Robot behavior data analysis method and system based on artificial intelligence

Abstract

The invention discloses a robot behavior data analysis method and system based on artificial intelligence, and belongs to the technical field of robot management. The system comprises a data acquisition module, a behavior analysis module, an efficiency management module and a data visualization module. The data acquisition module is used for acquiring a driving route map of a factory area and historical logs and behavior data of the robot; the behavior analysis module is used for predicting the transportation time of each robot, planning a transportation route according to the driving route map, and analyzing the difference between different robots so as to calculate a difference index; the efficiency management module establishes a co-transport combination through the difference index, calculates the efficiency index of each robot in the co-transport combination to determine a transport object, the transport object undertakes the transfer of the materials loaded by other robots in the co-transport combination and carries out collaborative transport, and the co-transport combination is updated in real time to continue to hand over the materials until the transport is completed; and the data visualization module displays the information of the transportation object through a data center visualization large screen.

Description

Technical Field

[0001] The present invention relates to the field of robot management technology, and in particular to a robot behavior data analysis method and system based on artificial intelligence. Background Art

[0002] Robots are an important part of modern intelligent manufacturing and automated logistics. With the advancement of Industry 4.0, enterprises are increasingly demanding production efficiency and flexibility. Robots can be used to automate handling, warehouse management, and material transportation, significantly reducing labor costs and human errors, while improving the response speed and operational safety of production lines. Studying robot behavior data can help us understand the behavior patterns of robots in different environments and tasks, and optimize robots based on these data to improve their performance and reliability.

[0003] At present, strong distributed and weak centralized control is usually adopted for the behavior management of robots. Most of the control rights are handed over to the robots for autonomous decision-making. The interaction between different robots is weak or even non-existent. This method has certain disadvantages. For example, 1. Each robot performs loading and transportation work according to its own set logic program. In terms of task volume optimization, it can only achieve local optimality but not global optimality. The transportation work of different robots may be repeated and cause redundant work. 2. Due to the large differences in work tasks between different robots, the conditions for each robot to trigger transportation and the location of each starting point will be more complex and diverse compared to the overall planning, which makes it impossible to uniformly meet the transportation needs of all robots by setting a simple auxiliary transportation mechanism, resulting in a waste of resources. Therefore, at this stage, a more efficient and intelligent robot transportation management and scheduling technology solution is needed to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a robot behavior data analysis method and system based on artificial intelligence to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solution: a robot behavior data analysis method based on artificial intelligence, the method comprising the following steps:

[0006] S100, collecting the driving route map of the factory area, as well as the historical logs and current behavior data of all robots in the factory, and analyzing the position of each robot in the driving route map.

[0007] S200, using a machine learning algorithm to predict the transportation time of each robot, planning the transportation route according to the driving route map, analyzing the differences in transportation time and transportation routes between different robots to calculate the difference index.

[0008] S300. Establish a cooperative transport combination according to the difference index, analyze and calculate the efficiency index of each robot in the cooperative transport combination to determine the transport object, and the transport object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation.

[0009] S400, automatically dissolve the cooperative transport combination that has completed material transfer and establish a new cooperative transport combination based on the difference index, continue to hand over materials until the transportation is completed, and display the information of the transportation object through the large visual screen of the data center.

[0010] In S100, the driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials. The historical log refers to the operation record of the robot. Each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state. The transportation information records the historical transportation routes and historical transportation speeds of the robot in the transportation state. The behavior data includes the robot's state, position, moving direction and mass of the loaded materials. The state includes the loading state and the transportation state. The position of each robot is analyzed and marked in the driving route map.

[0011] The loading state and transport state of the robot are used to describe the different working states of the robot. The loading state refers to the state of the robot when it is working, including operations such as grabbing, carrying and placing materials. In this state, the mass of the material loaded by the robot is increasing. The transport state refers to the state of the robot when it is performing a transport task. This means that the robot has loaded the material and is moving to another location. In this state, the mass of the material loaded by the robot does not change.

[0012] In S200, the specific steps are as follows:

[0013] S201. Obtain all operation records under the history log of each robot, analyze the change of the mass of the loaded materials in the loading information in each operation record, mark the time points corresponding to the minimum and maximum material mass in chronological order, and connect the adjacent marked time points to establish different time periods. In each time period, there is only one situation in which the material mass changes from small to large or from large to small, and the time corresponding to the minimum and maximum material mass is at the two ends of the time period.

[0014] S202, filter out the time period corresponding to the maximum material mass time point later than the minimum material mass time point and mark them, analyze the robot position change in each marked time period and draw the historical movement trajectory, establish a coordinate set for each historical movement trajectory, intercept m trajectory coordinates on the historical movement trajectory and put them into the corresponding coordinate set, substitute them into the formula to calculate the difference coefficient CX between different historical movement trajectories, the formula is as follows:

[0015]

[0016] In the formula, is the sth track coordinate in the coordinate set corresponding to the pth historical moving track, It is the sth trajectory coordinate in the coordinate set corresponding to the lth historical moving trajectory.

