A method, device, equipment and storage medium for operating upper and lower chassis of a robot

Abstract

The application discloses a robot upper and lower chassis operation method and device, equipment and storage medium, relates to the technical field of robot control, and comprises the following steps: obtaining target relative pose information between a robot and a target mobile chassis based on a preset pose sensing mode; the preset pose sensing mode is one or a combination of several of a depth point cloud sensing mode, a laser contour sensing mode and a visual mark sensing mode; determining whether the robot and the target mobile chassis currently satisfy a preset alignment condition according to the obtained target relative pose information; if the robot and the target mobile chassis do not satisfy the preset alignment condition, generating an alignment adjustment instruction based on the target relative pose information, and controlling the robot and / or the target mobile chassis to adjust the pose based on the alignment adjustment instruction until the robot and the target mobile chassis satisfy the preset alignment condition, and triggering the robot to perform an upper and lower chassis action. In this way, the success rate and safety of the robot upper and lower chassis action can be improved.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a method, apparatus, device, and storage medium for operating the upper and lower chassis of a robot. Background Technology

[0002] With the rapid development of robotics technology, collaborative operation scenarios involving humanoid robots, legged robots, and wheeled or tracked mobile chassis are becoming increasingly common. In practical tasks, robots often need to perform actions such as climbing onto or descending from a mobile chassis, which is a crucial preliminary step for performing complex tasks such as transportation, inspection, and relocation. Before such actions are executed, the relative pose between the robot and the target mobile chassis—including lateral deviation, longitudinal distance, relative angle, height difference, and the robot's own posture—directly determines the success rate and safety of subsequent actions. If the initial pose deviation is large, directly triggering the climbing or descending action can easily lead to foot landing point misalignment, failure to switch support phases, body tilting, or even overturning.

[0003] Currently, existing solutions primarily rely on manual handling or remote control to move the robot to a fixed starting position in front of the chassis before triggering a pre-set action script. While simple to implement, this method is highly dependent on operator experience and requires high initial docking accuracy, making it difficult to adapt to unstructured scenarios where the chassis's position, orientation, or height varies. Furthermore, some existing technologies focus on trajectory planning and control optimization for the chassis movements themselves, such as using pre-set action sequences, full-body control, or imitation learning strategies. However, they lack a unified mechanism for relative pose perception and collaborative alignment judgment before action triggering. When the relative state between the robot and the chassis does not meet the initial conditions of the action strategy, the system may still blindly trigger actions, reducing execution robustness and introducing safety hazards.

[0004] In summary, there is an urgent need to propose a collaborative alignment method for robots before they mount and dismount from the chassis, in order to effectively improve the success rate, safety, and environmental adaptability of the robot's chassis mounting and dismounting actions. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, device, and storage medium for operating a robot's upper and lower chassis, which can improve the success rate and safety of the robot's upper and lower chassis movements. The specific solution is as follows: In a first aspect, this application discloses a method for operating the upper and lower chassis of a robot, including: The robot acquires the target relative pose information between itself and the target mobile chassis based on a preset pose perception method; the preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual mark perception method. Based on the obtained target relative pose information, it is determined whether the robot and the target mobile chassis currently meet the preset alignment conditions; the preset alignment conditions include that when the robot performs the up and down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state fed back by the robot is a stable state that has completed standing, balancing or maintaining the initial posture; If the robot and the target mobile chassis do not meet the preset alignment conditions, an alignment adjustment command is generated based on the target relative pose information, and the robot is controlled based on the alignment adjustment command. Alternatively, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then the robot is triggered to execute the target mobile chassis's up and down chassis movement strategy.

[0006] Optionally, the step of obtaining the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method includes: The robot acquires three-dimensional point cloud data of the target mobile chassis using the depth point cloud perception method, and extracts the point cloud region corresponding to the target mobile chassis from the three-dimensional point cloud data to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region; the first relative pose information includes one or a combination of several of the following: height difference, longitudinal distance, lateral deviation, and relative angle. The laser scanning data or laser point cloud data of the target mobile chassis is acquired through the laser contour perception method, and the geometric contour features of the target mobile chassis are extracted from the laser scanning data or laser point cloud data to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features; the second relative pose information includes one or a combination of several of the following: relative distance, lateral deviation and relative angle. Images containing visual identifiers are acquired and the visual identifiers are identified using the visual identifier perception method. Based on the identification results of the visual identifiers, a third relative pose information between the robot and the target mobile chassis is determined. The third relative pose information includes the position of the robot relative to the target mobile chassis and / or its orientation. The target pose information is determined from the first relative pose information, the second relative pose information, and the third relative pose information; If the number of pose information in the target pose information is greater than 1, the target pose information is transformed to a unified coordinate system, and the consistency judgment and confidence evaluation of each relative pose information after transformation are performed in the unified coordinate system to obtain the consistency judgment result and confidence evaluation result. Based on the consistency judgment result and the confidence evaluation result, the relative pose information of each is weighted and fused to obtain the target relative pose information between the robot and the target mobile chassis.

[0007] Optionally, the step of acquiring three-dimensional point cloud data of the target mobile chassis through the depth point cloud perception method, and extracting the point cloud region corresponding to the target mobile chassis from the three-dimensional point cloud data, so as to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region, includes: Depth images of the target mobile chassis are acquired using a depth camera or RGB-D sensor, and the depth images are converted into three-dimensional point cloud data based on the camera's intrinsic parameters. The three-dimensional point cloud data is subjected to invalid depth point removal, outlier point removal, region smoothing or edge enhancement processing, and the point cloud region corresponding to the target mobile chassis is extracted from the processed three-dimensional point cloud data. Plane fitting, edge extraction, or contour fitting are performed on the point cloud region to determine the upper surface height, front edge position, left and right boundary positions, and chassis orientation of the target mobile chassis; Based on the height of the upper surface, the position of the front edge, the positions of the left and right boundaries, and the orientation of the chassis, the height difference, longitudinal distance, lateral deviation, and relative angle between the robot and the target mobile chassis are determined sequentially to obtain the first relative pose information.

[0008] Optionally, the step of acquiring laser scanning data or laser point cloud data of the target mobile chassis through the laser contour perception method, and extracting geometric contour features of the target mobile chassis from the laser scanning data or laser point cloud data, so as to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features, includes: The robot acquires laser scanning data or laser point cloud data of the target mobile chassis using a two-dimensional lidar, and determines the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis based on the distance abrupt change, line segment fitting, contour clustering, or known geometric dimensions of the target mobile chassis in the laser scanning data or laser point cloud data, so as to obtain the second relative pose information. Alternatively, the laser scanning data or laser point cloud data obtained by the three-dimensional lidar can be processed by removing ground points or filtering by height range to obtain the remaining point cloud. The remaining point cloud can then be clustered or its contour extracted to determine the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis, thereby obtaining the second relative pose information.

