Intelligent scheduling method and system for charging pile automated assembly line

Abstract

The present invention relates to the technical field of assembly line scheduling, and in particular to an intelligent scheduling method and system for an automated assembly line of a charging pile. The method comprises the following steps: obtaining a charging pile assembly production log; performing sliding window segmentation processing on the charging pile assembly production log, and performing staged assembly demand feature analysis, thereby obtaining assembly demand feature data at different stages; extracting the real-time production progress of each assembly line based on the charging pile assembly production log; predicting the demand-supply deviation of the real-time production progress of each assembly line according to the assembly demand feature data at different stages, and marking assembly lines with insufficient production capacity; collecting all-round real-time monitoring images of the assembly line; performing dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and performing process operation logic analysis to generate multi-stage process operation logic. The present invention realizes efficient and adaptive intelligent scheduling of assembly lines.

Description

Technical Field

[0001] The present invention relates to the technical field of assembly line scheduling, and in particular to an intelligent scheduling method and system for an automatic assembly line of charging piles. Background Art

[0002] In today's industrial automation field, charging piles are important equipment for charging electric vehicles, and the degree of automation of assembly lines in their production process is increasing. The automated assembly line of charging piles requires efficient scheduling methods to ensure the stable operation and efficient production of the production line. With the expansion of production scale and the continuous improvement of production efficiency requirements, traditional manual scheduling methods can no longer meet the needs of modern production.

[0003] During the operation of the charging pile automated assembly line, factors such as the process parameters, assembly sequence, and equipment status of each assembly link will affect the production efficiency and quality of the entire production line. In addition, external factors such as equipment failure, delayed parts supply, and personnel deployment also affect the operation of the assembly line. Intelligent scheduling methods are needed to achieve efficient operation of automated production.

[0004] Traditional assembly line scheduling methods usually rely on manually formulated static scheduling plans, which are difficult to adapt to the dynamic changes in the production environment and the needs of real-time adjustments. This method has the problems of low scheduling efficiency and poor adaptability. An intelligent scheduling method is needed to achieve dynamic optimization scheduling of assembly lines. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes an intelligent scheduling method and system for an automated assembly line of charging piles to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides an intelligent scheduling method for an automatic charging pile assembly line, comprising the following steps:

[0007] Step S1: Obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform staged assembly demand feature analysis to obtain assembly demand feature data at different stages;

[0008] Step S2: extracting the real-time production progress of each assembly line based on the charging pile assembly production log; predicting the demand-supply deviation of the real-time production progress of each assembly line according to the assembly demand characteristic data at different stages, and marking the assembly lines with insufficient production capacity;

[0009] Step S3: collecting all-round real-time monitoring images of the assembly line; performing dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and performing process operation logic analysis to generate multi-stage process operation logic;

[0010] Step S4: performing potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and performing dynamic avoidance path planning to generate a dynamic avoidance assembly path;

[0011] Step S5: obtaining a global scene graph of the assembly line based on the omnidirectional real-time monitoring image of the assembly line; performing dynamic behavior mapping processing on the global scene graph of the assembly line according to the dynamic avoidance assembly path to construct a dynamic assembly twin model;

[0012] Step S6: Based on the assembly line with insufficient production capacity, the dynamic assembly twin model is subjected to dynamic collaborative scheduling iterative analysis, and assembly line cluster collaborative control is performed to build a dynamic cluster scheduling model and execute the assembly line intelligent scheduling method operation.

[0013] The present invention understands the assembly demand characteristics of different stages through sliding window segmentation processing and staged assembly demand characteristic analysis, provides data support for subsequent scheduling, obtains assembly demand characteristic data of different stages to optimize the production plan of the assembly line, improves assembly efficiency and capacity utilization, predicts demand-supply deviation for each assembly line, and promptly discovers capacity shortages, which is conducive to adjusting the production plan to avoid production delays, marks assembly lines with capacity shortages for key attention and priority scheduling, and ensures that the production capacity of the assembly line reaches the optimal state. Dynamic behavior analysis and process operation logic analysis understand the real-time production situation of the assembly line, improve the visual monitoring and control capabilities of the production process, and generate multi-stage process operation logic to provide a basis for subsequent obstacle inference and path planning, dynamic avoidance Path planning avoids collisions and blockages during the assembly process, improves assembly efficiency and safety, handles potential obstacles and performs dynamic avoidance path planning to optimize the assembly process and reduce unexpected events and pauses in production. The dynamic assembly twin model combines real-time monitoring images and dynamic avoidance paths to reflect the status and operation of the assembly line in real time, providing accurate data support for intelligent scheduling. Dynamic behavior mapping processing associates monitoring data with the actual assembly process to improve scheduling accuracy and efficiency. Dynamic collaborative scheduling iterative analysis and assembly line cluster collaborative control realize collaborative operations and resource optimization between assembly lines, improve overall production efficiency, and build a dynamic cluster scheduling model to realize intelligent scheduling and optimization of assembly lines, ensuring efficient operation of the production process and capacity utilization.

[0014] Preferably, step S1 comprises the following steps:

[0015] Step S11: Obtain charging pile assembly production log;

[0016] Step S12: performing time stamp standardization processing on the charging pile assembly production log to obtain a time standardized production log;

[0017] Step S13: performing sliding window segmentation processing on the time-standardized production log, thereby obtaining multiple time window log segments;

[0018] Step S14: performing time series analysis on multiple time window log segments to generate multiple log segment time series feature data;

[0019] Step S15: performing phased assembly requirement feature analysis on multiple log segment time series feature data, thereby obtaining assembly requirement feature data at different stages.

[0020] The present invention obtains assembly production logs to obtain real-time production data, providing a basis for subsequent analysis and scheduling, standardizing timestamps to unify the formats of different time records, ensuring the accuracy and consistency of subsequent data processing, and dividing production logs into multiple time periods through time window segmentation processing to perform more fine-grained analysis and monitoring of the production process. Time series analysis extracts feature data within each time window to provide detailed information for subsequent demand feature analysis and assembly scheduling. By performing staged assembly demand feature analysis on time series feature data, the assembly demand features at different stages can be understood, the assembly process can be optimized, and production efficiency can be improved.

[0021] Preferably, step S2 comprises the following steps:

[0022] Step S21: performing assembly demand forecasting on the assembly demand characteristic data at different stages, thereby obtaining demand forecasting data for multiple time periods;

[0023] Step S22: extracting the real-time production progress of each assembly line based on the charging pile assembly production log;

[0024] Step S23: Calculate the production completion rate of each assembly line based on the real-time production progress to obtain the production completion rate of each assembly line;

[0025] Step S24: analyzing the production capacity characteristics of the assembly line process based on the production completion rate of each assembly line to obtain the production capacity characteristics of each assembly line;

[0026] Step S25: performing capacity matching calculation for the corresponding time period on the capacity characteristics of each assembly line based on multiple time window log segments to obtain the assembly line capacity for the corresponding time period;

[0027] Step S26: Based on the demand forecast data for multiple time periods, the demand-supply deviation of the assembly line capacity in the corresponding time period is predicted, and the assembly line with insufficient capacity is marked.

