Control Method, Device, Readable Storage Medium and Robot of a Robot

Abstract

The present invention provides a control method, device, readable storage medium and robot for a robot. The control method for the robot includes: obtaining the point cloud and contour box of a target object; selecting one or a combination of the point cloud and the contour box as path planning information; determining the driving path of the robot according to the path planning information; and controlling the robot to drive according to the driving path. Through the above technical solution, the present invention obtains the point cloud and contour box of the target object, and can select one or a combination of the point cloud and the contour box as path planning information according to actual needs, so as to quickly and efficiently plan the driving path of the robot, control the robot to drive according to the planned driving path, and further improve the obstacle avoidance effect during the driving process of the robot.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control, and more particularly, to a control method, device, readable storage medium, and robot for a robot. Background Art

[0002] The obstacle avoidance ability is an important indicator to measure the intelligence level of a floor cleaning robot. Currently, the main obstacle avoidance sensors of floor cleaning robots include infrared, line lasers, radars, etc. In the prior art, there is a problem of unstable obstacle avoidance effect for floor cleaning robots. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art or related technologies.

[0004] To this end, a first aspect of the present invention is to propose a control method for a robot.

[0005] A second aspect of the present invention is to propose a control device for a robot.

[0006] A third aspect of the present invention is to propose a control device for a robot.

[0007] A fourth aspect of the present invention is to propose a readable storage medium.

[0008] A fifth aspect of the present invention is to propose a robot.

[0009] In view of this, according to the first aspect of the present invention, a control method for a robot is proposed, including: obtaining the point cloud and bounding box of a target object; selecting one or a combination of the point cloud and bounding box as path planning information; determining the driving path of the robot according to the path planning information; and controlling the robot to drive according to the driving path.

[0010] In this technical solution, during the driving process of the robot, there may be a situation where the robot's progress is blocked by a target object. In response to the above situation, the robot is controlled to perform an avoidance action, and the point cloud and bounding box of the target object are obtained respectively. During the process of planning the driving path, the point cloud or bounding box can be used alone as path planning information, or the point cloud and bounding box can be used as path planning information. According to the determined path planning information, the driving path of the robot is planned, and then the robot is controlled to drive according to the planned driving path.

[0011] It should be noted that the target object includes obstacles of different types, shapes, and poses. For example, low objects (carpets, thresholds), irregular objects (wires, clothes), and specific types of obstacles (pet feces, liquid stains).

[0012] The robot can be a floor-sweeping robot, and an image acquisition device such as a camera is configured in the floor-sweeping robot.

[0013] The bounding box of the target object includes the smallest rectangle of the complete contour of the target object. The bounding box can be obtained by means of visual recognition. Specifically, an image including the target object is captured by an image acquisition device such as a camera, and the bounding box of the target object is obtained by recognizing the image.

[0014] The point cloud of the target object includes discrete point data of the specific contour of the target object. The point cloud of the target object is distributed inside the bounding box, and both the bounding box and the point cloud can be used as path planning information.

[0015] It can be understood that for the same target object, the point cloud of the target object is located inside the bounding box.

[0016] In the case where there is a large amount of point cloud inside the bounding box, it is determined that the difference in recognition effect between using the point cloud as path planning information and using the bounding box as path planning information is small at this time. To reduce the calculation amount, the bounding box can be used as path planning information.

[0017] In the case where the point cloud is concentrated in a certain area of the bounding box, that is, when the number of point clouds inside the bounding box is small, it is determined that the accuracy of path planning by only using the bounding box as path planning information is low at this time. Therefore, it can be selected to use the point cloud as path planning information, or use the point cloud and the bounding box as path planning information.

[0018] In the related art, the obstacle avoidance strategy of the floor-sweeping robot is based on a single piece of data for planning the driving path. For example, using only the bounding box of the obstacle as the basis for obstacle avoidance planning, during the process of identifying and avoiding a slender obstacle (wire), since the true contour of the obstacle is only the diagonal of the bounding box, it affects the cleaning coverage rate of the floor-sweeping robot.

[0019] The control method of the robot in this technical solution obtains the point cloud and the bounding box of the target object, and can select one or a combination of the point cloud and the bounding box as path planning information according to actual needs, so as to quickly and efficiently plan the driving path of the robot, control the robot to drive according to the planned driving path, and further improve the obstacle avoidance effect during the driving process of the robot.

[0020] According to the control method of the above-mentioned robot of the present invention, the following additional technical features may also be included:

[0021] In the above technical solution, selecting one or a combination of the point cloud and the bounding box as path planning information includes: determining the area ratio of the point cloud in the bounding box; selecting one or a combination of the point cloud and the bounding box as path planning information according to the area ratio.

[0022] In this technical solution, the data of the target object is processed through operations. Specifically, the area data of the point cloud and the contour box are processed through operations respectively. The area of the contour box and the distribution contour area of the point cloud are calculated, and further, the proportion of the distribution contour area of the point cloud in the area of the contour box is calculated. According to the value of the area proportion, one piece of data from the point cloud and the contour box or a data combination of the point cloud and the contour box is used as the path planning information for path planning.

[0023] In the process of determining the proportion of the area of the point cloud in the contour box, it is necessary to obtain the area of the point cloud and the area of the contour box. Among them, the acquired image is divided into several grids with the same area, and the area of a single grid is set as the unit area. The area of the point cloud is the area corresponding to the number of grids occupied by the point cloud, and the area of the contour box is the area corresponding to the number of grids included in the contour box. The proportion of the area of the point cloud in the contour box is calculated based on the area of the point cloud and the area of the contour box.

[0024] Specifically, in the case where the area proportion is relatively large, it is determined that the contour box of the target object can reflect the true contour of the target object, and then the contour box is used as the path planning information. In the case where the area proportion is relatively small, it is determined that the contour box of the recognized target object cannot reflect the true contour of the target object, and then the point cloud alone is used as the path planning information, or the point cloud and the contour box are used as the path planning information.

[0025] The control method of the robot in this technical solution can determine whether the contour box obtained through image recognition can reflect the true contour of the target object based on the proportion of the coverage area of the point cloud in the area of the contour box, and accordingly select one or a combination of the point cloud and the contour box as the path planning information. This further ensures the accuracy of path planning. When the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect but also further improve the cleaning effect of the sweeping robot.

[0026] In the above technical solution, according to the area proportion, one or a combination of the point cloud and the contour box is selected as the path planning information, including: in the case where the area proportion is greater than or equal to the first preset proportion, the contour box is used as the path planning information; in the case where the area proportion is less than or equal to the second preset proportion, the point cloud is used as the path planning information; in the case where the area proportion is greater than the second preset proportion and less than the first preset proportion, the point cloud and the contour box are used as the path planning information; where the first preset proportion is greater than the second preset proportion.

[0027] In this technical solution, the area ratio of the coverage area of the point cloud of the target object in the area of the contour box is calculated, and a first preset ratio and a second preset ratio are obtained. The ratio of the first preset ratio is greater than the ratio of the second preset ratio. The calculated area ratio is compared with the first preset ratio and the second preset ratio respectively, and according to the comparison results, path planning information is determined in the point cloud and the contour box.

[0028] Specifically, the quantitative relationship between the area ratio and the first preset ratio is judged. When it is detected that the area ratio is greater than or equal to the first preset ratio, it is considered that there is a large amount of point cloud in the contour box, and it is determined that the contour box obtained by image recognition is close to the true contour of the target object. At this time, the contour box is used as the path planning information.

[0029] When the area ratio is less than the first preset ratio, the quantitative relationship between the area ratio and the second preset ratio is judged again. When the area ratio is greater than the second preset ratio, it can be determined that there is a certain amount of point cloud in the contour box, and the contour box is cropped by the point cloud of the target object. The cropped contour box is close to the true contour of the target object. At this time, the contour box cropped according to the point cloud is used as the path planning information.

[0030] Under the condition that the comparison result shows that the area ratio is less than or equal to the second preset ratio, it is determined that there is less point cloud in the contour box, and it is determined that the point cloud obtained by image recognition is close to the true contour of the target object. At this time, the point cloud is used as the path planning information.

[0031] The control method of the robot in this technical solution sets the first preset ratio and the second preset ratio, compares the calculated area ratio with the first preset ratio and the second preset ratio respectively, determines whether the contour box obtained by image recognition can reflect the true contour of the target object, and accordingly selects one or a combination of the point cloud and the contour box as the path planning information. Further ensures the accuracy of path planning. When the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect, but also further improve the cleaning effect of the sweeping robot.

[0032] In the above technical solution, the value range of the first preset ratio is 15% to 35%; and / or the value range of the second preset ratio is 65% to 85%.

[0033] In this technical solution, the value ranges of the first preset ratio and the second preset ratio are set. Specifically, the value range of the first preset ratio is set to 15% to 35%, and / or the value range of the second preset ratio is set to 65% to 85%.

[0034] In the control method of the robot in the present invention, by setting the value ranges of the first preset ratio and the second preset ratio, the magnitude relationship between the first preset ratio and the second preset ratio is limited, ensuring the accuracy of the step of selecting the robot path planning information.

[0035] In the above technical solution, the path planning information includes point clouds or bounding boxes. According to the path planning information, the driving path of the robot is determined, including: determining the first boundary information of the point cloud or the bounding box; and determining the driving path according to the first boundary information.

[0036] In this technical solution, when the path planning information only includes a single point cloud or a single bounding box, the driving path is planned only according to the first boundary information of the corresponding point cloud or the first boundary information of the corresponding bounding box.

[0037] When the path planning information includes the point cloud of the target object, first determine the first boundary information of the point cloud of the target object, where the first boundary information includes the distribution contour edge information of the point cloud. When the path planning information includes the bounding box of the target object, first determine the first boundary information of the bounding box of the target object, where the first boundary information includes the boundary information of the bounding box. After determining the first boundary information, plan the driving path of the robot according to the first boundary information.