[0009] Optionally, the step of acquiring an image containing visual identifiers through the visual identifier perception method and identifying the visual identifiers to determine the third relative pose information between the robot and the target mobile chassis based on the identification result of the visual identifiers includes: Images containing the visual identifiers are acquired by a preset camera; the preset camera is mounted on the target mobile chassis, the robot's body, or an external sensing component. If the visual identifier is AprilTag, then the identifier boundary and corner points of the visual identifier in the image are detected, and the spatial pose of the visual identifier relative to the camera is determined based on the camera intrinsic parameters, the actual size of the visual identifier, and the position of the corner points in the image, so as to obtain the third relative pose information. If the visual identifier is a QR code or a pattern identifier, the visual identifier is determined based on the recognition result of the image, and the spatial pose of the visual identifier relative to the camera is determined based on the camera intrinsic parameters and the actual size or depth information of the visual identifier, so as to obtain the third relative pose information.

[0010] Optionally, determining whether the robot and the target mobile chassis currently meet preset alignment conditions based on the acquired target relative pose information includes: Determine whether the lateral deviation between the robot and the target mobile chassis in the target relative pose information is less than a first threshold; And / or, determine whether the longitudinal distance between the robot and the target mobile chassis in the target relative pose information is within a preset distance range; And / or, determine whether the relative angle between the robot and the target mobile chassis in the target relative pose information is less than a second threshold; And / or, determine whether the height difference between the robot and the target mobile chassis in the target relative pose information is within the height range where the chassis can be moved up and down; And / or, determine whether the robot's current posture meets the requirements for the starting posture of the upper and lower chassis movements; And / or, determine whether the current state fed back by the robot is a stable state; the stable state indicates that the robot has completed standing, balancing or maintaining the initial posture.

[0011] Optionally, the step of controlling the robot based on the alignment adjustment command, and / or adjusting the pose of the target mobile chassis until the robot and the target mobile chassis meet the preset alignment conditions, and then triggering the robot to execute the up-and-down chassis movement strategy of the target mobile chassis, includes: In the first stage, if the relative angle is not less than the second threshold, then based on the robot turning command in the alignment adjustment command and / or the target mobile chassis orientation adjustment command, control the robot and / or the target mobile chassis to perform orientation adjustment until the relative angle is less than the second threshold. In the second stage, if the lateral deviation is not less than the first threshold or the longitudinal distance is not within the preset distance range, then based on the robot lateral adjustment command, robot forward command, robot backward command in the alignment adjustment command, and / or the target mobile chassis position adjustment command, control the robot and / or the target mobile chassis to perform position adjustment until the lateral deviation is less than the first threshold and the longitudinal distance is within the preset distance range; In the third stage, it is determined whether the height difference is within the height range and whether the robot's current posture meets the starting posture requirements of the upper and lower chassis movements; If the robot's current posture does not meet the requirements of the starting posture of the upper and lower chassis movements, the robot's starting posture is adjusted based on the robot gait parameter adjustment instruction or the robot movement starting posture adjustment instruction in the alignment adjustment instruction. In the fourth stage, the current state fed back by the robot is obtained to confirm whether the robot is in the stable state. In the fifth stage, when the conditions corresponding to the first to the fourth stages are all met, the robot is triggered to execute the up and down chassis movement strategy relative to the target moving chassis. If the preset alignment conditions are not met within the preset time, preset number of adjustments, or preset adjustment distance, the upper and lower chassis action strategy will be stopped, and a re-sensing command, a repositioning command, a manual assistance prompt, a safety stop command, an exit from the upper and lower chassis mode command, or an alarm prompt command will be output.

[0012] Secondly, this application discloses a robot's upper and lower chassis operating device, comprising: The pose information acquisition module is used to acquire the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method; the preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual mark perception method. The condition judgment module is used to determine whether the robot and the target mobile chassis currently meet the preset alignment conditions based on the acquired target relative pose information. The preset alignment conditions include that when the robot performs the up and down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state reported by the robot is a stable state that has completed standing, balancing or maintaining the initial posture. The pose adjustment module is used to generate an alignment adjustment command based on the target relative pose information if the robot and the target mobile chassis do not meet the preset alignment conditions, and control the robot based on the alignment adjustment command, and / or, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then triggers the robot to execute the up and down chassis movement strategy of the target mobile chassis.

[0013] Thirdly, this application discloses an electronic device, including: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned method for operating the robot's upper and lower chassis.

[0014] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the aforementioned method for operating the robot's upper and lower chassis.

[0015] As can be seen, in this invention, target relative pose information between the robot and the target mobile chassis is obtained based on a preset pose perception method; the preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual identification perception method; based on the obtained target relative pose information, it is determined whether the robot and the target mobile chassis currently meet preset alignment conditions; the preset alignment conditions include that when the robot performs the up-and-down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state fed back by the robot is a stable state that has completed standing, balancing, or maintaining the initial posture; if the robot and the target mobile chassis do not meet the preset alignment conditions, an alignment adjustment command is generated based on the target relative pose information, and the robot is controlled based on the alignment adjustment command, and / or the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then the robot is triggered to execute the up-and-down chassis movement strategy of the target mobile chassis. In other words, the triggering conditions for the lifting and lowering of the chassis are changed from manual judgment based on experience to automatic judgment based on the target's relative pose information. Before the action is triggered, the robot and the target moving chassis are aligned in advance, avoiding the robot from directly executing the lifting and lowering of the chassis when the initial pose deviation is large, thus improving the success rate of the lifting and lowering of the chassis. Because corresponding preset alignment conditions are set, the risk of the robot's foot landing point deviation, support phase switching failure, posture instability, or even falling is reduced, thus improving safety. At the same time, the alignment adjustment command is automatically generated by the system based on the target's relative pose information, reducing the need for manual remote control and manual placement of the robot, and reducing the reliance on the operator's experience. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This application discloses a flowchart of a method for operating the upper and lower chassis of a robot. Figure 2 This application discloses a flowchart of a multi-sensor fusion process for acquiring relative state information. Figure 3 This application discloses a flowchart of a specific method for operating the upper and lower chassis of a robot. Figure 4This is a schematic diagram of the upper and lower chassis operating device of a robot disclosed in this application; Figure 5 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0018] 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 skilled in the art without creative effort are within the scope of protection of this application.

[0019] In scenarios where robots and mobile chassis work collaboratively, the stable execution of the robot's mounting and dismounting actions largely depends on the relative pose between the robot and the target mobile chassis at the moment the action is triggered. Therefore, this application will specifically introduce a method for a robot to mount and dismount from its chassis, which can improve the accuracy of robot movements by adjusting the relative pose.

[0020] See Figure 1 As shown in the figure, this application discloses a method for operating the upper and lower chassis of a robot, including: Step S11: Obtain the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method; the preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual mark perception method.

[0021] In this embodiment, the target relative pose information is used to characterize the spatial relationship between the robot and the target mobile chassis. Specifically, it may include one or a combination of several of the following: relative distance between the robot and the target mobile chassis, lateral deviation, longitudinal distance, relative angle, height difference, the robot's current posture, the robot's current feedback state, the position of the target mobile chassis, the orientation of the target mobile chassis, and the height of the target mobile chassis. Specifically, lateral deviation represents the robot's offset relative to the chassis centerline in the direction perpendicular to the upper and lower chassis; longitudinal distance represents the distance the robot travels along the upper and lower chassis direction to the front edge of the chassis; relative angle represents the angle between the robot's orientation and the target mobile chassis's orientation; and height difference represents the height difference between the target mobile chassis's bearing surface and the robot's current standing plane. These quantities correspond to the constraints on the upper and lower chassis's movements in the lateral, longitudinal, orientation, and vertical directions, and therefore can serve as the basis for subsequent judgments regarding whether preset alignment conditions are met.

