AI-based industrial robot control methods, systems, and storage media

Abstract

This invention relates to the field of industrial robot control technology, specifically disclosing an AI-based industrial robot control method, system, and storage medium. The method includes acquiring and statistically analyzing all robot behaviors across different production cycles; acquiring a work area map, converting robot behaviors into trajectories containing time periods, and inserting these trajectories into the work area map; performing time-domain discretization on the work area map to obtain a map set containing robot positions at each time point; determining the amount of AI resources needed at each time point based on the map set; and identifying the demand information at each time point based on the AI ​​resource amount to generate robot control commands. This invention statistically analyzes historical robot activity information across different production cycles of the same production task, determines the amount of AI resources needed at different times, and transforms the original high-performance, high-frequency application process into a dynamic, high-frequency application process, reducing the amount of AI resources needed for most of the time and lowering costs.

Description

Technical Field

[0001] This invention relates to the field of industrial robot control technology, specifically to an AI-based industrial robot control method, system, and storage medium. Background Technology

[0002] With the performance improvement of large-scale AI models, the application of artificial intelligence in industry has become an important means to improve the efficiency and competitiveness of manufacturing. AI has extremely strong information understanding and processing capabilities. It can autonomously identify information and make feedback. It is often combined with robots to control robots in industrial production scenarios, thereby providing assistance to industrial production activities. However, the application of AI requires a lot of resources, which increases costs. Many small businesses will not actively choose to apply AI. Therefore, how to reduce the cost of AI application and further increase the popularity of AI in industrial production activities is the technical problem that this invention aims to solve. Summary of the Invention

[0003] The purpose of this invention is to provide an AI-based industrial robot control method, system, and storage medium to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] An AI-based industrial robot control method, the method comprising:

[0006] Acquire and analyze all robot behaviors for the same production task across different production cycles; the robot behaviors include time period, starting device label, and ending device label.

[0007] Obtain the work area map, convert robot behavior into trajectories containing time periods, and insert them into the work area map; one production cycle corresponds to one work area map containing trajectories.

[0008] The work area map is discretized in the time domain to obtain a map set containing the robot's position at each time point;

[0009] The amount of AI resources at each time point is determined based on the map set. The amount of AI resources is related to the distribution of robot positions on the map at each time point, and is directly proportional to the total number of robot positions and the concentration of robot positions.

[0010] Based on the availability of AI resources, the system identifies the demand information at each moment, generates robot control commands, and simultaneously generates robot behavior.

[0011] As a further aspect of the present invention: the step of acquiring and statistically analyzing all robot behaviors for the same production task in different production cycles includes:

[0012] Set the starting point of the production cycle for any production task to time zero;

[0013] The robot is detected in real time within a preset distance threshold by a robot detection device built into the production equipment; the robot detection device includes a camera and a ranging radar.

[0014] When a robot exists within the preset distance threshold, establish a connection channel with the robot and query the device tag in the robot's tag register; when the device tag is empty, insert the tag of the current production device as the starting device tag into the robot.

[0015] When the device tag is not empty, read the device tag as the starting device tag, take the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period, and count the time period, starting device tag and ending device tag to obtain a robot behavior.

[0016] As a further aspect of the present invention: the step of acquiring and statistically analyzing all robot behaviors for the same production task in different production cycles further includes:

[0017] Signal transceivers are inserted into all production equipment and all robots as robot inspection instruments;

[0018] The distance between the production equipment and the robot is determined based on the signal transceivers built into the production equipment and the signal transceivers built into the robot; the determined distance contains time stamps.

[0019] The distance acquisition frequency for any production equipment and any robot is determined based on the distance between the production equipment and the robot; the distance acquisition frequency is inversely proportional to the distance.

[0020] For any robot, query its distance to at least three production devices at each time point, determine the robot's coordinates at that time point, and store them in the coordinate library.

[0021] As a further aspect of the present invention: the steps of acquiring the work area map, converting robot behavior into trajectories containing time periods, and inserting them into the work area map include:

[0022] Get the workspace map and query the workspace map scale;

[0023] Read all robot behaviors, and based on the scale, query the point in the work area map corresponding to the starting device label as the starting point, and query the point in the work area map corresponding to the ending device label as the ending point;

[0024] The navigation route is determined based on the starting point and the destination within the travel segment. Time period labels are inserted into the navigation route to obtain the trajectory.