[0038] Specifically, for different target objects, in the case where there is a small amount of point cloud in the bounding box, the point cloud data distributed at the edge in the point cloud of the target object can outline the edge contour of the target object, and the coordinate data of the edge contour outlined by the point cloud can be set as the first boundary information. In the case where there is a large amount of point cloud in the bounding box, the border of the bounding box of the target object can be recognized as the edge contour of the target object, and the coordinate data of the border of the bounding box can be set as the first boundary information.

[0039] The control method of the robot in this technical solution plans the driving path of the robot through the first boundary information of the point cloud or the bounding box of the target object, ensuring the accuracy of the step of planning the driving path of the robot.

[0040] In the above technical solution, the path planning information includes point clouds and bounding boxes. According to the path planning information, the driving path of the robot is determined, including: determining the second boundary information of the point cloud; determining the target area in the bounding box according to the second boundary information; determining the third boundary information of the target area; and determining the driving path according to the third boundary information.

[0041] In this technical solution, when the path planning information only includes point clouds and bounding boxes, the bounding box is cropped according to the second boundary information of the point cloud, and the driving path is planned according to the third boundary information of the target area obtained by the cropping.

[0042] Specifically, the robot is a floor-sweeping robot. When the proportion of the area of the point cloud in the contour box is greater than a second preset proportion and less than a first preset proportion, if the driving path is planned only based on the point cloud, it is easy to cause obstacle avoidance failure. If the driving path is planned only based on the contour box, it is easy to affect the cleaning effect of the floor-sweeping robot. Therefore, it is selected to crop the contour box according to the point cloud.

[0043] When the path planning information includes the point cloud and the contour box of the target object, first determine the second boundary information of the point cloud of the target object, and then crop the contour box according to the second boundary information.

[0044] It should be noted that the point cloud of the target object is distributed inside the contour box. When there is a certain number of point clouds in the contour box, the point cloud data distributed at the edge in the point cloud of the target object can outline part of the edge contour of the target object. Set the coordinate data of the edge contour outlined by the point cloud as the second boundary information, and then crop the contour box through the second boundary information. The border of the cropped contour box is closer to the edge contour of the target object. The coordinate data of the border of the contour box cropped according to the point cloud can be set as the third boundary information.

[0045] The control method of the robot in this technical solution determines the target area inside the contour box through the second boundary information of the point cloud of the target object, and then plans the driving path of the robot according to the third boundary information of the target area. When there is a certain number of point clouds in the contour box, the accuracy of the step of planning the driving path of the robot is ensured.

[0046] In the above technical solution, obtaining the point cloud and the contour box of the target object includes: obtaining a first image set, and each first image in the first image set includes the target object; identifying the first image set to determine the point cloud and the contour box.

[0047] In this technical solution, an image acquisition device is provided in the robot, which can acquire images including the target object during the driving process of the robot. To improve the accuracy of the obtained point cloud and contour box of the target object, in the step of identifying the point cloud and the contour box of the target object, multiple images including the target object need to be acquired. The set of these images is the first image set. Each image in the first image set includes the target object. By performing image processing on the first image set, the point cloud and the contour box of the target object can be accurately identified.

[0048] Specifically, during the movement of the robot, the image acquisition device starts to acquire images and simultaneously performs recognition on the acquired images. In the case where the target object is included in the detected image, the image is used as an image in the first image set. In the case where the target object is not included in the detected image, the image is filtered out. After the acquisition of the first image set is completed, each image in the first image set is subjected to recognition processing to identify the contour frame and point cloud of the target object in each image, and then smoothing operations are performed on the identified multiple contour frames and multiple point clouds to determine the contour frame and point cloud of the target object.

[0049] The control method of the robot in this technical solution obtains a first image set including multiple images of the shooting object including the target object, and performs image recognition processing on the multiple images in the first image set to identify the point cloud and contour frame of the target object, ensuring the accuracy of the identified point cloud and contour frame, and thus ensuring the accuracy of the robot path planning.

[0050] In the above technical solution, after identifying the first image set and determining the point cloud and contour frame of the target object, it further includes: in the case where the recognition effect of the point cloud and / or contour frame does not meet the preset conditions, obtaining a second image set, the acquisition time of the second image set is earlier than the acquisition time of the first image set, and each second image in the second image set includes the target object; identifying the second image set and determining the point cloud and contour frame.

[0051] In this technical solution, after identifying the point cloud and contour frame of the target object according to the first image set, when the recognition effect of the point cloud and / or contour frame does not reach the preset conditions, a second image set can be obtained to re-identify the contour frame and point cloud of the target object. Among them, the preset conditions include the border size standard of the contour frame, the quantity and position information of the point cloud are accurate. For example, the size of the contour frame is consistent with the size of the target object, and the position information of the point cloud is consistent with the actual coordinate information of the target object.

[0052] It should be noted that the second image set is a set of historical images saved separately when the first image set is acquired, and the second image set will be temporarily stored for a period of time for subsequent calls. During the acquisition of the first image set, situations such as camera occlusion or robot movement jolts may occur, resulting in poor image quality in the first image set, and further resulting in the situation where the identified contour frame and point cloud of the target object do not meet the preset conditions. In response to the occurrence of such situations, the second image set can be temporarily stored to ensure that the identified contour frame and point cloud can meet the preset conditions.

[0053] When the recognition result of the first image set of the control method of the robot in this technical solution is not good, by calling the temporarily stored second image set, when the second image set can be used, there is no need for secondary acquisition. While ensuring the recognition effect, the recognition efficiency of the contour frame and point cloud of the target object is also improved.

[0054] In the above technical solution, after recognizing the first image set and determining the point cloud and contour frame of the target object, it further includes: when the recognition effect of the point cloud and / or contour frame does not meet the preset conditions, determining the first position of the robot according to the first image set; controlling the robot to drive to the first position; during the driving process of the robot, acquiring a third image set, and each third image in the third image set includes the target object; recognizing the third image set to determine the point cloud and contour frame.

[0055] In this technical solution, when the contour frame and point cloud of the target object recognized according to the acquired first image set do not meet the preset conditions, the robot can be controlled to retreat to the first position, and the robot is re-controlled to move forward and re-acquire images, and the recognition process of the contour frame and point cloud is carried out again.

[0056] Specifically, record the first position of the robot in advance, and the first position is the position information when the first image set is acquired. When it is determined that the contour frame and point cloud of the recognized target object do not meet the preset conditions, control the robot to stop moving forward and return to the first position. During the return process, start acquiring images again at the same time to form a third image set. After the third image set is acquired, perform recognition processing on each image in the third image set, recognize the contour frame and point cloud of the target object in each image, and then perform a smoothing operation on the recognized multiple contour frames and multiple point clouds to determine the contour frame and point cloud of the target object.

[0057] In the control method of the robot in this technical solution, when the contour frame and point cloud of the recognized target object do not meet the preset conditions, by controlling the robot to retreat to the first position, acquiring the third image set at the same time, performing image recognition processing on the images in the third image set, and recognizing the contour frame and point cloud of the target object, it is ensured that the recognition effect of the point cloud and / or contour frame of the target object meets the preset conditions, thereby improving the accuracy of the robot's driving path.

[0058] In the above technical solution, determining the first position of the robot according to the first image set includes: determining the relative position information between the robot and the target object according to the first image set; determining the first position according to the relative position information.

[0059] In this technical solution, based on the first image set, the relative position relationship between the robot and the target object is identified, and relative position information is obtained, where the relative position information includes the distance information and the angle information between the robot and the target object. Then, according to the relative position information, the first position of the robot is determined.

[0060] Specifically, according to multiple images in the first image set, it can be judged the shooting angle when the camera on the robot shoots the target object when collecting the first image set, and the relative distance between the camera and the target object can also be judged. Based on the above parameters such as the shooting angle and the relative distance, with the target object as the reference, the first position of the robot can be determined.

[0061] In the control method of the robot in this technical solution, when the contour frame and the point cloud of the identified target object do not meet the preset conditions, according to multiple images in the first image set, the first position is determined, and the robot is controlled to return to the first position, and images are collected again, and the contour frame and the point cloud of the target object are re-identified, ensuring the recognition effect of the point cloud and / or the contour frame of the target object.

[0062] In the above technical solution, the first driving speed of the robot when driving to the first position is less than the second driving speed of the robot when driving along the driving path.

[0063] In this technical solution, the robot is controlled to drive forward at the second driving speed. When the contour frame and the point cloud of the identified target object do not meet the preset conditions, the robot is controlled to return to the first position, and during the process of returning to the first position, the robot is controlled to drive at the first driving speed. Among them, the first driving speed is less than the second driving speed.

[0064] The control method of the robot in this technical solution ensures that during the process of re-collecting images, the driving process of the robot is more stable by limiting the driving speed of the robot when returning to the first position, ensuring the image quality of the collected images, and further ensuring the recognition effect of the point cloud and / or the contour frame.

[0065] According to the second aspect of the present invention, a control device for a robot is proposed. The control device for the robot includes: an acquisition module for acquiring the point cloud and the contour frame of the target object; a selection module for selecting one or a combination of the point cloud and the contour frame as path planning information; a determination module for determining the driving path of the robot according to the path planning information; and a control module for controlling the robot to drive according to the driving path.

[0066] In this technical solution, during the driving process of the robot, the situation may occur that the robot's progress is blocked by a target object. In response to the above situation, the robot is controlled to perform an avoidance action, and the acquisition module respectively acquires the point cloud and the bounding box of the target object. During the process of planning the driving path, the selection module can separately use the point cloud or the bounding box as the path planning information, or use both the point cloud and the bounding box as the path planning information. Based on the determined path planning information, the determination module plans the driving path of the robot, and then the control module controls the robot to drive according to the planned driving path.