[0022] In this embodiment, the step of obtaining the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method includes: obtaining three-dimensional point cloud data of the target mobile chassis through the depth point cloud perception method, and extracting the point cloud region corresponding to the target mobile chassis from the three-dimensional point cloud data, so as to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region; the first relative pose information includes one or a combination of height difference, longitudinal distance, lateral deviation, and relative angle; obtaining laser scanning data or laser point cloud data of the target mobile chassis through the laser contour perception method, and extracting the geometric contour features of the target mobile chassis from the laser scanning data or laser point cloud data, so as to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features; the second relative pose information includes one or a combination of relative distance, lateral deviation, and relative angle. The robot is combined with the target mobile chassis. Images containing visual identifiers are acquired and identified using the visual identifier perception method. Based on the identification results of the visual identifiers, a third relative pose information between the robot and the target mobile chassis is determined. The third relative pose information includes the position and / or orientation of the robot relative to the target mobile chassis. Target pose information is determined from the first, second, and third relative pose information. If the number of pose information items in the target pose information is greater than 1, the target pose information is transformed to a unified coordinate system. Consistency judgment and confidence assessment are performed on each transformed relative pose information item in the unified coordinate system to obtain consistency judgment results and confidence assessment results. Based on the consistency judgment results and the confidence assessment results, each relative pose information item is weighted and fused to obtain the target relative pose information between the robot and the target mobile chassis.

[0023] It should be noted that the three sensing methods mentioned above each have their own strengths in measurement characteristics, making them suitable for combined use. Depth point cloud sensing can directly acquire the three-dimensional spatial information of the target's moving chassis, offering high accuracy in estimating height differences and the position of the chassis's upper surface. However, it is prone to depth loss when ambient light is too strong, the chassis surface is reflective, or there are obstructions. Laser contour sensing is insensitive to changes in illumination and has stable ranging capabilities, making it suitable for determining relative distances and lateral deviations at greater distances. However, it primarily acquires geometric contours, which can lead to mis-extraction when the chassis contour is close to surrounding objects. Visual marker sensing relies on the prior knowledge of the marker's size and pattern, offering high resolution in calculating relative angles, but requires the marker to be within the camera's field of view and unobstructed. Therefore, unifying the measurement results of the three methods into the same coordinate system and performing weighted fusion allows for supplementation by other sensing sources when a single sensing source is limited, resulting in more stable target relative pose information than any single sensing source. It should be noted that in actual operation, one or more sensing methods are actually used in the preset pose perception method to obtain the corresponding relative pose information respectively; when there are at least two sensing sources, the relative pose information is converted to a unified coordinate system and consistency judgment, confidence assessment and fusion are performed.

[0024] Furthermore, such as Figure 2 As shown, before fusing the first, second, and third relative pose information, it is necessary to transform each sensing result to a unified coordinate system. This unified coordinate system can be the robot's body coordinate system, the target mobile chassis coordinate system, the world coordinate system, or the starting coordinate system for the motion required to perform the chassis mounting and dismounting actions. Since the installation positions and orientations of each sensing component are different, their original measurement results are in their respective sensor coordinate systems. Only after coordinate unification can the relative distances, lateral deviations, longitudinal distances, relative angles, and height differences obtained from different sensing sources be comparable, and subsequent consistency judgments and weighted fusion become meaningful. After coordinate unification, a consistency judgment is performed on the relative distances, lateral deviations, longitudinal distances, relative angles, and height differences obtained from different sensing sources. When the differences between multiple sensing results are less than a preset threshold, the multiple sensing results are determined to be consistent, and the final target relative pose information can be obtained based on these multiple sensing results.

[0025] The step of acquiring 3D point cloud data of the target mobile chassis through the depth point cloud perception method and extracting the point cloud region corresponding to the target mobile chassis from the 3D point cloud data to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region includes: acquiring depth images of the target mobile chassis through a depth camera or RGB-D sensor, and converting the depth images into 3D point cloud data according to camera intrinsic parameters; performing invalid depth point removal, outlier removal, region smoothing or edge enhancement processing on the 3D point cloud data, and extracting the point cloud region corresponding to the target mobile chassis from the processed 3D point cloud data; performing plane fitting, edge extraction or contour fitting on the point cloud region to determine the upper surface height, front edge position, left and right boundary positions and chassis orientation of the target mobile chassis; and determining the height difference, longitudinal distance, lateral deviation and relative angle between the robot and the target mobile chassis sequentially according to the upper surface height, front edge position, left and right boundary positions and chassis orientation to obtain the first relative pose information.

[0026] Specifically, when extracting the point cloud region corresponding to the target mobile chassis from 3D point cloud data, the search range can be narrowed down based on a preset region of interest (ROI), and then the point cloud region corresponding to the target mobile chassis can be determined within that ROI. The point cloud region corresponding to the target mobile chassis can be determined by depth continuity, edge contour, planar height, color features, or target detection results. For example, the chassis bearing surface appears as a continuous region with a gentle depth change in the depth direction, while its height exhibits a significant abrupt change relative to the ground. Based on this, the chassis point cloud can be distinguished from the ground point cloud and background point cloud.

[0027] After obtaining the point cloud region corresponding to the target mobile chassis, plane fitting, edge extraction, or contour fitting are performed on this point cloud region to determine the height of the target mobile chassis's upper surface, the position of its front edge, the positions of its left and right boundaries, and its orientation. Correspondingly, the relative poses between the robot and the target mobile chassis can be derived as follows: estimate the chassis height based on the point cloud height of the target mobile chassis's upper surface, and subtract this height from the height of the robot's current standing plane to obtain the height difference; estimate the longitudinal distance between the robot and the target mobile chassis based on the position of the point cloud at the front edge of the chassis; estimate the lateral deviation of the robot relative to the chassis's centerline based on the chassis's left and right boundaries or the centerline determined by the left and right boundaries; estimate the orientation of the target mobile chassis relative to the robot based on the chassis edge direction or the main direction of the point cloud contour, thus obtaining the relative angle.

[0028] In addition, invalid depth point removal, outlier removal, region smoothing, or edge enhancement are performed on the 3D point cloud data to suppress the impact of environmental noise, surface reflection, local occlusion, or missing depth on relative pose estimation. Since plane fitting and edge extraction are sensitive to outliers, even a small number of erroneous depth points falling outside the chassis can cause the fitted plane to deviate or the extracted edge positions to shift. Therefore, performing the above preprocessing before extracting the point cloud region helps improve the stability of the first relative pose information.

[0029] The step of acquiring laser scanning data or laser point cloud data of the target mobile chassis through the laser contour perception method, and extracting the geometric contour features of the target mobile chassis from the laser scanning data or laser point cloud data to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features, includes: acquiring laser scanning data or laser point cloud data of the target mobile chassis through a two-dimensional lidar, and determining the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis based on the distance abrupt changes, line segment fitting, contour clustering, or known geometric dimensions of the target mobile chassis in the laser scanning data or laser point cloud data to obtain the second relative pose information; or, performing ground point removal or height range filtering on the laser scanning data or laser point cloud data acquired by a three-dimensional lidar to obtain the remaining point cloud, and performing clustering or contour extraction on the remaining point cloud to determine the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis to obtain the second relative pose information.