[0028] The present invention predicts the characteristic data of assembly demand at different stages, understands the assembly demand in future time periods in advance, is conducive to adjusting the production plan to meet the demand fluctuation, obtains the demand forecast data of multiple time periods to optimize the scheduling of assembly lines, improves the assembly efficiency and production response speed, extracts the production progress of the assembly line in real time, can timely understand the production status of each assembly line, and provide real-time data support for subsequent production analysis and scheduling, calculates the production completion rate of the assembly line to evaluate the production efficiency and progress of the assembly line, helps monitor the production progress, timely discovers production anomalies and takes measures, performs capacity characteristic analysis based on the production completion rate to understand the production capacity of each assembly line, and provides a basis for subsequent capacity matching and scheduling, and obtains the capacity of each assembly line in each time period by matching and calculating the capacity characteristics of different time periods, optimizes production scheduling and resource allocation, and predicts the demand and supply deviation for the assembly line capacity in the corresponding time period, can timely discover the capacity shortage, is conducive to adjusting the production plan to avoid production bottlenecks, marks the assembly line with capacity shortage for key attention and priority scheduling, and ensures that the production capacity of the assembly line reaches the optimal state.

[0029] Preferably, the specific steps of step S21 are:

[0030] Perform assembly line demand distribution statistics on the assembly demand characteristic data at different stages to obtain the demand characteristic data of each assembly line;

[0031] Perform periodic demand fluctuation analysis on the demand characteristic data of each assembly line to obtain the assembly demand fluctuation data of different periods;

[0032] The demand trend evolution is carried out on the assembly demand fluctuation data in different periods, so as to obtain the assembly demand trend evolution law;

[0033] Based on the evolution law of assembly demand trend, multi-period assembly demand forecasting is performed on time-standardized production logs to obtain demand forecast data for multiple periods.

[0034] The present invention understands the characteristics of assembly demand at different stages by statistically analyzing the distribution of assembly line demand, provides data support for formulating reasonable scheduling plans, obtains demand characteristic data for each assembly line, helps to reasonably allocate assembly resources, improves production efficiency and quickly responds to changes in demand, and periodic demand fluctuation analysis reveals the periodic change law of assembly demand, helps predict the load change of assembly lines, obtains assembly demand fluctuation data of different periods to formulate more effective production scheduling strategies, improves assembly line resource utilization, and demand trend evolution analysis predicts the trend development of assembly demand, helps adjust production plans to adapt to future changes in demand, obtains the assembly demand trend evolution law to timely adjust assembly line resource allocation and production plans, and multi-period assembly demand forecasting predicts the demand for each future period according to the trend law, helps to reasonably arrange production resources, and obtains demand forecast data for multiple time periods to optimize the production plan of the assembly line, thereby improving production efficiency and the punctuality of product delivery.

[0035] Preferably, step S3 specifically comprises the following steps:

[0036] Step S31: collecting all-round real-time monitoring images of the assembly line;

[0037] Step S32: performing global balanced brightness enhancement processing on the omnidirectional real-time monitoring image of the assembly line to generate a global enhanced monitoring image;

[0038] Step S33: performing dynamic behavior analysis on the global enhanced monitoring image to obtain dynamic behavior characteristic data of the assembly line;

[0039] Step S34: performing phased process identification on the dynamic behavior characteristic data of the assembly line, thereby obtaining process behavior characteristics of different stages;

[0040] Step S35: Perform process operation logic analysis on process behavior characteristics at different stages to generate multi-stage process operation logic.

[0041] The present invention helps to monitor the operation status of the assembly line in real time through real-time monitoring image acquisition, promptly discovers abnormal situations, provides data support for subsequent steps through image acquisition, and realizes comprehensive monitoring and analysis of the assembly line. Global enhancement processing improves the quality and clarity of the image, reduces noise and improves the recognizability of the image, and generates a globally enhanced monitoring image to more accurately analyze the dynamic behavior of the assembly line, providing better input for subsequent steps. Dynamic behavior analysis identifies various dynamic behaviors in the assembly line, including movement, operation, etc., so as to understand the real-time operation of the assembly line, obtain dynamic behavior feature data to monitor the operation status of the assembly line, promptly discover abnormalities or bottlenecks, and provide information support for intelligent scheduling. Process identification decomposes dynamic behavior data into different process stages, helps to understand the operation flow and characteristics of each stage, obtains process behavior characteristics to distinguish process behaviors at different stages, and provides a basis for subsequent operation logic analysis. Process operation logic analysis analyzes the relationship and sequence between process behaviors at different stages, helps to optimize the operation logic of the assembly line, generates multi-stage process operation logic to optimize the assembly process, improves the efficiency and accuracy of the assembly line, and thus achieves the goal of intelligent scheduling.

[0042] Preferably, the specific steps of step S4 are:

[0043] Step S41: Performing object visual detection processing on the global enhanced monitoring image, marking assembly equipment and component objects;

[0044] Step S42: performing optical flow movement tracking on the assembly equipment and component objects to generate an optical flow movement trajectory of the object;

[0045] Step S43: performing potential obstacle inference processing on the object optical flow movement trajectory to obtain obstacle nodes;

[0046] Step S44: predicting the operation time of each stage process based on the multi-stage process operation logic of the object optical flow movement trajectory, and extracting the predicted value of the process time of each stage;

[0047] Step S45: Perform dynamic avoidance path planning on obstacle nodes according to the predicted process time consumption value of each stage to generate a dynamic avoidance assembly path.

[0048] The present invention helps to mark assembly equipment and component objects through object visual detection processing, realizes the recognition and marking of objects on the assembly line, marks the execution of subsequent steps of assembly equipment and component objects, and provides basic data for assembly line status monitoring and optimization. Optical flow movement tracking generates optical flow movement trajectory of objects, helps to track the movement path of objects on the assembly line, generates optical flow movement trajectory to understand the movement of assembly equipment and component objects, and provides data support for subsequent obstacle inference. Obstacle inference processing is performed on the optical flow movement trajectory to identify potential obstacle nodes, and operation risks are discovered in advance. Obstacle node information is obtained to provide a basis for subsequent path planning, and dynamic avoidance strategy in intelligent scheduling is realized. The process operation time consumption is predicted for the optical flow movement trajectory to estimate the process time consumption of each stage, which helps to optimize the operation efficiency of the assembly line, extracts the predicted value of the stage process time consumption to realize more accurate assembly path planning, improves production efficiency and quality, and performs dynamic avoidance path planning according to the predicted value of the process time consumption to optimize the assembly path, avoids obstacle interference, and improves production efficiency. The dynamic avoidance assembly path is generated to realize real-time dynamic adjustment in intelligent scheduling, and ensures the smooth operation of the assembly line.

[0049] Preferably, the specific steps of step S43 are:

[0050] Perform trajectory density distribution analysis on the object's optical flow movement trajectory to generate trajectory density distribution data;

[0051] Mark the location points of all objects according to their optical flow movement trajectory;

[0052] The number of times all objects stay at their locations is counted along multiple trajectories to obtain the number of times the location points stay;

[0053] Calculating the track point dwell time for all the object stop positions to generate multiple track point dwell times;

[0054] Perform potential obstacle analysis on the trajectory density distribution data according to the number of stop times at the location point and the stop time of multiple trajectory points, and extract multiple potential obstacle trajectory points;

[0055] Based on multiple potential obstacle trajectory points, the global enhanced monitoring image is segmented into bounding boxes to extract the potential obstacle bounding box image;

[0056] Obstacle depth convolution is performed on the potential obstacle bounding box image to obtain the obstacle node.