[0017] On the historical movement trajectory, a sampling point is set at a fixed length in the moving direction. The number of sampling points and the fixed length are set by the management personnel according to the average length of the historical movement trajectory. The more sampling points and the shorter the fixed length, the higher the accuracy of the historical movement trajectory comparison.

[0018] S203. Set a coefficient threshold R, classify the marked time periods corresponding to the two historical movement trajectories whose difference coefficient is less than R into the same category, obtain the correlation data between the material quality and time in the marked time period, and use a machine learning algorithm to fit the correlation data of each type of marked time period respectively to obtain a relationship diagram between the material quality and time.

[0019] S204, obtaining the behavior data of each robot, analyzing the changes in position and moving direction and drawing the real-time moving trajectory, calculating the difference coefficient between the real-time moving trajectory and each historical moving trajectory, taking the historical moving trajectory LG with the smallest difference coefficient as the expected moving trajectory of the robot, and the relationship diagram corresponding to LG as the selected relationship diagram of the robot.

[0020] S205. Find the time point corresponding to the mass of the material currently loaded by the robot in the selected relationship diagram as the starting time point, and the time point with the maximum material mass as the ending time point. Calculate the difference between the starting time point and the ending time point and add the difference to the current time to obtain the transportation time of the corresponding robot. Calculate the difference index between different robots based on the transportation time.

[0021] The transport time refers to a predicted point in the future when the robot is expected to switch from the loading state to the moving state and start transporting the previously loaded materials.

[0022] In S205, the specific calculation steps of the difference index are as follows:

[0023] S205-1. The end point of the expected moving trajectory is used as the starting point YQ of the corresponding robot, and the transportation information in all the operation records of the robot is obtained. The historical transportation route starting from YQ is found, and the end point of the historical transportation route is used as the arrival point YZ of the robot, and the historical transportation speed is used as the current transportation speed of the robot. According to the positions of YQ and YZ of the robot, a shortest route is planned for each robot on the driving route map as the expected transportation route.

[0024] S205-2, set the prediction time set {t1, t2, ..., t u}, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t1,t2,...,t u}, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and substitute them into the formula to calculate the difference index between each two robot combinations:

[0025]

[0026] Where PX is the difference index, α is a constant greater than 1, and CX q is the difference coefficient between the transportation routes predicted by different robots, u is the number of all elements in the position set, is the kth predicted position coordinate in the corresponding position set of the dth robot, It is the kth predicted position coordinate in the corresponding position set of the jth robot.

[0027] Each time point in the predicted time set is later than the current time and the intervals between time points are the same. If multiple predicted times are earlier than the robot's transportation time, the robot may still be in the same position at different predicted times.

[0028] In S300, the specific steps are as follows:

[0029] S301, set the index threshold B, filter out robot combinations with difference index less than B, and establish cooperative operation combinations with these robot combinations in ascending order of difference index. One cooperative operation combination is established for every two robots, and the robots included in each cooperative operation combination are different. Get the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point.

[0030] S302, first set the meeting point HD n and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two The transportation route of each robot is calculated based on its transportation speed. xs, the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , substitute into the formula to calculate the RBT at each confluence point one and RBT two The efficiency index LZ formula is as follows:

[0031]

[0032] In the formula, is the expected transportation route of robot w, V w is the current transport speed of robot w, is the expected transportation route of robot e, V e is the current transport speed of robot e.

[0033] A waypoint refers to a specific location that must be passed when planning a transportation route. This location can be the end of the transportation route or an intermediate location.

[0034] S303, analyzing RBT at each confluence point one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding meeting point is taken as the handover point JJ; the transport object takes the shortest path re-planned based on JJ and the location of the other robot's arrival point as the waypoint as the current transport route. The transport object travels along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point from the handover point JJ ​​along the original route.

[0035] When one robot in the cooperative transport combination re-plans the transport route, the other robot does not need to re-plan. During the actual transportation process, only the robot that has re-planned the transport route will travel the entire journey, and the other robot only needs to reach the handover point to transfer the materials and then return along the original route.

[0036] In S400, after taking over the material transfer, the robot automatically disbands the original cooperative transport combination, calculates the difference index with other robots in real time, and continuously tries to re-establish the cooperative transport combination and re-plan the current transportation route according to the change of the difference index, and continues to take over or transfer materials until the transportation is completed. The location of each transport object and the quality of the loaded materials are displayed in real time on the data center visualization screen.

[0037] Both the robot that receives materials and the robot that transfers materials can continue to participate in the coordinated transport combination and reorganization. The robot that receives materials calculates the difference index based on the route after the handover point under the current transportation route, and the robot that transfers materials calculates the difference index based on the route returning from the handover point.