[0030] Specifically, when using a two-dimensional LiDAR, the front edge, side edge, or center position of the target mobile chassis can be extracted based on abrupt changes in distance at laser scanning points, line segment fitting, contour clustering, or the known geometric dimensions of the target mobile chassis. The relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis can then be determined based on these positions. Specifically, abrupt changes in distance correspond to the step change in the ranging value when the scanning beam switches from the background to the chassis surface, and can be used to determine the chassis boundary. Line segment fitting involves fitting scanning points falling on the same side of the chassis into a straight line, the position and direction of which correspond to the position and orientation of the chassis edge, respectively.

[0031] When using 3D LiDAR, ground points or height ranges can be filtered from the point cloud to remove ground objects and irrelevant objects significantly higher than the chassis. The remaining point cloud is then clustered or its contours extracted to obtain the spatial location, height, and orientation of the target mobile chassis. Furthermore, based on the known shape, edge structure, or size range of the target mobile chassis, geometric matching can be performed on the extracted point cloud contours. Objects whose extraction results do not match the known geometric features of the chassis are discarded, thus reducing the probability of misidentifying surrounding objects as the target mobile chassis and improving the stability of target mobile chassis identification.

[0032] In one optional implementation, the position and orientation of the target moving chassis can also be tracked based on multiple consecutive frames of LiDAR data. Since the edge positions extracted from a single frame point cloud fluctuate between frames due to the influence of scanning resolution and random noise, tracking the results of multiple frames and taking their changing trends can reduce the impact of single-frame point cloud noise on alignment judgment and avoid erroneous triggering or termination of alignment adjustment due to measurement jumps in individual frames.

[0033] The step of acquiring an image containing a visual identifier and identifying the visual identifier using the visual identifier perception method, and determining the third relative pose information between the robot and the target mobile chassis based on the identification result of the visual identifier, includes: acquiring an image containing the visual identifier using a preset camera; the preset camera is set on the target mobile chassis, the robot's body, or an external sensing component; if the visual identifier is an AprilTag, then the identifier boundary and corner points of the visual identifier in the image are detected, and the spatial pose of the visual identifier relative to the camera is determined based on the camera's intrinsic parameters, the actual size of the visual identifier, and the position of the corner points in the image, to obtain the third relative pose information; if the visual identifier is a QR code or a pattern identifier, then the visual identifier is determined based on the identification result of the image, and the spatial pose of the visual identifier relative to the camera is determined based on the camera's intrinsic parameters, the actual size or depth information of the visual identifier, to obtain the third relative pose information.

[0034] Specifically, the visual identifier can be placed on the robot's torso, hips, waist, back, or other predetermined parts of the robot body. After a camera mounted on the target mobile chassis, the robot body, or an external sensing component captures an image containing the visual identifier, the visual identifier is identified using a computer vision algorithm. In practice, the camera and visual identifier cannot both be placed on the robot body; the visual identifier can be placed on the chassis and the camera on the robot, or the visual identifier can be on the robot and the camera on the chassis / external sensing component.

[0035] When the visual identifier is AprilTag, the identifier number can be parsed simultaneously with the detection of identifier boundaries and corner points, enabling the differentiation of specific identification objects in situations involving multiple identifiers or multiple devices. Since the actual size of the identifier is known, and the coordinates of each corner point in the identifier's own coordinate system are also known, the spatial pose of the visual identifier relative to the camera can be determined by solving the projection relationship between the 3D point and the 2D pixel. Furthermore, since the visual identifier is set at a predetermined location on the robot's body, the installation relationship between the identifier's coordinate system and the robot's body coordinate system is known. Therefore, after obtaining the spatial pose of the identifier relative to the camera, the pose of the robot body relative to the camera can be calculated using this installation relationship. Similarly, the positional relationship between the camera and the target mobile chassis is also known or measurable, thereby determining the robot's position and orientation relative to the target mobile chassis.

[0036] When the visual identifier is a QR code or other pattern identifier, the image position, scale, orientation or corner position of the identifier can be determined based on the image recognition results, and the relative pose information between the robot and the target mobile chassis can be estimated by combining camera parameters, identifier size or depth information.

[0037] In one optional implementation, the confidence level of the visual sign recognition result can be determined based on the recognition area of ​​the visual sign, corner detection stability, reprojection error, recognition continuity, or occlusion degree. These parameters are directly related to the pose calculation accuracy: the smaller the recognition area, the greater the amplification factor of the pixel error in corner localization when converted to spatial pose; the larger the reprojection error, the lower the degree of agreement between the calculated pose and the actual observation. When the visual sign is occluded, the image is blurred, or the recognition confidence level is lower than a preset threshold, the weight of the visual sign recognition result in the fusion process can be reduced, or the perception results from a depth camera, RGB-D sensor, or LiDAR can be switched.

[0038] Furthermore, when fusing the various perception results, i.e., relative pose information, if there are differences between the results from different perception sources, different weights can be assigned to different perception results based on the availability status, recognition confidence, measurement stability, historical error, or preset priority of each perception source, and fusion can be performed based on these weights. For example, when the visual tag recognition is stable and not occluded, the recognition result of AprilTag or the visual tag can be used first to determine the relative angle between the robot and the target mobile chassis; when estimating the height or edge position of the target mobile chassis, the spatial point cloud results obtained by depth cameras, RGB-D sensors, or 3D LiDAR can be used first; when the outline of the target mobile chassis is clear but the visual tag is not visible, the outline extraction result of LiDAR can be used first to determine the relative distance and lateral deviation.

[0039] When the deviation between the output of a certain sensing source and the results of other sensing sources exceeds a preset range, or when the confidence level of that sensing source is lower than a preset threshold, the result of that sensing source can be marked as an abnormal result, and its weight can be reduced or the result can be directly removed during the fusion process. Through this method, even under conditions of sensor occlusion, recognition failure, missing depth, point cloud noise, or changes in ambient light, relatively stable target relative pose information between the robot and the target mobile chassis can still be obtained.

[0040] Step S12: Based on the obtained target relative pose information, determine whether the robot and the target mobile chassis currently meet the preset alignment conditions.

[0041] In this embodiment, the preset alignment conditions include that when the robot performs the up-and-down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and that the movement trajectory of the feet does not interfere with the target mobile chassis, and that the current state fed back by the robot is a stable state that has completed standing, balancing, or maintaining the initial posture. Specifically, determining whether the robot and the target mobile chassis currently meet preset alignment conditions based on the acquired target relative pose information includes: determining whether the lateral deviation between the robot and the target mobile chassis in the target relative pose information is less than a first threshold; and / or, determining whether the longitudinal distance between the robot and the target mobile chassis in the target relative pose information is within a preset distance range; and / or, determining whether the relative angle between the robot and the target mobile chassis in the target relative pose information is less than a second threshold; and / or, determining whether the height difference between the robot and the target mobile chassis in the target relative pose information is within the height range for executing chassis-mounting and chassis-mounting actions; and / or, determining whether the robot's current posture meets the starting posture requirements for chassis-mounting and chassis-mounting actions; and / or, determining whether the robot's current state feedback is a stable state; the stable state indicates that the robot has completed standing, balancing, or maintaining the starting posture.

