Robot execution business operation method, device and robot

Abstract

The present invention discloses a method, device and robot for a robot to perform business operations. The above method includes: during the robot navigation process, dynamically updating the map information established based on the world coordinate system; obtaining the current posture data of the robot, and obtaining the current position information of the robot in the world coordinate system based on the posture data; judging whether the current position information corresponds to the first position information in the world coordinate system on the path to be tracked, and if so, executing the business operation bound to the first position information, wherein one or more position information on the path to be tracked are respectively bound to corresponding business operations, and the one or more position information include the first position information, and the path to be tracked is bound to the map information. The above technical solution can improve the safety of the robot in the process of autonomous navigation, and more simply and practically realize the robot to perform a variety of work tasks.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method and device for a robot to perform business operations, and a robot. Background Art

[0002] With the rapid development of sensor technology, artificial intelligence theory, and computer technology, the field of robotics has been continuously studied, and various autonomous mobile robots with environmental perception, behavior control, dynamic decision-making, and human-computer interaction capabilities have been developed. Compared with traditional industrial robots, the biggest feature of autonomous mobile robots is that they can move freely in complex environments and complete a variety of work tasks.

[0003] In the related technology, in order to realize autonomous navigation of the robot along the predetermined route in a complex environment, a laser mapping method is usually adopted. The required map is constructed by correcting the laser map, and subsequent navigation and positioning work is performed according to the map established by the laser. However, the accuracy of laser mapping cannot meet high requirements. Laser navigation belongs to two-dimensional obstacle avoidance and the effect is poor. In addition, laser sensors are mostly deployed at the top, which increases the height of the fuselage. At the same time, there may be blind spots in the detection of low obstacles. Therefore, the robot may fall, roll over or collide in a complex environment.

[0004] Furthermore, in order to enable robots to complete various tasks while moving autonomously, they are usually programmed using software, which usually involves detailed coding to anticipate or attempt to anticipate every situation the robot may encounter. This approach is not only costly in terms of time, effort, and computer resources, but also limits the robot's capabilities.

[0005] Therefore, there is currently a lack of a robot navigation technology solution to address the problem of how to enable the robot to perform multiple work tasks more easily and practically while ensuring the safety of the robot's autonomous navigation process. Summary of the invention

[0006] The main purpose of the present invention is to disclose a method, device and robot for performing business operations by a robot, so as to at least solve the problem in the related technology of how to ensure the safety of the robot during autonomous navigation, so as to enable the robot to perform various work tasks more simply and practically. At present, there is still a lack of a robot navigation technology solution and other problems.

[0007] According to one aspect of the present invention, a method for a robot to perform business operations is provided.

[0008] The robot execution business operation method according to the present invention includes: during the robot navigation process, dynamically updating the map information established based on the world coordinate system; obtaining the current posture data of the above-mentioned robot, and according to the posture data, obtaining the current position information of the above-mentioned robot in the world coordinate system; judging whether the above-mentioned current position information corresponds to the first position information in the world coordinate system on the path to be tracked, and if so, executing the business operation bound to the above-mentioned first position information, wherein one or more position information on the above-mentioned path to be tracked are respectively bound to corresponding business operations, the above-mentioned one or more position information include the above-mentioned first position information, and the above-mentioned path to be tracked is bound to the above-mentioned map information.

[0009] According to another aspect of the present invention, a robot performing business operations device is provided.

[0010] The robot business operation execution device according to the present invention includes: a map update module, which is used to dynamically update the map information established based on the world coordinate system when the robot executes navigation; an acquisition module, which is used to obtain the current posture data of the above-mentioned robot, and obtain the current position information of the above-mentioned robot in the world coordinate system based on the posture data; an execution module, which is used to determine whether the above-mentioned current position information corresponds to the first position information in the world coordinate system on the path to be tracked, and if so, execute the business operation bound to the above-mentioned first position information, wherein one or more position information on the above-mentioned path to be tracked are respectively bound to corresponding business operations, and the one or more position information include the above-mentioned first position information, and the above-mentioned path to be tracked is bound to the above-mentioned map information.

