Robot grinding path optimization method based on industrial internet of things and MES linkage

Abstract

A robot polishing path optimization method based on industrial Internet of Things and MES linkage comprises the steps of deploying a multi-sensor network through the industrial Internet of Things to obtain real-time operation data of production line equipment, and preprocessing the real-time operation data by adopting an edge computing technology to construct a structured data stream; a reinforcement learning or heuristic algorithm is adopted to calculate an optimal task allocation scheme of a polishing process and upstream and downstream processes based on a structured data stream from the industrial Internet of Things in combination with order requirements, process requirements and historical operation data. And calling a self-adaptive path fine tuning algorithm to perform fine tuning on the offline programming path of the polishing robot in the process of executing the optimal task allocation scheme by the polishing robot so as to generate an optimal path. And the optimal path executed by the grinding robot is monitored in real time, real-time monitoring data are transmitted back to the MES system and the industrial Internet of Things platform, and historical operation data are analyzed through machine learning, so that self-adaptive iterative optimization is performed on the optimal path.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent industry and workpiece grinding, and particularly relates to a method for optimizing the grinding path of a robot based on the linkage between the industrial Internet of Things and MES. Background Art

[0002] With the continuous progress of industrial robot technology and the wide application of industrial robots in industries such as welding, cutting, engraving, grinding, and deburring, more and more manual grinding operations have been replaced by automated and unmanned grinding workstations, solving the occupational health problems and safety problems of operators, and the grinding quality and grinding efficiency have also been greatly improved.

[0003] In order to obtain higher grinding quality and grinding efficiency, strict requirements are imposed on the position, posture, and grinding path of the robot during the grinding process. At present, the programming of most grinding workstations belongs to the online teaching programming method, with very low programming efficiency. The teaching time for a single part is up to dozens of hours, and it is difficult to accurately control the position, posture, and grinding path of the robot. As a key technology in the application process of robots, offline programming has effectively improved the operation efficiency of robots and reduced the work difficulty of robot teaching personnel. However, the current offline programming technology for grinding workstations still has the following problems:

[0004] When the offline programming software plans the grinding path, it usually only considers the process requirements of a single robot or a single workpiece, while ignoring the overall production line scheduling, beat synchronization, and information interaction between other processes (such as stamping and detection) at the production line level, which is more likely to cause production line bottlenecks or resource conflicts and is more difficult to exert the overall efficiency of the "smart factory". Summary of the Invention

[0005] To solve the deficiencies in the prior art, the purpose of the present invention is to solve the above-mentioned defects, and further propose a method for optimizing the grinding path of a robot based on the linkage between the industrial Internet of Things and MES.

[0006] The present invention adopts the following technical solutions.

[0007] The first aspect of the present invention discloses a method for optimizing the grinding path of a robot based on the linkage between the industrial Internet of Things and MES, and the method includes:

[0008] Deploy a multi-sensor network through the industrial Internet of Things to obtain the real-time operation data of production line equipment, and use edge computing technology to preprocess the real-time operation data to construct a structured data stream;

[0009] Based on the structured data stream from the industrial Internet of Things, and combined with order requirements, process requirements, and historical operation data, use reinforcement learning or heuristic algorithms to calculate the optimal task allocation plan for the grinding process and upstream and downstream processes;

[0010] Call the adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the optimal task allocation plan by the grinding robot to generate an optimal path;

[0011] Perform real-time monitoring on the optimal path executed by the grinding robot, and transmit the real-time monitoring data back to the MES system and the industrial Internet of Things platform, and analyze the historical operation data through machine learning to perform adaptive iterative optimization on the optimal path;

[0012] Among them, the production line equipment includes a grinding robot, a stamping machine, and a detection device, the heuristic algorithm includes a genetic algorithm and a particle swarm priority algorithm, and the real-time monitoring data includes the position data, motion state data, and workpiece processing quality data of the grinding robot.

[0013] Furthermore, the industrial Internet of Things is used to deploy a multi-sensor network to obtain the real-time operation data of the production line equipment, and edge computing technology is used to preprocess the real-time operation data to construct a structured data stream, including:

[0014] Collect the original operation data from the production line equipment, and use the edge computing gateway to send the original operation data to the preprocessing unit to preprocess the original operation data through a filtering function to obtain digital preprocessed data;

[0015] Fuse the digital preprocessed data of different devices, add a unified time stamp according to the data collection time to form a time-series fused data vector, and construct the structured data stream based on the fused data vector;

[0016] Among them, the original operation data includes the motor temperature, grinding force, and end position of the grinding robot, and also includes the stamping pressure and die temperature of the stamping machine and the optical and dimensional sensor data of the detection device.