[0057] The present invention helps to understand the movement and distribution patterns of objects on the assembly line through trajectory density distribution analysis, provides basic data for subsequent analysis, generates trajectory density distribution data to identify dense trajectory areas on the assembly line, suggests potential obstacles or bottleneck locations, marks object stop locations to help determine the static locations of objects on the assembly line, provides key data points for subsequent analysis, identifies stop locations to understand the stop conditions of objects on the assembly line, provides support for obstacle analysis and path planning, counts the number of stop times for multiple trajectories to reveal the busyness of different locations on the assembly line, helps identify existing bottleneck areas, obtains stop number information to optimize the layout and process of the assembly line, improves production efficiency, calculates the stop time of trajectory points to understand the length of time objects stay at different locations, and helps evaluate the movement of the assembly line. The intelligent scheduling system can monitor and avoid obstacles in real time to ensure the smooth operation of the assembly line. The bounding box segmentation helps to accurately locate the position of potential obstacles and visually display the existence of obstacles more intuitively. The deep convolution inference and obstacle node recognition of the potential obstacle bounding box image are extracted. The obstacle deep convolution inference can more accurately identify obstacles and help the intelligent scheduling system to monitor and handle obstacles on the assembly line in real time. Obstacle node information is obtained, which provides an important reference for path planning and assembly line scheduling, and improves production efficiency and safety.

[0058] Preferably, the specific steps of step S5 are:

[0059] Step S51: performing global scene convolution recognition on the omnidirectional real-time monitoring image of the assembly line to obtain a global scene graph of the assembly line;

[0060] Step S52: performing three-dimensional point cloud modeling on the global scene graph of the assembly line to generate a global scene point cloud model;

[0061] Step S53: dynamically evolving the dynamic avoidance assembly path according to the process behavior characteristics of different stages to generate dynamic process behavior evolution data;

[0062] Step S54: Perform dynamic behavior mapping processing on the global scene point cloud model based on the dynamic process behavior evolution data to construct a dynamic assembly twin model.

[0063] The present invention uses global scene convolution recognition to quickly and accurately identify various objects, equipment and environmental information on the assembly line, provide data support for subsequent analysis, and obtain a global scene graph to help the system understand the layout and status of the entire assembly line, provide real-time monitoring and decision support for the intelligent scheduling system, and use three-dimensional point cloud modeling to convert the global scene of the assembly line into a digital three-dimensional model to provide richer information for subsequent data processing and analysis. The generation of a global scene point cloud model helps the system to more intuitively understand the spatial structure and layout of the assembly line, and provides a basis for intelligent scheduling and path planning. Through the analysis and evolution of process behavior characteristics, the assembly path is adjusted according to actual conditions, obstacles are avoided or the production process is optimized, and assembly efficiency is improved. Dynamic process behavior evolution data is generated to help the system achieve flexible path planning and scheduling, and adapt to changes in different production situations. Dynamic behavior mapping processing combines dynamic process behavior with the global scene point cloud model to achieve real-time updating and reflection of the assembly line status, and constructs a dynamic assembly twin model to help the system simulate and optimize the assembly process, and improve the intelligent scheduling system's understanding and response capabilities to the assembly line.

[0064] Preferably, the specific steps of step S6 are:

[0065] Step S61: performing assembly operation simulation on the dynamic assembly twin model based on the insufficient capacity assembly line to generate insufficient assembly line operation simulation data;

[0066] Step S62: Perform dynamic collaborative scheduling iterative analysis on the simulation data of the defective assembly line operation to generate multiple dynamic collaborative scheduling strategies;

[0067] Step S63: performing assembly collaborative operation efficiency evaluation on multiple dynamic collaborative scheduling strategies to generate operation efficiency evaluation values ​​of multiple strategies;

[0068] Step S64: extracting the optimal dynamic collaborative scheduling strategy based on the operation efficiency evaluation values ​​of multiple strategies;

[0069] Step S65: fine-tuning the dynamic parameters of the charging pile automated assembly line using the optimal dynamic collaborative scheduling strategy to generate a dynamic assembly fine-tuning engine;

[0070] Step S66: Perform assembly line cluster collaborative control on the dynamic assembly fine-tuning engine, build a dynamic cluster scheduling model, and execute the assembly line intelligent scheduling method operation.

[0071] The present invention simulates the situation of insufficient production capacity, systematically evaluates the bottlenecks and shortcomings of the assembly line, provides data support for subsequent optimization, generates simulation data of the insufficient assembly line operation, helps the system understand the operating performance of the assembly line under high load conditions, and provides a basis for scheduling optimization. Through dynamic collaborative scheduling iterative analysis, the system generates multiple scheduling strategies for the insufficient assembly line to improve assembly efficiency. Multiple dynamic collaborative scheduling strategies are generated to help the system select the optimal scheduling solution under different circumstances to cope with changes in actual operation. Through the efficiency evaluation of multiple strategies, the system compares the performance of different scheduling solutions and finds the optimal scheduling strategy. The operating efficiency evaluation values ​​of multiple strategies are generated to help the system quantitatively evaluate the advantages and disadvantages of various strategies and provide a basis for the final selection. For reference, the optimal dynamic collaborative scheduling strategy is extracted to ensure that the assembly line reaches the highest efficiency in actual operation and improve production efficiency. By carefully selecting the optimal strategy, the system optimizes the scheduling and operation of the assembly line, improves assembly efficiency and capacity utilization, and through dynamic parameter fine-tuning, the system adjusts the parameters of the assembly line according to the optimal strategy to further improve assembly efficiency and production quality. A dynamic assembly fine-tuning engine is generated to help the system achieve automated scheduling and optimization, and improve the flexibility and adaptability of the assembly line. Through cluster collaborative control and dynamic scheduling, the system realizes collaborative operation between multiple assembly lines, improves overall production efficiency and scheduling flexibility, and builds a dynamic cluster scheduling model to help the system overall plan and optimize the operation of the assembly line, and realize intelligent scheduling and management.

[0072] In this specification, an intelligent scheduling system for an automatic assembly line of charging piles is provided, which is used to execute the intelligent scheduling method for the automatic assembly line of charging piles as described above, including:

[0073] The demand feature module is used to obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform stage-by-stage assembly demand feature analysis to obtain assembly demand feature data at different stages;

[0074] The demand deviation prediction module is used to extract the real-time production progress of each assembly line based on the charging pile assembly production log; the demand supply deviation of the real-time production progress of each assembly line is predicted according to the assembly demand characteristic data at different stages, and the assembly lines with insufficient production capacity are marked;

[0075] The operation logic module is used to collect all-round real-time monitoring images of the assembly line; perform dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and perform process operation logic analysis to generate multi-stage process operation logic;

[0076] The path planning module is used to perform potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and to perform dynamic avoidance path planning to generate a dynamic avoidance assembly path;

[0077] The dynamic twin module is used to obtain the global scene graph of the assembly line based on the real-time monitoring image of the assembly line in all directions; the global scene graph of the assembly line is dynamically mapped according to the dynamic avoidance assembly path to build a dynamic assembly twin model;

[0078] The dynamic collaborative scheduling module is used to perform dynamic collaborative scheduling iterative analysis on the dynamic assembly twin model based on the assembly line with insufficient capacity, and to perform collaborative control of the assembly line cluster, build a dynamic cluster scheduling model, and execute assembly line intelligent scheduling method operations.