[0011] According to yet another aspect of the present invention, a robot is provided.

[0012] The robot according to the present invention comprises: a memory and a processor, wherein the memory is used to store computer-executable instructions; and the processor is used to execute the computer-executable instructions stored in the memory, so that the robot executes any one of the above methods.

[0013] According to the present invention, during the robot's navigation along the path to be tracked, the map information established based on the world coordinate system is dynamically updated, and a higher-precision map is constructed through real-time analysis and optimization. Combined with the high-precision map and the robot's positioning function, the robot can be prevented from running into dangerous areas, thereby improving the robot's safety. In addition, since the path to be tracked is bound to the above-mentioned map information, one or more position information on the path to be tracked is respectively bound to corresponding business operations (for example, turning, arm movements, spraying water, vacuuming, detouring, opening corresponding equipment, etc.), when the robot moves to a position bound to a business operation, the business operation corresponding to the position can be executed, thereby completing a variety of work tasks more intelligently during autonomous navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flowchart of a method for a robot to perform a business operation according to an embodiment of the present invention;

[0015] Figure 2 is a structural block diagram of a robot performing business operation according to an embodiment of the present invention;

[0016] Figure 3 is a structural block diagram of a robot according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The specific implementation of the present invention is described in detail below with reference to the accompanying drawings.

[0018] According to an embodiment of the present invention, a method for a robot to perform a business operation is provided.

[0019] Figure 1 FIG. 1 is a flow chart of a method for a robot to perform a business operation according to an embodiment of the present invention. Figure 1 As shown, the robot performs a business operation method including:

[0020] Step S101: During robot navigation, dynamically update the map information established based on the world coordinate system;

[0021] Step S103: obtaining current posture data of the robot, and obtaining current position information of the robot in the world coordinate system according to the posture data;

[0022] Step S105: Determine whether the above-mentioned current position information corresponds to the first position information in the world coordinate system on the path to be tracked. If so, execute the business operation bound to the above-mentioned first position information, wherein one or more position information on the above-mentioned path to be tracked are respectively bound to corresponding business operations, the above-mentioned one or more position information include the above-mentioned first position information, and the above-mentioned path to be tracked is bound to the above-mentioned map information.

[0023] use Figure 1The method shown in the figure dynamically updates the map information (e.g., three-dimensional point cloud map and / or semantic map, etc.) established based on the world coordinate system during the robot's navigation along the path to be tracked, and builds a higher-precision map through real-time analysis and optimization. Combined with the high-precision map and the robot's positioning function, the robot can be prevented from running into dangerous areas, thereby improving the robot's safety. In addition, since the path to be tracked is bound to the three-dimensional point cloud map and / or semantic map, one or more position information on the path to be tracked is respectively bound to corresponding business operations (e.g., turning, arm movements, spraying water, vacuuming, detouring, opening corresponding equipment, etc.). When the robot moves to a position bound to a business operation, it can execute the business operation corresponding to the position, thereby completing a variety of work tasks more intelligently during autonomous navigation.

[0024] Among them, the above-mentioned map information (for example, three-dimensional point cloud map and / or semantic map, etc.) is established based on the world coordinate system, which can determine the position coordinates of the point cloud data points in the three-dimensional point cloud map in the world coordinate system, or determine the position coordinates of each object in the semantic map (for example, table, vase, elevator, etc.) in the world coordinate system.

[0025] Preferably, before the robot navigates, it can also include: establishing a world coordinate system, and building a world map based on the world coordinate system; establishing a shallow semantic map based on multiple sensor data and super-point cloud information extracted by a deep learning algorithm, and binding the shallow semantic map to the world map.