[0017] Furthermore, the optimal task allocation plan for the grinding process and the upstream and downstream processes is calculated by using reinforcement learning or a heuristic algorithm based on the structured data stream from the industrial Internet of Things, combined with order requirements, process requirements, and historical operation data, including:

[0018] Parse the operation status, processing progress, and beat information of each device from the structured data stream, and use the MES system to parse the current task queue and the initial process requirements of each process according to the operation status, processing progress, and beat information of each device to generate a task demand vector;

[0019] Define the production scheduling objective as minimizing the overall processing time and waiting time, while introducing equipment load balancing constraints, constructing a production scheduling optimization model, and using the task demand vector as the input of the production scheduling optimization model to output the production scheduling objective function and the task quantity constraints of each device;

[0020] Among them, the operating status includes the operating status of the grinding robot, the stamping machine, and the inspection equipment. The operating status of the grinding robot includes temperature, grinding force, and displacement. The operating status of the stamping machine includes stamping pressure and die temperature. The operating status of the inspection equipment includes dimensional deviation. The processing progress is the percentage of the processing progress of the current workpiece. The cycle time information is the time required for the grinding robot to process a single piece. The task demand vector is jointly composed of the processing progress percentage of each current process, the process standards required for each process, and the real-time operating data obtained by each device.

[0021] Furthermore, the method of using reinforcement learning or heuristic algorithm based on the structured data stream from the industrial Internet of Things, and combining order requirements, process requirements, and historical operating data to calculate the optimal task allocation scheme for the grinding process and upstream and downstream processes also includes:

[0022] Use the heuristic algorithm to construct a scheduling vector, and at the same time generate an equipment allocation vector and a processing time allocation vector to solve the production scheduling objective function and obtain a local or global optimal solution of the production scheduling objective function;

[0023] Obtain the finally optimized process scheduling vector based on the local or global optimal solution of the production scheduling objective function;

[0024] Among them, the scheduling vector is composed of the processing sequence numbers of each workpiece. The equipment allocation vector is composed of the corresponding equipment allocated to each workpiece. The processing time allocation vector is composed of the expected processing time of each workpiece. The process scheduling vector is jointly composed of the scheduling vector, the equipment allocation vector, and the processing time allocation vector.

[0025] Furthermore, the method of calling the adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the optimal task allocation scheme by the grinding robot to generate an optimal path includes:

[0026] According to the position of the workpiece in the processing sequence and the equipment allocation information in the process scheduling vector, extract the initial grinding path corresponding to each workpiece from the offline programming database to obtain the defined offline path vector;

[0027] During the grinding process, real-time feedback data is collected through multiple sensors to form a real-time feedback vector. Based on the actual grinding quality data of the workpiece and the actual working state data of the equipment, correction is performed in combination with the off-line programming path, and an error function is established.

[0028] Among them, the off-line path vector is composed of the spatial coordinates of each sampling point in the initial grinding path, and the expression of the error function is:

[0029]

[0030] In the formula, E represents the average spatial error, which is used to reflect the deviation degree between the grinding path and the actual state. M is the total number of path sampling points, and Δx i , Δy i , Δz i are the real-time deviations of the i-th sampling point in the path in the X, Y, and Z axis directions respectively.

[0031] Furthermore, the call to the adaptive path fine-tuning algorithm fine-tunes the off-line programming path of the grinding robot during the process of the grinding robot executing the best task allocation plan to generate an optimal path, and further includes:

[0032] Based on the real-time feedback vector and the average spatial error, a path compensation vector defined by the deviation vectors of each sampling point in the initial grinding path is obtained, and a dynamic adjustment formula is called to calculate the corrected path, and the interpolation smoothing algorithm is used to smooth the corrected path to obtain an adjusted optimal path vector;

[0033] The optimal path vector is compared and verified with the actual equipment state through a virtual simulation platform or a digital twin system, and an optimal path for the grinding robot to execute is generated in combination with the defined verification index;

[0034] Among them, the dynamic adjustment formula is:

[0035] P * = P0 + k × ΔP

[0036] In the formula, P * is the corrected path, P0 is the initial path before correction, k is the compensation factor, and ΔP is the path compensation vector.