[0067] It should be noted that the target objects include obstacles of different types, shapes, and poses. For example, low objects (carpets, thresholds), irregular objects (wires, clothes), and specific classified obstacles (pet feces, liquid stains).

[0068] The robot can be selected as a sweeping robot, and image acquisition devices such as cameras are configured in the sweeping robot.

[0069] The bounding box of the target object includes the smallest rectangle of the complete contour of the target object. The bounding box can be obtained by means of visual recognition. Specifically, an image including the target object is captured by an image acquisition device such as a camera, and the bounding box of the target object is obtained by identifying the image.

[0070] The point cloud of the target object includes discrete point data of the specific contour of the target object. The point cloud of the target object is distributed inside the bounding box, and both the bounding box and the point cloud can be used as path planning information.

[0071] It can be understood that for the same target object, the point cloud of the target object is located inside the bounding box.

[0072] In the case where there is a large amount of point cloud inside the bounding box, it is determined that the recognition effect difference between using the point cloud as the path planning information and using the bounding box as the path planning information is small at this time. To reduce the calculation amount, the bounding box can be used as the path planning information.

[0073] In the case where the point cloud is concentrated in a certain area of the bounding box, that is, when the number of point clouds inside the bounding box is small, it is determined that the path planning accuracy of using only the bounding box as the path planning information is low at this time. Therefore, it can be selected to use the point cloud as the path planning information, or use both the point cloud and the bounding box as the path planning information.

[0074] In the related technology, the obstacle avoidance strategies of sweeping robots are all based on a single piece of data to plan the driving path. For example, using only the bounding box of the obstacle as the basis for obstacle avoidance planning. During the process of identifying and avoiding a long and thin obstacle (wire), since the true contour of the obstacle is only the diagonal of the bounding box, it affects the cleaning coverage rate of the sweeping robot.

[0075] In the control device of the robot in this technical solution, the acquisition module acquires the point cloud and contour box of the target object. The selection module can select one or a combination of the point cloud and contour box as the path planning information according to actual requirements. The determination module can thus quickly and efficiently plan the driving path of the robot, and the control module controls the robot to drive according to the planned driving path, further improving the obstacle avoidance effect during the driving of the robot.

[0076] According to the third aspect of the present invention, a control device for a robot is proposed, which includes a processor and a memory. A program or instruction is stored in the memory, and when the program or instruction is executed by the processor, the steps of the control method of the robot in any of the above technical solutions are implemented. Therefore, this control device has all the beneficial effects of the control method of the robot in any of the above technical solutions, and will not be elaborated here.

[0077] According to the fourth aspect of the present invention, a readable storage medium is proposed, on which a program or instruction is stored, and when the program or instruction is executed by the processor, the control method of the robot in any of the above technical solutions is implemented. Therefore, this readable storage medium has all the beneficial effects of the control method of the robot in any of the above technical solutions, and will not be elaborated here.

[0078] According to the fifth aspect of the present invention, a robot is proposed, which includes: the control device of the robot defined in the second aspect above, or the control device of the robot defined in the third aspect above, or the readable storage medium defined in the fourth aspect above. Therefore, it has all the beneficial technical effects of the control device of the robot defined in the second aspect above, or the control device of the robot defined in the third aspect above, or the readable storage medium defined in the fourth aspect above, and will not be elaborated here too much.

[0079] The additional aspects and advantages of the present invention will become obvious in the following description part, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, where:

[0081] Figure 1 FIG. 1 shows one of the flow diagrams of the control method of the robot in the first embodiment of the present invention;

[0082] Figure 2 FIG. 2 shows another flow diagram of the control method of the robot in the first embodiment of the present invention;

[0083] Figure 3 FIG. 3 shows yet another flow diagram of the control method of the robot in the first embodiment of the present invention;

[0084] Figure 4 Shows the fourth schematic flowchart of the control method of the robot in the first embodiment of the present invention;

[0085] Figure 5 Shows the fifth schematic flowchart of the control method of the robot in the first embodiment of the present invention;

[0086] Figure 6 Shows the sixth schematic flowchart of the control method of the robot in the first embodiment of the present invention;

[0087] Figure 7 Shows the seventh schematic flowchart of the control method of the robot in the first embodiment of the present invention;

[0088] Figure 8 Shows the eighth schematic flowchart of the control method of the robot in the first embodiment of the present invention;

[0089] Figure 9 Shows the first effect diagram of the control method of the robot in the first embodiment of the present invention;

[0090] Figure 10 Shows the second effect diagram of the control method of the robot in the first embodiment of the present invention;

[0091] Figure 11 Shows the third effect diagram of the control method of the robot in the first embodiment of the present invention;

[0092] Figure 12 Shows the structural block diagram of the control device of the robot in the second embodiment of the present invention;

[0093] Figure 13 Shows the structural block diagram of the control device of the robot in the third embodiment of the present invention. Detailed implementation manners

[0094] In order to be able to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

[0095] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the limitations of the specific embodiments disclosed below.

[0096] The following combines Figures 1 to 13, the control method, device, readable storage medium and robot of the robot provided by the embodiments of the present application are described in detail through specific embodiments and their application scenarios.

[0097] Embodiment 1:

[0098] As Figure 1 shown, in the first embodiment of the present invention, a control method for a robot is provided. The control method of the robot includes:

[0099] Step 102, obtain the point cloud and contour box of the target object;

[0100] Step 104, select one or a combination of the point cloud and the contour box as path planning information;

[0101] Step 106, determine the driving path of the robot according to the path planning information;

[0102] Step 108, control the robot to drive according to the driving path.

[0103] In this embodiment, during the driving process of the robot, there will be a situation where the target object hinders the robot's progress. In response to the above situation, the robot is controlled to perform an avoidance action, and the point cloud and contour box of the target object are obtained respectively. During the process of planning the driving path, the point cloud or the contour box can be used alone as the path planning information, or both the point cloud and the contour box can be used as the path planning information. According to the determined path planning information, the driving path of the robot is planned, and then the robot is controlled to drive according to the planned driving path.

[0104] It should be noted that the target object includes obstacles of different types, shapes and poses. For example, low objects (carpets, thresholds), irregular objects (wires, clothes), and specific classified obstacles (pet feces, liquid stains).

[0105] The robot can be a sweeping robot, and image acquisition devices such as cameras are configured in the sweeping robot.

[0106] The contour box of the target object includes the smallest rectangle of the complete contour of the target object. The contour box can be obtained by means of visual recognition. Specifically, an image including the target object is captured by an image acquisition device such as a camera, and the contour box of the target object is obtained by recognizing the image.

[0107] The point cloud of the target object includes discrete point data of the specific contour of the target object. The point cloud of the target object is distributed inside the contour box, and both the contour box and the point cloud can be used as path planning information.

[0108] It can be understood that for the same target object, the point cloud of the target object is located inside the contour box.

[0109] In the case where there are a large number of point clouds inside the contour box, it is determined that the difference in recognition effects between using the point cloud as path planning information and using the contour box as path planning information is small at this time. To reduce the computational load, the contour box can be used as path planning information.

[0110] In the case where the point clouds in the point cloud set are distributed in a certain area of the contour box, that is, when the number of point clouds inside the contour box is small, it is determined that the accuracy of path planning by only using the contour box as path planning information is low at this time. Therefore, it can be selected to use the point cloud as path planning information, or use both the point cloud and the contour box as path planning information.

[0111] In some embodiments, the target object is a building block toy on the ground. Due to the regular shape of the building block toy, when it is recognized that the point clouds of the building block toy are scattered inside the contour box and many point clouds are close to the edge position of the contour box, the obtained contour box is used as path planning information to plan the driving path of the robot, and then the robot is controlled to drive along the planned driving path to avoid the target object.

[0112] In some other embodiments, the target object is the power cord of a household appliance. The contour box of the power cord is large, but the point clouds are concentrated in a part of the contour box. At this time, if the contour box is used as path planning information to plan the driving path, the driving path will deviate greatly from the actual drivable path. Therefore, using only the point cloud as path planning information, or using both the point cloud and the contour box as path planning information can improve the accuracy of path planning. Specifically, when using both the point cloud and the contour box as path planning information, the contour box is cropped by the power supply, and then path planning is performed according to the cropped contour box.

[0113] In the related art, the obstacle avoidance strategies of sweeping robots are all based on a single piece of data to plan the driving path. For example, using only the contour box of the obstacle as the basis for obstacle avoidance planning. During the process of identifying and avoiding a slender obstacle (wire), since the true contour of the obstacle is only the diagonal of the contour box, it affects the cleaning coverage rate of the sweeping robot.

[0114] The control method of the robot in this embodiment obtains the point cloud and contour box of the target object, and can select one or a combination of the point cloud and the contour box as path planning information according to actual needs, so as to quickly and efficiently plan the driving path of the robot, control the robot to drive along the planned driving path, and further improve the obstacle avoidance effect during the driving process of the robot.

[0115] As Figure 2 shown, in any of the above embodiments, selecting one or a combination of the point cloud and the contour box as path planning information includes:

[0116] Step 202, determine the area ratio of the point cloud in the contour box;

[0117] Step 204, select one or a combination of the point cloud and the contour box as path planning information according to the area ratio.

[0118] In this embodiment, the data of the target object is processed by calculation, and the area data of the point cloud and the contour box are respectively processed by calculation. The area of the contour box and the distribution contour area of the point cloud are calculated, and further the area ratio of the distribution contour area of the point cloud in the area of the contour box is calculated. According to the value of the area ratio, one piece of data of the point cloud and the contour box or a data combination of the point cloud and the contour box is used as the path planning information for planning the path.