[0026] Among them, the above-mentioned super-point cloud information is point cloud information extracted by a deep learning algorithm. The point cloud information has stronger descriptive capabilities and can assist users or robots in better positioning and mapping.

[0027] Preferably, before the robot navigates, it can also include: establishing a world coordinate system, and constructing a world map based on the world coordinate system; establishing a deep semantic map based on the semantic information extracted by the deep learning algorithm, and binding the deep semantic map with the world map, wherein the deep semantic map includes: boundary information, category information.

[0028] During the preferred implementation, the world coordinate system of the robot is first established through the initial posture of the robot and the specific operating scenario, and a world map is constructed based on the world coordinate system; then, the scene is described through the received sensor data, such as image data, inertial sensor (IMU) data, odometer data, laser data and other data, and the received super point cloud (super point cloud) information extracted by the deep learning algorithm (including point cloud data with higher dimensional information), and a shallow semantic map is established. The shallow semantic map is then bound to the above-mentioned world map, and the feature information of the deep learning map can be used as shallow semantic information to assist in improving map construction and improve positioning accuracy, and the shallow semantic map is used to assist in positioning and mapping.

[0029] In the preferred implementation process, a deep semantic map can also be obtained based on the semantic information extracted by the deep learning algorithm, and the feature information based on the deep learning map is used as shallow semantic information to assist in improving map construction and improve positioning accuracy; and the deep semantic map is bound to the world map constructed based on the world coordinate system, wherein the deep semantic map includes, in addition to basic visual point cloud information, super point cloud information extracted by deep learning (point cloud information with more descriptive ability), and deep semantic information (which may include boundary point cloud information, boundary information of the deep semantic map, and cloud information of the deep semantic map). The cloud information of the clustered point cloud is fitted into a point cloud cluster set, and the centroid point of each of the above point cloud cluster sets is determined. . For example, the k-means clustering algorithm is used to determine the proximity of point cloud data points by calculating the distance between different point cloud data points, and similar point cloud data points will be placed in the same category. Specifically, you first need to select a k value, where k is the number of clusters given in advance. Then you need to randomly select the starting k cluster points (i.e., centroid points), and then calculate the distance of all points in the point cloud data set from the selected centroid points, and divide them into the category closest to their centroids. After completion, you need to calculate the average value of each cluster set and use this point as the new centroid. Repeat the above steps until the centroid remains unchanged or changes very little, and you get the final result.

[0030] It can be seen that based on the deep semantic information, the local point cloud map is established as a higher-dimensional and more descriptive centroid point. This method combines semantic information with traditional point cloud by clustering fitting. The cluster point is used as the center point (centroid point) of the local point cloud cluster set that meets the standard. The directionality of the local point cloud is extracted according to the boundary information. The local extraction range is divided according to the boundary information of the deep semantic map. The extracted point cloud is fitted into a high-dimensional point cloud cluster set (which can be called landmark information) by clustering. At the same time, the relative relationship between different high-dimensional landmark information is established based on the existing high-dimensional landmark information. In the actual robot operation process, the trajectory and positioning are optimized and corrected by observing one or more identical landmarks, thereby improving the overall positioning accuracy and map accuracy.

[0031] Get the semantic information of the current deep learning, determine the relevant semantic information (including object category information, such as elevators, plants, vases, etc.) identified by the current deep learning through time alignment, and transform the relevant semantic information to the world coordinate system through relative pose. It should be noted that if this part of the data overlaps with the previously constructed deep semantic map, then the information of this part will supplement the high-dimensional point cloud clustering set clustered in the deep semantic map (for example, the deep semantic map is a point cloud clustering set of item 1 with length x, width y, and height z, and the semantic information is represented as a blue seat, then the road sign information of this part will be updated to a blue seat with length x, width y, and height z). At the same time, this part of the map is a semantic map added on top of the basic point cloud map.