[0025] Insert the trajectory into the workspace map.

[0026] As a further aspect of the present invention: the step of discretizing the work area map in the time domain to obtain a map set containing the robot's position at each time point includes:

[0027] Create time points based on preset time steps;

[0028] Read the time points sequentially, compare the time points with the time periods of each trajectory, and select the predicted point in the trajectory when the time point belongs to the time period.

[0029] The robot corresponding to the trajectory is queried, and the predicted point is used as the robot's coordinates at that time.

[0030] Calculate the coordinates of all robots at the same time point, insert them into the work area map, and use them as the map at that time point;

[0031] Compile maps at all points in time to obtain a map atlas;

[0032] The process of selecting predicted points in the trajectory is as follows: calculating the position ratio of the predicted points, and locating the predicted points on the trajectory according to the position ratio; the starting point of the trajectory is taken as the zero-ratio position; the process of determining the position ratio is as follows:

[0033] In the formula, For positional proportions, For the current time point, and These are the left and right endpoints of the time period, respectively;

[0034] The step of discretizing the work area map in the time domain to obtain a map set containing the robot's position at each time point further includes:

[0035] Read the robot's coordinates at each moment from the coordinate library and update the predicted points.

[0036] As a further aspect of the present invention: the step of determining the amount of AI resources at each time point based on the map atlas includes:

[0037] For any given moment in the map, read all the coordinates in the map;

[0038] Calculate the coordinate concentration based on all coordinates;

[0039] The proportion of AI resources is determined based on the coordinate concentration and the total number of coordinates.

[0040] The AI ​​resource quantity is calculated by multiplying the AI ​​resource percentage by the preset total resource quantity.

[0041] The process for determining the proportion of AI resources is as follows:

[0042] In the formula, for The proportion of AI resources at any given moment For the first One production cycle The proportion of AI resources at any given moment; This represents the total number of production cycles. ; For the first One production cycle The total number of coordinates at any given time. The threshold is a preset threshold, which is the maximum number of robots. For the first One production cycle In the time map, the first The minimum distance between each coordinate and other coordinates.

[0043] The present invention also provides an AI-based industrial robot control system, the system comprising:

[0044] The behavior acquisition module is used to acquire and statistically analyze all robot behaviors for the same production task in different production cycles; the robot behaviors include time period, starting device label, and ending device label.

[0045] The trajectory generation and insertion module is used to acquire the work area map, convert robot behavior into trajectories containing time periods, and insert them into the work area map; one production cycle corresponds to one work area map containing a trajectory.

[0046] The discretization processing module is used to discretize the work area map in the time domain to obtain a map set containing the robot's position at each time point;

[0047] The resource quantity determination module is used to determine the AI ​​resource quantity at each time point based on the map set. The AI ​​resource quantity is related to the distribution of robot positions on the map at each time point, and is directly proportional to the total number of robot positions and the concentration of robot positions.

[0048] The demand identification and processing module is used to identify demand information at various times based on the amount of AI resources, generate robot control instructions, and synchronously generate robot behavior.

[0049] As a further aspect of the present invention: the behavior acquisition module includes:

[0050] The zero-time setting unit is used to set the start point of the production cycle of any production task to zero time.

[0051] A robot detection unit is used to detect robots within a preset distance threshold in real time using a robot detection instrument built into the production equipment; the robot detection instrument includes a camera and a ranging radar.

[0052] The tag query unit is used to establish a connection channel with the robot when the robot exists within the preset distance threshold, and to query the device tag in the robot's tag register; when the device tag is empty, the tag of the current production device is used as the starting device tag and inserted into the robot.

[0053] The tag application unit is used to read the device tag when it is not empty, use it as the starting device tag, use the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period, and count the time period, starting device tag and ending device tag to obtain a robot behavior.

[0054] As a further aspect of the present invention: the trajectory generation and insertion module includes:

[0055] The scale query unit is used to obtain the work area map and query the scale of the work area map.

