is continuously adjusted until the included angle is zero.

[0126] When the included angle is zero, the grinding disc of the floor grinding device is driven by the differential drive wheel to start grinding.

[0127] The mobile obstacle avoidance module 204 is used to control the moving speed and moving angle of the floor grinding device to avoid the obstacle when the floor grinding device moves, based on the posture information.

[0128] During the movement of the floor grinding device, the angular velocity and linear velocity of the device need to be deduced based on the aforementioned posture information and the planar dimensions of the device. The specific formulas are as follows:

[0129] v1 = ω1·r;

[0130] v2 = ω2·r;

[0131]

[0132] v = ω·R + ω·(Bb) / 2;

[0133]

[0134] In the formula, v1 and v2 are the linear velocities of the left and right wheels of the floor grinding device in the forward direction, respectively; ω1 and ω2 are the angular velocities of the left and right wheels of the floor grinding device in the forward direction, respectively; r is the radius of the differential drive wheel; ω is the angular velocity of the floor grinding device; B is the width of the floor grinding device; b is the minimum measurement distance between the side of the floor grinding device and the obstacle; R is the radius from the right wheel in the forward direction to the turning center radius; v is the linear velocity of the floor grinding device; θ is the direction angle between the forward direction of the floor grinding device and the obstacle; d1 and d2 are the minimum lateral distances between the front and rear sides of the floor grinding device and the obstacle in the ground area, respectively; and L is the length of the floor grinding device.

[0135] Based on the angular and linear velocities of the floor grinding device and the directional angle between the floor grinding device and obstacles in the ground area, the angular and linear velocities of the differential drive wheel are adjusted to change the direction of movement of the floor grinding device, so that the floor grinding device can accurately reach the area to be ground while avoiding obstacles.

[0136] In some embodiments, the grinding area division module 201 specifically comprises:

[0137] This application embodiment designs a floor grinding device, including a 3D laser scanner, an inertial measurement unit, several lidar sensors, a grinding disc, and several differential drive wheels. The 3D laser scanner is used to scan the ground point cloud coordinates and generate point cloud data of the area to be ground. The inertial measurement unit can determine the orientation and motion information of the floor grinding device and determine the current pose information of the device; the lidar sensors are used to collect the pose information of the floor grinding device relative to obstacles in the ground area in real time, assisting the floor grinding device in accurately avoiding obstacles during movement. The floor grinding device has a total of four lidar sensors, distributed at the front, back, left, and right of the device.

[0138] When the floor grinding device is stationary, a 3D laser scanner is used to scan the ground, and then the scanned data is transmitted to the control and analysis system to obtain point cloud data of the ground area. In this embodiment, the forward direction of the floor grinding device is taken as the longitudinal axis (y-axis) of the 3D laser scanner, the right side of the forward direction is taken as the transverse axis (x-axis) of the 3D laser scanner, and the vertical upward direction is taken as the vertical axis (z-axis) of the 3D laser scanner. The center point of the laser emitter of the 3D laser scanner is set as the origin.

[0139] Because the set ground elevation is located below the origin of the 3D laser scanner, data with vertical axis coordinates greater than or equal to (h-100mm) in the point cloud data of the ground area are first deleted. This means only point cloud data above 100mm of the ground elevation are retained, resulting in the first point cloud data. Here, h is the height of the set origin from the bottom surface of the ground grinding device, and the set ground elevation is 100mm. The first point cloud data is then preprocessed to improve data quality and reduce data volume. Specifically, bilateral filtering is used to denoise the first point cloud data, keeping it smooth and removing noise and outliers to improve data quality. Then, the denoised first point cloud data is downsampled using a uniform sampling method, maintaining an average point cloud spacing of 10mm to filter the data, thereby reducing the amount of point cloud data that needs to be stored and improving the speed of subsequent point cloud processing.

[0140] After completing the bilateral filtering and downsampling processing of the first point cloud data, feature recognition is performed on the optimized first point cloud data, and the area to be polished is determined by identifying the locally excessively high areas.

[0141] Specifically, for each point in the optimized first point cloud data, the normal vector of the corresponding local plane is determined. The dot product of the normal vectors of any two local planes is calculated, and identifiers 'a' and 'b' are created for the point clouds contained in the two local planes respectively. If the dot product of the normal vectors of any two local planes is 0, it indicates that the two local planes are perpendicular. After traversing all the local planes, the result of the dot product of the normal vectors is judged. If identifier 'a' > identifier 'b', the point cloud corresponding to identifier 'b' is deleted; if identifier 'a' < identifier 'b', the point cloud corresponding to identifier 'a' is deleted. Finally, the region with excessively high elevation can be found based on the retained point cloud.

[0142] The local plane is a plane obtained by fitting a point cloud with other point clouds within a preset adjacent range.