[0119] In the process of determining the area ratio of the point cloud in the contour box, it is necessary to obtain the point cloud area and the contour box area. Among them, the obtained image is divided into several grids with the same area, and the area of a single grid is set as the unit area. The point cloud area is the area corresponding to the number of grids occupied by the point cloud, and the contour box area is the area corresponding to the number of grids included in the contour box. The area ratio of the point cloud in the contour box is calculated through the point cloud area and the contour box area.

[0120] Specifically, in the case of a relatively large area ratio, it is determined that the contour box of the target object can reflect the true contour of the target object, and the contour box is used as the path planning information. In the case of a relatively small area ratio, it is determined that the contour box of the recognized target object cannot reflect the true contour of the target object, and the point cloud alone is used as the path planning information, or the point cloud and the contour box are used as the path planning information.

[0121] In some embodiments, the target object is the power cord of a household appliance. An image including the target object is obtained and divided into several grids with the same area. Among them, the point cloud of the target object occupies 9 grids in the image, and then the point cloud area is calculated to be 9 unit areas. The contour box of the target object contains 20 grids, and the contour box area is calculated to be 20 unit areas. The area ratio of the target object is calculated according to the point cloud area and the contour box area. According to the ratio of the area ratio, the point cloud of the target object is used as the path planning information for planning the path.

[0122] The control method of the robot in this embodiment can determine whether the contour box obtained by image recognition can reflect the true contour of the target object based on the area ratio of the coverage area of the point cloud in the area of the contour box, and accordingly select one or a combination of the point cloud and the contour box as the path planning information. Further, the accuracy of path planning is ensured. In the case where the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect, but also further improve the cleaning effect of the sweeping robot.

[0123] In any of the above embodiments, according to the area ratio, one or a combination of the point cloud and the contour box is selected as the path planning information, including:

[0124] When the area ratio is greater than or equal to the first preset ratio, the contour box is used as the path planning information;

[0125] When the area ratio is less than or equal to the second preset ratio, the point cloud is used as the path planning information;

[0126] When the area ratio is greater than the second preset ratio and less than the first preset ratio, the point cloud and the contour box are used as the path planning information;

[0127] Wherein, the first preset ratio is greater than the second preset ratio.

[0128] In this embodiment, the area ratio of the coverage area of the point cloud of the target object in the area of the contour box is calculated, the first preset ratio and the second preset ratio are obtained, and the ratio of the first preset ratio is greater than the ratio of the second preset ratio. The calculated area ratio is compared with the first preset ratio and the second preset ratio respectively, and according to the comparison results, the path planning information is determined from the point cloud and the contour box.

[0129] Specifically, the quantitative relationship between the area ratio and the first preset ratio is judged. When it is detected that the area ratio is greater than or equal to the first preset ratio, it is considered that there is a large amount of point cloud in the contour box, and it is determined that the contour box obtained by image recognition is close to the true contour of the target object. At this time, the contour box is used as the path planning information.

[0130] When the area ratio is less than the first preset ratio, the quantitative relationship between the area ratio and the second preset ratio is judged again. When the area ratio is greater than the second preset ratio, it can be determined that there is a certain amount of point cloud in the contour box, and the contour box is cropped by the point cloud of the target object. The cropped contour box is close to the true contour of the target object. At this time, the contour box cropped according to the point cloud is used as the path planning information.

[0131] Under the condition that the comparison result shows that the area ratio is less than or equal to the second preset ratio, it is determined that there is less point cloud in the contour box, and it is determined that the point cloud obtained by image recognition is close to the true contour of the target object. At this time, the point cloud is used as the path planning information.

[0132] In some embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is the power cord of a household appliance, the area of the contour frame of the target object is calculated to be 80 unit areas, the area of the distribution contour of the point cloud is calculated to be 20 unit areas, and the area proportion of the distribution contour of the point cloud in the contour frame area is calculated to be 25%. Since the area proportion is less than the first preset proportion, it is determined that there is less point cloud in the contour frame. It is determined that the point cloud obtained through image recognition is close to the true contour of the target object. At this time, the point cloud of the target object is used as the path planning information for path planning.

[0133] In some other embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is a toy block on the ground, the area of the contour frame of the target object is calculated to be 20 unit areas, the area of the distribution contour of the point cloud is calculated to be 18 unit areas, and the area proportion of the distribution contour of the point cloud in the contour frame area is calculated to be 90%. It is determined that there is a large amount of point cloud in the contour frame. It is determined that the contour frame obtained through image recognition is close to the true contour of the target object. At this time, the contour frame of the target object is used as the path planning information for path planning.

[0134] In some other embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is a threshold, the area of the contour frame of the target object is calculated to be 80 unit areas, the area of the distribution contour of the point cloud is calculated to be 40 unit areas, and the area proportion of the distribution contour of the point cloud in the contour frame area is calculated to be 50%. Since the area proportion is greater than the first preset proportion and at the same time less than the second preset proportion, it can be determined that there is a certain amount of point cloud in the contour frame. The contour frame is cropped by the point cloud of the target object. The cropped contour frame is close to the true contour of the target object. At this time, the contour frame cropped according to the point cloud is used as the path planning information.

[0135] The control method of the robot in this embodiment sets the first preset proportion and the second preset proportion, compares the calculated area proportion with the first preset proportion and the second preset proportion respectively, determines whether the contour frame obtained through image recognition can reflect the true contour of the target object, and accordingly selects one or a combination of the point cloud and the contour frame as the path planning information. This further ensures the accuracy of path planning. When the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect, but also further improve the cleaning effect of the sweeping robot.

[0136] In any of the above embodiments, the value range of the first preset proportion is 15% to 35%; and / or the value range of the second preset proportion is 65% to 85%.

[0137] In this embodiment, the value ranges of the first preset ratio and the second preset ratio are set respectively. Specifically, the first preset ratio is set, and the value range is set to be from 15% to 35%; the second preset ratio is set, and the value range is set to be from 65% to 85%.

[0138] In some embodiments, the first preset ratio is set to 30%, and the second preset ratio is set to 70%.

[0139] In the control method of the robot in this embodiment, by setting the first preset ratio and the second preset ratio, the size relationship between the first preset ratio and the second preset ratio is limited, ensuring the accuracy of the step of selecting the robot path planning information.

[0140] As Figure 3 shown, in any of the above embodiments, the path planning information includes point cloud or bounding box. According to the path planning information, the driving path of the robot is determined, including:

[0141] Step 302, determining the first boundary information of the point cloud or the bounding box;

[0142] Step 304, determining the driving path according to the first boundary information.

[0143] In this embodiment, when the path planning information only includes a single point cloud or a single bounding box, the driving path is planned only according to the first boundary information of the corresponding point cloud or the first boundary information of the corresponding bounding box.

[0144] When the path planning information includes the point cloud of the target object, first determine the first boundary information of the point cloud of the target object, where the first boundary information includes the distribution contour edge information of the point cloud. When the path planning information includes the bounding box of the target object, first determine the first boundary information of the bounding box of the target object, where the first boundary information includes the boundary information of the bounding box. After determining the first boundary information, plan the driving path of the robot according to the first boundary information.

[0145] Specifically, for different target objects, in the case where there is a small amount of point cloud in the bounding box, the point cloud data distributed at the edge in the point cloud of the target object can outline the edge contour of the target object, and the coordinate data of the edge contour outlined by the point cloud can be set as the first boundary information. In the case where there is a large amount of point cloud in the bounding box, the border of the bounding box of the target object can be regarded as the edge contour of the target object, and the coordinate data of the border of the bounding box can be set as the first boundary information.

[0146] As Figure 9As shown, when the area ratio of the point cloud 904 in the contour box 902 is greater than the first preset ratio, it is determined that a large amount of the point cloud 904 is contained within the contour box 902 of the target object. The border of the contour box 902 of the target object can be determined as the edge contour of the target object. The coordinate data of the border of the contour box 902 can be set as the first boundary information, and the driving path is planned based on the first boundary information determined according to the contour box 902. The planned driving path is located at an external position of the border of the contour box 902. It can be seen that part of the planned path is outside the contour box 902. That is, when the area ratio of the point cloud 904 in the contour box 902 is relatively large, planning the driving path according to the first boundary information confirmed by the contour box 902 is more accurate.

[0147] As Figure 10 shown, when the area ratio of the point cloud 1004 in the contour box 1002 is less than the second preset ratio, it is determined that a small amount of the point cloud 1004 is contained within the contour box 1002 of the target object. The point cloud 1004 data distributed at the edge in the point cloud 1004 of the target object can outline the edge contour of the target object. The coordinate data of the edge contour outlined by the point cloud 1004 can be set as the first boundary information. The driving path is planned based on the first boundary information determined according to the point cloud 1004, and the planned driving path is located at the edge position of the point cloud 1004. It can be seen that part of the planned path is within the contour box 1002. That is, when the area ratio of the point cloud 1004 in the contour box 1002 is relatively small, planning the driving path according to the first boundary information of the point cloud 1004 is more accurate.

[0148] The control method of the robot in this embodiment plans the driving path of the robot through the first boundary information of the point cloud or the contour box of the target object, ensuring the accuracy of the step of planning the driving path of the robot.

[0149] As Figure 4 shown, in any of the above embodiments, the path planning information includes the point cloud and the contour box. According to the path planning information, the driving path of the robot is determined, including:

[0150] Step 402, determining the second boundary information of the point cloud;

[0151] Step 404, determining the target area in the contour box according to the second boundary information;

[0152] Step 406, determining the third boundary information of the target area;

[0153] Step 408, determining the driving path according to the third boundary information.

[0154] In this embodiment, when the path planning information only includes the point cloud and the contour box, the contour box is cropped according to the second boundary information of the point cloud, and the driving path is planned according to the third boundary information of the target area obtained by cropping.