[0056] The start and end point determination unit is used to read all robot behaviors, query the point corresponding to the start device label in the work area map based on the scale, and use it as the start point; and query the point corresponding to the end device label in the work area map, and use it as the end point.

[0057] The generation unit is used to determine the navigation path in the traffic segment based on the start and end points, insert time period labels into the navigation path, and obtain the trajectory;

[0058] Insert unit, used to insert the trajectory into the work area map.

[0059] The present invention also provides a storage medium storing at least one line of program code, which, when loaded and executed by a processor, implements the AI-based industrial robot control method.

[0060] Compared with the prior art, the beneficial effects of the present invention are: the present invention statistically analyzes the historical activity information of the robot in different production cycles of the same production task, determines the amount of AI resources required at different times, transforms the original high-performance high-frequency application process into a dynamic high-frequency application process, reduces the amount of AI resources for most of the time, and reduces costs. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0062] Figure 1 This is a flowchart of an AI-based industrial robot control method.

[0063] Figure 2 This is the first sub-flowchart of an AI-based industrial robot control method.

[0064] Figure 3 This is the second sub-flowchart of the AI-based industrial robot control method.

[0065] Figure 4 This is the third sub-flowchart of the AI-based industrial robot control method.

[0066] Figure 5 This is the fourth sub-flowchart of the AI-based industrial robot control method.

[0067] Figure 6 This is a block diagram of the structure of an AI-based industrial robot control system. Detailed Implementation

[0068] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0069] Figure 1 This is a flowchart of an AI-based industrial robot control method. In this embodiment of the invention, an AI-based industrial robot control method includes:

[0070] Step S100: Obtain and statistically analyze all robot behaviors for the same production task in different production cycles; the robot behaviors include time period, starting device label, and ending device label;

[0071] The production process is a continuous repetitive process. The same production task may occur in multiple scenarios and may have multiple production cycles. Each production cycle represents the execution of the production task once. For any production task, each of its production cycles is taken as an analysis unit. The robot's operation in each production cycle is acquired and statistically analyzed as robot behavior. In this application, robot behavior is simplified to a starting device label, an ending device label, and a time period, indicating where the robot moves from and to. As for what specific work it performs, whether it is transportation or data forwarding, this application does not discuss it.

[0072] Step S200: Obtain the work area map, convert the robot behavior into a trajectory containing time periods, and insert it into the work area map; wherein, one production cycle corresponds to one work area map containing a trajectory.

[0073] The work area is the area where the robot works. To obtain a map of the work area, since the robot's behavior includes time periods, starting device labels, and ending device labels, we can obtain a route by taking the starting device label as the starting point and the ending device label as the ending point. We can then insert time periods into the route as the trajectory and insert them into the work area map.

[0074] It should be noted that the obtained route is generally contained within the passable area, that is, the passable road segment of the work area. It can be compared to generating a navigation path from the starting device to the ending device, which is generally a multi-segment broken line.

[0075] Step S300: Discretize the work area map in the time domain to obtain a map set containing the robot's position at each time point;

[0076] After generating the trajectory and inserting it into the work area map, the work area map contains a large number of trajectories with time periods. These time periods are not necessarily the same, making them difficult to analyze and process. For ease of processing, this application converts them into a set of maps containing the robot's position at each time, called a map set. The conversion process is called the time-domain discretization process, that is, determining the robot's position in the work area map at each time.

[0077] Step S400: Determine the amount of AI resources at each time point based on the map set; the amount of AI resources is related to the distribution of robot positions in the map at each time point, and is directly proportional to the total number of robot positions and the concentration of robot positions.

[0078] Once the atlas is obtained, it can be analyzed to determine the distribution of robots in a single production cycle. Based on the distribution of robots in a single production cycle, the amount of AI resources can be determined, thereby providing a more practical AI resource allocation scheme and reducing costs.

[0079] Specifically, the more robot locations there are, the more robots are working at that time, and the more AI resources are needed. The more concentrated the robot locations are, the more special situations may have occurred in one or two devices. In this case, a higher level of real-time understanding is required, and more AI resources need to be provided.