[0032] It can be seen that during the robot navigation process, map optimization operations are performed in real time. The main optimization contents include: the current map accuracy, the path trajectory bound in the map, and the semantic information in the map. Specifically, the higher-dimensional landmark information is used as important observation information used in the optimization to jointly optimize the entire map and the robot's trajectory and posture. The semantic information of the current environment, the robot's path trajectory information, the robot's various business operation information, and the robot's various status information are jointly recorded, and a higher-precision map is constructed through analysis and optimization, so that the robot can play a greater role in production and life.

[0033] Preferably, the robot navigation process may also include: extracting point cloud information based on the boundary information of the deep semantic map, fitting the extracted point cloud information into a point cloud cluster set by clustering, and determining the centroid points of each of the above point cloud cluster sets; calculating the Euclidean distance between the robot and other centroid points among the above multiple centroid points except for the above at least one centroid point according to the first position relationship between the multiple centroid points in the above semantic map and the Euclidean distance between the above robot and at least one centroid point among the above multiple centroid points; comparing the calculated Euclidean distance with the current actual distance information, and if the deviation is greater than the preset deviation threshold, executing corresponding decision processing to make the above robot return to the above path to be tracked.

[0034] In the preferred implementation process, point cloud information can be extracted based on the boundary information of the super point cloud information, and the extracted point cloud information can be fitted into a point cloud cluster set through clustering (for example, k-means clustering algorithm, etc.), and the centroid point of each of the above point cloud cluster sets is determined. As mentioned above, the stable result finally determined by the k-means clustering algorithm can be used as the centroid point of the point cloud cluster set. For example, there are three objects in the semantic map (which can be a coffee table, a chair, a stool, etc.), and each centroid point is taken as the central representative point of an object. According to the Euclidean distance of the robot from the centroid point 1 and the centroid point 2, and the positional relationship between the centroid points 1, 2, and 3, the Euclidean distance of the robot relative to the centroid point 3 can be calculated. If the calculated Euclidean distance is 3 meters, and according to the trajectory of the path to be tracked, the actual distance information should be 2.8 meters, and the offset of 0.2 meters is greater than the preset deviation threshold (for example, 0.1 meters), it is necessary to optimize based on these two distance parameters and all current observation information, and adjust the robot's navigation strategy according to the optimization results. For example, adjust the path according to the calculated deviation to make the above-mentioned robot return to the above-mentioned path to be tracked.

[0035] It can be seen from this that whether the robot deviates from the path to be tracked can be determined based on the positional relationship between each object in the semantic map. When it is determined that the robot deviates from the path to be tracked, the robot's autonomous navigation path can be adjusted in time to prevent the robot from running into dangerous areas, thereby ensuring the safety of the robot.

[0036] Preferably, before the robot navigates, the following processing may also be included: responding to the user's first operation instruction, for each of the one or more position information on the above-mentioned path to be tracked, binding the business operation corresponding to the position information (for example, turning, arm movement, spraying water, vacuuming, detouring, turning on the corresponding equipment, etc.), wherein the above-mentioned path to be tracked is a path pre-set by the user, a path for the user to pre-control the robot to perform a task, a path pre-planned by the robot, or a path after the last navigation update.

[0037] For example, when the user uses the remote control to control the robot to perform a task for the first time, at position 1 of the path to be tracked, there is a virtual wall or fragile object (e.g., a vase, etc.) set in the predetermined range in front, then the position 1 information (e.g., the coordinate position information in the world coordinate system) is bound to the detour business operation, and the above binding relationship is recorded and saved. In the subsequent autonomous navigation process of the robot, when it is determined that it has reached this position, the business operation bound to the above position is executed, and the detour operation is performed on the virtual wall or fragile object according to the preset detour strategy.

[0038] Preferably, after binding the business operation corresponding to the location information, it can also include: the robot determines the first binding relationship between the above-mentioned business operation and the corresponding scene in the above-mentioned map information; for the updated map information, the above-mentioned robot automatically adds or releases the current binding relationship between one or more location information on the path to be tracked and the corresponding business operation according to the above-mentioned first binding relationship.