[0038] A robot behavior data analysis system based on artificial intelligence, which includes a data acquisition module, a behavior analysis module, an efficiency management module and a data visualization module.

[0039] The data acquisition module is used to collect the driving route map of the factory area, as well as the historical logs and behavior data of the robot. The behavior analysis module is used to predict the transportation time of each robot, plan the transportation route according to the driving route map, and analyze the differences between different robots to calculate the difference index. The efficiency management module establishes a cooperative transport combination through the difference index, calculates the efficiency index of each robot in the cooperative transport combination to determine the transportation object, and the transportation object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation, and updates the cooperative transport combination in real time to continue to hand over materials until the transportation is completed. The data visualization module displays the information of the transportation object through the visualization screen of the data center.

[0040] The data collection module includes a behavior data collection unit, a history log collection unit and a map information collection unit.

[0041] The behavior data acquisition unit is used to collect the state, position, moving direction and mass of the loaded materials of each robot, and the state includes loading state and transportation state.

[0042] The historical log collection unit is used to collect the operation records of each robot. Each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state, and the transportation information records the historical transportation route and historical transportation speed of the robot in the transportation state.

[0043] The map information collection unit is used to collect a driving route map. The driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials.

[0044] The behavior analysis module includes a time prediction unit and a difference analysis unit.

[0045] The time prediction unit is used to predict the transportation time of each robot.

[0046] Firstly, the changes in the mass of the loaded materials in the loading information of each operation record of the robot are analyzed, and the time points corresponding to the minimum and maximum material mass are marked in chronological order. The adjacent marked time points are connected to establish different time periods. The time periods when the material mass changes from small to large are marked, and the changes in the robot position in each marked time period are analyzed and the historical movement trajectory is drawn. The difference coefficients between different historical movement trajectories are calculated respectively.

[0047] Secondly, the marked time periods are classified according to the difference coefficients between the historical movement trajectories, and the machine learning algorithm is used to fit the relationship between the material mass and time in each marked time period to obtain the relationship diagram. The real-time movement trajectory is drawn according to the robot's behavior data, and the difference coefficient between the real-time movement trajectory and each historical movement trajectory is calculated. The historical movement trajectory LG with the smallest difference coefficient is used as the robot's expected movement trajectory, and the relationship diagram corresponding to LG is used as the robot's selected relationship diagram.

[0048] Finally, in the selected relationship diagram, find the time point corresponding to the mass of the material currently loaded by the robot as the starting time point, and the time point with the maximum material mass as the ending time point. Calculate the difference between the starting time point and the ending time point and add the current time to get the transportation time of the corresponding robot.

[0049] The difference analysis unit is used to analyze the differences between different robots and calculate the difference index.

[0050] First, the end point of the expected moving trajectory is taken as the starting point YQ of the corresponding robot, and the historical transportation route starting from YQ is found in the transportation information in the operation record. The end point of the historical transportation route is taken as the arrival point YZ of the robot, and the historical transportation speed is taken as the current transportation speed of the robot.

[0051] Then, according to the YQ and YZ positions of the robot, a shortest route is planned for each robot on the driving route map as the estimated transportation route. Set the prediction time set {t1, t2, ..., t u}, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t1,t2,...,t u}, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and calculate the difference index between each two robot combinations according to the position set.

[0052] The efficiency management module includes a collaborative planning unit and a dynamic adjustment unit.

[0053] The collaborative planning unit is used to establish collaborative transport combinations and determine transport objects.

[0054] First, set the index threshold B, filter out robot combinations with difference index less than B, and create a cooperative operation combination for every two robots in the order of difference index from small to large; obtain the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point.

[0055] Secondly, first set the junction point HD n and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two Calculate the driving time T of the robot in the new transportation route. xs , the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , calculate the RBT at each junction one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding junction point is taken as the intersection point JJ.

[0056] Finally, the transport object re-plans the shortest path based on the location of JJ and the arrival point of another robot as the waypoint as the current transport route. The transport object travels along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point by the original route.

[0057] The dynamic adjustment unit is used to automatically dissolve the cooperative transport combination after taking over the material transfer, and calculate the difference index between the robots in real time to continuously try to re-establish the cooperative transport combination to continue taking over or transferring materials until the transportation is completed.

[0058] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0059] 1. Intelligent transportation prediction: In this application, the historical loading information of each robot is analyzed, and the difference coefficients between the historical movement trajectories are comprehensively calculated to screen out the relationship diagram that best suits the current situation. The transportation time of each robot is predicted based on the current loading material quality, which is more intelligent and reliable than traditional mechanical predictions.