[0037] Furthermore, the optimal path executed by the grinding robot is monitored in real time, and the real-time monitoring data is transmitted back to the MES system and the industrial Internet of Things platform, and the historical operation data is analyzed through machine learning to perform adaptive iterative optimization on the optimal path, including:

[0038] By performing real-time monitoring on the optimal path executed by the grinding robot, position data, motion state data, and workpiece processing quality data when the grinding robot executes the optimal path are obtained, so as to output a defined feedback data vector composed of the average path deviation, the standard deviation of the average grinding force, and the workpiece surface processing quality index;

[0039] By defining a position error vector with the optimal path executed by the grinding robot as the ideal path, calculating the path deviation between the actual execution path and the ideal path, so as to perform error evaluation on the feedback data vector.

[0040] Further, the real-time monitoring of the optimal path executed by the grinding robot, and the real-time monitoring data is transmitted back to the MES system and the industrial Internet of Things platform, and the historical operation data is analyzed by machine learning to perform adaptive iterative optimization on the optimal path, and further includes:

[0041] Adjust the compensation factor in the dynamic adjustment formula according to the magnitude of the path deviation between the actual execution path and the ideal path, and obtain the recommended re-calibration trigger condition according to the system response situation, output the optimization recommendation vector, and at the same time transmit the feedback data vector and the optimization recommendation vector back to the MES system and the industrial Internet of Things platform;

[0042] Wherein, the optimization recommendation vector is jointly composed of the change amount of the compensation factor recommended to be adjusted, the feed speed adjustment recommended to be modified, and the recommended re-calibration time or trigger condition.

[0043] A second aspect of the present invention discloses a terminal, including a processor and a storage medium; characterized in that:

[0044] The storage medium is used for storing instructions;

[0045] The processor is used for operating according to the instructions to execute the steps of the method described in the first aspect.

[0046] A third aspect of the present invention discloses a computer-readable storage medium, on which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0047] The beneficial effects of the present invention are that, compared with the prior art, the present invention has the following advantages:

[0048] (1) By deploying a multi-sensor network through the industrial Internet of Things (IIoT), collecting real-time status data of the robot grinding unit and each workstation on the production line (such as stamping, inspection, assembly), and performing data preprocessing with edge computing technology, the network load can be effectively reduced, and partial intelligent decision-making can be performed locally.

[0049] (2) Calculate the optimal task allocation plan using reinforcement learning or heuristic algorithms (such as genetic algorithms and particle swarm optimization), which can ensure the coordinated operation of the grinding process with upstream and downstream processes (stamping, assembly, and inspection), thereby reducing production bottlenecks.

[0050] (3) Combine the scheduling results of the MES system to perform secondary optimization on the offline programming path of the grinding robot, which can ensure that the path selection takes into account the production rhythm, production line layout, and equipment status. That is, adopt an adaptive path adjustment algorithm to perform fine-tuning according to real-time data during the robot's execution process, such as avoiding bottlenecks in other processes and adjusting the feed speed.

[0051] (4) After the grinding robot executes the task, transmit key process parameters (such as grinding force, time, quality inspection results, etc.) back to the MES and IIoT platforms, forming a closed-loop feedback of the production process. Analyze historical data through machine learning to further optimize the grinding strategy and achieve adaptive iterative optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of a robot grinding path optimization method based on the linkage between industrial Internet of Things and MES. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0054] As Figure 1 shown, in one embodiment, a robot grinding path optimization method based on the linkage between industrial Internet of Things and MES includes the following steps:

[0055] Step S110, deploy a multi-sensor network through the industrial Internet of Things to obtain the real-time operation data of the production line equipment, and use edge computing technology to preprocess the real-time operation data to construct a structured data stream.

[0056] Among them, the production line equipment includes a grinding robot, a stamping machine, and inspection equipment.

[0057] In some embodiments, for the robot grinding path optimization method provided by the present invention, step S110 specifically includes the following steps:

[0058] Step S111: Collect the original operation data from the production line equipment, and use the edge computing gateway to send the original operation data to the preprocessing unit to preprocess the original operation data through a filtering function to obtain digitized preprocessed data.

[0059] Step S112: Fuse the digitized preprocessed data of different devices, add a unified timestamp according to the data collection time to form a time-sequenced fused data vector, and construct a structured data stream based on the fused data vector.