[0039] During the preferred implementation process, the robot can also learn the binding relationship between the position and the business operation, autonomously establish a new binding relationship between the position and the business operation, or delete the existing binding relationship between the position and the business operation. For example, during the robot's autonomous navigation process, it continuously optimizes and updates the current semantic map. Due to changes in the environment, when other fragile items (for example, vases, etc.) appear in the current scene, the robot determines the item classification set to which it belongs based on the category information of the item. When it is determined that the item belongs to the fragile item category, the robot automatically adds a binding relationship between position 2 and the detour operation at position 2 before reaching the fragile item in accordance with the previously learned binding relationship, and records and saves the binding relationship. During the next autonomous navigation process, when the robot reaches position 2, a detour operation is triggered, and the detour operation is executed, and local optimization and adjustment are performed on the path to be tracked. In the specific implementation process, the above detour strategy can be to detour around the object at a specified distance value. The centroid of the point cloud cluster set corresponding to the object can be used as the center, and the corresponding specified distance value can be calculated in combination with the size information of the object in the shallow semantic map (for example, length, width, height, etc.). If the object is large in volume, the calculated distance value is relatively large. On the contrary, if the object is small in volume, the calculated distance value is relatively small. After that, the robot maintains the distance from the above centroid according to the calculated specified distance value and performs a detour operation. In the preferred implementation process, the optimized path can also be used as a new path to be tracked, and re-bound to the map information as the path to be tracked for the next robot navigation.

[0040] In addition, the robot learns the binding relationship between the position and the business operation. In addition to autonomously establishing the binding relationship between the new position and the business operation, it can also autonomously delete (release) the binding relationship between the current position and the business operation. For example, during the robot's autonomous navigation process, the current semantic map is continuously optimized and updated. Due to changes in the environment, when the robot detects that the fragile object disappears in the current scene, it will automatically release the binding relationship between the above-mentioned position 2 and the detour operation. In the next autonomous navigation process, when the robot reaches position 2, it will no longer perform the detour operation, and will re-perform local optimization adjustments on the path to be tracked, and use the optimized path as the new path to be tracked, and re-bind it to the three-dimensional point cloud map and / or semantic map as the path to be tracked for the next robot navigation.

[0041] Preferably, before the robot navigates, the following processing may also be included: in response to the user's second operation instruction, for each position information of at least one position information on the above-mentioned path to be tracked, the state information corresponding to the position information (for example, slipping state, water shortage state, power shortage state, etc.) is respectively bound.

[0042] Preferably, after binding the status information corresponding to the position information, it may also include: for the second position information bound with the status information, in the process of the robot moving towards the above-mentioned second position information, when the distance between the above-mentioned robot and the above-mentioned second position information is less than a predetermined distance threshold, executing corresponding decision processing in advance according to the above-mentioned status information.

[0043] In the preferred implementation process, when the user uses the remote control to control the robot to perform a task, the robot status information and the position information corresponding to the status information are obtained in real time. For example, when it is determined that the wheeled robot runs to position 3, it slips, which indicates that the robot is in a special state at position 3, and a binding relationship between position 3 and the slip status information is established. During the next robot autonomous navigation process, when the robot runs to the corresponding position before position 3, it will adjust the parameters adapted to the special scene according to the existing status information, so that the robot can still have a high positioning accuracy and smooth operation in the special scene. For example, in the event of a robot slip, the credibility of the observation data of the sensor (such as the odometer, etc.) that is more affected by the slip state is reduced, and the credibility of the observation data of the sensor (such as the IMU, etc.) that is less affected by the slip state is improved.

[0044] Preferably, after binding the status information corresponding to the location information, the following processing may also be included: the above-mentioned robot determines the second binding relationship between the above-mentioned status information and the corresponding scene in the above-mentioned map information; for the updated map information, the above-mentioned robot automatically adds or releases the current binding relationship between one or more location information on the path to be tracked and the corresponding status information according to the above-mentioned second binding relationship.