[0060] 2. Dynamic transportation planning: In this application, according to the predicted position of each robot at different times according to the transportation time and the expected transportation route, it is analyzed whether there is a coincidence in time and space to dynamically establish a cooperative transportation combination, set the handover points and transportation objects to achieve collaborative transportation, and compared with the existing technology, it makes more efficient and full use of transportation resources.

[0061] In summary, compared with traditional technologies, the present invention has the advantages of intelligent transportation prediction and dynamic transportation planning, and can improve the efficiency of robot transportation management and scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0063] Figure 1 It is a flow chart of the robot behavior data analysis method based on artificial intelligence of the present invention;

[0064] Figure 2 It is a structural schematic diagram of the robot behavior data analysis system based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] See also Figure 1 The present invention provides a robot behavior data analysis method based on artificial intelligence, the method comprising the following steps:

[0067] S100, collecting the driving route map of the factory area, as well as the historical logs and current behavior data of all robots in the factory, and analyzing the position of each robot in the driving route map.

[0068] S200, using a machine learning algorithm to predict the transportation time of each robot, planning the transportation route according to the driving route map, analyzing the differences in transportation time and transportation routes between different robots to calculate the difference index.

[0069] S300. Establish a cooperative transport combination according to the difference index, analyze and calculate the efficiency index of each robot in the cooperative transport combination to determine the transport object, and the transport object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation.

[0070] S400, automatically dissolve the cooperative transport combination that has completed material transfer and establish a new cooperative transport combination based on the difference index, continue to hand over materials until the transportation is completed, and display the information of the transportation object through the large visual screen of the data center.

[0071] In S100, the driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials. The historical log refers to the operation record of the robot. Each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state. The transportation information records the historical transportation routes and historical transportation speeds of the robot in the transportation state. The behavior data includes the robot's state, position, moving direction and mass of the loaded materials. The state includes the loading state and the transportation state. The position of each robot is analyzed and marked in the driving route map.

[0072] The loading state and transport state of the robot are used to describe the different working states of the robot. The loading state refers to the state of the robot when it is working, including operations such as grabbing, carrying and placing materials. In this state, the mass of the material loaded by the robot is increasing. The transport state refers to the state of the robot when it is performing a transport task. This means that the robot has loaded the material and is moving to another location. In this state, the mass of the material loaded by the robot does not change.

[0073] In S200, the specific steps are as follows:

[0074] S201. Obtain all operation records under the history log of each robot, analyze the change of the mass of the loaded materials in the loading information in each operation record, mark the time points corresponding to the minimum and maximum material mass in chronological order, and connect the adjacent marked time points to establish different time periods. In each time period, there is only one situation in which the material mass changes from small to large or from large to small, and the time corresponding to the minimum and maximum material mass is at the two ends of the time period.

[0075] S202, filter out the time period corresponding to the maximum material mass time point later than the minimum material mass time point and mark them, analyze the robot position change in each marked time period and draw the historical movement trajectory, establish a coordinate set for each historical movement trajectory, intercept m trajectory coordinates on the historical movement trajectory and put them into the corresponding coordinate set, substitute them into the formula to calculate the difference coefficient CX between different historical movement trajectories, the formula is as follows:

[0076]

[0077] In the formula, is the sth track coordinate in the coordinate set corresponding to the pth historical moving track, It is the sth trajectory coordinate in the coordinate set corresponding to the lth historical moving trajectory.

[0078] On the historical movement trajectory, a sampling point is set at a fixed length in the moving direction. The number of sampling points and the fixed length are set by the management personnel according to the average length of the historical movement trajectory. The more sampling points and the shorter the fixed length, the higher the accuracy of the historical movement trajectory comparison.

[0079] S203. Set a coefficient threshold R, classify the marked time periods corresponding to the two historical movement trajectories whose difference coefficient is less than R into the same category, obtain the correlation data between the material quality and time in the marked time period, and use a machine learning algorithm to fit the correlation data of each type of marked time period respectively to obtain a relationship diagram between the material quality and time.

[0080] S204, obtaining the behavior data of each robot, analyzing the changes in position and moving direction and drawing the real-time moving trajectory, calculating the difference coefficient between the real-time moving trajectory and each historical moving trajectory, taking the historical moving trajectory LG with the smallest difference coefficient as the expected moving trajectory of the robot, and the relationship diagram corresponding to LG as the selected relationship diagram of the robot.

[0081] S205. Find the time point corresponding to the mass of the material currently loaded by the robot in the selected relationship diagram as the starting time point, and the time point with the maximum material mass as the ending time point. Calculate the difference between the starting time point and the ending time point and add the difference to the current time to obtain the transportation time of the corresponding robot. Calculate the difference index between different robots based on the transportation time.

[0082] The transport time refers to a predicted point in the future when the robot is expected to switch from the loading state to the moving state and start transporting the previously loaded materials.