[0079] The present invention obtains the charging pile assembly production log and performs feature analysis, so that the system understands the demand characteristics of the assembly stage, optimizes the assembly process and resource allocation, and obtains the assembly demand feature data at different stages to help the system better plan the production plan and scheduling tasks, improve production efficiency and quality, and through comparative analysis of real-time production progress and demand feature data, the system predicts the demand and supply deviation of the assembly line, and timely marks the assembly line with insufficient production capacity to further optimize scheduling and resource allocation. The predicted demand deviation helps the system to adjust the plan in time during the production process, reduce production interruptions and improve the operating efficiency of the assembly line. Through all-round real-time monitoring images and dynamic behavior analysis, the system monitors the operating status of the assembly line in real time, analyzes the process logic, promptly discovers problems and optimizes the production process, and generates multi-stage process operation logic to help the system better plan the production process and improve production efficiency. Rate and quality. According to the process operation logic and obstacle inference processing, the system generates a dynamic avoidance assembly path to improve the operation efficiency and safety of the assembly line. Dynamic avoidance path planning helps the system avoid collisions and optimize assembly routes during the production process, reduce production risks and improve assembly efficiency. Based on real-time monitoring images, a dynamic assembly twin model is constructed. The system realizes real-time global scene mapping and dynamic behavior analysis of the assembly line and optimizes the assembly process. The dynamic twin model helps the system better understand the operation status of the assembly line, guides scheduling decisions, and improves production efficiency and assembly quality. Through dynamic collaborative scheduling iterative analysis and cluster collaborative control, the system optimizes the scheduling plan for assembly lines with insufficient production capacity, realizes intelligent scheduling and management of assembly lines, and constructs a dynamic cluster scheduling model to help the system comprehensively consider the collaborative operations between assembly lines and improve the overall production efficiency and flexibility of assembly lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A schematic flow chart of the steps of an intelligent scheduling method for an automated assembly line of charging piles according to the present invention;

[0081] Figure 2 Detailed implementation flow chart of step S1;

[0082] Figure 3 Detailed implementation flow chart of step S2;

[0083] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0084] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0085] The present application example provides an intelligent scheduling method and system for an automated assembly line of charging piles. The execution subjects of the intelligent scheduling method and system for automated assembly line of charging piles include but are not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0086] See also Figures 1 to 4 The present invention provides an intelligent scheduling method for an automatic assembly line of charging piles, and the intelligent scheduling method for an automatic assembly line of charging piles comprises the following steps:

[0087] Step S1: Obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform staged assembly demand feature analysis to obtain assembly demand feature data at different stages;

[0088] Step S2: extracting the real-time production progress of each assembly line based on the charging pile assembly production log; predicting the demand-supply deviation of the real-time production progress of each assembly line according to the assembly demand characteristic data at different stages, and marking the assembly lines with insufficient production capacity;

[0089] Step S3: collecting all-round real-time monitoring images of the assembly line; performing dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and performing process operation logic analysis to generate multi-stage process operation logic;

[0090] Step S4: performing potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and performing dynamic avoidance path planning to generate a dynamic avoidance assembly path;

[0091] Step S5: obtaining a global scene graph of the assembly line based on the omnidirectional real-time monitoring image of the assembly line; performing dynamic behavior mapping processing on the global scene graph of the assembly line according to the dynamic avoidance assembly path to construct a dynamic assembly twin model;

[0092] Step S6: Based on the assembly line with insufficient production capacity, the dynamic assembly twin model is subjected to dynamic collaborative scheduling iterative analysis, and assembly line cluster collaborative control is performed to build a dynamic cluster scheduling model and execute the assembly line intelligent scheduling method operation.

[0093] The present invention understands the assembly demand characteristics of different stages through sliding window segmentation processing and staged assembly demand characteristic analysis, provides data support for subsequent scheduling, obtains assembly demand characteristic data of different stages to optimize the production plan of the assembly line, improves assembly efficiency and capacity utilization, predicts demand-supply deviation for each assembly line, and promptly discovers capacity shortages, which is conducive to adjusting the production plan to avoid production delays, marks assembly lines with capacity shortages for key attention and priority scheduling, and ensures that the production capacity of the assembly line reaches the optimal state. Dynamic behavior analysis and process operation logic analysis understand the real-time production situation of the assembly line, improve the visual monitoring and control capabilities of the production process, and generate multi-stage process operation logic to provide a basis for subsequent obstacle inference and path planning, dynamic avoidance Path planning avoids collisions and blockages during the assembly process, improves assembly efficiency and safety, handles potential obstacles and performs dynamic avoidance path planning to optimize the assembly process and reduce unexpected events and pauses in production. The dynamic assembly twin model combines real-time monitoring images and dynamic avoidance paths to reflect the status and operation of the assembly line in real time, providing accurate data support for intelligent scheduling. Dynamic behavior mapping processing associates monitoring data with the actual assembly process to improve scheduling accuracy and efficiency. Dynamic collaborative scheduling iterative analysis and assembly line cluster collaborative control realize collaborative operations and resource optimization between assembly lines, improve overall production efficiency, and build a dynamic cluster scheduling model to realize intelligent scheduling and optimization of assembly lines, ensuring efficient operation of the production process and capacity utilization.

[0094] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of an intelligent scheduling method for an automatic assembly line of charging piles according to the present invention. In this example, the steps of the method include:

[0095] Step S1: Obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform staged assembly demand feature analysis to obtain assembly demand feature data at different stages;

[0096] In this embodiment, original production log data is collected from the charging pile assembly production line. The log contains various event records in the assembly process. A time window length is set, such as 1 hour, and a sliding window is cut on the entire production log to obtain log segments in multiple time windows. For each time window log segment, the time series characteristics such as the occurrence time and frequency of various events are extracted to form a data set describing the time series characteristics of the log segment. The time series characteristics of each time window log segment are analyzed to identify the assembly requirement characteristics of different production stages, and form a data set describing different assembly requirement characteristics.

[0097] Step S2: extracting the real-time production progress of each assembly line based on the charging pile assembly production log; predicting the demand-supply deviation of the real-time production progress of each assembly line according to the assembly demand characteristic data at different stages, and marking the assembly lines with insufficient production capacity;

[0098] In this embodiment, the actual production progress of each assembly line in the time period is calculated based on the events occurring in each time window, and the obtained assembly demand feature data is compared with the extracted real-time production progress to predict the demand-supply deviation of each assembly line in the current period. The prediction results are analyzed to identify those assembly lines whose production progress cannot meet the current demand and mark them as assembly lines with insufficient production capacity.

[0099] Step S3: collecting all-round real-time monitoring images of the assembly line; performing dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and performing process operation logic analysis to generate multi-stage process operation logic;

[0100] In this embodiment, multiple cameras are deployed at key positions of the assembly line to collect all-round image data of the assembly process in real time to ensure that the dynamic changes of the assembly line can be fully captured. Computer vision and deep learning technology are used to analyze and interpret the collected real-time image data, identify the dynamic behavior characteristics of each process in the assembly operation, and combine the obtained dynamic behavior characteristics to analyze the logical relationship and timing dependency between the processes, establish a model that describes the operation logic of the entire process of the assembly line, generate the process operation logic models of the corresponding stages, and form a data set that describes the changes in the processes of the entire production cycle of the assembly line.

[0101] Step S4: performing potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and performing dynamic avoidance path planning to generate a dynamic avoidance assembly path;

[0102] In this embodiment, the generated multi-stage process operation logic model is used to fuse the collected all-round monitoring images, identify potential obstacles that appear in the current production stage, and use an algorithm to optimize the dynamic path of the assembly line based on the inferred potential obstacle information to generate the best dynamic avoidance assembly path. The planned dynamic avoidance path is fed back to the assembly operation execution unit to guide the assembly robot to execute the optimized dynamic avoidance path.

