A final assembly method and system for an automobile production line

Abstract

The present invention discloses a method and system for final assembly of an automobile production line, the method comprising: optimizing the assembly sequence of components by using historical production data based on the current model and the existing parts inventory; assembling components on the automobile production line according to the optimized assembly sequence, monitoring the working status and progress of each assembly station; performing multi-level real-time monitoring of the work operations of workers at the assembly station, and generating real-time safety scores; when potential safety hazards or irregular operations are monitored, automatically notifying relevant staff through alarms and starting the recording function at the same time according to the safety score, and saving image data before and after the incident; after the assembly is completed, establishing a data-driven quality prediction model to automatically correct potential quality problems according to the prediction results. By using the embodiments of the present invention, it is possible to optimize the assembly process, strengthen worker safety monitoring, and improve the quality control level, and realize a comprehensive intelligent transformation from the assembly process to quality control.

Description

Technical Field

[0001] The invention belongs to the technical field of automobile assembly, and in particular to an assembly method and system for an automobile production line. Background Art

[0002] With the continuous development of the global automotive industry and the intensification of market competition, the assembly efficiency and product quality of automobile production lines have become increasingly critical. The final assembly process of automobiles not only involves multiple complex processes, but also requires the coordination of the effective combination of different types of components and materials to ensure the performance and safety of the final product. Therefore, how to optimize the production process and improve the safety of workers' operations and assembly quality has become an urgent problem for automobile manufacturers.

[0003] Traditional methods of final assembly of automobiles often rely on experienced operators and fixed assembly sequences. This method shows certain limitations when faced with different models and new technical requirements. First, the assembly sequence of components on the production line is usually based on manual judgment and experience, resulting in large differences in the efficiency and rationality of the assembly process in different production batches. As the complexity of the model increases, the fixed assembly sequence is often difficult to adapt to the rapidly changing needs of the automotive industry, which in turn affects production efficiency. Summary of the invention

[0004] The purpose of the present invention is to provide a final assembly method and system for an automobile production line to address the deficiencies in the prior art, optimize the assembly process, strengthen worker safety monitoring and improve the quality control level, and achieve a comprehensive intelligent transformation from assembly process to quality control.

[0005] An embodiment of the present application provides a final assembly method for an automobile production line, the method comprising:

[0006] Using historical production data based on the current model and existing parts inventory, a heuristic algorithm is used to optimize the assembly sequence of components;

[0007] Assemble components on the automobile production line according to the optimized component assembly sequence and monitor the working status and progress of each assembly station;

[0008] Combined with image recognition and behavior analysis, the work operations of assembly station workers are monitored in real time at multiple levels to generate real-time safety scores. Electronic billboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time.

[0009] When potential safety hazards or irregular operations are detected, the relevant staff will be automatically notified through an alarm according to the safety score, and the recording function will be activated to save the image data before and after the incident;

[0010] After assembly is completed, a data-driven quality prediction model is established by integrating data from each link on the production line. Based on the prediction results, potential quality problems can be automatically corrected to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0011] Optionally, the optimizing the assembly sequence of components by using a heuristic algorithm based on historical production data of the current vehicle model and existing parts inventory conditions includes:

[0012] Extract the assembly data of the target vehicle model from the historical production database, and obtain the current parts inventory data, clean and format the assembly data and the current parts inventory data, remove outliers and incomplete data items, and ensure the accuracy and consistency of the data, wherein the assembly data includes assembly sequence, time consumption, output and common problem information, and the parts inventory data includes inventory quantity, supplier delivery cycle, and parts importance level information;

[0013] Determining optimization goals, wherein the optimization goals include minimizing assembly time, maximizing production line efficiency, and improving assembly quality, wherein weight factors between the goals are determined according to the enterprise production strategy;

[0014] The assembly order is encoded into a chromosome structure, and the sequence encoding is used to represent the assembly order of each part. The fitness function is constructed to transform the multi-objective optimization into a single fitness value. The weighted summation method is used to combine the deviation between the actual value and the expected value of each objective to evaluate the quality of each chromosome.

[0015] Define heuristic crossover operations, combine the characteristics of the current assembly line station layout, select stations that can operate in parallel for crossover, reduce the overall assembly time, and implement heuristic mutation strategies. For assembly steps that are known to often cause bottlenecks in historical data, introduce local disturbances and dynamically adjust their positions in the sequence in order to try out a better solution;

[0016] The simulated annealing algorithm is used as a local search strategy to fine-tune the current optimal solution to avoid falling into the local optimum and obtain the optimal component assembly sequence.

[0017] Optionally, assembling components on an automobile production line according to the optimized component assembly sequence and monitoring the working status and progress of each assembly station include:

[0018] After receiving the optimized assembly sequence of components, the assembly tasks are assigned to the corresponding assembly station workstations using the automated production scheduling system, and a task queue is generated for each workstation according to the assembly sequence list to ensure that the work is carried out in sequence, wherein the task queue contains the assembly sequence and related requirements of each component;

[0019] Automatically check the availability of required tools and parts at each workstation before starting assembly, use the material tracking system to ensure that all required parts are in place, activate the required tools according to the assembly task, and set the workstation status to ready;

[0020] Collect progress data of the assembly process in real time, update the work status identification of each workstation in real time, and record time data in the enterprise resource planning system, wherein the progress data includes the timestamps of the start, progress and completion of the assembly, and the work status identification includes in progress, paused, and completed;

[0021] Data analysis algorithms are used to analyze workstation progress data in real time, compare it with the expected assembly rhythm, identify progress deviations or abnormal conditions, and feed back progress deviations or abnormal conditions to workshop managers through the alarm system.

[0022] Optionally, the combination of image recognition and behavior analysis can be used to conduct multi-level real-time monitoring of the work operations of workers at the assembly station to generate a real-time safety score, including:

[0023] The multiple sensors arranged at the assembly station capture the worker's movement trajectory, workstation environment and behavior details. The multiple sensors include high-definition cameras, depth cameras, motion sensors and environmental sensors. Image data and non-image data are captured synchronously, and the time consistency of multi-source data is ensured by timestamp alignment.

[0024] Input the real-time collected image data into the trained behavior recognition model to identify the operation steps and postures performed by the workers, and automatically classify the worker's behavior into normal operation, potential risk operation or illegal operation, wherein the behavior recognition model is trained based on the labeled worker operation video image dataset;

[0025] The safety score of each dimension is dynamically calculated by using the results of behavioral pattern analysis and combining it with environmental sensor data. According to the safety score of each dimension, a real-time total safety score of the workers' operation at the workstation is comprehensively generated, wherein the dimensions include: action accuracy, operation standardization, posture safety and environmental adaptability. Action accuracy: evaluates the matching degree between the workers' actions and standard operations, and the safety score ranges from 0 to 1, where 1 indicates a perfect match; Operation standardization: based on the degree to which the workers comply with safety processes and steps, the safety score ranges from 0 to 1, where 1 indicates full compliance; Posture safety score: evaluates whether the workers' posture meets ergonomic standards, and the safety score ranges from 0 to 1, where 1 indicates full compliance; Environmental adaptability score: based on environmental sensor data, evaluates the impact of the workstation environment on work operations, and the safety score ranges from 0 to 1, where 1 indicates that the environment is completely suitable.