[0083] In S205, the specific calculation steps of the difference index are as follows:

[0084] S205-1. The end point of the expected moving trajectory is used as the starting point YQ of the corresponding robot, and the transportation information in all the operation records of the robot is obtained. The historical transportation route starting from YQ is found, and the end point of the historical transportation route is used as the arrival point YZ of the robot, and the historical transportation speed is used as the current transportation speed of the robot. According to the positions of YQ and YZ of the robot, a shortest route is planned for each robot on the driving route map as the expected transportation route.

[0085] S205-2, set the prediction time set {t1, t2, ..., t u}, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t1,t2,...,t u}, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and substitute them into the formula to calculate the difference index between each two robot combinations:

[0086]

[0087] Where PX is the difference index, α is a constant greater than 1, and CX q is the difference coefficient between the transportation routes predicted by different robots, u is the number of all elements in the position set, is the kth predicted position coordinate in the corresponding position set of the dth robot, It is the kth predicted position coordinate in the corresponding position set of the jth robot.

[0088] Each time point in the predicted time set is later than the current time and the intervals between time points are the same. If multiple predicted times are earlier than the robot's transportation time, the robot may still be in the same position at different predicted times.

[0089] In S300, the specific steps are as follows:

[0090] S301, set the index threshold B, filter out robot combinations with difference index less than B, and establish cooperative operation combinations with these robot combinations in ascending order of difference index. One cooperative operation combination is established for every two robots, and the robots included in each cooperative operation combination are different. Get the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point.

[0091] S302, first set the meeting point HDn and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two The transportation route of each robot is calculated based on its transportation speed. xs , the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , substitute into the formula to calculate the RBT at each confluence point one and RBT two The efficiency index LZ formula is as follows:

[0092]

[0093] In the formula, is the expected transportation route of robot w, V w is the current transport speed of robot w, is the expected transportation route of robot e, V e is the current transport speed of robot e.

[0094] A waypoint refers to a specific location that must be passed when planning a transportation route. This location can be the end of the transportation route or an intermediate location.

[0095] S303, analyzing RBT at each confluence point one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding meeting point is taken as the handover point JJ; the transport object takes the shortest path re-planned based on JJ and the location of the other robot's arrival point as the waypoint as the current transport route. The transport object travels along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point from the handover point JJ ​​along the original route.

[0096] When one robot in the cooperative transport combination re-plans the transport route, the other robot does not need to re-plan. During the actual transportation process, only the robot that has re-planned the transport route will travel the entire journey, and the other robot only needs to reach the handover point to transfer the materials and then return along the original route.

[0097] In S400, after taking over the material transfer, the robot automatically disbands the original cooperative transport combination, calculates the difference index with other robots in real time, and continuously tries to re-establish the cooperative transport combination and re-plan the current transportation route according to the change of the difference index, and continues to take over or transfer materials until the transportation is completed. The location of each transport object and the quality of the loaded materials are displayed in real time on the data center visualization screen.

[0098] Both the robot that receives materials and the robot that transfers materials can continue to participate in the coordinated transport combination and reorganization. The robot that receives materials calculates the difference index based on the route after the handover point under the current transportation route, and the robot that transfers materials calculates the difference index based on the route returning from the handover point.

[0099] See also Figure 2 The present invention provides a robot behavior data analysis system based on artificial intelligence, which includes a data acquisition module, a behavior analysis module, an efficiency management module and a data visualization module.

[0100] The data acquisition module is used to collect the driving route map of the factory area, as well as the historical logs and behavior data of the robot. The behavior analysis module is used to predict the transportation time of each robot, plan the transportation route according to the driving route map, and analyze the differences between different robots to calculate the difference index. The efficiency management module establishes a cooperative transport combination through the difference index, calculates the efficiency index of each robot in the cooperative transport combination to determine the transportation object, and the transportation object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation, and updates the cooperative transport combination in real time to continue to hand over materials until the transportation is completed. The data visualization module displays the information of the transportation object through the visualization screen of the data center.

[0101] The data collection module includes a behavior data collection unit, a history log collection unit and a map information collection unit.

[0102] The behavior data acquisition unit is used to collect the state, position, moving direction and mass of the loaded materials of each robot, and the state includes loading state and transportation state.

[0103] The historical log collection unit is used to collect the operation records of each robot. Each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state, and the transportation information records the historical transportation route and historical transportation speed of the robot in the transportation state.

[0104] The map information collection unit is used to collect a driving route map. The driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials.

[0105] The behavior analysis module includes a time prediction unit and a difference analysis unit.

[0106] The time prediction unit is used to predict the transportation time of each robot.