[0143] Optionally, in this embodiment of the application, the adjacent range is set to a range of 100mm × 100mm centered on the location of a point cloud.

[0144] Then, the median ordinate of the ground point cloud is calculated using its ordinate, and a horizontal plane with z = median ordinate is constructed. The ordinates of the points in the ground point cloud are then evaluated using this horizontal plane. If a point cloud's ordinate is greater than its median ordinate, that point cloud is retained. If there are n consecutive point clouds with ordinates greater than their median ordinates within a preset size range, the corresponding horizontal coordinate range and center coordinates are determined based on these n point clouds. A corresponding point cloud index is then constructed for this horizontal coordinate range and center coordinates to facilitate subsequent path planning for the floor grinding device to move towards the area to be ground.

[0145] By retaining the range of horizontal coordinates where the ordinate of the point cloud is greater than the median ordinate, and the corresponding center coordinates, the corresponding areas to be polished with local protrusions are obtained. Using the center coordinates, the subsequent floor polishing device can automatically set its path based on the point cloud index, and the device will move towards each area to be polished and perform polishing. The range of horizontal coordinates is composed of the range of horizontal and vertical coordinates of the areas to be polished, that is, determined by the maximum / minimum horizontal coordinates and maximum / minimum vertical coordinates of the n point clouds whose ordinates are greater than the median ordinate.

[0146] For example, assuming the median value of the ordinate is mean, then constructing a horizontal plane with z = mean, parallel to the planes constructed along the y and x axes, yields a horizontal coordinate range {x...} min ,,x max} and {y min y max The center coordinates corresponding to the horizontal coordinate range are [x mean y mean ].

[0147] In some embodiments, the area polishing module 203 specifically comprises:

[0148] Before path planning, the timestamps and coordinate systems of the inertial measurement unit (IMU), 3D laser scanner, and LiDAR are synchronized. Because the IMU, 3D laser scanner, and LiDAR use different coordinate systems during data acquisition, the acquired data needs to be unified in terms of coordinate system and timestamps to ensure real-time updates of the floor grinding device's pose information, providing data support for precise pose control.

[0149] In this embodiment, the coordinate systems of the 3D laser scanner and the inertial measurement unit (IMU) are provided. For the 3D laser scanner's coordinate system, the forward direction of the floor grinding device is taken as the y-axis, the right side of the forward direction is the x-axis, and the vertically upward direction is the z-axis. For the IMU's coordinate system, the forward direction of the floor grinding device is the x-axis, the right side of the forward direction is the y-axis, the vertically upward direction relative to the ground is the z-axis, and the center of the IMU is the origin on the left. Because the positions of the origins are different, the coordinate systems of the 3D laser scanner and the IMU are also different.

[0150] Because the inertial measurement unit (IMU) and the 3D laser scanner have different measurement frequencies and spatial coordinate systems, spatiotemporal calibration is required so that the IMU and the 3D laser scanner can describe the same object at the same time and place.

[0151] Specifically, hardware synchronization is achieved by providing UTC timestamps to the 3D laser scanner through the connection of the inertial measurement unit (IMU) and the NMEA protocol. The NMEA protocol, short for National Marine Electronics Association, is the unified standard protocol for current GPS navigation devices. After timestamp synchronization, the pose calibration of the IMU and the 3D laser scanner in space needs to be completed. For the same object, its coordinates differ in different coordinate systems; therefore, it is necessary to describe the transformation process of the object in different coordinate systems and determine the external relationships between the coordinate systems, i.e., to complete the calculation of external parameters.

[0152] For the spatial calibration between the coordinates of the inertial measurement unit (IMU) and the 3D laser scanner, this embodiment employs a hand-eye calibration method. The hand (represented here as the IMU) and the eye (represented here as the 3D laser scanner) are fixed on the same carrier. When the carrier moves, the pose changes of the "hand" and "eye" satisfy certain constraints. The coordinate transformation relationship between the "hand" and "eye" can be solved using statically indeterminate equations. According to the hand-eye calibration principle, although the coordinates of the same object differ in different coordinate systems, its position in world coordinates is fixed. Therefore, solving the extrinsic parameter equation between the two coordinate systems in the form AX = XB completes the coordinate system unification.

[0153] The same process can be used to unify the coordinate system of the lidar, thereby achieving spatiotemporal synchronization of the inertial measurement unit, the 3D laser scanner, and the lidar.

[0154] After completing spatiotemporal synchronization, the position of the inertial measurement unit is projected onto the coordinate system of the 3D laser scanner. Based on the gyroscope and accelerometer built into the inertial measurement unit, the pose of the floor grinding device in the point cloud can be determined, and the current pose information of the floor grinding device in the ground point cloud can be obtained.