[0080] Step S500: Identify the demand information at each moment based on the amount of AI resources, generate robot control instructions, and simultaneously generate robot behavior;

[0081] Finally, based on the amount of AI resources, the demand information at each moment is identified, and robot control instructions are generated to control the robot to perform related tasks. During the process of performing related tasks, the robot will also generate new robot behaviors, which can be acquired synchronously.

[0082] It is worth mentioning that the process of identifying demand information at various times based on the amount of AI resources is not complicated. Existing AI modules have extremely strong understanding capabilities. When receiving demand information, they can be converted into control commands. The difference lies in the response time. The more AI resources the AI ​​module uses, the shorter the response time and the faster the response.

[0083] Figure 2 The first sub-process flowchart of the AI-based industrial robot control method includes the step of acquiring and statistically analyzing all robot behaviors for the same production task in different production cycles:

[0084] Step S101: Set the starting point of the production cycle for any production task to time zero;

[0085] Step S102: The robot is detected in real time within a preset distance threshold by a robot detector built into the production equipment; the robot detector includes a camera and a ranging radar.

[0086] Step S103: When a robot exists within the preset distance threshold, establish a connection channel with the robot and query the device tag in the robot's tag register; when the device tag is empty, insert the tag of the current production device as the starting device tag into the robot.

[0087] Step S104: When the device tag is not empty, read the device tag as the starting device tag, take the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period; count the time period, the starting device tag and the ending device tag to obtain a robot behavior.

[0088] In one example of the technical solution of this invention, the starting point of the production cycle of any production task is set as time zero, so that all production cycles have the same (or very similar) time scale. Correspondingly, the subsequent time tags are relative to time zero. A robot detector is installed on each production device. The robot detector can be a camera or a ranging radar. The robot detector works in real time and detects whether there is a robot in the surrounding area (within a preset distance threshold). The distance threshold is generally a numerical value representing a radius. Correspondingly, the surrounding area is also a spherical range. The spherical range also means that even if the robot is a flying drone, the technical solution of this application can still be applied, and the application range is extremely wide.

[0089] When a robot is present within the preset distance threshold, a connection channel is established with the robot. The device tag in the robot's tag register is queried. The tag register stores the unique tag of the production device and can only store one tag. When the device tag is empty, the tag of the current production device is inserted into the robot as the starting device tag. Generally, when the robot departs from its storage location, a starting tag is inserted into the tag register. When the device tag is not empty, it is read, indicating that the robot is moving from that device to the current device. This tag is then used as the starting device tag, and the tag of the current production device is used as the ending device tag. At the same time, the tag register has a time recording module that records the insertion and reading times of the device tag. The insertion and reading times of the device tag are queried, the time period is calculated, and finally, the time period, the starting device tag, and the ending device tag are statistically analyzed to obtain a robot behavior.

[0090] It should be noted that the robot inspection device of each production equipment runs in real time. It can detect each robot in real time, thus ensuring that every operation of the robot is detected by the production equipment. In addition, the robot's movement is also broken down into multiple point-to-point movements under this architecture.

[0091] As a preferred embodiment of the technical solution of the present invention, the step of acquiring and statistically analyzing all robot behaviors of the same production task in different production cycles further includes:

[0092] Signal transceivers are inserted into all production equipment and all robots as robot inspection instruments;

[0093] The distance between the production equipment and the robot is determined based on the signal transceivers built into the production equipment and the signal transceivers built into the robot; the determined distance contains time stamps.

[0094] The distance acquisition frequency for any production equipment and any robot is determined based on the distance between the production equipment and the robot; the distance acquisition frequency is inversely proportional to the distance.

[0095] For any robot, query its distance to at least three production devices at each time point, determine the robot's coordinates at that time point, and store them in the coordinate library.

[0096] In one embodiment of the technical solution of this invention, an additional solution is provided, specifically an optimization of the robot detection instrument. The original robot detection instrument only detects robots within a limited range. In the above technical solution, the robot detection instrument is limited to a signal transceiver installed in each production device and robot, enabling rapid and highly accurate measurement of the distance between any production device and robot. The distance between the production device and the robot is determined based on the signal transceivers built into the production devices and the robot, and the determined distance contains a time stamp. For any robot, its distance to at least three production devices is queried at various times. Since the positions of the production devices are fixed, the robot's coordinates at that time can be determined by combining the distances between the robot and the three production devices, and stored in a coordinate database. The function of this process is to obtain the actual position of the robot at different times.