[0155] Specifically, the robot is a sweeping robot. When the area ratio of the point cloud in the contour box is greater than the second preset ratio and less than the first preset ratio, if the driving path is planned only according to the point cloud, it is easy to cause obstacle avoidance failure. If the driving path is planned only according to the contour box, it is easy to affect the cleaning effect of the sweeping robot. Therefore, it is selected to crop the contour box according to the point cloud.

[0156] When the path planning information includes the point cloud and the contour box of the target object, first determine the second boundary information of the point cloud of the target object, and then crop the contour box according to the second boundary information.

[0157] It should be noted that the point cloud of the target object is distributed inside the contour box. When there is a certain number of point clouds in the contour box, the point cloud data distributed at the edge in the point cloud of the target object can outline part of the edge contour of the target object. Set the coordinate data of the edge contour outlined by the point cloud as the second boundary information, and then crop the contour box through the second boundary information. The border of the cropped contour box is closer to the edge contour of the target object. The coordinate data of the border of the contour box cropped according to the point cloud can be set as the third boundary information.

[0158] As Figure 11 shown, when the area ratio of the point cloud 1104 in the contour box 1102 is less than the first preset ratio and greater than the second preset ratio, it is determined that there is a certain number of point clouds 1104 in the contour box 1102 of the target object. The point cloud 1104 data distributed at the edge in the point cloud 1104 of the target object can outline part of the edge contour of the target object. Set the coordinate data of the edge contour outlined by the point cloud 1104 as the second boundary information, and then crop the contour box 1102 through the second boundary information. The border of the cropped contour box 1102 is closer to the edge contour of the target object. The coordinate data of the border of the cropped contour box 1102 can be set as the first boundary information. Plan the path according to the first boundary information jointly determined by the point cloud 1104 and the contour box 1102. Part of the planned driving path is located at the edge position of the point cloud 1104, and part is located outside the border of the contour box 1102. It can be seen that when there is a certain number of point clouds 1104 in the contour box 1102, planning the driving path according to the third boundary information jointly confirmed by the point cloud 1104 and the contour box 1102 is more accurate.

[0159] In the control method of the robot in this embodiment, the target area inside the contour box is determined through the second boundary information of the point cloud of the target object, and then the driving path of the robot is planned according to the third boundary information of the target area. When a certain number of point clouds are included in the contour box, the accuracy of the step of planning the driving path of the robot is ensured.

[0160] As Figure 5 shown, in any of the above embodiments, obtaining the point cloud and contour box of the target object includes:

[0161] Step 502, obtaining a first image set, where each first image in the first image set includes the target object;

[0162] Step 504, identifying the first image set to determine the point cloud and contour box.

[0163] In this embodiment, an image acquisition device is provided in the robot, which can acquire images including the target object during the driving process of the robot. To improve the accuracy of the point cloud and contour box of the target object obtained, in the step of identifying the point cloud and contour box of the target object, multiple images including the target object need to be acquired. The set of these images is the first image set. Among them, each image in the first image set includes the target object. By performing image processing on the first image set, the point cloud and contour box of the target object can be accurately identified.

[0164] Specifically, during the driving process of the robot, the image acquisition device starts to acquire images and simultaneously identifies the acquired images. When it is detected that the image includes the target object, the image is used as an image in the first image set. When it is detected that the image does not include the target object, the image is filtered out. After the first image set is acquired, each image in the first image set is identified and processed to identify the contour box and point cloud of the target object in each image, and then a smoothing operation is performed on the identified multiple contour boxes and multiple point clouds to determine the contour box and point cloud of the target object.

[0165] The control method of the robot in this embodiment ensures the accuracy of identifying the point cloud and contour box by obtaining a first image set including multiple images of the shooting object including the target object and performing image recognition processing on the multiple images in the first image set, and further ensures the accuracy of the robot path planning.

[0166] As Figure 6 shown, in any of the above embodiments, after identifying the first image set and determining the point cloud and contour box of the target object, it further includes:

[0167] Step 602: When the recognition effect of the point cloud and / or the contour box does not meet the preset conditions, obtain a second image set. The acquisition time of the second image set is earlier than that of the first image set, and each second image in the second image set includes the target object.

[0168] Step 604: Recognize the second image set and determine the point cloud and the contour box.

[0169] In this embodiment, after recognizing the point cloud and the contour box of the target object according to the first image set, if the recognition effect of the point cloud and / or the contour box does not reach the preset conditions, a second image set can be obtained to re-recognize the contour box and the point cloud of the target object. The preset conditions include the border size standard of the contour box and the accurate quantity and position information of the point cloud. For example, the size of the contour box is consistent with the size of the target object, and the position information of the point cloud is consistent with the actual coordinate information of the target object.

[0170] It should be noted that the second image set is another saved historical image set when collecting the first image set. The second image set will be temporarily stored for a period of time for subsequent calls. During the process of collecting the first image set, situations such as camera occlusion or robot driving jolts may occur, resulting in poor image quality in the first image set, and further resulting in the situation where the recognized contour box and point cloud of the target object do not meet the preset conditions. In response to such situations, the second image set can be temporarily stored to ensure that the recognized contour box and point cloud can meet the preset conditions.

[0171] The control method of the robot in this embodiment, when the recognition result of the first image set is not good, by calling the temporarily stored second image set, without secondary acquisition when the second image set can be used, not only ensures the recognition effect, but also improves the recognition efficiency of the contour box and the point cloud of the target object.

[0172] As Figure 7 shown, in any of the above embodiments, after recognizing the first image set and determining the point cloud and the contour box of the target object, it further includes:

[0173] Step 702: When the recognition effect of the point cloud and / or the contour box does not meet the preset conditions, determine the first position of the robot according to the first image set.

[0174] Step 704: Control the robot to drive to the first position.

[0175] Step 706: During the process of the robot driving, obtain a third image set. Each third image in the third image set includes the target object.

[0176] Step 708: Recognize the third image set and determine the point cloud and the contour box.

[0177] In this embodiment, when the contour box and point cloud of the identified target object do not meet the preset conditions according to the acquired first image set, the robot can be controlled to retreat to the first position, and then the robot is re-controlled to move forward and re-acquire images, and the processes of identifying the contour box and point cloud are carried out again.

[0178] Specifically, record the first position of the robot in advance. The first position is the position information when the first image set is acquired. When it is determined that the contour box and point cloud of the identified target object do not meet the preset conditions, the robot is controlled to stop moving forward and return to the first position. During the return process, image acquisition is restarted simultaneously to form a third image set. After the third image set is acquired, each image in the third image set is subjected to recognition processing to identify the contour box and point cloud of the target object in each image, and then smoothing operations are performed on the identified multiple contour boxes and multiple point clouds to determine the contour box and point cloud of the target object.

[0179] In the control method of the robot in this embodiment, when the point cloud and / or contour box of the identified target object do not meet the preset conditions, by controlling the robot to retreat to the first position and simultaneously acquiring the third image set, image recognition processing is performed on the images in the third image set to identify the contour box and point cloud of the target object, ensuring that the recognition effect of the point cloud and / or contour box of the target object meets the preset conditions, thereby improving the accuracy of the robot's driving path.

[0180] As Figure 8 shown, in any of the above embodiments, determining the first position of the robot according to the first image set includes:

[0181] Step 802, determining the relative position information between the robot and the target object according to the first image set;

[0182] Step 804, determining the first position according to the relative position information.

[0183] In this embodiment, based on the first image set, the relative position relationship between the robot and the target object is identified, and the relative position information is obtained. The relative position information includes the distance information and angle information between the robot and the target object. Then, according to the relative position information, the first position of the robot is determined.

[0184] Specifically, according to multiple images in the first image set, it is possible to judge the shooting angle when the camera on the robot shoots the target object when the first image set is acquired, and it is also possible to judge the relative distance between the camera and the target object. Based on the above parameters such as the shooting angle and relative distance, with the target object as the reference, the first position of the robot can be determined.

[0185] In the control method of the robot in this embodiment, when the contour frame and point cloud of the recognized target object do not meet the preset conditions, based on multiple images in the first image set, the first position is determined, and the robot is controlled to retreat to the first position, re-collect images, and re-recognize the contour frame and point cloud of the target object, ensuring the recognition effect of the point cloud and / or contour frame of the target object.

[0186] In any of the above embodiments, the first traveling speed of the robot traveling to the first position is less than the second traveling speed of the robot traveling along the traveling path.

[0187] In this embodiment, the robot is controlled to travel forward at the second traveling speed. When the contour frame and point cloud of the recognized target object do not meet the preset conditions, the robot is controlled to return to the first position, and during the process of returning to the first position, the robot is controlled to travel at the first traveling speed. Among them, the first traveling speed is less than the second traveling speed.

[0188] The control method of the robot in this embodiment ensures that the traveling process of the robot is more stable during the process of re-collecting images by limiting the traveling speed of the robot returning to the first position, ensuring the image quality of the collected images, and further ensuring the recognition effect of the point cloud and / or contour frame.

[0189] Embodiment 2:

[0190] As Figure 12 shown, in the second embodiment of the present invention, a control device for a robot is provided. The control device 1200 of the robot includes:

[0191] An acquisition module 1202, configured to acquire the point cloud and contour frame of the target object;

[0192] A selection module 1204, configured to select one or a combination of the point cloud and the contour frame as path planning information;

[0193] A determination module 1206, configured to determine the traveling path of the robot according to the path planning information;

[0194] A control module 1208, configured to control the robot to travel according to the traveling path.