[0045] In the preferred implementation process, the robot can also learn the binding relationship between the position and the state information, autonomously establish a new binding relationship between the position and the state information, or delete the binding relationship between the existing position and the state information. For example, during the autonomous navigation process of the robot, the current semantic map is continuously optimized and updated. Due to changes in the environment, new grooves appear in the current scene, and the wheels of the wheeled robot are stuck between the grooves. When the robot determines that a slipping state has occurred, the robot automatically adds the binding relationship between the current position 4 and the slipping state information according to the previously learned binding relationship, and records and saves the binding relationship. During the next autonomous navigation process, when the robot reaches position 4, the credibility of the observation data of the sensor (such as the odometer, etc.) that is more affected by the slipping state is reduced, and the credibility of the observation data of the sensor (such as the IMU, etc.) that is less affected by the slipping state is increased. Alternatively, other corresponding decision strategies can also be executed, such as accelerating or decelerating backward to escape.

[0046] In addition, the robot learns the binding relationship between position and state information. In addition to autonomously establishing a new binding relationship between position and state information, it can also autonomously delete (release) the binding relationship between the current position and state information. For example, during the robot's autonomous navigation, the current semantic map is continuously optimized and updated. Due to changes in the environment, when the robot detects that the groove disappears in the current scene, it will automatically release the binding relationship between the above position 4 and the slip state information. In the next autonomous navigation process, when the robot reaches position 4, the credibility of the observation data of the above sensors (such as odometers, etc.) that are greatly affected by the slip state is improved.

[0047] In summary, based on the three-dimensional point cloud map or multi-layer semantic map (first-layer world map, second-layer shallow semantic map, third-layer deep semantic map) created by the world coordinate system, when the user pre-controls the robot to perform tasks (for example, controlling the robot by remote control, etc.), after obtaining all the robot status and observation information, the robot's operation trajectory and observation information are corrected and fused, and bound to the robot's current map. At the same time, the navigation information, semantic information, robot business operation information, and status information are fused and bound to the current robot's operation trajectory to obtain a more accurate robot state estimation result and a robot positioning and obstacle avoidance map.

[0048] For the above-mentioned map that integrates multiple information, the robot optimizes the current map in real time during autonomous navigation. The main optimizations include: the accuracy of the currently created map, the trajectory posture recorded in the map, and the semantic information in the map. Specifically, the higher-dimensional landmark information (point cloud cluster set) is used as important observation information used in the optimization to jointly optimize the entire map and the robot's trajectory posture. The other information in the map is corrected through the correction results of the robot's running posture trajectory and the relative relationship between the recorded landmarks, states and the robot's posture.

[0049] After creating a map that integrates multiple information and optimizing the map, the map can be stored. The corresponding map can be selected according to the scene and used for the next navigation and positioning work of the robot.

[0050] According to an embodiment of the present invention, a robot performing business operation device is provided.

[0051] Figure 2 1 is a structural block diagram of a robot performing business operation according to an embodiment of the present invention. Figure 2 As shown, the robot execution business operation device includes: a map update module 20, which is used to dynamically update the map information established based on the world coordinate system when the robot executes navigation; an acquisition module 22, which is used to obtain the current posture data of the above-mentioned robot, and obtain the current position information of the above-mentioned robot in the world coordinate system based on the posture data; an execution module 24, which is used to determine whether the above-mentioned current position information corresponds to the first position information in the world coordinate system on the path to be tracked, and if so, execute the business operation bound to the above-mentioned first position information, wherein one or more position information on the above-mentioned path to be tracked are respectively bound to corresponding business operations, the one or more position information include the above-mentioned first position information, and the above-mentioned path to be tracked is bound to the above-mentioned map information.