[0026] Another embodiment of the present application provides a final assembly system for an automobile production line, the system comprising:

[0027] An optimization module that uses historical production data based on the current model and existing parts inventory to optimize the assembly sequence of components using a heuristic algorithm;

[0028] The assembly module is used to assemble components on the automobile production line according to the optimized component assembly sequence and monitor the working status and progress of each assembly station;

[0029] The monitoring module combines image recognition and behavior analysis to conduct multi-level real-time monitoring of the work operations of assembly station workers and generate real-time safety scores. Electronic billboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time.

[0030] A notification module is used to automatically notify relevant staff members through an alarm when potential safety hazards or irregular operations are detected according to the safety score, and to start the recording function to save the image data before and after the incident;

[0031] The correction module is used to establish a data-driven quality prediction model by integrating data from various links on the production line after assembly is completed. It can automatically correct potential quality problems based on the prediction results to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0032] Yet another embodiment of the present application provides a storage medium, wherein the storage medium stores a computer program, wherein the computer program is configured to execute any of the above methods when running.

[0033] Yet another embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute any of the methods described above.

[0034] Compared with the prior art, the present invention provides a final assembly method for an automobile production line, which optimizes the assembly sequence of components by using a heuristic algorithm based on historical production data of the current model and the existing parts inventory; assembles components on the automobile production line according to the optimized assembly sequence, and monitors the working status and progress of each assembly station; combines image recognition and behavior analysis to conduct multi-level real-time monitoring of the work operations of workers at the assembly station, and generates a real-time safety score; when potential safety hazards or irregular operations are monitored, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is started to save the image data before and after the incident; after the assembly is completed, a data-driven quality prediction model is established by integrating the data of each link on the production line, so as to automatically correct potential quality problems according to the prediction results, and ensure that the assembly quality of each vehicle off the line meets the standards, thereby optimizing the assembly process, strengthening worker safety monitoring and improving the quality control level, and realizing a comprehensive intelligent transformation from assembly process to quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A hardware structure block diagram of a computer terminal for a final assembly method of an automobile production line provided by an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of a flow chart of an assembly method for an automobile production line provided by an embodiment of the present invention;

[0037] Figure 3 A schematic structural diagram of an assembly system for an automobile production line provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, but should not be construed as limiting the present invention.

[0039] The embodiment of the present invention first provides a final assembly method for an automobile production line. The method can be applied to electronic equipment, such as a computer terminal, specifically a common computer, etc.

[0040] The following describes it in detail by taking running on a computer terminal as an example. Figure 1 The hardware structure block diagram of a computer terminal of an assembly method of an automobile production line provided by an embodiment of the present invention. Figure 1 As shown, the computer terminal may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Optionally, the computer terminal may also include a transmission device 106 for communication functions and an input and output device 108. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0041] The memory 104 can be used to store software programs and modules of application software, such as program instructions / modules corresponding to the final assembly method of the automobile production line in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implementing the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0042] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] See also Figure 2 , an embodiment of the present invention provides a final assembly method for an automobile production line, which may include the following steps:

[0044] S201, uses a heuristic algorithm to optimize the assembly sequence of components based on historical production data and existing parts inventory based on the current model;

[0045] By utilizing the historical production data of current models and the existing parts inventory, a heuristic algorithm is used to optimize the assembly sequence of components. The core of this process is to analyze and integrate historical data to formulate the optimal assembly sequence in a scientific way. This can not only significantly improve assembly efficiency, reduce time waste between processes, ensure the proper allocation of parts required for each workstation, but also reduce assembly quality problems caused by human errors. Ultimately, this optimization will enable companies to respond flexibly to market changes and improve the overall production responsiveness and competitiveness.

[0046] Specifically, the assembly data of the target vehicle model can be extracted from the historical production database, and the current parts inventory data can be obtained. The assembly data and the current parts inventory data can be cleaned and formatted to remove abnormal values ​​and incomplete data items to ensure the accuracy and consistency of the data. The assembly data includes assembly sequence, time consumption, output and common problem information, and the parts inventory data includes inventory quantity, supplier delivery cycle, and parts importance level information.

[0047] Extract relevant information from historical production and current inventory data to provide the necessary basic data for assembly sequence optimization. This includes time consumption during assembly, production output, and common problems that occur, which can help identify factors that affect assembly efficiency. By cleaning and formatting data, ensure data accuracy and consistency to avoid optimization failures due to incorrect data.

[0048] Assembly data: refers to all data information related to vehicle assembly, including the assembly order of each component (such as which component is installed first), time consumption (the time required to assemble a certain component), output (the number of vehicles assembled per unit time) and information on common problems (such as historical assembly errors); Parts inventory data: including the number of currently available parts, the supplier delivery cycle of each part (the time required from order to delivery) and the importance level of parts (such as key components and accessories).

[0049] First, you can connect to the historical production database to extract data and use defined SQL query statements to obtain relevant assembly data of the target model (such as assembly sequence, time consumption, output, etc.) and current parts inventory data (such as inventory quantity, supplier delivery cycle, etc.). Next, establish a data preprocessing process, input the extracted data into the data cleaning module, apply statistical analysis methods such as Z-score or IQR (interquartile range) to identify and remove outliers, and process missing data, using the mean or median to fill in missing values. Then, through standardization or normalization, data of different dimensions are converted into a consistent format to ensure the accuracy and consistency of the data in subsequent analysis.

[0050] Determining optimization goals, wherein the optimization goals include minimizing assembly time, maximizing production line efficiency, and improving assembly quality, wherein weight factors between the goals are determined according to the enterprise production strategy;

[0051] Before optimizing the assembly sequence of components, it is crucial to clarify the different optimization objectives. Each objective has its specific measurement criteria, and there may be conflicts between different objectives. Therefore, by setting weight factors for each objective, a comprehensive optimization solution can be achieved so that multiple objectives can be considered simultaneously.

[0052] Optimization goal: refers to the specific results that you hope to achieve during the optimization process, such as minimizing assembly time (shortening the total time to complete assembly), maximizing production line efficiency (maximizing the workload of each workstation), and improving assembly quality (ensuring that the assembled vehicle meets the standards); Weight factor: In multi-objective optimization, the importance of each goal may be different. By assigning weights to each goal (for example, 40% for time, 40% for efficiency, and 20% for quality), you can better balance the various requirements.

[0053] The opinions of relevant departments such as production, quality and logistics can be collected to clarify the content and importance of each optimization goal. The Delphi method is used to conduct multiple rounds of anonymous surveys to collect professionals' views on each goal, and then comprehensively evaluate the priority of each goal. On this basis, a weight factor is set for each goal. For example, the minimum assembly time can be set to 0.4, the maximum production line efficiency can be set to 0.4, and the improvement of assembly quality can be set to 0.2 to ensure that the final weight distribution reflects the company's strategic goals.