[0060] Among them, the original operation data includes the motor temperature, grinding force, and end position of the grinding robot, as well as the stamping pressure and die temperature of the stamping machine and the optical and dimensional sensor data of the detection equipment.

[0061] Step S120: Use reinforcement learning or heuristic algorithms based on the structured data stream from the industrial Internet of Things, and combine order requirements, process requirements, and historical operation data to calculate the optimal task allocation plan for the grinding process and upstream and downstream processes.

[0062] Among them, the heuristic algorithms include genetic algorithms and particle swarm optimization algorithms.

[0063] In some embodiments, for the robot grinding path optimization method provided by the present invention based on the linkage between the industrial Internet of Things and MES, step S120 specifically includes the following steps:

[0064] Step S121: Parse the operation status, processing progress, and beat information of each device from the structured data stream, and use the MES system to parse the current task queue and the initial process requirements of each process according to the operation status, processing progress, and beat information of each device to generate a task demand vector.

[0065] Step S122: Define the production scheduling objective as minimizing the overall processing time and waiting time, introduce the equipment load balance constraint, construct a production scheduling optimization model, and use the task demand vector as the input of the production scheduling optimization model to output the production scheduling objective function and the task quantity constraints of each device.

[0066] Among them, the operation status includes the operation status of the grinding robot, the stamping machine, and the detection equipment. The operation status of the grinding robot includes temperature, grinding force, and displacement. The operation status of the stamping machine includes stamping pressure and die temperature. The operation status of the detection equipment includes dimensional deviation. The processing progress is the percentage of the processing progress of the current workpiece. The beat information is the time for the grinding robot to process a single piece. The task demand vector is jointly composed of the processing progress percentage of each current process, the process standards required for each process to be completed, and the real-time operation data obtained by each device.

[0067] In some embodiments, for the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention, step S120 specifically further includes the following steps:

[0068] Step S123, construct a scheduling vector using a heuristic algorithm, and at the same time generate an equipment allocation vector and a processing time allocation vector to solve the production scheduling objective function and obtain a local or global optimal solution of the production scheduling objective function.

[0069] Step S124, obtain the defined finally optimized process scheduling vector based on the local or global optimal solution of the production scheduling objective function.

[0070] Among them, the scheduling vector is composed of the processing sequence numbers of each workpiece, the equipment allocation vector is composed of the corresponding equipment allocated to each workpiece, the processing time allocation vector is composed of the expected processing time of each workpiece, and the process scheduling vector is jointly composed of the scheduling vector, the equipment allocation vector, and the processing time allocation vector.

[0071] Step S130, call the adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the best task allocation scheme by the grinding robot to generate an optimal path.

[0072] In some embodiments, for the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention, step S130 specifically includes the following steps:

[0073] Step S131, according to the position of the workpiece in the processing sequence and the equipment allocation information in the process scheduling vector, extract the initial grinding path corresponding to each workpiece from the offline programming database to obtain the defined offline path vector.

[0074] Step S132, collect real-time feedback data through multiple sensors during the grinding process to form a real-time feedback vector, and correct it in combination with the offline programming path according to the actual grinding quality data of the workpiece and the actual working state data of the equipment, and establish an error function.

[0075] Among them, the offline path vector is composed of the spatial coordinates of each sampling point in the initial grinding path, and the expression of the error function is:

[0076]

[0077] In the formula, E represents the average spatial error, which is used to reflect the deviation degree between the grinding path and the actual state, M is the total number of path sampling points, and Δx i 、Δy i 、Δz i are the real-time deviations of the i-th sampling point on the path in the X, Y, and Z axis directions respectively.

[0078] In some embodiments, for the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention, step S130 specifically further includes the following steps:

[0079] Step S133: Based on the real-time feedback vector and the average spatial error, obtain the path compensation vector defined by the deviation vectors of each sampling point in the initial grinding path, and call the dynamic adjustment formula to calculate the corrected path, and use the interpolation smoothing algorithm to smooth the corrected path to obtain the adjusted optimal path vector.

[0080] Step S134: Compare and verify the optimal path vector with the actual equipment status through a virtual simulation platform or a digital twin system, and generate the optimal path for the grinding robot to execute in combination with the defined verification index.

[0081] Wherein, the dynamic adjustment formula is:

[0082] P * = P0 + k × ΔP

[0083] In the formula, P * is the corrected path, P0 is the initial path before correction, k is the compensation factor, and ΔP is the path compensation vector.