[0103] Step S5: obtaining a global scene graph of the assembly line based on the omnidirectional real-time monitoring image of the assembly line; performing dynamic behavior mapping processing on the global scene graph of the assembly line according to the dynamic avoidance assembly path to construct a dynamic assembly twin model;

[0104] In this embodiment, the collected multi-angle monitoring images are used to generate a three-dimensional model describing the global scene of the entire assembly line through image fusion and three-dimensional reconstruction technology. The generated dynamic avoidance assembly path is mapped to the constructed global scene graph of the assembly line, so that the three-dimensional model can dynamically reflect the real-time changes of the assembly operation. The dynamically mapped assembly line scene model is combined with the obtained multi-stage process operation logic to establish a twin model that can reflect the dynamic state of the assembly line in real time.

[0105] Step S6: Based on the assembly line with insufficient production capacity, the dynamic assembly twin model is subjected to dynamic collaborative scheduling iterative analysis, and assembly line cluster collaborative control is performed to build a dynamic cluster scheduling model and execute the assembly line intelligent scheduling method operation.

[0106] In this embodiment, the marked information of assembly lines with insufficient capacity is utilized to perform dynamic optimization scheduling analysis on these links with insufficient capacity in the constructed dynamic assembly twin model. Based on the dynamic scheduling analysis results, the entire assembly line cluster is collaboratively controlled, including process rescheduling, resource reallocation and other measures. The dynamic scheduling strategies and control methods are modeled and packaged to form a set of dynamic cluster scheduling models to provide support for subsequent intelligent scheduling of assembly lines. The constructed dynamic cluster scheduling model is applied to the actual assembly line operation environment to perform intelligent dynamic scheduling and optimization of the assembly line.

[0107] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0108] Step S11: Obtain charging pile assembly production log;

[0109] Step S12: performing time stamp standardization processing on the charging pile assembly production log to obtain a time standardized production log;

[0110] Step S13: performing sliding window segmentation processing on the time-standardized production log, thereby obtaining multiple time window log segments;

[0111] Step S14: performing time series analysis on multiple time window log segments to generate multiple log segment time series feature data;

[0112] Step S15: performing phased assembly requirement feature analysis on multiple log segment time series feature data, thereby obtaining assembly requirement feature data at different stages.

[0113] In this embodiment, original production log data is collected from the charging pile assembly production line. The log contains various event records in the assembly process. The timestamp information in the log is checked to ensure that the time format is unified. If there is a deviation, time standardization is performed to ensure that all log events have standardized time tags. The size of the sliding window (such as 1 hour, 1 day, etc.) and the step size (such as 30 minutes, 1 hour, etc.) are determined. The entire production log is cut by a sliding window to obtain log segments in multiple time windows. For each time window log segment, the time series characteristics such as the occurrence time and frequency of various events are extracted to form a data set describing the time series characteristics of the log segment. The time series characteristics of the log segments in each time window are analyzed to identify the assembly demand characteristics of different production stages, such as: the characteristics of different time periods such as the morning peak, the lunch break, and the evening peak. The staged assembly demand characteristics are analyzed on the time series characteristic data to identify the demand characteristics of different stages.

[0114] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0115] Step S21: performing assembly demand forecasting on the assembly demand characteristic data at different stages, thereby obtaining demand forecasting data for multiple time periods;

[0116] Step S22: extracting the real-time production progress of each assembly line based on the charging pile assembly production log;

[0117] Step S23: Calculate the production completion rate of each assembly line based on the real-time production progress to obtain the production completion rate of each assembly line;

[0118] Step S24: analyzing the production capacity characteristics of the assembly line process based on the production completion rate of each assembly line to obtain the production capacity characteristics of each assembly line;

[0119] Step S25: performing capacity matching calculation for the corresponding time period on the capacity characteristics of each assembly line based on multiple time window log segments to obtain the assembly line capacity for the corresponding time period;

[0120] Step S26: Based on the demand forecast data for multiple time periods, the demand-supply deviation of the assembly line capacity in the corresponding time period is predicted, and the assembly line with insufficient capacity is marked.

[0121] In this embodiment, the assembly demand feature data of different stages are sorted to ensure that the data is complete and in a consistent format. A suitable demand forecasting model is selected, such as time series analysis (ARIMA, SARIMA), regression analysis or machine learning model (such as random forest, XGBoost). The historical assembly demand feature data is used to train the model, and the parameters are adjusted to improve the prediction accuracy. The trained model is used to predict the assembly demand for multiple time periods in the future to generate demand forecasting data. The production records of each assembly line are extracted from the charging pile assembly production log, including the completed process, production time, etc., and the real-time progress is calculated: Based on the extracted data, the real-time production progress of each assembly line is calculated, usually expressed as a percentage. The completion rate of the real-time production progress of each assembly line is calculated to obtain the production completion rate of each assembly line, and the indicators of production capacity characteristics are determined, such as capacity utilization, bottleneck processes, average production time, etc. The production completion rate of each assembly line is analyzed over time, and the production rhythm, production capacity and other characteristics of each assembly line are extracted. The time period that needs to be matched is determined, which usually corresponds to the time period of demand forecast. The capacity characteristics of each assembly line are calculated for time period matching, such as by summing or weighted averaging, comparing the demand forecast data and assembly line capacity data for each time period, identifying assembly lines with insufficient capacity, predicting time periods and assembly lines with capacity gaps, and marking these key assembly lines with insufficient capacity.

[0122] In this embodiment, the specific steps of step S21 are:

[0123] Perform assembly line demand distribution statistics on the assembly demand characteristic data at different stages to obtain the demand characteristic data of each assembly line;

[0124] Perform periodic demand fluctuation analysis on the demand characteristic data of each assembly line to obtain the assembly demand fluctuation data of different periods;

[0125] The demand trend evolution is carried out on the assembly demand fluctuation data in different periods, so as to obtain the assembly demand trend evolution law;

[0126] Based on the evolution law of assembly demand trend, multi-period assembly demand forecasting is performed on time-standardized production logs to obtain demand forecast data for multiple periods.

[0127] In this embodiment, the assembly demand characteristic data of different stages are sorted to ensure that the data is complete and in a consistent format. Appropriate statistical methods are selected, such as frequency distribution, mean and standard deviation calculation, etc., to perform statistics on the demand characteristic data of each assembly line, calculate the distribution of demand, determine the analysis period, such as day, week, month, etc., to ensure that it matches the timestamp of the demand characteristic data, select an appropriate fluctuation analysis method, such as standard deviation, coefficient of variation or moving average, etc., perform periodic fluctuation analysis on the demand characteristic data of each assembly line, identify the demand fluctuations in different periods, select an appropriate trend evolution model, such as linear regression, exponential smoothing or time series analysis method, input the demand fluctuation data of different periods into the selected trend evolution model, perform demand trend evolution calculation, identify the evolution law and trend of demand, select an appropriate demand forecasting model, such as ARIMA, SARIMA, machine learning regression model, etc., use historical assembly demand trend data to train the forecasting model, ensure that the model can capture the law of demand evolution, input the time information in the time-standardized production log into the trained model, and perform multi-period demand forecasting.

[0128] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0129] Step S31: collecting all-round real-time monitoring images of the assembly line;

[0130] Step S32: performing global balanced brightness enhancement processing on the omnidirectional real-time monitoring image of the assembly line to generate a global enhanced monitoring image;

[0131] Step S33: performing dynamic behavior analysis on the global enhanced monitoring image to obtain dynamic behavior characteristic data of the assembly line;

[0132] Step S34: performing phased process identification on the dynamic behavior characteristic data of the assembly line, thereby obtaining process behavior characteristics of different stages;

[0133] Step S35: Perform process operation logic analysis on process behavior characteristics at different stages to generate multi-stage process operation logic.