[0054] The assembly order is encoded into a chromosome structure, and the sequence encoding is used to represent the assembly order of each part. The fitness function is constructed to transform the multi-objective optimization into a single fitness value. The weighted summation method is used to combine the deviation between the actual value and the expected value of each objective to evaluate the quality of each chromosome.

[0055] The assembly sequence is converted into the chromosome structure used in the genetic algorithm, so that the algorithm can search and optimize effectively. The fitness function is used to evaluate the performance of the current assembly sequence on the optimization goal, and multiple goals are combined into a comprehensive evaluation value through the weighted summation method, which is convenient for comparison and selection.

[0056] Chromosome structure: In genetic algorithms, chromosomes are data structures used to represent solutions, usually encoding a set of parameters or variables. Here, the assembly order is converted into a chromosome, where each gene represents the assembly position of a part; Fitness function: A function that evaluates the quality of the current chromosome, usually considering multiple objectives and merging them into a single value through a weighted summation method for easy comparison.

[0057] The assembly order is converted into a chromosome using a sequence encoding method, where each part represents its assembly order at a specific position in the chromosome. Next, a fitness function is constructed, by calculating the deviation between the actual value and the expected value of each target (assembly time, production line efficiency, and assembly quality), and applying the weighted summation method to integrate them into a comprehensive fitness value. For example, the fitness value Fitness can be calculated using the following formula:

[0058] Among them, w_1, w_2, and w_3 are corresponding weight factors. Finally, the fitness value is used to evaluate and select chromosomes to generate a new assembly order.

[0059] Define heuristic crossover operations, combine the characteristics of the current assembly line station layout, select stations that can operate in parallel for crossover, reduce the overall assembly time, and implement heuristic mutation strategies. For assembly steps that are known to often cause bottlenecks in historical data, introduce local disturbances and dynamically adjust their positions in the sequence in order to try out a better solution;

[0060] Crossover and mutation are important components of genetic algorithms. Through these operations, new solution spaces can be explored to try to generate a better assembly sequence. Combined with the actual workstation layout, the heuristic crossover operation can effectively shorten the assembly time, while the mutation strategy provides the flexibility to adapt to environmental changes, especially for the bottleneck steps identified in historical data.

[0061] Heuristic crossover operation: In genetic algorithms, crossover operation is used to generate new solutions, selecting two parent chromosomes to combine to form a child chromosome. Combined with the workstation layout, selecting workstations that can operate in parallel can reduce the waiting time caused by process dependence; mutation strategy: introduce randomness to change the assembly order of certain parts to explore new solution space and avoid the optimization process falling into local optimality.

[0062] When selecting a crossover operation, based on the current station layout, stations that can be operated in parallel are identified, and two chromosomes are randomly selected from them for crossover operations to generate new chromosomes. For example, crossover can be performed between parts that meet the conditions for parallel operation to ensure that the newly generated assembly sequence is still feasible. At the same time, a mutation strategy is implemented to randomly perturb assembly steps that have repeatedly appeared as bottlenecks in historical data, and slightly adjust their positions in the sequence to explore possible better solutions. This method can effectively improve the coverage of the solution space and explore potential excellent assembly sequences.

[0063] The simulated annealing algorithm is used as a local search strategy to fine-tune the current optimal solution to avoid falling into the local optimum and obtain the optimal component assembly sequence.

[0064] As a global search algorithm, simulated annealing helps avoid falling into the local optimal solution, making the optimization process more thorough. By making fine adjustments based on the current optimal solution, the assembly sequence of components can be further improved to ensure the final assembly efficiency and quality.

[0065] Based on the current optimal solution, an initial temperature and cooling rate are set to randomly generate neighborhood solutions. According to the acceptance criteria of simulated annealing, the fitness of the new solution is calculated and compared with the fitness of the current solution. If the fitness of the new solution is better, the new solution is accepted; if it is not better, the solution is accepted according to a certain probability to avoid falling into the local optimum. After multiple iterations, the temperature is lowered to gradually converge, and finally a deeply optimized component assembly sequence is obtained to ensure the highest efficiency and assembly quality of the production line.

[0066] S202, assembling components on the automobile production line according to the optimized component assembly sequence, and monitoring the working status and progress of each assembly station;

[0067] Improve assembly efficiency and quality on the automotive production line through optimized assembly sequence. The optimized component assembly sequence is based on the analysis of historical production data and existing parts inventory to determine the best assembly process.

[0068] Specifically, after receiving the optimized component assembly sequence, the automated production scheduling system can be used to assign assembly tasks to corresponding assembly station workstations, and a task queue can be generated for each workstation according to the assembly sequence list to ensure that the work is carried out in sequence, wherein the task queue contains the assembly sequence and related requirements of each component;

[0069] This step involves using the production scheduling system to automatically convert the optimized assembly sequence into specific work tasks and assign them to each workstation. The task queue lists in detail the assembly time, sequence and special requirements of each component. Automatic task allocation reduces manual intervention and can respond to production needs more quickly; the generation of the task queue ensures that the assembly sequence between workstations can be strictly followed, thereby avoiding assembly problems caused by incorrect sequence.

[0070] The automated production scheduling system is an integrated management software that aims to optimize the task allocation, resource management and real-time monitoring of the production process, making the work in the production line more efficient and orderly. For example, in the automobile manufacturing industry, the automated production scheduling system can automatically assign the assembly tasks of various parts of the car to different workstations based on real-time production data to ensure that production is carried out according to the predetermined plan. Assembly tasks refer to specific work items that need to be performed during the production process, usually including steps such as installation, connection, and debugging of specific parts. In automobile production, an assembly task can be connecting the engine to the body, which requires clear steps, tools and time limits. The task queue refers to an orderly arranged list of tasks that contain tasks and their related requirements that need to be performed at various workstations on the production line. In the assembly process of electric vehicles, the task queue can include tasks such as installing batteries, connecting cables, and installing wheels, each of which has its own specific order and requirements.

[0071] First, the assembly sequence of components generated by the optimization algorithm can be received and stored in a database in a structured form to ensure the accessibility of assembly sequence and part information. Information of all assembly stations, including the capacity, current workload and available tools of each station, is retrieved in real time to ensure the rationality of resource allocation. The priority of each assembly task is evaluated based on the part characteristics and station capacity in the optimized assembly sequence. The priority can be determined based on the importance of the part, the complexity of the assembly and the availability of the station. Next, the assembly tasks are assigned to the corresponding stations, with priority given to stations with matching capabilities and lower current loads. The assembly sequence and requirements of each task are clearly recorded in the task queue for subsequent tracking. A task queue is generated for each station to ensure that the tasks of each station are arranged in the predetermined assembly sequence. The task queue will include detailed information of each component, such as assembly time, required tools and execution order. After the task queue is generated, the system will send a notification to the workshop manager to ensure that all relevant personnel are aware of the current task allocation and can view and adjust the task queue at any time.