[0155] Another reason for unifying the coordinate system of the inertial measurement unit and the 3D laser scanner is that sometimes the position information determined by a single instrument is not accurate enough. In order to ensure that the position and orientation of the instrument can be updated in real time, it is necessary to combine the point cloud data collected by the inertial measurement unit and the 3D laser scanner to determine the position.

[0156] For areas with localized protrusions to be polished, the angle between the floor polishing device and the area to be polished, as well as the direction of movement, are obtained based on the center coordinates of the area and the current pose information. Then, based on the angle, the floor polishing device is controlled to move towards the center coordinates, and the direction of movement is continuously adjusted to avoid obstacles, until the angle is zero. This indicates that the floor polishing device has reached its center position and polishing can begin.

[0157] When the floor grinding device moves to the central coordinate, the grinding disc is driven to begin grinding, and the differential drive wheel moves back and forth and left and right to complete the grinding. During the grinding process, the moving distance of the floor grinding device should be at least equal to the following formula to achieve full coverage of the area to be ground:

[0158] max{max{x mean -x min x max -x mean}, max{y mean -y min y max -y mean}};

[0159] After completing one grinding cycle, the floor grinding device is driven to the next grinding area by using the stored horizontal coordinate range and center coordinates of the area to be ground.

[0160] After all the areas to be polished have been polished, the ground area is scanned again using a 3D laser scanner to obtain the maximum vertical coordinates of the scanned point cloud. Then, the maximum vertical coordinates of the scanned point cloud are compared with the median vertical coordinates of the point cloud data. When the maximum vertical coordinates are equal to the median vertical coordinates, it indicates that there are no locally protruding areas, and therefore the polishing is complete.

[0161] Implementing the embodiments of this application has the following beneficial effects:

[0162] This application embodiment uses a 3D laser rangefinder to collect point cloud data of the area to be polished, then filters and extracts the ground point cloud to reduce storage requirements and speed up subsequent point cloud processing. Based on the vertical coordinates, areas with excessively high elevations are accurately identified and polished, enabling precise identification of the polishing area in complex terrain. The corresponding center coordinates are also located to facilitate path planning for the movement of the floor polishing device towards the polishing area. Furthermore, a lidar sensor is used to detect the distance between obstacles during the movement of the floor polishing device, ensuring perfect obstacle avoidance. Additionally, an inertial measurement unit accurately measures the current pose information of the floor polishing device to control its movement direction, achieving precise path planning during the polishing process.

[0163] Third Embodiment

[0164] Furthermore, in order to execute the pose control method for floor grinding corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects, Figure 8 A structural diagram of a posture control system for floor grinding is provided. For ease of explanation, only the parts relevant to this embodiment are shown. The posture control system for floor grinding provided in this embodiment includes:

[0165] The floor grinding device M1 is used to grind up excessively high areas on the floor.

[0166] The posture control device M2 for floor grinding is used to implement the above method embodiment.

[0167] The floor grinding device M1 includes a three-dimensional laser scanner, an inertial measurement unit, several lidars (located at the front, rear, left, and right ends of the floor grinding device), a grinding disc, and several differential drive wheels.

[0168] The three-dimensional laser scanner is used to scan the ground area and generate point cloud data of the area to be polished when the floor polishing device is stationary.

[0169] The inertial measurement unit is used to determine the current pose information of the floor grinding device;

[0170] The lidar is used to collect real-time positional information of the floor grinding device relative to obstacles in the ground area;

[0171] The grinding disc is used to grind up localized areas that are too high within the area to be ground.

[0172] The differential drive wheel is used to drive the floor grinding device to move.

[0173] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A posture control method for floor grinding, characterized in that, include: Data extraction is performed on the collected point cloud data of the ground area to obtain the ground point cloud. The ground point cloud is then filtered according to the median of the vertical coordinate to divide it into several areas to be polished and the center coordinates of the areas to be polished. LiDAR is used to measure distance within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area; Based on the current pose information of the floor grinding device on the ground point cloud, the movement direction of the floor grinding device is controlled by the center coordinates until the floor grinding device reaches the area to be ground and then begins grinding. When the floor grinding device moves, the moving speed and moving angle of the floor grinding device are controlled according to the position information to avoid the obstacle.

2. The posture control method for floor grinding according to claim 1, characterized in that, The process of extracting ground point clouds from the collected point cloud data of the ground area specifically involves: When the floor grinding device is stationary, the ground area is scanned by a 3D laser scanner to obtain point cloud data of the ground area; Delete the point cloud data that is lower than the preset ground elevation to obtain the first point cloud data; The first point cloud data is sequentially subjected to bilateral filtering and downsampling. Then, the dot product of the normal vectors of the local plane corresponding to each point cloud in the downsampled first point cloud data is calculated. The local plane is a plane obtained by fitting a point cloud with other point clouds within a preset adjacent range. After obtaining the dot product results of the normal vectors of all the local planes, the ground point cloud is selected from the first point cloud data after downsampling by comparing the magnitude of the dot product results.