[0042] It should be noted that the above judgment conditions correspond to the feasibility requirements of the upper and lower chassis movements in different directions. A lateral deviation less than the first threshold ensures that the robot's foot, after being lifted and extended forward, lands within the effective support area of ​​the target mobile chassis's bearing surface. A longitudinal distance within a preset range is necessary because if the longitudinal distance is too large, the robot cannot reach the edge of the chassis after lifting its leg, while if the longitudinal distance is too small, the robot is prone to interference with the chassis during leg lifting or forward shift of the center of gravity. In other words, the target landing point of the foot is located within the target support area of ​​the corresponding bearing surface of the target mobile chassis; and the foot's movement trajectory or foot movement envelope does not interfere with the edge, contour, or obstacle area of ​​the target mobile chassis. A relative angle less than the second threshold ensures that the robot's torso orientation is basically consistent with the vertical direction of the target mobile chassis, avoiding a significant asymmetrical distribution of the foot landing points on the chassis. A height difference within the height range for executing upper and lower chassis movements is determined by the robot's own leg lifting height and joint range of motion. The robot's current state feedback can be directly returned by the robot control system. When the robot is in a stable state, it means that the robot has completed standing, balancing, or maintaining the initial posture, which can serve as a prerequisite for triggering the chassis movement strategy. If the robot is not in a stable state, such as when it is still walking, adjusting its center of gravity, or when its balance control has not yet converged, the chassis movement strategy will not be triggered.

[0043] Specifically, the robot's current posture can be obtained by the robot's own state estimation module, and the stable state can be output by the robot control system, balance control module, or motion state machine. The current posture and the stable state can serve as the basis for determining whether to trigger the upper and lower chassis motion strategy. The method for obtaining the current posture, the method for determining the stable state, and the corresponding judgment thresholds can be set according to the robot model, robot control system, and the upper and lower chassis motion strategy adopted; this application does not impose specific limitations on these aspects.

[0044] In addition, the first threshold, the preset distance range, the second threshold, and the height range mentioned above can all be set and adjusted according to the size of the robot, the height of the target mobile chassis, the requirements of the adopted action strategy, or the specific task scenario. This application does not limit their specific values.

[0045] Step S13: If the robot and the target mobile chassis do not meet the preset alignment conditions, an alignment adjustment command is generated based on the target relative pose information, and the robot is controlled based on the alignment adjustment command. And / or, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then the robot is triggered to execute the up and down chassis movement strategy of the target mobile chassis.

[0046] In this embodiment, the step of controlling the robot based on the alignment adjustment command, and / or adjusting the pose of the target mobile chassis until the robot and the target mobile chassis meet the preset alignment conditions, and then triggering the robot to execute the up-and-down chassis movement strategy of the target mobile chassis, includes: In the first stage, if the relative angle is not less than the second threshold, then based on the robot turning command in the alignment adjustment command, and / or the target mobile chassis orientation adjustment command, controlling the robot, and / or the target mobile chassis to perform orientation adjustment until the relative angle is less than the second threshold; In the second stage, if the lateral deviation is not less than the first threshold or the longitudinal distance is not within the preset distance range, then based on the robot lateral adjustment command, robot forward command, robot backward command in the alignment adjustment command, and / or the target mobile chassis position adjustment command, controlling the robot, and / or the target mobile chassis to perform position adjustment until the lateral deviation is less than the first threshold and the longitudinal distance is not within the preset distance range, controlling the robot, and / or the target mobile chassis to perform position adjustment until the lateral deviation is less than the first threshold and the longitudinal distance is less than the second threshold. The distance is within the preset distance range; in the third stage, it is determined whether the height difference is within the height range and whether the robot's current posture meets the starting posture requirements for the upper and lower chassis movement; if the robot's current posture does not meet the starting posture requirements for the upper and lower chassis movement, the robot is controlled to adjust its starting posture based on the robot gait parameter adjustment instruction or the robot movement starting posture adjustment instruction in the alignment adjustment instruction; in the fourth stage, the current state fed back by the robot is obtained to confirm whether the robot is in the stable state; in the fifth stage, when the conditions corresponding to the first to the fourth stages are all met, the robot is triggered to execute the upper and lower chassis movement strategy relative to the target moving chassis; if the preset alignment conditions are still not met within the preset time, preset number of adjustments, or preset adjustment distance range, the upper and lower chassis movement strategy is stopped from being triggered, and a re-sensing instruction, a repositioning instruction, a manual assistance prompt, a safety stop instruction, an exit upper and lower chassis mode instruction, or an alarm prompt instruction is output.

[0047] It should be noted that the order of the five stages mentioned above is not arbitrary. The relative angle is adjusted before the position because the lateral deviation and longitudinal distance between the robot and the target moving chassis are usually defined in a coordinate system based on the chassis orientation. If the position is adjusted before the relative angle converges, the previously adjusted lateral deviation and longitudinal distance will change again after the robot's orientation changes, causing repeated position adjustments. Adjusting the position after the relative angle converges ensures that the position adjustment result will not be invalidated by subsequent orientation changes. Similarly, the confirmation of height difference and initial posture is arranged after position alignment because the robot's standing plane and torso posture may change during movement, and confirmation is only meaningful after the position is basically stable. The robot's current state feedback is arranged in the last stage because posture adjustment itself will disrupt the robot's original balance state, and the action triggering condition is only truly met after it converges back to a stable state.

[0048] Specifically, the alignment adjustment commands may include at least one of the following: robot forward command, robot backward command, robot lateral adjustment command, robot turning command, target mobile chassis position adjustment command, target mobile chassis orientation adjustment command, robot gait parameter adjustment command, robot motion start posture adjustment command, and upper and lower chassis motion strategy switching command. When the lateral deviation exceeds a first threshold, a robot lateral adjustment command can be generated to control the robot to perform lateral fine-tuning, or a target mobile chassis position adjustment command can be generated to control the target mobile chassis to adjust its own position; when the relative angle exceeds a second threshold, a robot turning command, a target mobile chassis orientation adjustment command, or a coordinated turning command of the robot and the target mobile chassis can be generated; when the longitudinal distance is not within a preset distance range, a robot forward command or a robot backward command can be generated, or the target mobile chassis can be controlled to move closer to or away from the robot.

[0049] When the robot needs to adjust its starting posture, it can also generate robot gait parameter adjustment commands or robot starting posture adjustment commands. Specifically, the robot gait parameter adjustment commands can include at least one of the following: stride length adjustment command, stride width adjustment command, foot landing point adjustment command, torso posture adjustment command, foot lift height adjustment command, and movement speed adjustment command. For example, when there is only a small deviation in longitudinal distance, reducing the stride length allows the robot to approach the target position with a smaller step, avoiding oscillations around the target position due to excessively large strides; when improved landing accuracy is required, the movement speed can be reduced and the foot lift height adjusted accordingly.

[0050] Additionally, depending on the current scenario, adjustments can be made by the robot alone, by the target mobile chassis alone, or by both the robot and the target mobile chassis together. For example, when the robot has sufficient space to move and its current state allows for movement, the robot's position and orientation can be adjusted first. When the robot is not suitable for movement or needs to maintain a stable state, the target mobile chassis's position and orientation can be adjusted first. When the deviation between the robot and the target mobile chassis is large, both the robot and the target mobile chassis can be adjusted collaboratively, with the two sharing the adjustment amount to shorten the time required to complete the alignment.