[0072] Automatically check the availability of required tools and parts at each workstation before starting assembly, use the material tracking system to ensure that all required parts are in place, activate the required tools according to the assembly task, and set the workstation status to ready;

[0073] This step ensures that tools and materials are ready for each workstation before assembly work begins. The material tracking system is used to obtain the status of parts and tools in real time to avoid production delays due to missing equipment. This check ensures that all key tools and materials are in place, greatly reducing interruptions caused by insufficient materials; activating task-related tools allows workers to quickly enter the operating state and improve work efficiency. The material tracking system is a technical means, usually through barcodes, QR codes or RFID tags, to track and manage materials in production in real time to ensure that the required materials and tools are in the right place. In the production of electronic products, the material tracking system can track the inventory status of parts in real time to ensure that all components are available at a certain moment in the production line.

[0074] Based on the generated task queue, the system automatically generates a list of tools and parts required for each workstation, including tool names, specifications, quantities and detailed information of parts. The material tracking system is used to query the current inventory status in real time and confirm the availability of the required tools and parts. Through RFID tags or barcode scanning, the system can quickly identify the location of inventory materials. Based on the generated list, the system checks the availability of the tools and parts required for each workstation. If a tool or part is missing, the system will issue an alarm to notify the relevant personnel to replenish it. For available tools, the system will automatically activate the corresponding equipment, such as starting the automated tool or preparing it to the designated workstation, to ensure that the worker can start the work quickly. After confirming that all tools and parts are in place, the system updates the workstation status to "Ready" and displays the status of the workstation and the tasks to be completed to the worker through the electronic board or mobile device.

[0075] Collect progress data of the assembly process in real time, update the work status identification of each workstation in real time, and record time data in the enterprise resource planning system, wherein the progress data includes the timestamps of the start, progress and completion of the assembly, and the work status identification includes in progress, paused, and completed;

[0076] Through the real-time data collection system, various data in the assembly process are monitored, including the start time, elapsed time and completion time of the assembly, and this information is updated in real time to the enterprise resource planning system. This ensures that workshop managers can obtain the status of each workstation in a timely manner to facilitate quick decision-making; the recorded time data provides an important basis for subsequent production analysis, efficiency improvement and problem tracking.

[0077] The workstation status identification is an identifier used to indicate the current operating status of each workstation, usually including different states such as in progress, paused, and completed, so as to facilitate monitoring and management of the production process. On the automated assembly line, if a workstation is displayed as "in progress", it means that the workstation is assembling products, while if it is displayed as "paused", it means that the workstation has temporarily stopped working for some reason. The enterprise resource planning system is an integrated management software used for the comprehensive management of information of various departments of an enterprise (such as finance, production, inventory management, etc.), to promote the optimal allocation and efficient use of resources. In automobile manufacturing companies, the ERP system can help manage sales orders, production progress, inventory quantities, etc., improve overall operational efficiency, and realize data sharing.

[0078] After the assembly begins, the progress data of the assembly process is collected in real time through sensors and data acquisition devices deployed at each workstation. These devices may include motion sensors, time recorders, and visual monitoring systems. The start time of the assembly is recorded at each workstation, and the end time of the worker's operation is automatically monitored to generate the completion timestamp of the assembly. All timestamp data will be stored in the database in real time. Through data analysis, the system will update the work status identification of each workstation in real time, marking it as "in progress" when the assembly is in progress, "paused" if it is suspended for some reason, and "completed" when it is completed. This status information will also be synchronized and updated in the enterprise resource planning (ERP) system. Finally, the system summarizes the progress data of all workstations, compares it with the scheduled production rhythm, and generates a progress report to provide a basis for subsequent analysis.

[0079] Data analysis algorithms are used to analyze workstation progress data in real time, compare it with the expected assembly rhythm, identify progress deviations or abnormal conditions, and feed back progress deviations or abnormal conditions to workshop managers through the alarm system.

[0080] In this step, through real-time analysis of progress data, algorithms are used to determine whether the actual progress meets the preset production rhythm. If deviations are found, the system will automatically alarm and notify management personnel. By monitoring the progress, important deviations can be identified at the early stage of the problem, and timely measures can be taken to adjust the production plan; the automatic alarm mechanism enables workshop managers to respond to problems quickly and reduce potential production losses.

[0081] Schedule data refers to the actual time, completion status and other information recorded in each link of the production process, which is usually used to compare the scheduled production schedule. Schedule data can include the comparison of the start time and completion time of each process with the estimated time, which helps to evaluate production efficiency. Schedule deviation refers to the difference between the actual production progress and the planned progress, reflecting the operating efficiency and accuracy of the production process. If an assembly station is expected to complete a task within 8 hours, but the actual completion time is 10 hours, there is a 2-hour schedule deviation.

[0082] The progress data collected from each workstation is preprocessed, including data cleaning and formatting, to ensure data accuracy and consistency. Real-time data analysis algorithms (such as time series analysis) are used to dynamically analyze the assembly progress of each workstation and compare it with the preset production rhythm. The algorithm can quickly identify deviations that exceed the set range. Based on the set threshold standard, when the assembly progress of the workstation deviates significantly from expectations, the system will automatically trigger abnormal state detection and record relevant data. When a progress deviation or abnormal state is detected, the system will immediately notify the workshop management personnel through the alarm system, including detailed information of the current workstation and the degree of deviation, so that relevant personnel can take corrective measures in time.

[0083] The above steps aim to ensure efficient operation and management of the production line and improve the efficiency and quality of the overall assembly process.

[0084] S203, combining image recognition and behavior analysis, to conduct multi-level real-time monitoring of the work operations of workers at the assembly stations and generate real-time safety scores, wherein electronic signboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time;

[0085] This step aims to monitor and score the workers at the assembly station in real time through modern technology to ensure the safety and standardization of the assembly process. Combining image recognition and behavior analysis, it is possible to analyze in real time whether the workers' operating actions meet safety standards, and generate a safety score based on the monitoring results. The core of this process lies in the application of multiple sensors, including high-definition cameras, depth cameras, motion sensors, etc. These devices can fully capture the workers' movement trajectories, workstation environments, and behavioral details. Subsequently, the collected image data is input into the trained behavior recognition model to automatically identify the workers' operating steps and classify them. In addition, the generated real-time safety score is also fed back to the workers in real time through electronic billboards or handheld devices to ensure that workers can understand the safety status of individuals and assembly stations in a timely manner during work.

[0086] Specifically, multiple sensors arranged at the assembly station can be used to capture the worker's movement trajectory, workstation environment, and behavior details. The multiple sensors include high-definition cameras, depth cameras, motion sensors, and environmental sensors. Image data and non-image data are captured synchronously, and the temporal consistency of multi-source data is ensured by timestamp alignment.