[0134] In this embodiment, multiple surveillance cameras are arranged at key positions of the assembly line to collect all-round monitoring image data of the assembly line in real time, and the collected images are preprocessed, such as denoising and image format conversion. Select a suitable brightness enhancement algorithm, such as histogram equalization or adaptive histogram equalization (CLAHE). Perform brightness enhancement processing on each image to generate a global enhanced monitoring image to improve the visibility and contrast of the image. Select a suitable dynamic behavior analysis algorithm, such as optical flow method, motion detection or deep learning model (such as YOLO, OpenPose, etc.). Perform dynamic behavior analysis on the global enhanced monitoring image to extract dynamic behavior feature data of the assembly line, such as movement direction, speed, frequency, etc. According to the production process, define the stages of each process, such as assembly, inspection, packaging, etc. Select a suitable process recognition algorithm, such as pattern recognition, classification algorithm or rule-based detection. Analyze the dynamic behavior feature data to identify the process behavior features of different stages. Select a suitable operation logic analysis method, such as flowchart analysis, state transition analysis or logistic regression. Analyze the process behavior features of different stages and analyze the timing relationship and logical dependency between them.

[0135] In this embodiment, step S4 includes the following steps:

[0136] Step S41: Performing object visual detection processing on the global enhanced monitoring image, marking assembly equipment and component objects;

[0137] Step S42: performing optical flow movement tracking on the assembly equipment and component objects to generate an optical flow movement trajectory of the object;

[0138] Step S43: performing potential obstacle inference processing on the object optical flow movement trajectory to obtain obstacle nodes;

[0139] Step S44: predicting the operation time of each stage process based on the multi-stage process operation logic of the object optical flow movement trajectory, and extracting the predicted value of the process time of each stage;

[0140] Step S45: Perform dynamic avoidance path planning on obstacle nodes according to the predicted process time consumption value of each stage to generate a dynamic avoidance assembly path.

[0141] In this embodiment, a suitable object detection algorithm is selected, such as YOLO, Faster R-CNN or SSD, to quickly and accurately identify assembly equipment and parts. A deep learning model is used to ensure that the model is fully trained and can identify equipment and parts on the assembly line. The global enhanced monitoring image is input into the detection model, and object visual detection processing is performed to mark assembly equipment and parts objects. A suitable optical flow tracking algorithm is selected, such as the Lucas-Kanade algorithm or Dense Optical Flow, perform optical flow tracking on the marked assembly equipment and parts objects, record their motion trajectories in continuous frames, organize the tracked motion information into object optical flow movement trajectories, save the position changes of each object over time, select a suitable obstacle detection model, including semantic segmentation based on deep learning or traditional shape detection, use the object optical flow movement trajectory data, analyze the object's motion pattern, infer the obstacle position, organize the inferred potential obstacles into obstacle nodes, record their position and feature information, select a suitable time-consuming prediction model, such as linear regression, time series analysis, etc., organize the data of object optical flow movement trajectory and multi-stage process operation logic to ensure data consistency, input data into the time-consuming prediction model, calculate the operation time of each process stage, and select a suitable dynamic avoidance path planning algorithm, such as A* algorithm, Dijkstra algorithm or Rapidly-exploring Random Tree (RRT), input obstacle nodes and process time prediction values, calculate the dynamic avoidance path, consider the real-time and safety of the path, ensure that the planned path is effective and efficient, verify the generated dynamic avoidance assembly path, and ensure its feasibility in actual operation.

[0142] In this embodiment, the specific steps of step S43 are:

[0143] Perform trajectory density distribution analysis on the object's optical flow movement trajectory to generate trajectory density distribution data;

[0144] Mark the location points of all objects according to their optical flow movement trajectory;

[0145] The number of times all objects stay at their locations is counted along multiple trajectories to obtain the number of times the location points stay;

[0146] Calculating the track point dwell time for all the object stop positions to generate multiple track point dwell times;

[0147] Perform potential obstacle analysis on the trajectory density distribution data according to the number of stop times at the location point and the stop time of multiple trajectory points, and extract multiple potential obstacle trajectory points;

[0148] Based on multiple potential obstacle trajectory points, the global enhanced monitoring image is segmented into bounding boxes to extract the potential obstacle bounding box image;

[0149] Obstacle depth convolution is performed on the potential obstacle bounding box image to obtain the obstacle node.

[0150] In this embodiment, the optical flow movement trajectory data of the object is sorted to ensure that the position change information of each object is complete, and a suitable trajectory density estimation method is selected, such as kernel density estimation (KDE) or a histogram method. The selected method is applied to perform density distribution analysis on the trajectory of the object, and a high-density area is identified. The standard of the stay position is determined, such as a point whose position change is less than a certain threshold within a specific time. According to the defined standard, the stay position points of all objects are extracted from the optical flow movement trajectory of the object, and all marked stay position points are sorted into a data set, and the coordinates and related information of each stay point are recorded. A statistical method for determining the number of stays, such as a simple count or a count based on a time window, is used. Statistics are collected for each stop location, the number of stops at each location is recorded, and the method for calculating the stop time is determined, such as the time difference from entering the stop point to leaving the stop point. Each stop location is calculated, and the stop time of each point is recorded. The criteria for potential obstacles are determined, such as points with high stop times and long stop times. According to the set criteria, the trajectory density distribution data is analyzed to extract potential obstacle trajectory points that meet the conditions. An appropriate bounding box generation algorithm is selected, such as connected component analysis or edge detection. According to the potential obstacle trajectory points, the corresponding bounding boxes are generated in the global enhanced monitoring image. An appropriate deep learning model, such as YOLO, Faster R-CNN or SSD, is selected for obstacle detection. If necessary, the deep learning model is trained using the labeled data containing obstacles. The potential obstacle bounding box image is input into the trained model to perform deep convolution inference of obstacles. The obstacle nodes identified in the inference results are organized into a data set, and the location information and features of each obstacle are recorded.

[0151] In this embodiment, the specific steps of step S5 are:

[0152] Step S51: performing global scene convolution recognition on the omnidirectional real-time monitoring image of the assembly line to obtain a global scene graph of the assembly line;

[0153] Step S52: performing three-dimensional point cloud modeling on the global scene graph of the assembly line to generate a global scene point cloud model;

[0154] Step S53: dynamically evolving the dynamic avoidance assembly path according to the process behavior characteristics of different stages to generate dynamic process behavior evolution data;

[0155] Step S54: Perform dynamic behavior mapping processing on the global scene point cloud model based on the dynamic process behavior evolution data to construct a dynamic assembly twin model.

[0156] In this embodiment, the collected real-time monitoring images are preprocessed, including denoising and image enhancement, to improve recognition accuracy. A suitable convolutional neural network (CNN) architecture, such as ResNet, VGG or a custom model, is selected for global scene recognition. Model training (if necessary): The selected model is trained using a labeled data set to ensure that it can accurately recognize the global scene of the assembly line. The processed monitoring images are input into the trained CNN model for global scene convolution recognition, and the global scene graph of the assembly line is output: Select a suitable point cloud generation method, such as structured light, laser scanning, or image-based reconstruction method, use the selected method to convert the global scene graph into three-dimensional point cloud data, record the spatial coordinates of each point, optimize the generated point cloud, including denoising, downsampling, and normal vector calculation to improve the quality of the point cloud, extract the process behavior characteristics of different stages from the previous analysis, ensure that the data is complete and the format is consistent, select a suitable dynamic process evolution model, such as a state machine model or a rule-based model, perform evolution calculations of the dynamic avoidance assembly path based on the extracted process behavior characteristics, generate dynamic process behavior evolution data, select a suitable dynamic behavior mapping method, such as mapping based on physical simulation or a deep learning-based method, map the dynamic process behavior evolution data with the global scene point cloud model, construct a dynamic assembly twin model, simulate the dynamic behavior of the assembly line, verify the accuracy and effectiveness of the constructed assembly twin model, and ensure that it can truly reflect the dynamic changes of the assembly process.