[0097] It is worth mentioning that the signal transceiver itself is also an energy-consuming component. If the signal transceiver is used in real time, the energy consumption will be very high. For production equipment connected to a power source, there are only some cost issues, while for robots, the issue is battery life. In this regard, this application provides a dynamic application frequency for the signal transceiver. The distance acquisition frequency of any production equipment and any robot is determined according to the distance between the production equipment and the robot. The greater the distance, the lower the distance acquisition frequency, thereby converting the real-time acquisition process into an intermittent acquisition process, resulting in lower energy consumption.

[0098] Figure 3 The second sub-flowchart of the AI-based industrial robot control method includes the steps of acquiring a work area map, converting robot behavior into trajectories containing time periods, and inserting the work area map:

[0099] Step S201: Obtain the work area map and query the scale of the work area map;

[0100] Step S202: Read all robot behaviors, and based on the scale, query the point in the work area map corresponding to the starting device label as the starting point, and query the point in the work area map corresponding to the ending device label as the ending point;

[0101] Step S203: Determine the navigation path in the traffic segment based on the starting point and the destination, insert time period labels into the navigation path, and obtain the trajectory;

[0102] Step S204: Insert the trajectory into the work area map.

[0103] In one example of the technical solution of this invention, the process of generating and inserting the trajectory is described in detail: obtaining the work area map, querying the scale of the work area map, reading all robot behaviors, querying the point corresponding to the starting device label in the work area map based on the scale, using it as the starting point, querying the point corresponding to the ending device label in the work area map, using it as the ending point, determining the navigation path in the passage segment based on the starting point and the ending point, inserting time period labels into the navigation path to obtain the trajectory, and finally inserting the trajectory into the work area map.

[0104] Figure 4 The third sub-flowchart of the AI-based industrial robot control method includes the step of discretizing the work area map in the time domain to obtain a map set containing the robot's position at each time point:

[0105] Step S301: Create time points according to the preset time step;

[0106] Step S302: Read the time points sequentially, compare the time points with the time periods of each trajectory, and select the predicted point in the trajectory when the time point belongs to the time period.

[0107] Step S303: Query the robot corresponding to the trajectory and use the predicted point as the robot's coordinates at that moment;

[0108] Step S304: Calculate the coordinates of all robots at the same time point and insert them into the work area map as the map for that time point;

[0109] Step S305: Compile maps at all time points to obtain a map atlas;

[0110] In one example of the technical solution of this invention, the time-domain discretization process is described. Time points are created according to a preset time step, for example, a time point is created every 10 seconds. The time point is also a time point relative to time zero. The time points are read sequentially and compared with the time periods of each trajectory. When the time point belongs to a time period, a predicted point is selected in the trajectory. The predicted point is a position on the trajectory and is displayed in the form of coordinates. Then, the robot corresponding to the trajectory is queried, and the predicted point is used as the robot's coordinates at that time point. Then, for any time point, the coordinates of all robots at the same time point are counted and inserted into the work area map as the map of that time point. At this time, one time point corresponds to one map. By counting the maps of all time points, a map set is obtained.

[0111] The process of selecting predicted points in the trajectory is as follows: calculating the position ratio of the predicted points, and locating the predicted points on the trajectory according to the position ratio; the starting point of the trajectory is taken as the zero-ratio position; the process of determining the position ratio is as follows:

[0112] In the formula, For positional proportions, For the current time point, and These represent the left and right endpoints of the time period, respectively.

[0113] It should be noted that the step of discretizing the work area map in the time domain to obtain the map set containing the robot's position at each time point further includes:

[0114] Read the robot's coordinates at each moment from the coordinate library and update the predicted points.

[0115] In one example of the technical solution of this invention, if the robot detector uses a signal transceiver installed in each production device and robot, it can obtain the coordinates of the robot at each time point. That is, the coordinate library contains the coordinates of some robots at each time point. This is the actual position, which is much more accurate than the prediction result. At this time, using these actual positions as the predicted points at each time point has extremely high accuracy.