[0051] The aforementioned chassis-mounting motion strategy may specifically include at least one of the following: a preset chassis-mounting motion strategy, a chassis-mounting motion strategy trained using reinforcement learning, a chassis-mounting motion strategy trained using a combination of imitation learning and reinforcement learning, an assisted chassis-mounting motion strategy, an autonomous chassis-mounting motion strategy, a fixed-height chassis-mounting motion strategy, and a variable-height chassis-mounting motion strategy. A suitable motion strategy can be selected from these strategies based on the height of the target mobile chassis, the robot's current posture, the presence of an assistant, the orientation of the chassis-mounting, and the current alignment state. For example, when the height of the target mobile chassis is fixed and the relative pose between the robot and the target mobile chassis meets the initial conditions of the fixed-height motion strategy, the fixed-height chassis-mounting motion strategy can be triggered; when the height of the target mobile chassis is not fixed, the variable-height chassis-mounting motion strategy can be triggered; and when a person assists the robot in mounting and dismounting, an assisted chassis-mounting motion strategy can be triggered.

[0052] In addition, a safety judgment mechanism is set up during the collaborative alignment process. If the preset alignment conditions are not met within the preset time, preset number of adjustments, or preset adjustment distance, the upper and lower chassis movement strategy will be stopped, and corresponding processing instructions will be output. For example, if the relative angle between the robot and the target moving chassis still exceeds the second threshold after multiple adjustments, it means that it is difficult to complete the alignment by its own adjustment in the current scenario. At this time, the upper and lower chassis movement will be stopped, and a prompt will be made that repositioning or manual assistance is needed. This can prevent the robot from forcibly triggering the movement when the conditions for action execution are not met, which could cause danger.

[0053] As can be seen, in this embodiment, as Figure 3As shown, the robot obtains target relative pose information between itself and the target mobile chassis based on a preset pose perception method. The preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual identifier perception method. Based on the obtained target relative pose information, it is determined whether the robot and the target mobile chassis currently meet preset alignment conditions. If the robot and the target mobile chassis do not meet the preset alignment conditions, an alignment adjustment command is generated based on the target relative pose information, and the robot is controlled based on the alignment adjustment command. And / or, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then the robot is triggered to execute the target mobile chassis's up and down chassis movement strategy. In other words, by introducing a collaborative alignment step before the upper and lower chassis movements are triggered, the execution conditions for the upper and lower chassis movements are transformed from manual judgment based on experience to automatic judgment based on the target's relative pose information. This avoids the robot directly executing the upper and lower chassis movements when the initial pose deviation is large, thus improving the success rate of the upper and lower chassis movements. Since lateral deviation, longitudinal distance, relative angle, height difference, initial posture of the movement, and the robot's feedback of maintaining a stable state are all included in the preconditions for action triggering, the risk of the robot failing to land, losing its posture, or falling is reduced accordingly. The alignment adjustment command is automatically generated by the system based on the target's relative pose information, reducing the need for manual remote control and manual placement of the robot. At the same time, because the target's relative pose... The information is obtained by fusing the perception results from depth point cloud perception, laser contour perception, and visual mark perception. It can still continue to perform collaborative alignment even when some perception sources are occluded or fail. In addition, the alignment process can be performed by the robot, the target moving chassis, or both. It can also be combined with various upper and lower chassis action strategies, such as preset action strategies, reinforcement learning strategies, imitation learning and reinforcement learning combined strategies, auxiliary mode strategies, and autonomous mode strategies, making it widely applicable. When alignment fails or the robot does not report maintaining a stable state, the system does not trigger the upper and lower chassis action strategies and can output a safety stop command or manual assistance prompts, thereby avoiding the triggering of dangerous actions.

[0054] refer to Figure 4 As shown in the figure, this application also discloses a robot's upper and lower chassis operation device, including: The pose information acquisition module 11 is used to acquire the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method; the preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method and visual mark perception method. The condition judgment module 12 is used to determine whether the robot and the target mobile chassis currently meet the preset alignment conditions based on the acquired target relative pose information. The preset alignment conditions include that when the robot performs the up and down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state fed back by the robot is a stable state that has completed standing, balancing or maintaining the initial posture. The pose adjustment module 13 is used to generate an alignment adjustment command based on the target relative pose information if the robot and the target mobile chassis do not meet the preset alignment conditions, and control the robot based on the alignment adjustment command, and / or, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then triggers the robot to execute the target mobile chassis up and down chassis movement strategy.

[0055] As can be seen, in this embodiment, the triggering condition for the upper and lower chassis movements is changed from manual experience-based judgment to automatic judgment based on the target's relative pose information. Before the movement is triggered, the robot and the target moving chassis are aligned in advance, avoiding the robot from directly executing the upper and lower chassis movements when the initial pose deviation is large. The success rate of the upper and lower chassis movements is correspondingly improved. Due to the setting of corresponding preset alignment conditions, the risk of the robot experiencing foot landing point deviation, failure to switch support phases, posture instability, or even falling is reduced, thus improving safety. At the same time, the alignment adjustment command is automatically generated by the system based on the target's relative pose information, reducing the need for manual remote control and manual placement of the robot, and reducing the reliance on the operator's experience.

[0056] In some specific embodiments, the pose information acquisition module 11 may specifically include: The first pose information acquisition submodule is used to acquire the three-dimensional point cloud data of the target mobile chassis through the depth point cloud perception method, and extract the point cloud region corresponding to the target mobile chassis from the three-dimensional point cloud data, so as to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region; the first relative pose information includes one or a combination of several of the following: height difference, longitudinal distance, lateral deviation and relative angle. The second pose information acquisition submodule is used to acquire laser scanning data or laser point cloud data of the target mobile chassis through the laser contour perception method, and extract the geometric contour features of the target mobile chassis from the laser scanning data or laser point cloud data, so as to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features; the second relative pose information includes one or a combination of several of the following: relative distance, lateral deviation and relative angle. The third pose information acquisition submodule is used to acquire images containing visual symbols through the visual symbol perception method and identify the visual symbols, so as to determine the third relative pose information between the robot and the target mobile chassis based on the identification result of the visual symbols; the third relative pose information includes the position of the robot relative to the target mobile chassis, and / or, orientation. An information determination unit is configured to determine target pose information from the first relative pose information, the second relative pose information, and the third relative pose information; An information evaluation unit is used to convert the target pose information to a unified coordinate system if the number of pose information in the target pose information is greater than 1, and to perform consistency judgment and confidence evaluation on each relative pose information after conversion in the unified coordinate system to obtain consistency judgment results and confidence evaluation results. The information fusion unit is used to perform weighted fusion of each of the relative pose information based on the consistency judgment result and the confidence evaluation result to obtain the target relative pose information between the robot and the target mobile chassis.

[0057] In some specific embodiments, the first pose information acquisition submodule may specifically include: The first image acquisition unit is used to acquire depth images of the target mobile chassis through a depth camera or RGB-D sensor, and convert the depth images into three-dimensional point cloud data according to the camera intrinsic parameters; The region extraction unit is used to perform invalid depth point removal, outlier point removal, region smoothing or edge enhancement processing on the three-dimensional point cloud data, and extract the point cloud region corresponding to the target mobile chassis from the processed three-dimensional point cloud data. The region fitting unit is used to perform planar fitting, edge extraction, or contour fitting on the point cloud region to determine the upper surface height, front edge position, left and right boundary positions, and chassis orientation of the target mobile chassis. The first pose information acquisition unit is used to determine the height difference, longitudinal distance, lateral deviation and relative angle between the robot and the target mobile chassis in sequence according to the upper surface height, the front edge position, the left and right boundary positions and the chassis orientation, so as to obtain the first relative pose information.