[0107] Firstly, the changes in the mass of the loaded materials in the loading information of each operation record of the robot are analyzed, and the time points corresponding to the minimum and maximum material mass are marked in chronological order. The adjacent marked time points are connected to establish different time periods. The time periods when the material mass changes from small to large are marked, and the changes in the robot position in each marked time period are analyzed and the historical movement trajectory is drawn. The difference coefficients between different historical movement trajectories are calculated respectively.

[0108] Secondly, the marked time periods are classified according to the difference coefficients between the historical movement trajectories, and the machine learning algorithm is used to fit the relationship between the material mass and time in each marked time period to obtain the relationship diagram. The real-time movement trajectory is drawn according to the robot's behavior data, and the difference coefficient between the real-time movement trajectory and each historical movement trajectory is calculated. The historical movement trajectory LG with the smallest difference coefficient is used as the robot's expected movement trajectory, and the relationship diagram corresponding to LG is used as the robot's selected relationship diagram.

[0109] Finally, in the selected relationship diagram, find the time point corresponding to the mass of the material currently loaded by the robot as the starting time point, and the time point with the maximum material mass as the ending time point. Calculate the difference between the starting time point and the ending time point and add the current time to get the transportation time of the corresponding robot.

[0110] The difference analysis unit is used to analyze the differences between different robots and calculate the difference index.

[0111] First, the end point of the expected moving trajectory is taken as the starting point YQ of the corresponding robot, and the historical transportation route starting from YQ is found in the transportation information in the operation record. The end point of the historical transportation route is taken as the arrival point YZ of the robot, and the historical transportation speed is taken as the current transportation speed of the robot.

[0112] Then, according to the YQ and YZ positions of the robot, a shortest route is planned for each robot on the driving route map as the estimated transportation route. Set the prediction time set {t1, t2, ..., t u}, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t1,t2,...,t u}, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and calculate the difference index between each two robot combinations according to the position set.

[0113] The efficiency management module includes a collaborative planning unit and a dynamic adjustment unit.

[0114] The collaborative planning unit is used to establish collaborative transport combinations and determine transport objects.

[0115] First, set the index threshold B, filter out robot combinations with difference index less than B, and create a cooperative operation combination for every two robots in the order of difference index from small to large; obtain the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point.

[0116] Secondly, first set the junction point HD n and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two Calculate the driving time T of the robot in the new transportation route. xs , the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , calculate the RBT at each junction one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding junction point is taken as the intersection point JJ.

[0117] Finally, the transport object re-plans the shortest path based on the location of JJ and the arrival point of another robot as the waypoint as the current transport route. The transport object travels along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point by the original route.

[0118] The dynamic adjustment unit is used to automatically dissolve the cooperative transport combination after taking over the material transfer, and calculate the difference index between the robots in real time to continuously try to re-establish the cooperative transport combination to continue taking over or transferring materials until the transportation is completed.

[0119] Embodiment 1:

[0120] Assume that there are two historical moving trajectories A1 and A2, and their trajectory coordinates are A1 trajectory coordinates: (2, 3), (3, 4), (4, 5); A2 trajectory coordinates: (3, 2), (3, 3), (3, 4); Substitute them into the formula to calculate the difference coefficient between A1 and A2 respectively:

[0121] Coefficient of variation:

[0122] Then the difference index between A1 and A2 is 3.83.

[0123] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0124] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Robot behavior data analysis method based on artificial intelligence, Features: The method comprises the following steps: S100, collecting the driving route map of the factory area, as well as the historical logs and current behavior data of all robots in the factory, and analyzing the position of each robot in the driving route map; S200, using a machine learning algorithm to predict the transportation time of each robot, planning the transportation route according to the driving route map, analyzing the differences in transportation time and transportation routes between different robots to calculate a difference index; S300, establishing a cooperative transport combination according to the difference index, analyzing and calculating the efficiency index of each robot in the cooperative transport combination to determine the transport object, and the transport object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation; S400, automatically dissolve the cooperative transport combination that has completed material transfer and establish a new cooperative transport combination based on the difference index, continue to hand over materials until the transportation is completed, and display the information of the transportation object through the large visual screen of the data center.

2. The robot behavior data analysis method based on artificial intelligence according to claim 1, Features: In S100, the driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials; the historical log refers to the operation record of the robot, each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state, and the transportation information records the historical transportation routes and historical transportation speeds of the robot in the transportation state; the behavior data includes the robot's state, position, moving direction and mass of the loaded materials. The state includes the loading state and the transportation state. The position of each robot is analyzed and marked in the driving route map.