[0195] In this embodiment, during the movement of the robot, the situation may occur that the robot's progress is blocked by a target object. In response to the above situation, the robot is controlled to perform an avoidance action, and the acquisition module 1202 acquires the point cloud and the bounding box of the target object respectively. During the process of planning the driving path, the selection module 1204 can separately use the point cloud or the bounding box as the path planning information, or use both the point cloud and the bounding box as the path planning information. According to the determined path planning information, the determination module 1206 plans the driving path of the robot, and then the control module 1208 controls the robot to drive according to the planned driving path.

[0196] It should be noted that the target object includes obstacles of different types, shapes, and poses. For example, low objects (carpets, thresholds), irregular objects (wires, clothes), and specific classified obstacles (pet feces, liquid stains).

[0197] The robot can be selected as a sweeping robot, and image acquisition devices such as cameras are configured in the sweeping robot.

[0198] The bounding box of the target object includes the smallest rectangle of the complete contour of the target object. The bounding box can be obtained by means of visual recognition. Specifically, an image including the target object is captured by an image acquisition device such as a camera, and the bounding box of the target object is obtained by identifying the image.

[0199] The point cloud of the target object includes discrete point data of the specific contour of the target object. The point cloud of the target object is distributed inside the bounding box, and both the bounding box and the point cloud can be used as path planning information.

[0200] It can be understood that for the same target object, the point cloud of the target object is located inside the bounding box.

[0201] In the case where there is a large amount of point cloud inside the bounding box, it is determined that the recognition effect difference between using the point cloud as the path planning information and using the bounding box as the path planning information is small at this time. To reduce the computational amount, the bounding box can be used as the path planning information.

[0202] In the case where the point cloud is concentrated in a certain area of the bounding box, that is, when the number of point clouds inside the bounding box is small, it is determined that the path planning accuracy of using only the bounding box as the path planning information is low at this time. Therefore, it can be selected to use the point cloud as the path planning information, or use both the point cloud and the bounding box as the path planning information.

[0203] In some embodiments, the target object is a building block toy on the ground. Since the shape of the building block toy is regular, when it is recognized that the point cloud of the building block toy is scattered within the contour box and a large number of point clouds are close to the edge position of the contour box, the obtained contour box is used as path planning information to plan the driving path of the robot, and then the robot is controlled to drive along the planned driving path to avoid the target object.

[0204] In some other embodiments, the target object is the power cord of a household appliance. The contour box of the power cord is large, but the point cloud is concentrated in a part of the contour box. At this time, if the contour box is used as path planning information to plan the driving path, the driving path will deviate greatly from the actual drivable path. Therefore, using only the point cloud as path planning information, or using both the point cloud and the contour box as path planning information can improve the accuracy of path planning. Specifically, when using both the point cloud and the contour box as path planning information, the contour box is cropped by the power supply, and then path planning is performed according to the cropped contour box.

[0205] In the related art, the obstacle avoidance strategies of floor cleaning robots are all based on a single piece of data to plan the driving path. For example, only the contour box of the obstacle is used as the basis for obstacle avoidance planning. During the process of identifying and avoiding an obstacle (such as a wire), since the true contour of the obstacle is only the diagonal of the contour box, it affects the cleaning coverage rate of the floor cleaning robot.

[0206] The control device of the robot in this embodiment obtains the point cloud and contour box of the target object through the acquisition module 1202. The selection module 1204 can select one or a combination of the point cloud and the contour box as path planning information according to actual needs. The determination module 1206 can thus quickly and efficiently plan the driving path of the robot, and the control module 1208 controls the robot to drive along the planned driving path, further improving the obstacle avoidance effect during the driving process of the robot.

[0207] In any of the above embodiments, the control device 1200 of the robot includes:

[0208] A determination module 1206, configured to determine the area ratio of the point cloud in the contour box;

[0209] A selection module 1204, configured to select one or a combination of the point cloud and the contour box as path planning information according to the area ratio.

[0210] In this embodiment, the determination module 1206 performs arithmetic processing on the data of the target object, performs arithmetic processing on the area data of the point cloud and the contour box respectively, calculates the area of the contour box and the distribution contour area of the point cloud, and further calculates the area ratio of the distribution contour area of the point cloud in the area of the contour box. According to the value of the area ratio, the selection module 1204 selects one of the point cloud and the contour box or a combination of the data of the point cloud and the contour box as the path planning information for planning the path.

[0211] In the process of determining the area ratio of the point cloud in the contour box, it is necessary to obtain the area of the point cloud and the area of the contour box. Among them, the obtained image is divided into several grids with the same area, the area of a single grid is set as the unit area, the area of the point cloud is the area corresponding to the number of grids occupied by the point cloud, and the area of the contour box is the area corresponding to the number of grids included in the contour box. The area ratio of the point cloud in the contour box is calculated through the area of the point cloud and the area of the contour box.

[0212] Specifically, in the case of a relatively large area ratio, it is determined that the contour box of the target object can reflect the true contour of the target object, and the contour box is used as the path planning information. In the case of a relatively small area ratio, it is determined that the contour box of the recognized target object cannot reflect the true contour of the target object, and the point cloud alone is used as the path planning information, or the point cloud and the contour box are used as the path planning information.

[0213] In some embodiments, the target object is the power cord of a household appliance. An image including the target object is obtained and divided into several grids with the same area. Among them, the point cloud of the target object occupies 9 grids in the image, and then the area of the point cloud is calculated as 9 unit areas. The contour box of the target object contains 20 grids, and the area of the contour box is calculated as 20 unit areas. The area ratio of the target object is calculated according to the area of the point cloud and the area of the contour box. According to the ratio of the area ratio, the point cloud of the target object is used as the path planning information for planning the path.

[0214] The control device of the robot in this embodiment can determine whether the contour box obtained by image recognition can reflect the true contour of the target object according to the area ratio of the coverage area of the point cloud in the area of the contour box by the determination module 1206. The selection module 1204 accordingly selects one or a combination of the point cloud and the contour box as the path planning information. Further ensures the accuracy of path planning. When the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect, but also further improve the cleaning effect of the sweeping robot.

[0215] In any of the above embodiments, the control device 1200 of the robot includes:

[0216] A selection module 1204, configured to use the bounding box as path planning information when the area ratio is greater than or equal to a first preset ratio;

[0217] A selection module 1204, configured to use the point cloud as path planning information when the area ratio is less than or equal to a second preset ratio;

[0218] A selection module 1204, configured to use the point cloud and the bounding box as path planning information when the area ratio is greater than the second preset ratio and less than the first preset ratio;

[0219] Wherein, the first preset ratio is greater than the second preset ratio.

[0220] In this embodiment, the area ratio of the coverage area of the point cloud of the target object in the area of the bounding box is calculated, and the first preset ratio and the second preset ratio are obtained. The ratio of the first preset ratio is greater than the ratio of the second preset ratio. The calculated area ratio is compared with the first preset ratio and the second preset ratio respectively. The selection module 1204 determines the path planning information from the point cloud and the bounding box according to the comparison result.

[0221] Specifically, the quantitative relationship between the area ratio and the first preset ratio is judged. When it is detected that the area ratio is greater than or equal to the first preset ratio, it is considered that there is a large amount of point cloud in the bounding box, and it is determined that the bounding box obtained by image recognition is close to the true contour of the target object. At this time, the bounding box is used as the path planning information.

[0222] When the area ratio is less than the first preset ratio, the quantitative relationship between the area ratio and the second preset ratio is judged again. When the area ratio is greater than the second preset ratio, it can be determined that there is a certain amount of point cloud in the bounding box, and the bounding box is cropped by the point cloud of the target object. The cropped bounding box is close to the true contour of the target object. At this time, the bounding box cropped according to the point cloud is used as the path planning information.

[0223] Under the condition that the comparison result shows that the area ratio is less than or equal to the second preset ratio, it is determined that there is less point cloud in the bounding box, and it is determined that the point cloud obtained by image recognition is close to the true contour of the target object. At this time, the point cloud is used as the path planning information.

[0224] In some embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is the power cord of a household appliance, the area of the contour frame of the target object is calculated to be 80 unit areas, the area of the distribution contour of the point cloud is calculated to be 20 unit areas, and the area proportion of the distribution contour area of the point cloud in the contour frame area is calculated to be 25%. Since the area proportion is less than the first preset proportion, it is determined that there is less point cloud in the contour frame, and it is determined that the point cloud obtained by image recognition is close to the true contour of the target object. At this time, the point cloud of the target object is used as the path planning information for path planning.

[0225] In some other embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is a toy block on the ground, the area of the contour frame of the target object is calculated to be 20 unit areas, the area of the distribution contour of the point cloud is calculated to be 18 unit areas, and the area proportion of the distribution contour area of the point cloud in the contour frame area is calculated to be 90%. It is determined that there is a large amount of point cloud in the contour frame, and it is determined that the contour frame obtained by image recognition is close to the true contour of the target object. At this time, the contour frame of the target object is used as the path planning information for path planning.

[0226] In some other embodiments, the first preset proportion is set to 30%, the second preset proportion is set to 70%, the target object is a threshold, the area of the contour frame of the target object is calculated to be 80 unit areas, the area of the distribution contour of the point cloud is calculated to be 40 unit areas, and the area proportion of the distribution contour area of the point cloud in the contour frame area is calculated to be 50%. Since the area proportion is greater than the first preset proportion and at the same time less than the second preset proportion, it can be determined that there is a certain amount of point cloud in the contour frame. The contour frame is cropped by the point cloud of the target object, and the cropped contour frame is close to the true contour of the target object. At this time, the contour frame cropped according to the point cloud is used as the path planning information.

[0227] The control device of the robot in this embodiment sets the first preset proportion and the second preset proportion, compares the calculated area proportion with the first preset proportion and the second preset proportion respectively, determines whether the contour frame obtained by image recognition can reflect the true contour of the target object, and selects module 1204 and accordingly selects one or a combination of the point cloud and the contour frame as the path planning information. This further ensures the accuracy of path planning. In the case where the robot is a sweeping robot, it can not only ensure the obstacle avoidance effect, but also further improve the cleaning effect of the sweeping robot.