[0052] use Figure 2 In the device shown, during the robot's navigation along the path to be tracked, the map update module 20 dynamically updates the map information established based on the world coordinate system, and builds a higher-precision map through real-time analysis and optimization. Combined with the high-precision map and the robot's positioning function, the robot can be prevented from running into dangerous areas, thereby improving the robot's safety. In addition, since the path to be tracked is bound to the above-mentioned map information, one or more position information on the path to be tracked is respectively bound to corresponding business operations (for example, turning, detouring, or executing water spraying and other businesses). When the robot moves to a position bound to a business operation, the execution module 24 can execute the business operation corresponding to the position, thereby completing a variety of work tasks more intelligently during the autonomous navigation process.

[0053] It should be noted that the preferred implementation of the combination of the modules in the above robot execution business operation device can be specifically referred to Figure 1 The corresponding descriptions and effects in the embodiments can be understood and will not be repeated here.

[0054] According to an embodiment of the present invention, a robot is provided.

[0055] Figure 3 is a structural block diagram of a robot according to an embodiment of the present invention. Figure 3 As shown, the robot according to the present invention includes: a memory 30 and a processor 32, the memory 30 is used to store computer execution instructions; the processor 32 is used to execute the computer execution instructions stored in the memory, so that the robot executes the robot execution business operation method provided in the above embodiment.

[0056] The processor 32 may be a central processing unit (CPU). The processor 52 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0057] The memory 30, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the robot repositioning method in the embodiment of the present invention. The processor executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory.

[0058] The memory 30 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 30 may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0059] The one or more modules are stored in the memory 30, and when executed by the processor 32, the following is performed: Figure 1 The robot in the illustrated embodiment executes the business operation method.

[0060] For details of the above robots, please refer to Figure 1 The corresponding related descriptions and effects in the illustrated embodiments can be understood and will not be repeated here.

[0061] In summary, with the above implementation provided by the present invention, the semantic information of the current environment, the path trajectory information of the robot, the business operation information of the robot, and the state information of the robot are jointly recorded and bound. By analyzing and optimizing to build a higher-precision map, the business operation corresponding to the position can be executed, so that a variety of work tasks can be completed more intelligently during the autonomous navigation process, so that the robot can play a greater role in production and life. By integrating a map with multiple information, the robot navigation system can make more efficient plans based on historical information, and correct the map based on semantic information, so that the positioning is more accurate; at the same time, it can avoid the robot from running into dangerous areas during operation according to semantic information to ensure safety. In addition, the robot has a learning function. According to the binding relationship between the position and the business operation determined by the user controlling the robot to perform a task or the robot's autonomous navigation process, the robot can automatically add or release the binding relationship between one or more position information on the path to be tracked and the corresponding business operation. At the same time, according to the binding relationship between the position and the state information determined by the user controlling the robot to perform a task or the robot's navigation process, the robot can automatically add or release the binding relationship between one or more position information on the path to be tracked and the corresponding state information, making the robot more intelligent and humane.

[0062] The above disclosures are only several specific embodiments of the present invention; however, the present invention is not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A robot execution business operation method, characterized in that: include: Before the robot navigates, in response to a second operation instruction of the user, for each piece of position information in at least one piece of position information on the path to be tracked, respectively bind the state information corresponding to the piece of position information; During robot navigation, the map information based on the world coordinate system is dynamically updated; Acquire current posture data of the robot, and acquire current position information of the robot in a world coordinate system according to the posture data; Determine whether the current position information corresponds to first position information in the world coordinate system on the path to be tracked, and if so, execute a business operation bound to the first position information, wherein one or more position information on the path to be tracked are respectively bound to corresponding business operations, the one or more position information include the first position information, and the path to be tracked is bound to the map information; For the second position information bound with the state information, when the robot moves toward the second position information and the distance between the robot and the second position information is less than a predetermined distance threshold, corresponding decision processing is performed in advance according to the state information; Point cloud information is extracted based on boundary information of the deep semantic map, the extracted point cloud information is fitted into a point cloud cluster set by clustering, the centroid point of each point cloud cluster set is determined, and the Euclidean distance between the robot and other centroid points among the multiple centroid points except the at least one centroid point is calculated according to the first position relationship between the multiple centroid points in the semantic map and the Euclidean distance between the robot and at least one centroid point among the multiple centroid points, the calculated Euclidean distance is compared with the current actual distance information, and if the deviation is greater than a preset deviation threshold, corresponding decision processing is performed to make the robot return to the path to be tracked.