[0087] A large number of sensors are arranged to collect every action and operating environment made by workers during the assembly process. These sensors include high-definition cameras to capture workers' visual actions, depth cameras to identify distance and spatial relationships, motion sensors to record workers' dynamic behaviors, and environmental sensors to obtain environmental data of workstations. The image data and non-image data generated by these sensors will be synchronized based on timestamps to ensure the temporal consistency of the data during analysis.

[0088] Multiple sensors are arranged in the assembly station, including high-definition cameras placed above the workstation to obtain a panoramic view, depth cameras installed at key operating positions to capture distance and spatial information, and motion sensors within the worker's range of motion to detect specific movements. At the same time, environmental sensors monitor environmental factors such as temperature, humidity, and light. Ensure that each sensor can cover all important operating areas in the workstation to fully capture worker behavior. When all sensors start running, they collect data in real time. The high-definition camera is responsible for obtaining image data of the worker's movements, the depth camera records three-dimensional coordinate information, and the motion sensor monitors the worker's movement trajectory and strength. The environmental sensor captures environmental data around the workstation in parallel. This process will continue to ensure that all operations of the worker throughout the assembly process are covered. Each sensor will generate a timestamp at the same time when collecting data, ensuring that all data can be marked as data within the same time period. The precise alignment of this timestamp is an important basis for subsequent data analysis. It can provide consistency for the fusion of multi-source data, so that data at the same time point can be accurately associated. The data from different sensors are processed synchronously through algorithms. Align the image data with the data collected from the depth camera, motion sensor, and environmental sensor to ensure that they can be associated with each other in the same time sequence. For example, interpolation methods are used to adjust the differences between timestamps so that all data at a specific time point can be merged together to form a complete time series data set.

[0089] Input the real-time collected image data into the trained behavior recognition model to identify the operation steps and postures performed by the workers, and automatically classify the worker's behavior into normal operation, potential risk operation or illegal operation, wherein the behavior recognition model is trained based on the labeled worker operation video image dataset;

[0090] The image data collected in real time is input into a trained behavior recognition model, which can identify the specific operation steps and postures performed by the worker. By classifying the worker's operation (such as normal, potential risk or illegal operation), the system can quickly feedback whether the worker's behavior meets safety standards. The behavior recognition model is trained based on a well-labeled operation video image dataset and has high accuracy and reliability.

[0091] First, select a suitable deep learning model (such as a convolutional neural network, CNN) and train it using labeled worker operation video datasets. These datasets cover a variety of standard operations, potential risk operations, and illegal operations. Through repeated training and adjustment of hyperparameters, the accuracy and robustness of the model are optimized to ensure that the model can effectively identify and classify workers' behaviors. Before the real-time collected image data is input into the behavior recognition model, the data needs to be preprocessed. This includes image scaling, cropping, and normalization to ensure that the input data is consistent with the data format during training. In addition, image enhancement is required to improve the model's adaptability to different working environments and lighting conditions. The processed image data is analyzed in real time using the trained model. The model analyzes the worker's actions frame by frame and classifies the actions according to pre-set thresholds. Each identified action will be labeled as "normal operation", "potential risk operation", or "illegal operation".

[0092] The safety score of each dimension is dynamically calculated by using the results of behavioral pattern analysis and combining it with environmental sensor data. According to the safety score of each dimension, a real-time total safety score of the workers' operation at the workstation is comprehensively generated, wherein the dimensions include: action accuracy, operation standardization, posture safety and environmental adaptability. Action accuracy: evaluates the matching degree between the workers' actions and standard operations, and the safety score ranges from 0 to 1, where 1 indicates a perfect match; Operation standardization: based on the degree to which the workers comply with safety processes and steps, the safety score ranges from 0 to 1, where 1 indicates full compliance; Posture safety score: evaluates whether the workers' posture meets ergonomic standards, and the safety score ranges from 0 to 1, where 1 indicates full compliance; Environmental adaptability score: based on environmental sensor data, evaluates the impact of the workstation environment on work operations, and the safety score ranges from 0 to 1, where 1 indicates that the environment is completely suitable.

[0093] By analyzing the worker's behavior patterns and combining them with environmental sensor data, the worker's operation safety score is dynamically calculated. These safety scores are evaluated in multiple dimensions, including motion accuracy, operational standardization, posture safety, and environmental adaptability. The scores of each dimension will be combined to form a real-time overall safety score. The worker's behavior is scored in detail and each dimension is evaluated according to preset standards. Motion accuracy mainly measures the compatibility of the worker's movements with standard operations, operational standardization evaluates the degree to which the worker follows the procedures during the operation, and the posture safety score focuses on whether the worker's working posture meets ergonomic requirements. The environmental adaptability score analyzes the impact of the workstation environment on work operations through environmental sensor data.

[0094] First, the dimensions of the safety score are clearly defined: motion accuracy, operational standardization, posture safety, and environmental adaptability. Combine the behavior recognition results collected in the previous step with the data from the environmental sensors to monitor each dimension in real time. Set quantitative evaluation standards for each dimension. For example, motion accuracy can be obtained by calculating the match between the worker's operation and the standard process, operational standardization is evaluated by checking the steps followed by the worker in the execution process, posture safety is scored based on angle measurement and ergonomic specifications, and environmental adaptability is evaluated by associating environmental sensor data (such as temperature, humidity, and light). Based on the acquired data, the corresponding safety score is dynamically calculated for each dimension through an algorithm. The score range for each dimension is set to 0 to 1, where 1 indicates full compliance with the standard. The weighted average method can be used to combine the data of each dimension to ensure that the importance of each dimension is reflected in the overall safety score. The safety scores of each dimension are weighted and combined to obtain the real-time total safety score of the worker's operation at the workstation. The system feeds this score back to the monitoring center and the worker's personal device in real time to ensure that relevant personnel are aware of the current safety status at any time.

[0095] Among them, a calculation model is designed for the scoring of each dimension to ensure that the collected real-time data can be accurately evaluated. For example, the accuracy of the action can be designed to achieve scoring by comparing the similarity between the worker's operation trajectory and the preset standard trajectory (dynamic time warping algorithm can be used). The threshold of each evaluation criterion is set to ensure that all scores are between 0 and 1. For example, if the deviation between the worker's posture and the standard posture is within an acceptable range, the posture safety score is 1; if the deviation is large, the corresponding score is reduced. Based on the real-time data of environmental sensors (such as temperature, humidity, noise, etc.), reasonable thresholds are set to determine whether the environment restricts the worker's operation. For example, when the temperature is too high or the humidity is too high, the score will drop accordingly. Finally, the scores of each dimension are integrated to generate a comprehensive safety score, which is ensured to be passed to the monitoring system and the worker's visualization equipment to facilitate real-time monitoring and necessary improvements.

[0096] S204, when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is activated to save the image data before and after the incident;

[0097] This step is to respond to safety hazards or irregular operations on the production line in a timely manner through an automated monitoring system to ensure the safety of workers and the smooth progress of production. Specifically, when the system detects that the worker's operation safety score is lower than the preset threshold, it means that there is a potential risk, and the system will immediately trigger the early warning mechanism. This not only involves notifying relevant staff to intervene in a timely manner, but also provides a basis for subsequent investigations and problem analysis by recording visual data during the process, so as to enhance safety management and improve production processes.