[0084] Step S140: Real-time monitor the optimal path executed by the grinding robot, and transmit the real-time monitoring data back to the MES system and the industrial Internet of Things platform, and analyze the historical operation data through machine learning to adaptively iteratively optimize the optimal path.

[0085] Wherein, the real-time monitoring data includes the position data, motion state data, and workpiece processing quality data of the grinding robot.

[0086] In some embodiments, for the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention, step S140 specifically includes the following steps:

[0087] Step S141: Obtain the position data, motion state data, and workpiece processing quality data of the grinding robot when executing the optimal path through real-time monitoring of the optimal path executed by the grinding robot, so as to output the feedback data vector defined by the average path deviation, the standard deviation of the average grinding force, and the workpiece surface processing quality index.

[0088] Step S142: Define the position error vector with the optimal path executed by the grinding robot as the ideal path, calculate the path deviation between the actual execution path and the ideal path, so as to evaluate the error of the feedback data vector.

[0089] In some embodiments, the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention, step S140 specifically further includes the following steps:

[0090] Step S143, adjust the compensation factor in the dynamic adjustment formula according to the path deviation between the actual execution path and the ideal path, and obtain the recommended re-calibration trigger condition according to the system response situation, output the optimization recommendation vector, and at the same time transmit the feedback data vector and the optimization recommendation vector back to the MES system and the industrial Internet of Things platform.

[0091] Among them, the optimization recommendation vector is jointly composed of the change amount of the compensation factor recommended for adjustment, the feed speed adjustment recommended for modification, and the recommended re-calibration time or trigger condition.

[0092] It should be noted that MES (Manufacturing Execution System) is a manufacturing execution system, and its main functions include production scheduling, work order management, quality tracking, and real-time data collection, etc.

[0093] In a specific embodiment, the robot grinding path optimization method based on the linkage between industrial Internet of Things and MES provided by the present invention includes steps 1 to 4:

[0094] Step 1, industrial Internet of Things data collection and real-time working condition monitoring.

[0095] Deploy a multi-sensor network through the industrial Internet of Things (IIoT) to collect real-time status data of the robot grinding unit and each workstation on the production line (such as stamping, inspection, assembly), and use edge computing technology to preprocess the data to reduce network load and perform partial intelligent decision-making locally.

[0096] Specifically, it includes steps 1.1 to 1.4:

[0097] Step 1.1, equipment data collection and deployment.

[0098] Collect the original operation data from the production line equipment (grinding robot, stamping machine, inspection equipment), including:

[0099] Grinding robot: Install sensors to collect data, such as motor temperature, grinding force, and end displacement. For example, the temperature sensor collects data in real time (75°C), the force sensor measures the grinding force (12 N), and the position encoder outputs the displacement (0.2 m).

[0100] Stamping machine: Install a pressure sensor and a die temperature sensor. For example, the pressure sensor collects the stamping pressure (1200 kN), and the temperature sensor displays the die temperature (80°C).

[0101] Measuring equipment: Optical and dimensional inspection sensors are arranged. For example, the measuring equipment outputs a dimensional deviation (0.05 mm).

[0102] Step 1.2, Edge computing data preprocessing.

[0103] Filter and correct the collected raw sensor data to form digitized preprocessed data.

[0104] First, use the edge computing gateway to transfer the raw sensor data into the preprocessing module, and apply the filtering function to obtain the preprocessing vector X filt , and its expression is:

[0105] X filt = [f1(x1), f2(x2), f3(x3)]

[0106] In the formula, x1 is the raw temperature data, unit °C, x2 is the raw grinding force or stamping pressure, unit N or kN, x3 is the raw displacement or dimensional deviation, unit m or mm, f1, f2, f3 are the filtering functions of temperature, force and displacement data respectively, and the filtering algorithm includes moving average or Kalman filtering, and its parameters are determined by the actual sensor characteristics.

[0107] Step 1.3, Data fusion and time synchronization.

[0108] Fuse the preprocessed data of different devices and add timestamps to achieve multi-device data synchronization.

[0109] First, for the digitized preprocessed data X filt from each device, add a unified timestamp t according to the acquisition time to form a time-sequenced fusion data vector Y, and its expression is:

[0110] Y = [t, S, M]

[0111] In the formula, t is the timestamp, in the format of YYYY-MM-DDThh:mm:ss, for example, "2025-03-08T10:00:00". S is the structured data stream vector, which is a vector used to characterize the device status, including the key status data of each device, that is, the grinding robot (temperature, grinding force, displacement), the stamping machine (stamping pressure, die temperature), and the measuring equipment (dimensional deviation). M is the processing progress information, indicating the percentage of the current workpiece processed.