[0157] In this embodiment, the specific steps of step S6 are:

[0158] Step S61: performing assembly operation simulation on the dynamic assembly twin model based on the insufficient capacity assembly line to generate insufficient assembly line operation simulation data;

[0159] Step S62: Perform dynamic collaborative scheduling iterative analysis on the simulation data of the defective assembly line operation to generate multiple dynamic collaborative scheduling strategies;

[0160] Step S63: performing assembly collaborative operation efficiency evaluation on multiple dynamic collaborative scheduling strategies to generate operation efficiency evaluation values ​​of multiple strategies;

[0161] Step S64: extracting the optimal dynamic collaborative scheduling strategy based on the operation efficiency evaluation values ​​of multiple strategies;

[0162] Step S65: fine-tuning the dynamic parameters of the charging pile automated assembly line using the optimal dynamic collaborative scheduling strategy to generate a dynamic assembly fine-tuning engine;

[0163] Step S66: Perform assembly line cluster collaborative control on the dynamic assembly fine-tuning engine, build a dynamic cluster scheduling model, and execute the assembly line intelligent scheduling method operation.

[0164] In this embodiment, relevant data of the assembly line with insufficient production capacity, such as process time, resource allocation, and equipment status, are collected, simulation parameters are set in the dynamic assembly twin model, the simulation time range and operating conditions are defined, the assembly operation simulation is started, the assembly process is monitored in real time, and various operating data, including production capacity, efficiency, and delay, are recorded. A variety of dynamic collaborative scheduling strategy frameworks are designed, and different scheduling principles and algorithms (such as priority scheduling, shortest job priority, etc.) are considered. Based on the operation simulation data, the effects of each scheduling strategy are iteratively analyzed, and a variety of dynamic collaborative scheduling strategies are generated. Evaluation indicators, such as assembly cycle time, resource utilization, output quality, etc., are determined. The operation efficiency of each dynamic collaborative scheduling strategy is evaluated, and the evaluation value of each strategy is calculated. The operation efficiency evaluation values ​​of multiple strategies are compared to identify The most efficient strategy is to confirm the optimal dynamic collaborative scheduling strategy through quantitative indicators and feasibility analysis, and determine the parameters that need to be fine-tuned, such as equipment operating speed, process time, and resource allocation. According to the optimal dynamic collaborative scheduling strategy, the parameters of the charging pile automation assembly line are dynamically adjusted, and the fine-tuned parameters and strategies are integrated to generate a dynamic assembly fine-tuning engine for real-time adjustment and optimization. A dynamic cluster scheduling model is designed, considering the collaborative relationship and resource sharing between each assembly line. The dynamic assembly fine-tuning engine is used to collaboratively control the assembly line cluster to ensure the coordinated operation of each line. Intelligent scheduling algorithms (such as genetic algorithms, deep learning optimization, etc.) are applied to execute assembly line scheduling tasks, monitor real-time operating status, collect the results of intelligent scheduling, analyze and optimize, and provide a basis for subsequent scheduling.

[0165] In this embodiment, an intelligent scheduling system for an automatic assembly line of charging piles is provided, which is used to execute the intelligent scheduling method for the automatic assembly line of charging piles as described above, including:

[0166] The demand feature module is used to obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform stage-by-stage assembly demand feature analysis to obtain assembly demand feature data at different stages;

[0167] The demand deviation prediction module is used to extract the real-time production progress of each assembly line based on the charging pile assembly production log; the demand supply deviation of the real-time production progress of each assembly line is predicted according to the assembly demand characteristic data at different stages, and the assembly lines with insufficient production capacity are marked;

[0168] The operation logic module is used to collect all-round real-time monitoring images of the assembly line; perform dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and perform process operation logic analysis to generate multi-stage process operation logic;

[0169] The path planning module is used to perform potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and to perform dynamic avoidance path planning to generate a dynamic avoidance assembly path;

[0170] The dynamic twin module is used to obtain the global scene graph of the assembly line based on the real-time monitoring image of the assembly line in all directions; the global scene graph of the assembly line is dynamically mapped according to the dynamic avoidance assembly path to build a dynamic assembly twin model;

[0171] The dynamic collaborative scheduling module is used to perform dynamic collaborative scheduling iterative analysis on the dynamic assembly twin model based on the assembly line with insufficient capacity, and to perform collaborative control of the assembly line cluster, build a dynamic cluster scheduling model, and execute assembly line intelligent scheduling method operations.

[0172] The present invention obtains the charging pile assembly production log and performs feature analysis, so that the system understands the demand characteristics of the assembly stage, optimizes the assembly process and resource allocation, and obtains the assembly demand feature data at different stages to help the system better plan the production plan and scheduling tasks, improve production efficiency and quality, and through comparative analysis of real-time production progress and demand feature data, the system predicts the demand and supply deviation of the assembly line, and timely marks the assembly line with insufficient production capacity to further optimize scheduling and resource allocation. The predicted demand deviation helps the system to adjust the plan in time during the production process, reduce production interruptions and improve the operating efficiency of the assembly line. Through all-round real-time monitoring images and dynamic behavior analysis, the system monitors the operating status of the assembly line in real time, analyzes the process logic, promptly discovers problems and optimizes the production process, and generates multi-stage process operation logic to help the system better plan the production process and improve production efficiency. Rate and quality. According to the process operation logic and obstacle inference processing, the system generates a dynamic avoidance assembly path to improve the operation efficiency and safety of the assembly line. Dynamic avoidance path planning helps the system avoid collisions and optimize assembly routes during the production process, reduce production risks and improve assembly efficiency. Based on real-time monitoring images, a dynamic assembly twin model is constructed. The system realizes real-time global scene mapping and dynamic behavior analysis of the assembly line and optimizes the assembly process. The dynamic twin model helps the system better understand the operation status of the assembly line, guides scheduling decisions, and improves production efficiency and assembly quality. Through dynamic collaborative scheduling iterative analysis and cluster collaborative control, the system optimizes the scheduling plan for assembly lines with insufficient production capacity, realizes intelligent scheduling and management of assembly lines, and constructs a dynamic cluster scheduling model to help the system comprehensively consider the collaborative operations between assembly lines and improve the overall production efficiency and flexibility of assembly lines.

[0173] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0174] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent scheduling method for an automated assembly line of charging piles, characterized in that: The following steps are involved: Step S1: Obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform staged assembly demand feature analysis to obtain assembly demand feature data at different stages; Step S2: extracting the real-time production progress of each assembly line based on the charging pile assembly production log; predicting the demand-supply deviation of the real-time production progress of each assembly line according to the assembly demand characteristic data at different stages, and marking the assembly lines with insufficient production capacity; Step S3: collecting all-round real-time monitoring images of the assembly line; performing dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and performing process operation logic analysis to generate multi-stage process operation logic; Step S4: performing potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and performing dynamic avoidance path planning to generate a dynamic avoidance assembly path; Step S5: obtaining a global scene map of the assembly line based on the omnidirectional real-time monitoring image of the assembly line; Perform dynamic behavior mapping on the global scene graph of the assembly line according to the dynamic avoidance assembly path to build a dynamic assembly twin model; Step S6: Based on the under-capacity assembly line, the dynamic assembly twin model is subjected to dynamic collaborative scheduling iterative analysis, and assembly line cluster collaborative control is performed to build a dynamic cluster scheduling model and execute the assembly line intelligent scheduling method operation; Among them, the specific steps of step S2 are: Step S21: performing assembly demand forecasting on the assembly demand characteristic data at different stages, thereby obtaining demand forecasting data for multiple time periods; Step S22: extracting the real-time production progress of each assembly line based on the charging pile assembly production log; Step S23: Calculate the production completion rate of each assembly line based on the real-time production progress to obtain the production completion rate of each assembly line; Step S24: analyzing the production capacity characteristics of the assembly line process based on the production completion rate of each assembly line to obtain the production capacity characteristics of each assembly line; Step S25: performing capacity matching calculation for the corresponding time period on the capacity characteristics of each assembly line based on multiple time window log segments to obtain the assembly line capacity for the corresponding time period; Step S26: Based on the demand forecast data for multiple time periods, the demand-supply deviation of the assembly line capacity in the corresponding time period is predicted, and the assembly line with insufficient capacity is marked.