[0098] Specifically, the real-time safety score can be continuously monitored, and when the score is lower than the preset safety threshold (for example, 0.6), the potential risk alarm mechanism is triggered. The threshold setting can be dynamically adjusted based on historical safety data and the company's safety management strategy. A multi-level alarm system can be designed to use a combination of audio and visual signals for alarm notifications. Audio alarms can be set to different tones and volumes to distinguish different levels of safety hazards. For example, low tones are used for minor hazards and high tones are used for serious hazards. Instant notifications are sent to relevant personnel (such as workshop supervisors, safety officers) through the company's internal communication system (such as instant messaging software or dedicated applications). The notification should include the location, time, type of hazard and necessary response measures.

[0099] The moment a potential safety hazard is detected, the system automatically starts the recording module of the image data. This module should include the following functions: Short-term recording: At the moment the alarm is triggered, the system starts recording the current image data and goes back several minutes (for example, 5 minutes) to capture the behavioral details before and after the incident; Data storage format: Image data is compressed and saved using an efficient video encoding format (such as H.264) to save storage space and facilitate subsequent queries; Data tagging: Each saved image data segment must be accompanied by a timestamp and event tag so that it can be quickly located during subsequent retrieval.

[0100] After receiving the alarm, relevant staff can view real-time and historical video data through the system interface to quickly determine the cause of the incident, the worker's operating specifications and environmental factors. After the post-analysis is completed, the system can automatically generate a report containing the incident, the monitored safety score changes, surveillance video clips and subsequent improvement measures for management reference and decision-making.

[0101] Through the above specific implementation steps, the safety of the production line can be effectively guaranteed, potential safety hazards can be dealt with in a timely manner, and strong data support can be provided for future safety management.

[0102] S205, after the assembly is completed, a data-driven quality prediction model is established by integrating the data from each link on the production line. Based on the prediction results, potential quality problems can be automatically corrected to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0103] By using intelligent data analysis and machine learning technology, we can establish an efficient quality prediction model by deeply analyzing and integrating the data collected from each link of the production line. This model is designed to evaluate assembly quality in real time, identify potential quality problems, and then automatically adjust production processes or equipment parameters to ensure that every vehicle that finally rolls off the production line meets strict quality standards. This process not only improves production efficiency, but also provides a data-driven scientific basis for quality control.

[0104] Specifically, after assembly is completed, relevant data generated in each link is collected, including assembly time, parts quality inspection data, equipment operating parameters, and workers' operating behaviors during the production process. The collected data is cleaned and standardized to ensure compatibility and availability between various data sources. Outliers and missing data items are removed, and interpolation is used to fill in the gaps to ensure data integrity.

[0105] Extract key features from the fused data through data mining technology. These features may include assembly time deviation, part quality score, equipment failure rate, etc. Use statistical test methods such as chi-square test or correlation coefficient analysis to select features that are highly correlated with quality prediction for subsequent modeling. According to the preprocessed data, select a suitable machine learning algorithm (such as random forest, support vector machine or gradient boosting machine) for model training. These algorithms can effectively handle complex nonlinear relationships and have strong generalization capabilities. Divide the samples into training sets and test sets, use cross-validation methods to tune the model hyperparameters, and evaluate the prediction effect of the model. Finally, select the best performing model for application.

[0106] Deploy the trained model to the production line to monitor key quality indicators in the assembly process in real time. When the model detects potential quality issues (such as the predicted quality score is lower than the threshold), the system generates an alarm and initiates a correction mechanism. Automatically adjust the operating parameters of related equipment (such as adjusting the tightening force, the frequency of part replacement) or optimize the assembly line operation sequence through PLC (programmable logic controller) or other industrial control systems. Continuously collect new data and update the model so that the quality prediction can adapt to changes in the production line and maintain the accuracy of the model. After each production cycle, the system automatically generates a quality report, which includes the prediction results, the actual quality status and the effect analysis of the adjustment measures to help management make decisions. Using ensemble learning methods, the newly collected data is fed back to the model for retraining to continuously improve the prediction accuracy and reliability of the model.

[0107] Through the above specific implementation steps, a data-driven quality prediction model can be efficiently established to ensure that the assembly quality of every vehicle off the production line meets the standards, while providing strong data support for the optimization and improvement of the production line.

[0108] It can be seen that the heuristic algorithm is used to optimize the component assembly sequence by utilizing the historical production data based on the current model and the existing parts inventory; according to the optimized component assembly sequence, the components are assembled on the automobile production line, and the working status and progress of each assembly station are monitored; combined with image recognition and behavior analysis, the work operations of the workers at the assembly station are monitored in real time at multiple levels to generate a real-time safety score; when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is started to save the image data before and after the incident; after the assembly is completed, a data-driven quality prediction model is established by integrating the data of each link on the production line, so as to automatically correct potential quality problems according to the prediction results, and ensure that the assembly quality of each vehicle off the line meets the standards, thereby optimizing the assembly process, strengthening worker safety monitoring and improving the quality control level, and realizing a comprehensive intelligent transformation from assembly process to quality control.

[0109] Another embodiment of the present invention provides an assembly system for an automobile production line. Figure 3 , the system may include:

[0110] The optimization module 301 is used to optimize the assembly sequence of components by using a heuristic algorithm based on the historical production data of the current vehicle model and the existing parts inventory;

[0111] An assembly module 302, for assembling components on the automobile production line according to an optimized component assembly sequence, and monitoring the working status and progress of each assembly station;

[0112] Monitoring module 303, used to combine image recognition and behavior analysis to conduct multi-level real-time monitoring of the work operations of assembly station workers and generate real-time safety scores, wherein electronic billboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time;

[0113] The notification module 304 is used to automatically notify relevant staff members through an alarm according to the safety score when potential safety hazards or irregular operations are detected, and to start the recording function to save the image data before and after the incident;

[0114] The correction module 305 is used to establish a data-driven quality prediction model by integrating the data of each link on the production line after the assembly is completed, so as to automatically correct potential quality problems according to the prediction results to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0115] It can be seen that the heuristic algorithm is used to optimize the component assembly sequence by utilizing the historical production data based on the current model and the existing parts inventory; according to the optimized component assembly sequence, the components are assembled on the automobile production line, and the working status and progress of each assembly station are monitored; combined with image recognition and behavior analysis, the work operations of the workers at the assembly station are monitored in real time at multiple levels to generate a real-time safety score; when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is started to save the image data before and after the incident; after the assembly is completed, a data-driven quality prediction model is established by integrating the data of each link on the production line, so as to automatically correct potential quality problems according to the prediction results, and ensure that the assembly quality of each vehicle off the line meets the standards, thereby optimizing the assembly process, strengthening worker safety monitoring and improving the quality control level, and realizing a comprehensive intelligent transformation from assembly process to quality control.