2. The method according to claim 1, characterized in that Before robot navigation, it also includes: Establishing a world coordinate system, and constructing a world map based on the world coordinate system; Based on multiple sensor data and combined with super point cloud information extracted by a deep learning algorithm, a shallow semantic map is established, and the shallow semantic map is bound to the world map.

3. The method according to claim 1, characterized in that: Before robot navigation, it also includes: Establishing a world coordinate system, and constructing a world map based on the world coordinate system; A deep semantic map is established based on semantic information extracted by a deep learning algorithm, and the deep semantic map is bound to the world map, wherein the deep semantic map includes: boundary information and category information.

4. The method according to claim 1, characterized in that: Before robot navigation, it also includes: In response to the user's first operation instruction, for each of the one or more position information on the path to be tracked, the business operation corresponding to the position information is respectively bound, wherein the path to be tracked is a path pre-set by the user, a path for the user to pre-control the robot to perform a task, a path pre-planned by the robot, or a path after the last navigation update.

5. The method according to claim 4, characterized in that After binding the business operation corresponding to the location information, it also includes: The robot determines a first binding relationship between the business operation and the corresponding scene in the map information; For the updated map information, the robot automatically adds or releases the current binding relationship between one or more location information on the path to be tracked and the corresponding business operation according to the first binding relationship.

6. The method according to claim 1, characterized in that After binding the status information corresponding to the location information, it also includes: The robot determines a second binding relationship between the state information and a corresponding scene in the map information; For the updated map information, the robot automatically adds or releases the current binding relationship between one or more position information and corresponding state information on the path to be tracked according to the second binding relationship.

7. A robot-operated business operation device, characterized in that: include: Before the robot navigates, in response to a second operation instruction from the user, for each piece of at least one piece of position information on the path to be tracked, a module for binding the state information corresponding to the piece of position information; A map update module is used to dynamically update the map information established based on the world coordinate system when the robot performs navigation; An acquisition module, used to acquire current posture data of the robot, and acquire current position information of the robot in a world coordinate system according to the posture data; an execution module, configured to determine whether the current position information corresponds to first position information in a world coordinate system on a path to be tracked, and if so, execute a business operation bound to the first position information, wherein one or more position information on the path to be tracked are respectively bound to corresponding business operations, the one or more position information include the first position information, and the path to be tracked is bound to the map information; The execution module is further configured to, for the second position information bound with the state information, execute corresponding decision processing in advance according to the state information when the distance between the robot and the second position information is less than a predetermined distance threshold during the process of the robot moving toward the second position information; The execution module is also used to extract point cloud information based on the boundary information of the deep semantic map, fit the extracted point cloud information into a point cloud cluster set by clustering, determine the centroid point of each point cloud cluster set, calculate the Euclidean distance between the robot and other centroid points among the multiple centroid points except the at least one centroid point according to the first position relationship between the multiple centroid points in the semantic map and the Euclidean distance between the robot and at least one centroid point among the multiple centroid points, compare the calculated Euclidean distance with the current actual distance information, and if the deviation is greater than a preset deviation threshold, execute corresponding decision processing to make the robot return to the path to be tracked.

8. A robot comprising: A memory and a processor, characterized in that: The memory is used to store computer-executable instructions; The processor is used to execute the computer-executable instructions stored in the memory, so that the robot performs the method according to any one of claims 1 to 6.

Images