2. The intelligent scheduling method for the charging pile automated assembly line according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: Obtain charging pile assembly production log; Step S12: performing time stamp standardization processing on the charging pile assembly production log to obtain a time standardized production log; Step S13: performing sliding window segmentation processing on the time-standardized production log, thereby obtaining multiple time window log segments; Step S14: performing time series analysis on multiple time window log segments to generate multiple log segment time series feature data; Step S15: performing phased assembly requirement feature analysis on multiple log segment time series feature data, thereby obtaining assembly requirement feature data at different stages.

3. The intelligent scheduling method for the charging pile automated assembly line according to claim 2 is characterized in that: The specific steps of step S21 are: Perform assembly line demand distribution statistics on the assembly demand characteristic data at different stages to obtain the demand characteristic data of each assembly line; Perform periodic demand fluctuation analysis on the demand characteristic data of each assembly line to obtain the assembly demand fluctuation data of different periods; The demand trend evolution is carried out on the assembly demand fluctuation data in different periods, so as to obtain the assembly demand trend evolution law; Based on the evolution law of assembly demand trend, multi-period assembly demand forecasting is performed on time-standardized production logs to obtain demand forecast data for multiple periods.

4. The intelligent scheduling method for the charging pile automated assembly line according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: collecting all-round real-time monitoring images of the assembly line; Step S32: performing global balanced brightness enhancement processing on the omnidirectional real-time monitoring image of the assembly line to generate a global enhanced monitoring image; Step S33: performing dynamic behavior analysis on the global enhanced monitoring image to obtain dynamic behavior characteristic data of the assembly line; Step S34: performing phased process identification on the dynamic behavior characteristic data of the assembly line, thereby obtaining process behavior characteristics of different stages; Step S35: Perform process operation logic analysis on process behavior characteristics at different stages to generate multi-stage process operation logic.

5. The intelligent scheduling method for the charging pile automated assembly line according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: Performing object visual detection processing on the global enhanced monitoring image, marking assembly equipment and component objects; Step S42: performing optical flow movement tracking on the assembly equipment and component objects to generate an optical flow movement trajectory of the object; Step S43: performing potential obstacle inference processing on the object optical flow movement trajectory to obtain obstacle nodes; Step S44: based on the multi-stage process operation logic, the object optical flow movement trajectory is subjected to a stage process operation time prediction, and the process time prediction value of each stage is extracted; Step S45: Perform dynamic avoidance path planning for obstacle nodes according to the predicted process time consumption value of each stage to generate a dynamic avoidance assembly path.

6. The intelligent scheduling method for the charging pile automated assembly line according to claim 5, characterized in that: The specific steps of step S43 are: Perform trajectory density distribution analysis on the object's optical flow movement trajectory to generate trajectory density distribution data; Mark the location points of all objects according to their optical flow movement trajectory; The number of times all objects stay at their locations is counted along multiple trajectories to obtain the number of times the location points stay; Calculating the track point dwell time for all the object stop positions to generate multiple track point dwell times; Perform potential obstacle analysis on the trajectory density distribution data according to the number of stop times at the location point and the stop time of multiple trajectory points, and extract multiple potential obstacle trajectory points; Based on multiple potential obstacle trajectory points, the global enhanced monitoring image is segmented into bounding boxes to extract the potential obstacle bounding box image; Obstacle depth convolution is performed on the potential obstacle bounding box image to obtain the obstacle node.

7. The intelligent scheduling method for the charging pile automated assembly line according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing global scene convolution recognition on the omnidirectional real-time monitoring image of the assembly line to obtain a global scene graph of the assembly line; Step S52: performing three-dimensional point cloud modeling on the global scene graph of the assembly line to generate a global scene point cloud model; Step S53: dynamically evolving the dynamic avoidance assembly path according to the process behavior characteristics of different stages to generate dynamic process behavior evolution data; Step S54: Perform dynamic behavior mapping processing on the global scene point cloud model based on the dynamic process behavior evolution data to construct a dynamic assembly twin model.

8. The intelligent scheduling method for the charging pile automated assembly line according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: performing assembly operation simulation on the dynamic assembly twin model based on the insufficient capacity assembly line to generate insufficient assembly line operation simulation data; Step S62: Perform dynamic collaborative scheduling iterative analysis on the simulation data of the defective assembly line operation to generate multiple dynamic collaborative scheduling strategies; Step S63: performing assembly collaborative operation efficiency evaluation on multiple dynamic collaborative scheduling strategies to generate operation efficiency evaluation values ​​of multiple strategies; Step S64: extracting the optimal dynamic collaborative scheduling strategy based on the operation efficiency evaluation values ​​of multiple strategies; Step S65: fine-tuning the dynamic parameters of the charging pile automated assembly line using the optimal dynamic collaborative scheduling strategy to generate a dynamic assembly fine-tuning engine; Step S66: Perform assembly line cluster collaborative control on the dynamic assembly fine-tuning engine, build a dynamic cluster scheduling model, and execute the assembly line intelligent scheduling method operation.

9. An intelligent scheduling system for an automated assembly line of charging piles, characterized in that: The intelligent scheduling method for executing the charging pile automated assembly line according to claim 1 comprises: The demand feature module is used to obtain the charging pile assembly production log; perform sliding window segmentation processing on the charging pile assembly production log, and perform stage-by-stage assembly demand feature analysis to obtain assembly demand feature data at different stages; The demand deviation prediction module is used to extract the real-time production progress of each assembly line based on the charging pile assembly production log; the demand supply deviation of the real-time production progress of each assembly line is predicted according to the assembly demand characteristic data at different stages, and the assembly lines with insufficient production capacity are marked; The operation logic module is used to collect all-round real-time monitoring images of the assembly line; perform dynamic behavior analysis on the all-round real-time monitoring images of the assembly line, and perform process operation logic analysis to generate multi-stage process operation logic; The path planning module is used to perform potential obstacle inference processing on the global enhanced monitoring image according to the multi-stage process operation logic, and to perform dynamic avoidance path planning to generate a dynamic avoidance assembly path; The dynamic twin module is used to obtain the global scene graph of the assembly line based on the real-time monitoring image of the assembly line in all directions; the global scene graph of the assembly line is dynamically mapped according to the dynamic avoidance assembly path to build a dynamic assembly twin model; The dynamic collaborative scheduling module is used to perform dynamic collaborative scheduling iterative analysis on the dynamic assembly twin model based on the assembly line with insufficient capacity, and to perform collaborative control of the assembly line cluster, build a dynamic cluster scheduling model, and execute assembly line intelligent scheduling method operations.

Images