3. The robot behavior data analysis method based on artificial intelligence according to claim 2, Features: In S200, the specific steps are as follows: S201, obtaining all operation records under the historical log of each robot, analyzing the change of the mass of the loaded materials in the loading information in each operation record, marking the time points corresponding to the minimum and maximum mass of the materials in chronological order, and connecting adjacent marked time points to establish different time periods; S202, filter out the time period corresponding to the maximum material mass time point later than the minimum material mass time point and mark them, analyze the robot position change in each marked time period and draw the historical movement trajectory, establish a coordinate set for each historical movement trajectory, intercept m trajectory coordinates on the historical movement trajectory and put them into the corresponding coordinate set, substitute them into the formula to calculate the difference coefficient CX between different historical movement trajectories, the formula is as follows: In the formula, is the sth track coordinate in the coordinate set corresponding to the pth historical moving track, is the sth trajectory coordinate in the coordinate set corresponding to the lth historical moving trajectory; S203, setting a coefficient threshold R, classifying the marked time periods corresponding to the two historical movement trajectories whose difference coefficient is less than R into the same category, obtaining the correlation data between the material quality and time in the marked time period, and using a machine learning algorithm to fit the correlation data of each type of marked time period respectively, to obtain a relationship diagram between the material quality and time; S204, obtaining the behavior data of each robot, analyzing the changes in the position and movement direction and drawing the real-time movement trajectory, calculating the difference coefficient between the real-time movement trajectory and each historical movement trajectory, taking the historical movement trajectory LG with the smallest difference coefficient as the expected movement trajectory of the robot, and the relationship diagram corresponding to LG as the selected relationship diagram of the robot; S205. Find the time point corresponding to the mass of the material currently loaded by the robot in the selected relationship diagram as the starting time point, and the time point with the maximum material mass as the ending time point. Calculate the difference between the starting time point and the ending time point and add the difference to the current time to obtain the transportation time of the corresponding robot. Calculate the difference index between different robots based on the transportation time.

4. The robot behavior data analysis method based on artificial intelligence according to claim 3, Features: In S205, the specific calculation steps of the difference index are as follows: S205-1. The end point of the expected moving trajectory is used as the starting point YQ of the corresponding robot, and the transportation information in all the operation records of the robot is obtained. The historical transportation route starting from YQ is found, and the end point of the historical transportation route is used as the arrival point YZ of the robot, and the historical transportation speed is used as the current transportation speed of the robot; according to the positions of YQ and YZ of the robot, a shortest route is planned for each robot on the driving route map as the expected transportation route; S205-2. Set the prediction time set {t 1 ,t 2 ,...,t u }, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t 1 ,t 2 ,...,t u }, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and substitute them into the formula to calculate the difference index between each two robot combinations: Where PX is the difference index, α is a constant greater than 1, and CX q is the difference coefficient between the transportation routes predicted by different robots, u is the number of all elements in the position set, is the kth predicted position coordinate in the corresponding position set of the dth robot, It is the kth predicted position coordinate in the corresponding position set of the jth robot.

5. The robot behavior data analysis method based on artificial intelligence according to claim 3, Features: In S300, the specific steps are as follows: S301, set the index threshold B, filter out robot combinations with difference index less than B, and establish cooperative operation combinations with these robot combinations in ascending order of difference index, with each two robots forming a cooperative operation combination, and the robots included in each cooperative operation combination are different; obtain the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point; S302, first set the meeting point HD n and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two The transportation route of each robot is calculated according to its transportation speed. xs , the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , substitute into the formula to calculate the RBT at each confluence point one and RBT two The efficiency index LZ formula is as follows: In the formula, is the expected transportation route of robot w, V w is the current transport speed of robot w, is the expected transportation route of robot e, V e is the current transport speed of robot e; S303, analyzing RBT at each confluence point one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding junction point is taken as the intersection point JJ; The transport object takes the shortest path re-planned based on the location of JJ and the arrival point of another robot as the waypoint as the current transport route. The transport object travels along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point from the handover point JJ ​​along the original route.

6. The robot behavior data analysis method based on artificial intelligence according to claim 5, Features: In S400, after taking over the material transfer, the robot automatically disbands the original cooperative transport combination, calculates the difference index with other robots in real time, and continuously tries to re-establish the cooperative transport combination and re-plan the current transportation route according to the changes in the difference index, and continues to take over or transfer materials until the transportation is completed; the location of each transport object and the quality of the loaded materials are displayed in real time through the data center's large visualization screen.

7. Robot behavior data analysis system based on artificial intelligence, Features: The system includes a data acquisition module, a behavior analysis module, an efficiency management module and a data visualization module; The data acquisition module is used to collect the driving route map of the factory area, as well as the historical logs and behavior data of the robot; the behavior analysis module is used to predict the transportation time of each robot, plan the transportation route according to the driving route map, and analyze the differences between different robots to calculate the difference index; the efficiency management module establishes a cooperative transport combination through the difference index, calculates the efficiency index of each robot in the cooperative transport combination to determine the transportation object, and the transportation object undertakes the transfer of materials loaded by other robots in the cooperative transport combination and cooperates in transportation, and updates the cooperative transport combination in real time to continue to hand over materials until the transportation is completed; the data visualization module displays the information of the transportation object through the large visualization screen of the data center.