[0228] In any of the above embodiments, the value range of the first preset proportion is 15% to 35%; and / or the value range of the second preset proportion is 65% to 85%.

[0229] In this embodiment, the value ranges of the first preset ratio and the second preset ratio are respectively set. Specifically, the first preset ratio is set, and the value range is set to be from 15% to 35%; the second preset ratio is set, and the value range is set to be from 65% to 85%.

[0230] In some embodiments, the first preset ratio is set to 30%, and the second preset ratio is set to 70%.

[0231] In this embodiment, the control device of the robot sets the first preset ratio and the second preset ratio, limits the size relationship between the first preset ratio and the second preset ratio, and ensures the accuracy of the step of selecting the robot path planning information.

[0232] In any of the above embodiments, the control device of the robot includes:

[0233] A determination module 1206, configured to determine the first boundary information of the point cloud or the contour box;

[0234] A determination module 1206, configured to determine the driving path according to the first boundary information.

[0235] In this embodiment, when the path planning information only includes a single point cloud or a single contour box, the determination module 1206 plans the driving path only according to the first boundary information of the corresponding point cloud or the first boundary information of the corresponding contour box.

[0236] When the path planning information includes the point cloud of the target object, the determination module 1206 first determines the first boundary information of the point cloud of the target object, where the first boundary information includes the distribution contour edge information of the point cloud. When the path planning information includes the contour box of the target object, the determination module 1206 first determines the first boundary information of the contour box of the target object, where the first boundary information includes the boundary information of the contour box. After determining the first boundary information, the driving path of the robot is planned according to the first boundary information.

[0237] Specifically, for different target objects, in the case where there is a small amount of point cloud in the contour box, the point cloud data distributed at the edge in the point cloud of the target object can outline the edge contour of the target object, and the determination module 1206 can set the coordinate data of the edge contour outlined by the point cloud as the first boundary information. In the case where there is a large amount of point cloud in the contour box, the border of the contour box of the target object can be recognized as the edge contour of the target object, and the determination module 1206 can set the coordinate data of the border of the contour box as the first boundary information.

[0238] Such as Figure 9As shown, when the area ratio of the point cloud in the contour box is greater than the first preset ratio, it is determined that there is a large amount of point cloud within the contour box of the target object. The border of the contour box of the target object can be recognized as the edge contour of the target object. The coordinate data of the border of the contour box can be set as the first boundary information. Based on the first boundary information determined according to the contour box, a driving path is planned, and the planned driving path is located outside the border of the contour box. It can be seen that part of the planned path is outside the contour box. That is, when the area ratio of the point cloud in the contour box is relatively large, planning the driving path according to the first boundary information confirmed by the contour box is more accurate.

[0239] As Figure 10 shown, when the area ratio of the point cloud in the contour box is less than the second preset ratio, it is determined that there is a small amount of point cloud within the contour box of the target object. The point cloud data distributed at the edge in the point cloud of the target object can outline the edge contour of the target object. The coordinate data of the edge contour outlined by the point cloud can be set as the first boundary information. Based on the first boundary information determined according to the point cloud, a driving path is planned, and the planned driving path is located at the edge position of the point cloud. It can be seen that part of the planned path is within the contour box. That is, when the area ratio of the point cloud in the contour box is relatively small, planning the driving path according to the first boundary information of the point cloud is more accurate.

[0240] The control device of the robot in this embodiment determines the driving path of the robot through the first boundary information of the point cloud or the contour box of the target object, ensuring the accuracy of the step of planning the driving path of the robot.

[0241] In any of the above embodiments, the path planning information includes the point cloud and the contour box. According to the path planning information, determining the driving path of the robot includes:

[0242] Determination module 1206 is configured to determine the second boundary information of the point cloud;

[0243] Determination module 1206 is configured to determine the target area in the contour box according to the second boundary information;

[0244] Determination module 1206 is configured to determine the third boundary information of the target area;

[0245] Determination module 1206 is configured to determine the driving path according to the third boundary information.

[0246] In this embodiment, when the path planning information only includes the point cloud and the contour box, the contour box is cropped according to the second boundary information of the point cloud, and determination module 1206 plans the driving path according to the third boundary information of the cropped target area.

[0247] Specifically, the robot is a floor-sweeping robot. When the area ratio of the point cloud in the contour box is greater than the second preset ratio and less than the first preset ratio, if the driving path is planned only based on the point cloud, it is easy to cause obstacle avoidance failure. If the driving path is planned only based on the contour box, it is easy to affect the cleaning effect of the floor-sweeping robot. Therefore, it is selected to crop the contour box according to the point cloud.

[0248] When the path planning information includes the point cloud and the contour box of the target object, first determine the second boundary information of the point cloud of the target object, and then crop the contour box according to the second boundary information.

[0249] It should be noted that the point cloud of the target object is distributed inside the contour box. When there is a certain number of point clouds in the contour box, the point cloud data distributed at the edge in the point cloud of the target object can outline part of the edge contour of the target object. Set the coordinate data of the edge contour outlined by the point cloud as the second boundary information, and then crop the contour box through the second boundary information. The border of the cropped contour box is closer to the edge contour of the target object. The coordinate data of the border of the contour box cropped according to the point cloud can be set as the third boundary information.

[0250] As Figure 11 shown, when the area ratio of the point cloud in the contour box is less than the first preset ratio and greater than the second preset ratio, it is determined that there is a certain number of point clouds in the contour box of the target object. The point cloud data distributed at the edge in the point cloud of the target object can outline part of the edge contour of the target object. Set the coordinate data of the edge contour outlined by the point cloud as the second boundary information, and then crop the contour box through the second boundary information. The border of the cropped contour box is closer to the edge contour of the target object. The coordinate data of the border of the cropped contour box can be set as the first boundary information. Plan the driving path according to the first boundary information jointly determined by the point cloud and the contour box. Part of the planned driving path is located at the edge position of the point cloud, and part is located outside the border of the contour box. It can be seen that when there is a certain number of point clouds in the contour box, planning the driving path according to the third boundary information jointly confirmed by the point cloud and the contour box is more accurate.

[0251] The control device of the robot in this embodiment determines the target area inside the contour box through the second boundary information of the point cloud of the target object, and then plans the driving path of the robot according to the third boundary information of the target area, ensuring the accuracy of the step of planning the driving path of the robot when there is a certain number of point clouds in the contour box.

[0252] In any of the above embodiments, the control device 1200 of the robot includes:

[0253] An acquisition module 1202, configured to acquire a first image set, where each first image in the first image set includes a target object;

[0254] A determination module 1206, configured to identify the first image set and determine a point cloud and a bounding box.

[0255] In this embodiment, an image acquisition device is provided in the robot, which can acquire images including the target object during the driving process of the robot. To improve the accuracy of the point cloud and the bounding box of the acquired target object, in the step of identifying the point cloud and the bounding box of the target object, multiple images including the target object need to be acquired. The set of these images is the first image set. Each image in the first image set includes the target object. By performing image processing on the first image set, the point cloud and the bounding box of the target object can be accurately identified.

[0256] Specifically, during the driving process of the robot, the image acquisition device starts to acquire images and simultaneously performs recognition on the acquired images. When it is detected that the image includes the target object, the image is used as an image in the first image set. When it is detected that the image does not include the target object, the image is filtered out. After the first image set is acquired, each image in the first image set is subjected to recognition processing to identify the bounding box and the point cloud of the target object in each image, and then a smoothing operation is performed on the identified multiple bounding boxes and multiple point clouds to determine the bounding box and the point cloud of the target object.

[0257] The control device of the robot in this embodiment acquires, through the acquisition module 1202, a first image set including multiple images with the shooting object being the target object, and the determination module 1206 performs image recognition processing on the multiple images in the first image set to identify the point cloud and the bounding box of the target object, ensuring the accuracy of the identified point cloud and bounding box, and further ensuring the accuracy of the robot path planning.

[0258] In any of the above embodiments, the control device 1200 of the robot includes:

[0259] An acquisition module 1202, configured to acquire a second image set when the recognition effect of the point cloud and / or the bounding box does not meet the preset conditions. The acquisition time of the second image set is earlier than the acquisition time of the first image set, and each second image in the second image set includes the target object;

[0260] A determination module 1206, configured to identify the second image set and determine a point cloud and a bounding box.

[0261] In this embodiment, after identifying the point cloud and the contour box of the target object according to the first image set, if the recognition effect of the point cloud and / or the contour box fails to meet the preset conditions, the acquisition module 1202 can acquire a second image set and re-identify the contour box and the point cloud of the target object. The preset conditions include the border size standard of the contour box and the accurate quantity and position information of the point cloud. For example, the size of the contour box is consistent with the size of the target object, and the position information of the point cloud is consistent with the actual coordinate information of the target object.

[0262] It should be noted that the second image set is a collection of historical images saved separately when the first image set is acquired. The second image set will be temporarily stored for a period of time for subsequent calls. During the process of acquiring the first image set, situations such as camera occlusion or robot driving bumps may occur, resulting in poor image quality in the first image set, and further resulting in the situation where the recognized contour box and point cloud of the target object do not meet the preset conditions. In response to the occurrence of such situations, the second image set can be temporarily stored to ensure that the recognized contour box and point cloud can meet the preset conditions.

[0263] When the recognition result of the first image set by the control device of the robot in this embodiment is not good, the acquisition module 1202 calls the temporarily stored second image set. When the second image set can be used, there is no need for secondary acquisition, which not only ensures the recognition effect but also improves the recognition efficiency of the contour box and point cloud of the target object.