[0116] An embodiment of the present invention further provides a storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0117] Specifically, in this embodiment, the above storage medium may be configured to store a computer program for performing the following steps:

[0118] S201, uses a heuristic algorithm to optimize the assembly sequence of components based on historical production data and existing parts inventory based on the current model;

[0119] S202, assembling components on the automobile production line according to the optimized component assembly sequence, and monitoring the working status and progress of each assembly station;

[0120] S203, combining image recognition and behavior analysis, to conduct multi-level real-time monitoring of the work operations of workers at the assembly stations and generate real-time safety scores, wherein electronic signboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time;

[0121] S204, when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is activated to save the image data before and after the incident;

[0122] S205, after the assembly is completed, a data-driven quality prediction model is established by integrating the data from each link on the production line. Based on the prediction results, potential quality problems can be automatically corrected to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0123] It can be seen that the heuristic algorithm is used to optimize the component assembly sequence by utilizing the historical production data based on the current model and the existing parts inventory; according to the optimized component assembly sequence, the components are assembled on the automobile production line, and the working status and progress of each assembly station are monitored; combined with image recognition and behavior analysis, the work operations of the workers at the assembly station are monitored in real time at multiple levels to generate a real-time safety score; when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is started to save the image data before and after the incident; after the assembly is completed, a data-driven quality prediction model is established by integrating the data of each link on the production line, so as to automatically correct potential quality problems according to the prediction results, and ensure that the assembly quality of each vehicle off the line meets the standards, thereby optimizing the assembly process, strengthening worker safety monitoring and improving the quality control level, and realizing a comprehensive intelligent transformation from assembly process to quality control.

[0124] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0125] Specifically, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0126] Specifically, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0127] S201, uses a heuristic algorithm to optimize the assembly sequence of components based on historical production data and existing parts inventory based on the current model;

[0128] S202, assembling components on the automobile production line according to the optimized component assembly sequence, and monitoring the working status and progress of each assembly station;

[0129] S203, combining image recognition and behavior analysis, to conduct multi-level real-time monitoring of the work operations of workers at the assembly stations and generate real-time safety scores, wherein electronic signboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time;

[0130] S204, when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is activated to save the image data before and after the incident;

[0131] S205, after the assembly is completed, a data-driven quality prediction model is established by integrating the data from each link on the production line. Based on the prediction results, potential quality problems can be automatically corrected to ensure that the assembly quality of each vehicle off the production line meets the standards.

[0132] It can be seen that the heuristic algorithm is used to optimize the component assembly sequence by utilizing the historical production data based on the current model and the existing parts inventory; according to the optimized component assembly sequence, the components are assembled on the automobile production line, and the working status and progress of each assembly station are monitored; combined with image recognition and behavior analysis, the work operations of the workers at the assembly station are monitored in real time at multiple levels to generate a real-time safety score; when potential safety hazards or irregular operations are detected, the relevant staff are automatically notified through an alarm according to the safety score, and the recording function is started to save the image data before and after the incident; after the assembly is completed, a data-driven quality prediction model is established by integrating the data of each link on the production line, so as to automatically correct potential quality problems according to the prediction results, and ensure that the assembly quality of each vehicle off the line meets the standards, thereby optimizing the assembly process, strengthening worker safety monitoring and improving the quality control level, and realizing a comprehensive intelligent transformation from assembly process to quality control.

[0133] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings. The above is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the drawings. Any changes made according to the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the description and drawings, should be within the protection scope of the present invention.

Claims

1. A final assembly method for an automobile production line, characterized in that: The method comprises: Using historical production data based on the current model and existing parts inventory, a heuristic algorithm is used to optimize the assembly sequence of components; Assemble components on the automobile production line according to the optimized component assembly sequence and monitor the working status and progress of each assembly station; Combined with image recognition and behavior analysis, the work operations of assembly station workers are monitored in real time at multiple levels to generate real-time safety scores. Electronic billboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time. When potential safety hazards or irregular operations are detected, the relevant staff will be automatically notified through an alarm according to the safety score, and the recording function will be activated to save the image data before and after the incident; After assembly is completed, a data-driven quality prediction model is established by integrating data from various links on the production line. Based on the prediction results, potential quality problems can be automatically corrected to ensure that the assembly quality of each vehicle off the production line meets the standards. The method uses a heuristic algorithm to optimize the assembly sequence of components based on the historical production data of the current vehicle model and the existing parts inventory, including: Extract the assembly data of the target vehicle model from the historical production database, and obtain the current parts inventory data, clean and format the assembly data and the current parts inventory data, remove outliers and incomplete data items, and ensure the accuracy and consistency of the data, wherein the assembly data includes assembly sequence, time consumption, output and common problem information, and the parts inventory data includes inventory quantity, supplier delivery cycle, and parts importance level information; Determining optimization goals, wherein the optimization goals include minimizing assembly time, maximizing production line efficiency, and improving assembly quality, wherein weight factors between the goals are determined according to the enterprise production strategy; The assembly order is encoded into a chromosome structure, and the sequence encoding is used to represent the assembly order of each part. The fitness function is constructed to transform the multi-objective optimization into a single fitness value. The weighted summation method is used to combine the deviation between the actual value and the expected value of each objective to evaluate the quality of each chromosome. Define heuristic crossover operations, combine the characteristics of the current assembly line station layout, select stations that can operate in parallel for crossover, reduce the overall assembly time, and implement heuristic mutation strategies. For assembly steps that are known to often cause bottlenecks in historical data, introduce local disturbances and dynamically adjust their positions in the sequence in order to try out a better solution; Using simulated annealing algorithm as a local search strategy, fine-tune the current optimal solution to avoid falling into the local optimum and obtain the optimal component assembly sequence; The fitness function calculates the deviation between the actual value and the expected value of each target and integrates it into a comprehensive fitness value by using the weighted summation method; the calculation of the fitness value Fitness includes: ； Among them, w_1, w_2, w_3 are corresponding weight factors; the chromosomes are evaluated and selected by the fitness value to generate a new assembly order.

2. The method according to claim 1, characterized in that The method of assembling components on the automobile production line according to the optimized component assembly sequence and monitoring the working status and progress of each assembly station includes: After receiving the optimized assembly sequence of components, the assembly tasks are assigned to the corresponding assembly station workstations using the automated production scheduling system, and a task queue is generated for each workstation according to the assembly sequence list to ensure that the work is carried out in sequence, wherein the task queue contains the assembly sequence and related requirements of each component; Automatically check the availability of required tools and parts at each workstation before starting assembly, use the material tracking system to ensure that all required parts are in place, activate the required tools according to the assembly task, and set the workstation status to ready; Collect progress data of the assembly process in real time, update the work status identification of each workstation in real time, and record time data in the enterprise resource planning system, wherein the progress data includes the timestamps of the start, progress and completion of the assembly, and the work status identification includes in progress, paused, and completed; Data analysis algorithms are used to analyze workstation progress data in real time, compare it with the expected assembly rhythm, identify progress deviations or abnormal conditions, and feed back progress deviations or abnormal conditions to workshop managers through the alarm system.