3. The posture control method for floor grinding according to claim 1, characterized in that, The ground point cloud is then filtered based on the median of the vertical coordinates to divide it into several areas to be polished and the center coordinates of these areas. Specifically: The median of the ordinate is calculated from the ordinate of the ground point cloud, and a horizontal plane equal to the median of the ordinate is constructed. If there are three consecutive points in the ground point cloud with ordinates greater than the median ordinate within a preset size range, then the corresponding area to be polished is determined based on the three consecutive points with ordinates greater than the median ordinate, and the corresponding center coordinate is determined based on the horizontal coordinate range of the area to be polished; wherein the horizontal coordinate range is composed of the horizontal coordinate range and the vertical coordinate range of the area to be polished.

4. The posture control method for floor grinding according to claim 1, characterized in that, The method of using lidar to measure distance within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area is as follows: Using lidar located at both ends of the floor grinding device, the minimum lateral distance between the front and rear sides of the floor grinding device and obstacles in the ground area, as well as the directional angle between the floor grinding device and obstacles in the ground area are measured respectively.

5. The posture control method for floor grinding according to claim 1, characterized in that, The process involves controlling the movement direction of the floor grinding device based on the current pose information of the collected floor grinding device on the ground point cloud, using the center coordinates, until the floor grinding device reaches the area to be ground and grinding begins. Specifically: After synchronizing the timestamps and coordinate systems of the inertial measurement unit, the 3D laser scanner, and the lidar, the position of the inertial measurement unit is projected onto the coordinate system of the 3D laser scanner to obtain the current pose information; Based on the center coordinates and the current pose information, the angle between the floor grinding device and the area to be ground, as well as the direction of movement, are obtained. Based on the included angle, the floor grinding device is controlled to move toward the center coordinate and the direction of movement is continuously adjusted until the included angle is zero. When the included angle is zero, the grinding disc of the floor grinding device is driven by the differential drive wheel to start grinding.

6. The posture control method for floor grinding according to claim 1, characterized in that, When the floor grinding device moves, its moving speed and angle are controlled according to the position information to avoid the obstacle, specifically: Based on the pose information and the planar dimensions of the floor grinding device, the angular velocity and linear velocity of the floor grinding device are obtained; Adjust the angular and linear velocities of the differential drive wheel based on the angular and linear velocities of the floor grinding device and the directional angle between the floor grinding device and obstacles in the ground area.

7. The posture control method for floor grinding according to any one of claims 1 to 6, characterized in that, Also includes: After all the areas to be polished have been polished, the ground area is scanned again using a 3D laser scanner to obtain the maximum vertical coordinates of the point cloud in this scan. The maximum vertical coordinate of the scanned point cloud is compared with the median vertical coordinate of the point cloud data. The polishing process ends when the maximum vertical coordinate is equal to the median vertical coordinate.

8. A posture control device for floor grinding, characterized in that, include: The module includes a grinding area division module, a pose information acquisition module, an area grinding module, and a movement obstacle avoidance module. The grinding area division module is used to extract ground point cloud data from the collected ground area, and to filter the ground point cloud according to the median of the vertical coordinate to divide it into several grinding areas and the center coordinates of the grinding areas. The pose information acquisition module is used to measure distances using lidar within the ground area to obtain the pose information of the floor grinding device relative to obstacles within the ground area; The area grinding module is used to control the movement direction of the floor grinding device through the center coordinates based on the current pose information of the floor grinding device on the ground point cloud, until the floor grinding device reaches the area to be ground and then starts grinding. The mobile obstacle avoidance module is used to control the moving speed and moving angle of the floor grinding device to avoid obstacles when the floor grinding device moves, based on the posture information.

9. A posture control system for floor grinding, characterized in that, include: A posture control device and a floor grinding device for floor grinding. The posture control device for floor grinding is used to perform the posture control method for floor grinding as described in any one of claims 1 to 7. The floor grinding device is used to grind up excessively high areas on the floor.

10. The posture control system for floor grinding according to claim 9, characterized in that, The floor grinding device is specifically: The floor grinding device includes a three-dimensional laser scanner, an inertial measurement unit, several lidar sensors, a grinding disc, and several differential drive wheels; The three-dimensional laser scanner is used to scan the ground area and generate point cloud data of the area to be polished when the floor polishing device is stationary. The inertial measurement unit is used to determine the current pose information of the floor grinding device; The lidar is used to collect real-time positional information of the floor grinding device relative to obstacles in the ground area; The grinding disc is used to grind up localized areas that are too high within the area to be ground. The differential drive wheel is used to drive the floor grinding device to move.

Images