[0264] In any of the above embodiments, the control device 1200 of the robot includes:

[0265] A determination module 1206, configured to determine the first position of the robot according to the first image set when the recognition effect of the point cloud and / or the contour box does not meet the preset conditions;

[0266] A control module 1208, configured to control the robot to drive to the first position;

[0267] An acquisition module 1202, configured to acquire a third image set during the driving process of the robot, and each third image in the third image set includes the target object;

[0268] A determination module 1206, configured to recognize the third image set and determine the point cloud and the contour box.

[0269] In this embodiment, when the recognized contour box and point cloud of the target object do not meet the preset conditions according to the acquired first image set, the robot can be controlled to retreat to the first position, and then the robot is re-controlled to move forward and re-acquire images, and the process of re-identifying the contour box and the point cloud is carried out.

[0270] Specifically, record the first position of the robot in advance. The first position is the position information when the first image set is collected. In the case where it is determined that the contour box and point cloud of the target object do not meet the preset conditions, control the robot to stop moving forward and return to the first position. During the return process, start collecting images again at the same time to form a third image set. After the third image set is collected, perform recognition processing on each image in the third image set, recognize the contour box and point cloud of the target object in each image, and then perform a smoothing operation on the recognized multiple contour boxes and multiple point clouds to determine the contour box and point cloud of the target object.

[0271] In the control device of the robot in this embodiment, when the contour box and point cloud of the recognized target object do not meet the preset conditions, by controlling the robot to return to the first position, collecting the third image set at the same time, and performing image recognition processing on the images in the third image set to recognize the contour box and point cloud of the target object, it is ensured that the recognition effect of the point cloud and / or contour box of the target object reaches the preset conditions, thereby improving the accuracy of the robot's driving path.

[0272] In any of the above embodiments, the control device 1200 of the robot includes:

[0273] A determination module 1206, configured to determine the relative position information between the robot and the target object according to the first image set;

[0274] The determination module 1206 is configured to determine the first position according to the relative position information.

[0275] In this embodiment, the determination module 1206 identifies the relative position relationship between the robot and the target object based on the first image set, and obtains the relative position information, where the relative position information includes the distance information and angle information between the robot and the target object. Then, according to the relative position information, the first position of the robot is determined.

[0276] Specifically, according to multiple images in the first image set, it is possible to judge the shooting angle when the camera on the robot shoots the target object when the first image set is collected, and it is also possible to judge the relative distance between the camera and the target object. Based on the above parameters such as the shooting angle and relative distance, with the target object as the reference, the first position of the robot can be determined.

[0277] In the control device of the robot in this embodiment, when the contour box and point cloud of the recognized target object do not meet the preset conditions, the determination module 1206 determines the first position according to multiple images in the first image set, controls the robot to return to the first position, re-collects images, and re-recognizes the contour box and point cloud of the target object, ensuring the recognition effect of the point cloud and / or contour box of the target object.

[0278] In any of the above embodiments, the first traveling speed of the robot when traveling to the first position is less than the second traveling speed of the robot when traveling along the traveling path.

[0279] In this embodiment, the robot is controlled to travel forward at the second traveling speed. When the contour box and point cloud of the recognized target object do not meet the preset conditions, the robot is controlled to return to the first position, and during the process of returning to the first position, the robot is controlled to travel at the first traveling speed. Among them, the first traveling speed is less than the second traveling speed.

[0280] The control device of the robot in this embodiment ensures a more stable traveling process of the robot during the process of re-acquiring images by limiting the traveling speed of the robot when returning to the first position, thereby ensuring the image quality of the acquired images and further ensuring the recognition effect of the point cloud and / or the contour box.

[0281] Embodiment Three:

[0282] As Figure 13 shown, in the third embodiment of the present invention, a control device for a robot is provided. The control device 1300 of the robot includes a processor 1302 and a memory 1304. A program or instruction is stored in the memory 1304, and when the program or instruction is executed by the processor 1302, the steps of the control method for the robot in any of the above technical solutions are implemented. Therefore, the robot has all the beneficial effects of the control method for the robot in any of the above technical solutions, which will not be elaborated here.

[0283] Embodiment Four:

[0284] In the fourth embodiment of the present invention, a readable storage medium is provided, on which a program is stored. When the program is executed by a processor, the control method for the robot in any of the above embodiments is implemented, and thus has all the beneficial technical effects of the control method for the robot in any of the above embodiments.

[0285] Among them, the readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.

[0286] Embodiment Five:

[0287] In the fifth embodiment of the present invention, a robot is provided, including: the control device for the robot in any of the above embodiments, and / or the readable storage medium in any of the above embodiments, and thus has all the beneficial technical effects of the control device for the robot in any of the above embodiments, and / or the readable storage medium in any of the above embodiments, which will not be elaborated here too much.

[0288] It should be clear that in the claims, the specification and the drawings of the present invention, the term "a plurality of" means two or more, unless otherwise clearly defined. The orientation or positional relationship indicated by terms such as "upper", "lower", etc. is based on the orientation or positional relationship shown in the drawings. This is only for more convenient description of the present invention and to make the description process simpler, rather than to indicate or imply that the device or element referred to must have the specific orientation described, be constructed and operated in a specific orientation. Therefore, these descriptions should not be construed as limitations on the present invention; terms such as "connection", "installation", "fixation", etc. should all be understood in a broad sense. For example, "connection" can be a fixed connection between multiple objects, a detachable connection between multiple objects, or an integral connection; it can be a direct connection between multiple objects, or an indirect connection between multiple objects through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances of the above data.

[0289] In the claims, the specification and the drawings of the present invention, the description of terms such as "an embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In the claims, the specification and the drawings of the present invention, the schematic representation of the above terms does not necessarily refer to the same embodiment or instance. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0290] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A control method for a robot, characterized in that, Including: Obtain the point cloud and bounding box of the target object; Select one or a combination of the point cloud and the bounding box as path planning information; Determine the driving path of the robot according to the path planning information; Control the robot to drive according to the driving path; The step of selecting one or a combination of the point cloud and the bounding box as path planning information includes: Determine the area ratio of the point cloud in the bounding box; Select one or a combination of the point cloud and the bounding box as path planning information according to the area ratio; The step of selecting one or a combination of the point cloud and the bounding box as path planning information according to the area ratio includes: When the area ratio is greater than or equal to a first preset ratio, use the bounding box as the path planning information; When the area ratio is less than or equal to a second preset ratio, use the point cloud as the path planning information; When the area ratio is greater than the second preset ratio and less than the first preset ratio, use the point cloud and the bounding box as path planning information; Wherein, the first preset ratio is greater than the second preset ratio.

2. The control method of the robot according to claim 1, characterized in that When the path planning information includes the point cloud or the bounding box, the step of determining the driving path of the robot according to the path planning information includes: Determine the first boundary information of the point cloud or the bounding box; Determine the driving path according to the first boundary information.

3. The control method of the robot according to claim 1, characterized in that, When the path planning information includes the point cloud and the bounding box, the step of determining the driving path of the robot according to the path planning information includes: Determine the second boundary information of the point cloud; Determine the target area in the bounding box according to the second boundary information; Determine the third boundary information of the target area; Determine the driving path according to the third boundary information.

4. The control method of the robot according to claim 1, characterized in that, The step of obtaining the point cloud and bounding box of the target object includes: Obtain a first image set, where each first image in the first image set includes the target object; Identify the first image set to determine the point cloud and the bounding box.

5. The control method of the robot according to claim 4, wherein After identifying the first image set to determine the point cloud and the bounding box of the target object, it further includes: When the recognition effect of the point cloud and / or the bounding box does not meet the preset conditions, obtain a second image set, the acquisition time of the second image set is earlier than the acquisition time of the first image set, and each second image in the second image set includes the target object; Identify the second image set to determine the point cloud and the bounding box.

6. The control method of the robot according to claim 4, characterized in that, After identifying the first image set to determine the point cloud and the bounding box of the target object, it further includes: When the recognition effect of the point cloud and / or the bounding box does not meet the preset conditions, determine the first position of the robot according to the first image set; Control the robot to drive to the first position; During the driving of the robot, obtain a third image set, where each third image in the third image set includes the target object; Identify the third image set to determine the point cloud and the bounding box.

7. The control method of the robot according to claim 6, wherein Determining the first position of the robot according to the first image set includes: Determining the relative position information between the robot and the target object according to the first image set; Determining the first position according to the relative position information.

8. The control method of the robot according to claim 6, wherein The first traveling speed at which the robot travels to the first position is less than the second traveling speed at which the robot travels along the traveling path.

9. A control device for a robot, characterized in that, Including: An acquisition module for acquiring the point cloud and the contour box of the target object; A selection module for selecting one or a combination of the point cloud and the contour box as path planning information; A determination module for determining the traveling path of the robot according to the path planning information; A control module for controlling the robot to travel along the traveling path; The determination module is used to determine the area ratio of the point cloud in the contour box; The selection module is used to select one or a combination of the point cloud and the contour box as path planning information according to the area ratio; when the area ratio is greater than or equal to a first preset ratio, the contour box is used as the path planning information; when the area ratio is less than or equal to a second preset ratio, the point cloud is used as the path planning information; When the area ratio is greater than the second preset ratio and less than the first preset ratio, the point cloud and the contour box are used as path planning information; wherein, the first preset ratio is greater than the second preset ratio.

10. A control device for a robot, characterized in that, Including: A processor; A memory in which a program or instruction is stored, and the processor implements the steps of the method according to any one of claims 1 to 8 when executing the program or instruction in the memory.

11. A readable storage medium, characterized in that, The program or instruction is stored on the readable storage medium, and the steps of the method according to any one of claims 1 to 8 are implemented when the program or instruction is executed by the processor.

12. A robot, characterized in that, Including: The control device of the robot according to claim 9 or 10; Or The readable storage medium according to claim 11.

Images