[0058] In some specific embodiments, the second pose information acquisition submodule may specifically include: The first pose determination unit is used to acquire laser scanning data or laser point cloud data of the target mobile chassis through a two-dimensional lidar, and determine the relative distance, lateral deviation and relative angle between the robot and the target mobile chassis based on the distance abrupt change, line segment fitting, contour clustering or the known geometric dimensions of the target mobile chassis in the laser scanning data or the laser point cloud data, so as to obtain the second relative pose information. The second pose determination unit is used to remove ground points or filter height ranges from the laser scanning data or laser point cloud data obtained by the three-dimensional lidar to obtain the remaining point cloud, and to cluster or extract the contours of the remaining point cloud to determine the relative distance, lateral deviation and relative angle between the robot and the target mobile chassis, so as to obtain the second relative pose information.

[0059] In some specific embodiments, the third pose information acquisition submodule may specifically include: The second image acquisition unit is used to acquire images containing the visual identifiers through a preset camera; the preset camera is installed on the target mobile chassis, the robot body, or an external sensing component. The third pose determination unit is used to detect the identifier boundary and corner points of the visual identifier in the image if the visual identifier is AprilTag, and determine the spatial pose of the visual identifier relative to the camera based on the camera intrinsic parameters, the actual size of the visual identifier and the position of the corner points in the image, so as to obtain the third relative pose information. The fourth pose determination unit is used to determine the visual identifier based on the recognition result of the image if the visual identifier is a QR code or a pattern identifier, and to determine the spatial pose of the visual identifier relative to the camera based on the camera intrinsic parameters and the actual size or depth information of the visual identifier, so as to obtain the third relative pose information.

[0060] In some specific embodiments, the condition judgment module 12 may specifically include: The first condition judgment unit is used to determine whether the lateral deviation between the robot and the target mobile chassis in the target relative pose information is less than a first threshold. The second condition judgment unit is used to determine whether the longitudinal distance between the robot and the target mobile chassis in the target relative pose information is within a preset distance range; The third condition judgment unit is used to determine whether the relative angle between the robot and the target mobile chassis in the target relative pose information is less than the second threshold. The fourth condition judgment unit is used to determine whether the height difference between the robot and the target mobile chassis in the target relative pose information is within the height range where the chassis can be moved up and down. The fifth condition judgment unit is used to determine whether the robot's current posture meets the requirements for the starting posture of the upper and lower chassis movements; The sixth condition judgment unit is used to determine whether the current state fed back by the robot is a stable state; the stable state indicates that the robot has completed standing, balancing or maintaining the initial posture.

[0061] In some specific embodiments, the pose adjustment module 13 may specifically include: The first posture adjustment unit is used in the first stage to control the robot and / or the target mobile chassis to adjust their orientation based on the robot turning instruction in the alignment adjustment instruction and / or the target mobile chassis orientation adjustment instruction, if the relative angle is not less than the second threshold, until the relative angle is less than the second threshold. The second pose adjustment unit is used in the second stage to control the robot and / or the target mobile chassis to adjust their positions based on the robot lateral adjustment command, robot forward command, robot backward command, and / or the target mobile chassis position adjustment command in the alignment adjustment command, if the lateral deviation is not less than the first threshold or the longitudinal distance is not within the preset distance range, until the lateral deviation is less than the first threshold and the longitudinal distance is within the preset distance range. The adjustment action judgment unit is used in the third stage to determine whether the height difference is within the height range and whether the robot's current posture meets the starting posture requirements of the upper and lower chassis actions. The third pose adjustment unit is used to control the robot to adjust the starting pose of the action based on the robot gait parameter adjustment instruction or the robot action starting pose adjustment instruction in the alignment adjustment instruction if the current pose of the robot does not meet the starting pose requirements of the upper and lower chassis actions. The fourth pose adjustment unit is used to obtain the current state fed back by the robot in the fourth stage to confirm whether the robot is in the stable state. The fifth pose adjustment unit is used to trigger the robot to execute a chassis up and down movement strategy relative to the target moving chassis when the conditions corresponding to the first stage to the fourth stage are all met in the fifth stage. The strategy stop triggering unit is used to stop triggering the upper and lower chassis action strategy if the preset alignment conditions are not met within a preset time, preset number of adjustments, or preset adjustment distance, and outputs a re-sensing command, a repositioning command, a manual assistance prompt, a safety stop command, an exit from the upper and lower chassis mode command, or an alarm prompt command.

[0062] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0063] Figure 5 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the robot's upper and lower chassis operation method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0064] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0065] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0066] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the robot's upper and lower chassis operation method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0067] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned method for manipulating the robot's upper and lower chassis. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0068] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0069] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0070] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0071] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0072] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for operating the upper and lower chassis of a robot, characterized in that, include: The robot acquires the target relative pose information between itself and the target mobile chassis based on a preset pose perception method. The preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual mark perception method. Based on the obtained target relative pose information, it is determined whether the robot and the target mobile chassis currently meet the preset alignment conditions; the preset alignment conditions include that when the robot performs the up and down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state fed back by the robot is a stable state that has completed standing, balancing or maintaining the initial posture; If the robot and the target mobile chassis do not meet the preset alignment conditions, an alignment adjustment command is generated based on the target relative pose information, and the robot is controlled based on the alignment adjustment command. Alternatively, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then the robot is triggered to execute the target mobile chassis's up and down chassis movement strategy.

2. The method for operating the upper and lower chassis of a robot according to claim 1, characterized in that, The acquisition of target relative pose information between the robot and the target mobile chassis based on a preset pose perception method includes: The robot acquires three-dimensional point cloud data of the target mobile chassis using the depth point cloud perception method, and extracts the point cloud region corresponding to the target mobile chassis from the three-dimensional point cloud data to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region; the first relative pose information includes one or a combination of several of the following: height difference, longitudinal distance, lateral deviation, and relative angle. The laser scanning data or laser point cloud data of the target mobile chassis is acquired through the laser contour perception method, and the geometric contour features of the target mobile chassis are extracted from the laser scanning data or laser point cloud data to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features; the second relative pose information includes one or a combination of several of the following: relative distance, lateral deviation and relative angle. Images containing visual identifiers are acquired and the visual identifiers are identified using the visual identifier perception method. Based on the identification results of the visual identifiers, a third relative pose information between the robot and the target mobile chassis is determined. The third relative pose information includes the position of the robot relative to the target mobile chassis and / or its orientation. The target pose information is determined from the first relative pose information, the second relative pose information, and the third relative pose information; If the number of pose information in the target pose information is greater than 1, the target pose information is transformed to a unified coordinate system, and the consistency judgment and confidence evaluation of each relative pose information after transformation are performed in the unified coordinate system to obtain the consistency judgment result and confidence evaluation result. Based on the consistency judgment result and the confidence evaluation result, the relative pose information of each is weighted and fused to obtain the target relative pose information between the robot and the target mobile chassis.