[0116] It is conceivable that the higher the sampling frequency, the more locations the robot acquires, and the more predicted points it obtains.

[0117] Figure 5 This is the fourth sub-process flowchart of the AI-based industrial robot control method. The step of determining the amount of AI resources at each time step based on the map atlas includes:

[0118] Step S401: For any given moment in the map, read all the coordinates in the map;

[0119] Step S402: Calculate the coordinate concentration based on all coordinates;

[0120] Step S403: Determine the proportion of AI resources based on the coordinate concentration and the total number of coordinates;

[0121] Step S404: Multiply the AI ​​resource ratio by the preset total resource amount to calculate the AI ​​resource quantity.

[0122] In one example of the technical solution of this invention, the application process of the atlas is described. For the map at any given time, all coordinates in the map are read. All coordinates reflect the distribution of the robots. The coordinate concentration is calculated based on all coordinates. The proportion of AI resources is determined based on the coordinate concentration and the total number of coordinates. The proportion of AI resources is a percentage. The AI ​​resource proportion is multiplied by the preset total amount of resources to calculate the amount of AI resources.

[0123] The process for determining the proportion of AI resources is as follows:

[0124] In the formula, for The proportion of AI resources at any given moment For the first One production cycle The proportion of AI resources at any given moment; This represents the total number of production cycles. ; For the first One production cycle The total number of coordinates at any given time. The threshold is a preset threshold, which is the maximum number of robots. For the first One production cycle In the time map, the first The minimum distance between each coordinate and other coordinates.

[0125] The process of determining the proportion of AI resources is influenced by all historical data, that is, all production cycles. Specifically, for the resource proportion at a certain moment, the resource proportions at the same moment across all production cycles are first calculated, and then the average is calculated as the final value. The proportion of AI resources at any given moment.

[0126] Specifically, for the first One production cycle The proportion of AI resources at any given moment This term represents the concentration degree. It calculates the minimum distance between any coordinate and all other coordinates, and then calculates the mean of the minimum distances of all coordinates. The larger the mean, the smaller the concentration degree; the smaller the distance, the larger the concentration degree. This value is within the range of zero to one. Simultaneously, it adds... This item, Take a limiting value that ensures the term is also within the range of zero to one, and that it is proportional to the total number of coordinates; furthermore, The overall concentration is also directly proportional to the total number of coordinates. The final result is that the greater the concentration and the more coordinates there are, the higher the proportion of AI resources.

[0127] Figure 6This is a block diagram illustrating the structural composition of an AI-based industrial robot control system. In this embodiment of the invention, an AI-based industrial robot control system 10 includes:

[0128] The behavior acquisition module 11 is used to acquire and statistically analyze all robot behaviors of the same production task in different production cycles; the robot behaviors include time period, starting device label and ending device label.

[0129] The trajectory generation and insertion module 12 is used to acquire the work area map, convert the robot behavior into a trajectory containing time periods, and insert it into the work area map; wherein, one production cycle corresponds to one work area map containing the trajectory.

[0130] Discretization module 13 is used to discretize the work area map in the time domain to obtain a map set containing the robot's position at each time.

[0131] The resource quantity determination module 14 is used to determine the AI ​​resource quantity at each time according to the map set; the AI ​​resource quantity is related to the distribution of robot positions in the map at each time, and is directly proportional to the total number of robot positions and the concentration of robot positions.

[0132] The demand identification and processing module 15 is used to identify demand information at each moment based on the amount of AI resources, generate robot control instructions, and synchronously generate robot behavior.

[0133] Furthermore, the behavior acquisition module 11 includes:

[0134] The zero-time setting unit is used to set the start point of the production cycle of any production task to zero time.

[0135] A robot detection unit is used to detect robots within a preset distance threshold in real time using a robot detection instrument built into the production equipment; the robot detection instrument includes a camera and a ranging radar.

[0136] The tag query unit is used to establish a connection channel with the robot when the robot exists within the preset distance threshold, and to query the device tag in the robot's tag register; when the device tag is empty, the tag of the current production device is used as the starting device tag and inserted into the robot.