8. The robot behavior data analysis system based on artificial intelligence according to claim 7, Features: The data collection module includes a behavior data collection unit, a history log collection unit and a map information collection unit; The behavior data acquisition unit is used to collect the status, position, moving direction and mass of the loaded materials of each robot. The status includes loading status and transportation status. The historical log collection unit is used to collect the operation records of each robot; each operation record includes loading information and transportation information. The loading information records the mass and location of the loaded materials at different times when the robot is in the loading state, and the transportation information records the historical transportation route and historical transportation speed of the robot in the transportation state. The map information collection unit is used to collect a driving route map. The driving route map refers to a map used to display the drivable paths of all robots in the factory when transporting materials.

9. The robot behavior data analysis system based on artificial intelligence according to claim 8, Features: The behavior analysis module includes a time prediction unit and a difference analysis unit; The time prediction unit is used to predict the transportation time of each robot; First, the changes in the mass of the loaded materials in the loading information of each operation record of the robot are analyzed, and the time points corresponding to the minimum and maximum mass of the materials are marked in chronological order. The adjacent marked time points are connected to establish different time periods; the time periods when the mass of the materials changes from small to large are marked, and the changes in the position of the robot in each marked time period are analyzed and the historical movement trajectory is drawn, and the difference coefficients between different historical movement trajectories are calculated respectively; Secondly, the marked time periods are classified according to the difference coefficients between the historical movement trajectories, and the relationship diagrams of the material mass and time in each marked time period are fitted by a machine learning algorithm; the real-time movement trajectory is drawn according to the robot's behavior data, and the difference coefficient between the real-time movement trajectory and each historical movement trajectory is calculated. The historical movement trajectory LG with the smallest difference coefficient is used as the robot's expected movement trajectory, and the relationship diagram corresponding to LG is used as the robot's selected relationship diagram; Finally, in the selected relationship diagram, find the time point corresponding to the mass of the material currently loaded by the robot as the starting time point, and the time point with the maximum mass of the material as the ending time point. Calculate the difference between the starting time point and the ending time point and add the current time to get the transportation time of the corresponding robot; The difference analysis unit is used to analyze the differences between different robots and calculate the difference index; First, the end point of the expected moving trajectory is used as the starting point YQ of the corresponding robot. The historical transportation route starting from YQ is found in the transportation information in the operation record. The end point of the historical transportation route is used as the arrival point YZ of the robot, and the historical transportation speed is used as the current transportation speed of the robot. Then, according to the YQ and YZ positions of the robot, a shortest route is planned for each robot on the driving route map as the expected transportation route; Set the prediction time set {t 1 ,t 2 ,...,t u }, analyze the robot's travel time from the transportation time along the expected transportation route at the current transportation speed at {t 1 ,t 2 ,...,t u }, establish a position set for each robot, put these predicted position coordinates into the corresponding position set, and calculate the difference index between each two robot combinations according to the position set.

10. The robot behavior data analysis system based on artificial intelligence according to claim 9, Features: The efficiency management module includes a collaborative planning unit and a dynamic adjustment unit; The collaborative planning unit is used to establish collaborative transport combinations and determine transport objects; First, set the index threshold B, filter out robot combinations with difference index less than B, and create a cooperative operation combination for every two robots in the order of difference index from small to large; obtain the RBT of the robots in the cooperative operation combination one and robot RBT two The predicted position coordinates at the same time in the two position sets are mapped to the driving route map for the shortest route connection. According to RBT one and RBT two The transport speed is marked on each line at the junction point; Secondly, first set the junction point HD n and RBT two The location of the arrival point is used as a waypoint to re-plan the RBT one The transport route, and then the meeting point HD n and RBT one The location of the arrival point is used as a waypoint to re-plan the RBT two transportation routes; Calculate the driving time T of the robot in the new transport route respectively xs , the travel time of the shortest route from the arrival point to the departure point on the new transport route is T gs , and the driving time T of the other robot returning to the original route after reaching the meeting point fh , calculate the RBT at each junction one and RBT two efficiency index, select the highest efficiency index LZ max The robot is used as the transportation object, LZ max The corresponding junction point is taken as the intersection point JJ; Finally, the transport object re-plans the shortest path based on the location of JJ and the other robot's arrival point as the waypoint as the current transport route. The transport object drives along the current transport route and undertakes the material transfer of another robot in the cooperative transport combination at the handover point JJ. After completing the material transfer, the other robot returns to the starting point along the original route. The dynamic adjustment unit is used to automatically dissolve the cooperative transport combination after taking over the material transfer, and calculate the difference index between the robots in real time to continuously try to re-establish the cooperative transport combination to continue taking over or transferring materials until the transportation is completed.

Images