3. The method for operating the upper and lower chassis of a robot according to claim 2, characterized in that, The step of acquiring 3D point cloud data of the target mobile chassis through the depth point cloud perception method, and extracting the point cloud region corresponding to the target mobile chassis from the 3D point cloud data, so as to determine the first relative pose information between the robot and the target mobile chassis based on the point cloud region, includes: Depth images of the target mobile chassis are acquired using a depth camera or RGB-D sensor, and the depth images are converted into three-dimensional point cloud data based on the camera's intrinsic parameters. The three-dimensional point cloud data is subjected to invalid depth point removal, outlier point removal, region smoothing or edge enhancement processing, and the point cloud region corresponding to the target mobile chassis is extracted from the processed three-dimensional point cloud data. Plane fitting, edge extraction, or contour fitting are performed on the point cloud region to determine the upper surface height, front edge position, left and right boundary positions, and chassis orientation of the target mobile chassis; Based on the height of the upper surface, the position of the front edge, the positions of the left and right boundaries, and the orientation of the chassis, the height difference, longitudinal distance, lateral deviation, and relative angle between the robot and the target mobile chassis are determined sequentially to obtain the first relative pose information.

4. The method for operating the upper and lower chassis of a robot according to claim 2, characterized in that, The step of acquiring laser scanning data or laser point cloud data of the target mobile chassis through the laser contour perception method, and extracting the geometric contour features of the target mobile chassis from the laser scanning data or laser point cloud data, to determine the second relative pose information between the robot and the target mobile chassis based on the geometric contour features, includes: The robot acquires laser scanning data or laser point cloud data of the target mobile chassis using a two-dimensional lidar, and determines the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis based on the distance abrupt change, line segment fitting, contour clustering, or known geometric dimensions of the target mobile chassis in the laser scanning data or laser point cloud data, so as to obtain the second relative pose information. Alternatively, the laser scanning data or laser point cloud data obtained by the three-dimensional lidar can be processed by removing ground points or filtering by height range to obtain the remaining point cloud. The remaining point cloud can then be clustered or its contour extracted to determine the relative distance, lateral deviation, and relative angle between the robot and the target mobile chassis, thereby obtaining the second relative pose information.

5. The method for operating the upper and lower chassis of a robot according to claim 2, characterized in that, The step of acquiring images containing visual identifiers through the visual identifier perception method and identifying the visual identifiers, and determining the third relative pose information between the robot and the target mobile chassis based on the identification result of the visual identifiers, includes: Images containing the visual identifiers are acquired by a preset camera; the preset camera is mounted on the target mobile chassis, the robot's body, or an external sensing component. If the visual identifier is AprilTag, then the identifier boundary and corner points of the visual identifier in the image are detected, and the spatial pose of the visual identifier relative to the camera is determined based on the camera intrinsic parameters, the actual size of the visual identifier, and the position of the corner points in the image, so as to obtain the third relative pose information. If the visual identifier is a QR code or a pattern identifier, the visual identifier is determined based on the recognition result of the image, and the spatial pose of the visual identifier relative to the camera is determined based on the camera intrinsic parameters and the actual size or depth information of the visual identifier, so as to obtain the third relative pose information.

6. The method for operating the upper and lower chassis of a robot according to any one of claims 2 to 5, characterized in that, The step of determining whether the robot and the target mobile chassis currently meet the preset alignment conditions based on the acquired target relative pose information includes: Determine whether the lateral deviation between the robot and the target mobile chassis in the target relative pose information is less than a first threshold. And / or, determine whether the longitudinal distance between the robot and the target mobile chassis in the target relative pose information is within a preset distance range; And / or, determine whether the relative angle between the robot and the target mobile chassis in the target relative pose information is less than a second threshold; And / or, determine whether the height difference between the robot and the target mobile chassis in the target relative pose information is within the height range where the chassis can be moved up and down; And / or, determine whether the robot's current posture meets the requirements for the starting posture of the upper and lower chassis movements; And / or, determine whether the current state fed back by the robot is a stable state; the stable state indicates that the robot has completed standing, balancing or maintaining the initial posture.

7. The method for operating the upper and lower chassis of a robot according to claim 6, characterized in that, The step of controlling the robot based on the alignment adjustment command, and / or adjusting the pose of the target mobile chassis until the robot and the target mobile chassis meet the preset alignment conditions, triggering the robot to execute the up-and-down chassis movement strategy of the target mobile chassis, includes: In the first stage, if the relative angle is not less than the second threshold, then based on the robot turning command in the alignment adjustment command and / or the target mobile chassis orientation adjustment command, control the robot and / or the target mobile chassis to perform orientation adjustment until the relative angle is less than the second threshold. In the second stage, if the lateral deviation is not less than the first threshold or the longitudinal distance is not within the preset distance range, then based on the robot lateral adjustment command, robot forward command, robot backward command in the alignment adjustment command, and / or the target mobile chassis position adjustment command, control the robot and / or the target mobile chassis to perform position adjustment until the lateral deviation is less than the first threshold and the longitudinal distance is within the preset distance range; In the third stage, it is determined whether the height difference is within the height range and whether the robot's current posture meets the starting posture requirements of the upper and lower chassis movements; If the robot's current posture does not meet the requirements of the starting posture of the upper and lower chassis movements, the robot's starting posture is adjusted based on the robot gait parameter adjustment instruction or the robot movement starting posture adjustment instruction in the alignment adjustment instruction. In the fourth stage, the current state fed back by the robot is obtained to confirm whether the robot is in the stable state. In the fifth stage, when the conditions corresponding to the first to the fourth stages are all met, the robot is triggered to execute the up and down chassis movement strategy relative to the target moving chassis. If the preset alignment conditions are not met within the preset time, preset number of adjustments, or preset adjustment distance, the upper and lower chassis action strategy will be stopped, and a re-sensing command, a repositioning command, a manual assistance prompt, a safety stop command, an exit from the upper and lower chassis mode command, or an alarm prompt command will be output.

8. A robot's upper and lower chassis operating device, characterized in that, include: The pose information acquisition module is used to acquire the target relative pose information between the robot and the target mobile chassis based on a preset pose perception method. The preset pose perception method is one or a combination of several of the following: depth point cloud perception method, laser contour perception method, and visual mark perception method. The condition judgment module is used to determine whether the robot and the target mobile chassis currently meet the preset alignment conditions based on the acquired target relative pose information. The preset alignment conditions include that when the robot performs the up and down chassis movement, its feet can land in the target support area of ​​the corresponding bearing surface of the target mobile chassis, and the movement trajectory of the feet does not interfere with the target mobile chassis, and the current state reported by the robot is a stable state that has completed standing, balancing or maintaining the initial posture. The pose adjustment module is used to generate an alignment adjustment command based on the target relative pose information if the robot and the target mobile chassis do not meet the preset alignment conditions, and control the robot based on the alignment adjustment command, and / or, the target mobile chassis adjusts its pose until the robot and the target mobile chassis meet the preset alignment conditions, and then triggers the robot to execute the up and down chassis movement strategy of the target mobile chassis.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method of operating the upper and lower chassis of the robot as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the upper and lower chassis operation method of the robot as described in any one of claims 1 to 7.

Images