[0137] The tag application unit is used to read the device tag when it is not empty, use it as the starting device tag, use the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period, and count the time period, starting device tag and ending device tag to obtain a robot behavior.

[0138] Specifically, the trajectory generation and insertion module 12 includes:

[0139] The scale query unit is used to obtain the work area map and query the scale of the work area map.

[0140] The start and end point determination unit is used to read all robot behaviors, query the point corresponding to the start device label in the work area map based on the scale, and use it as the start point; and query the point corresponding to the end device label in the work area map, and use it as the end point.

[0141] The generation unit is used to determine the navigation path in the traffic segment based on the start and end points, insert time period labels into the navigation path, and obtain the trajectory;

[0142] Insert unit, used to insert the trajectory into the work area map.

[0143] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI-based industrial robot control method, characterized in that, The method includes: Acquire and analyze all robot behaviors for the same production task across different production cycles; the robot behaviors include time period, starting device label, and ending device label. Obtain the work area map, convert robot behavior into trajectories containing time periods, and insert them into the work area map; one production cycle corresponds to one work area map containing trajectories. The work area map is discretized in the time domain to obtain a map set containing the robot's position at each time point; The amount of AI resources at each time point is determined based on the map set. The amount of AI resources is related to the distribution of robot positions on the map at each time point, and is directly proportional to the total number of robot positions and the concentration of robot positions. Based on the amount of AI resources, the demand information at each moment is identified to generate robot control instructions and simultaneously generate robot behavior. The step of determining the amount of AI resources at each time point based on the map atlas includes: For any given moment in the map, read all the coordinates in the map; Calculate the coordinate concentration based on all coordinates; The proportion of AI resources is determined based on the coordinate concentration and the total number of coordinates. The AI ​​resource quantity is calculated by multiplying the AI ​​resource percentage by the preset total resource quantity. The process for determining the proportion of AI resources is as follows: In the formula, for The proportion of AI resources at any given moment For the first One production cycle The proportion of AI resources at any given moment; This represents the total number of production cycles. ; For the first One production cycle The total number of coordinates at any given time. The threshold is a preset threshold, which is the maximum number of robots. For the first One production cycle In the time map, the first The minimum distance between each coordinate and other coordinates.

2. The AI-based industrial robot control method according to claim 1, characterized in that, The steps for acquiring and statistically analyzing all robot behaviors for the same production task across different production cycles include: Set the starting point of the production cycle for any production task to time zero; The robot is detected in real time within a preset distance threshold by a robot detection device built into the production equipment; the robot detection device includes a camera and a ranging radar. When a robot exists within the preset distance threshold, establish a connection channel with the robot and query the device tag in the robot's tag register; when the device tag is empty, insert the tag of the current production device as the starting device tag into the robot. When the device tag is not empty, read the device tag as the starting device tag, take the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period, and count the time period, starting device tag and ending device tag to obtain a robot behavior.

3. The AI-based industrial robot control method according to claim 2, characterized in that, The step of acquiring and statistically analyzing all robot behaviors for the same production task in different production cycles also includes: Signal transceivers are inserted into all production equipment and all robots as robot inspection instruments; The distance between the production equipment and the robot is determined based on the signal transceivers built into the production equipment and the signal transceivers built into the robot; the determined distance contains time stamps. The distance acquisition frequency for any production equipment and any robot is determined based on the distance between the production equipment and the robot; the distance acquisition frequency is inversely proportional to the distance. For any robot, query its distance to at least three production devices at each time point, determine the robot's coordinates at that time point, and store them in the coordinate library.

4. The AI-based industrial robot control method according to claim 1, characterized in that, The steps of acquiring the work area map, converting robot behavior into trajectories containing time periods, and inserting them into the work area map include: Get the workspace map and query the workspace map scale; Read all robot behaviors, and based on the scale, query the point in the work area map corresponding to the starting device label as the starting point, and query the point in the work area map corresponding to the ending device label as the ending point; The navigation route is determined based on the starting point and the destination within the travel segment. Time period labels are inserted into the navigation route to obtain the trajectory. Insert the trajectory into the workspace map.