3. The method according to claim 2, characterized in that The combination of image recognition and behavior analysis conducts multi-level real-time monitoring of the work operations of assembly station workers and generates real-time safety scores, including: The multiple sensors arranged at the assembly station capture the worker's movement trajectory, workstation environment and behavior details. The multiple sensors include high-definition cameras, depth cameras, motion sensors and environmental sensors. Image data and non-image data are captured synchronously, and the time consistency of multi-source data is ensured by timestamp alignment. Input the real-time collected image data into the trained behavior recognition model to identify the operation steps and postures performed by the workers, and automatically classify the worker's behavior into normal operation, potential risk operation or illegal operation, wherein the behavior recognition model is trained based on the labeled worker operation video image dataset; The safety score of each dimension is dynamically calculated by using the results of behavioral pattern analysis and combining it with environmental sensor data. According to the safety score of each dimension, a real-time total safety score of the workers' operation at the workstation is comprehensively generated, wherein the dimensions include: action accuracy, operation standardization, posture safety and environmental adaptability. Action accuracy: evaluates the matching degree between the workers' actions and standard operations, and the safety score ranges from 0 to 1, where 1 indicates a perfect match; Operation standardization: based on the degree to which the workers comply with safety processes and steps, the safety score ranges from 0 to 1, where 1 indicates full compliance; Posture safety score: evaluates whether the workers' posture meets ergonomic standards, and the safety score ranges from 0 to 1, where 1 indicates full compliance; Environmental adaptability score: based on environmental sensor data, evaluates the impact of the workstation environment on work operations, and the safety score ranges from 0 to 1, where 1 indicates that the environment is completely suitable.

4. A final assembly system for an automobile production line, characterized in that: The system comprises: An optimization module that uses historical production data based on the current model and existing parts inventory to optimize the assembly sequence of components using a heuristic algorithm; The assembly module is used to assemble components on the automobile production line according to the optimized component assembly sequence and monitor the working status and progress of each assembly station; The monitoring module combines image recognition and behavior analysis to conduct multi-level real-time monitoring of the work operations of assembly station workers and generate real-time safety scores. Electronic billboards or handheld devices are used to display the assembly tasks, work status and progress of each assembly station to workers in real time. A notification module is used to automatically notify relevant staff members through an alarm when potential safety hazards or irregular operations are detected according to the safety score, and to start the recording function to save the image data before and after the incident; The correction module is used to establish a data-driven quality prediction model by integrating data from various links on the production line after assembly is completed. It can automatically correct potential quality problems based on the prediction results to ensure that the assembly quality of each vehicle off the production line meets the standards; The optimization module is specifically used to: extract assembly data of the target vehicle model from the historical production database, obtain current parts inventory data, clean and format the assembly data and current parts inventory data, remove abnormal values ​​and incomplete data items, and ensure the accuracy and consistency of the data, wherein the assembly data includes assembly sequence, time consumption, output and common problem information, and the parts inventory data includes inventory quantity, supplier delivery cycle, and parts importance level information; Determining optimization goals, wherein the optimization goals include minimizing assembly time, maximizing production line efficiency, and improving assembly quality, wherein weight factors between the goals are determined according to the enterprise production strategy; The assembly order is encoded into a chromosome structure, and the sequence encoding is used to represent the assembly order of each part. The fitness function is constructed to transform the multi-objective optimization into a single fitness value. The weighted summation method is used to combine the deviation between the actual value and the expected value of each objective to evaluate the quality of each chromosome. Define heuristic crossover operations, combine the characteristics of the current assembly line station layout, select stations that can operate in parallel for crossover, reduce the overall assembly time, and implement heuristic mutation strategies. For assembly steps that are known to often cause bottlenecks in historical data, introduce local disturbances and dynamically adjust their positions in the sequence in order to try out a better solution; Using simulated annealing algorithm as a local search strategy, fine-tune the current optimal solution to avoid falling into the local optimum and obtain the optimal component assembly sequence; The fitness function calculates the deviation between the actual value and the expected value of each target and integrates it into a comprehensive fitness value by using the weighted summation method; the calculation of the fitness value Fitness includes: ； Among them, w_1, w_2, w_3 are corresponding weight factors; the chromosomes are evaluated and selected by the fitness value to generate a new assembly order.

5. The system according to claim 4, characterized in that The assembly module is specifically used for: After receiving the optimized assembly sequence of components, the assembly tasks are assigned to the corresponding assembly station workstations using the automated production scheduling system, and a task queue is generated for each workstation according to the assembly sequence list to ensure that the work is carried out in sequence, wherein the task queue contains the assembly sequence and related requirements of each component; Automatically check the availability of required tools and parts at each workstation before starting assembly, use the material tracking system to ensure that all required parts are in place, activate the required tools according to the assembly task, and set the workstation status to ready; Collect progress data of the assembly process in real time, update the work status identification of each workstation in real time, and record time data in the enterprise resource planning system, wherein the progress data includes the timestamps of the start, progress and completion of the assembly, and the work status identification includes in progress, paused, and completed; Data analysis algorithms are used to analyze workstation progress data in real time, compare it with the expected assembly rhythm, identify progress deviations or abnormal conditions, and feed back progress deviations or abnormal conditions to workshop managers through the alarm system.

6. The system according to claim 5, characterized in that The monitoring module is specifically used for: The multiple sensors arranged at the assembly station capture the worker's movement trajectory, workstation environment and behavior details. The multiple sensors include high-definition cameras, depth cameras, motion sensors and environmental sensors. Image data and non-image data are captured synchronously, and the time consistency of multi-source data is ensured by timestamp alignment. Input the real-time collected image data into the trained behavior recognition model to identify the operation steps and postures performed by the workers, and automatically classify the worker's behavior into normal operation, potential risk operation or illegal operation, wherein the behavior recognition model is trained based on the labeled worker operation video image dataset; The safety score of each dimension is dynamically calculated by using the results of behavioral pattern analysis and combining it with environmental sensor data. According to the safety score of each dimension, a real-time total safety score of the workers' operation at the workstation is comprehensively generated, wherein the dimensions include: action accuracy, operation standardization, posture safety and environmental adaptability. Action accuracy: evaluates the matching degree between the workers' actions and standard operations, and the safety score ranges from 0 to 1, where 1 indicates a perfect match; Operation standardization: based on the degree to which the workers comply with safety processes and steps, the safety score ranges from 0 to 1, where 1 indicates full compliance; Posture safety score: evaluates whether the workers' posture meets ergonomic standards, and the safety score ranges from 0 to 1, where 1 indicates full compliance; Environmental adaptability score: based on environmental sensor data, evaluates the impact of the workstation environment on work operations, and the safety score ranges from 0 to 1, where 1 indicates that the environment is completely suitable.

7. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 3 when executed.

8. An electronic device, comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 3.

Images