5. The AI-based industrial robot control method according to claim 1, characterized in that, The step of discretizing the work area map in the time domain to obtain a map set containing the robot's position at each time point includes: Create time points based on preset time steps; Read the time points sequentially, compare the time points with the time periods of each trajectory, and select the predicted point in the trajectory when the time point belongs to the time period. The robot corresponding to the trajectory is queried, and the predicted point is used as the robot's coordinates at that time. Calculate the coordinates of all robots at the same time point, insert them into the work area map, and use them as the map at that time point; Compile maps at all points in time to obtain a map atlas; The process of selecting predicted points in the trajectory is as follows: calculating the position ratio of the predicted points, and locating the predicted points on the trajectory according to the position ratio; the starting point of the trajectory is taken as the zero-ratio position; the process of determining the position ratio is as follows: In the formula, For positional proportions, For the current time point, and These are the left and right endpoints of the time period, respectively; The step of discretizing the work area map in the time domain to obtain a map set containing the robot's position at each time point further includes: Read the robot's coordinates at each moment from the coordinate library and update the predicted points.

6. An AI-based industrial robot control system, characterized in that, The system includes: The behavior acquisition module is used to acquire and statistically analyze all robot behaviors for the same production task in different production cycles; the robot behaviors include time period, starting device label, and ending device label. The trajectory generation and insertion module is used to acquire the work area map, convert robot behavior into trajectories containing time periods, and insert them into the work area map; one production cycle corresponds to one work area map containing a trajectory. The discretization processing module is used to discretize the work area map in the time domain to obtain a map set containing the robot's position at each time point; The resource quantity determination module is used to determine the AI ​​resource quantity at each time point based on the map set. The AI ​​resource quantity is related to the distribution of robot positions on the map at each time point, and is directly proportional to the total number of robot positions and the concentration of robot positions. The demand identification and processing module is used to identify demand information at various times based on the amount of AI resources, generate robot control instructions, and synchronously generate robot behavior. The determination of AI resource quantity at each time point based on the map atlas includes: For any given moment in the map, read all the coordinates in the map; Calculate the coordinate concentration based on all coordinates; The proportion of AI resources is determined based on the coordinate concentration and the total number of coordinates. The AI ​​resource quantity is calculated by multiplying the AI ​​resource percentage by the preset total resource quantity. The process for determining the proportion of AI resources is as follows: In the formula, for The proportion of AI resources at any given moment For the first One production cycle The proportion of AI resources at any given moment; This represents the total number of production cycles. ; For the first One production cycle The total number of coordinates at any given time. The threshold is a preset threshold, which is the maximum number of robots. For the first One production cycle In the time map, the first The minimum distance between each coordinate and other coordinates.

7. The AI-based industrial robot control system according to claim 6, characterized in that, The behavior acquisition module includes: The zero-time setting unit is used to set the start point of the production cycle of any production task to zero time. A robot detection unit is used to detect robots within a preset distance threshold in real time using a robot detection instrument built into the production equipment; the robot detection instrument includes a camera and a ranging radar. The tag query unit is used to establish a connection channel with the robot when the robot exists within the preset distance threshold, and to query the device tag in the robot's tag register; when the device tag is empty, the tag of the current production device is used as the starting device tag and inserted into the robot. The tag application unit is used to read the device tag when it is not empty, use it as the starting device tag, use the tag of the current production device as the ending device tag, query the insertion time and reading time of the device tag, calculate the time period, and count the time period, starting device tag and ending device tag to obtain a robot behavior.

8. The AI-based industrial robot control system according to claim 6, characterized in that, The trajectory generation and insertion module includes: The scale query unit is used to obtain the work area map and query the scale of the work area map. The start and end point determination unit is used to read all robot behaviors, query the point corresponding to the start device label in the work area map based on the scale, and use it as the start point; and query the point corresponding to the end device label in the work area map, and use it as the end point. The generation unit is used to determine the navigation path in the traffic segment based on the start and end points, insert time period labels into the navigation path, and obtain the trajectory; Insert unit, used to insert the trajectory into the work area map.

9. A storage medium, characterized in that, The storage medium stores at least one piece of program code, which, when loaded and executed by the processor, implements the AI-based industrial robot control method as described in any one of claims 1-5.

Images