[0112] Step 1.4, Generate structured data stream.

[0113] Construct a structured data stream according to the fusion data vector Y for use in subsequent process scheduling.

[0114] First, define the structured data stream vector S, and its expression is:

[0115] S = [D, P, B

[0116] where D is the device status data, including the status of the grinding robot, the stamping machine, and the inspection equipment; P is the processing progress information, taken from M; B is the beat information, representing the single-piece processing time of the equipment. For example, each workpiece processed by the grinding robot takes 45 seconds.

[0117] Step 2, production scheduling optimization of the MES system.

[0118] The MES system receives the data stream from the IIoT, combines the order requirements, process requirements, and historical data, and optimizes the production scheduling at the production line level. Reinforcement learning or heuristic algorithms (such as genetic algorithms, particle swarm optimization) are used to calculate the optimal task allocation scheme to ensure the coordinated operation of the grinding process with the upstream and downstream processes (stamping, assembly, inspection), and reduce production bottlenecks.

[0119] Specifically, it includes Steps 2.1 to 2.4:

[0120] Step 2.1, structured data parsing and task requirement extraction.

[0121] Parse the device status, processing progress, and beat information from the structured data stream vector S generated in Step 1 to provide initial parameters for production scheduling.

[0122] First, input the structured data stream vector S = [D, P, B]. According to the data in S, the MES system will parse the current task queue and the initial process requirements of each process, and generate a task requirement vector T rep :

[0123] T rep = [p, c, s]

[0124] where p is the processing progress percentage of each current workpiece, c is the process standard required to be completed for each process, such as surface finish requirements, dimensional tolerances, and s is the real-time status information of each device.

[0125] Step 2.2, establish a production scheduling objective function.

[0126] Construct a production scheduling optimization model considering processing time, waiting time, and equipment load balance.

[0127] First, define the production scheduling objective as minimizing the overall processing and waiting time, and its expression is:

[0128]

[0129] where F is the optimization objective function of production scheduling, N is the total number of current workpieces to be processed, t p，i$t_{i}$ is the actual processing time of the $i$-th workpiece, in seconds, which is determined by the process requirements of each workpiece and the actual capacity of the equipment. Initially, a step value of 45 - 60 s can be taken. w，i $w_{i}$ is the waiting time of the $i$-th workpiece (such as equipment idle waiting or tool change time), in seconds, and is calculated based on the equipment load situation. For example, when the current equipment load exceeds 80%, the additional waiting time can be set to 5 - 10 s.

[0130] Meanwhile, introduce the equipment load balance constraint. For the number of tasks $n_{j}$ assigned to each equipment $j$ (such as grinding robots, stamping machines, inspection equipment) j should satisfy:

[0131] $\max(n_{j}) - \min(n_{j}) \leq \Delta n$ j where $\Delta n$ is the allowable difference in the number of tasks. For example, $n_{j1} - n_{j2}$ j represents the difference between different numbers of tasks $n_{j1}$

[0132] and $n_{j2}$, which is used to ensure scheduling balance. j - $n_{j2}$ i j and $n_{j2}$ i is used to ensure scheduling balance.

[0133] Step 2.3, scheduling optimization algorithm and scheduling plan generation.

[0134] Using the optimization algorithm, generate a preliminary process schedule according to the optimization objective function $F$ of production scheduling and task requirements.

[0135] First, select a heuristic algorithm (such as genetic algorithm or particle swarm algorithm) to solve the production scheduling objective function $F$, construct a scheduling vector $Q = [q_{1}, q_{2},..., q_{i},..., q_{N}]$, where $q_{i}$ i is the processing sequence number of the $i$-th workpiece. For example, $q_{1} = 3$ means the third workpiece is given priority for processing. At the same time, generate an equipment allocation vector $R = (r_{1}, r_{2},..., r_{i},..., r_{N})$, where $r_{i}$ N represents that the $i$-th workpiece is assigned to a certain equipment (such as "1" represents grinding robot A, "2" represents grinding robot B). Subsequently, generate a processing time allocation vector $L = [l_{1}, l_{2},..., l_{i},..., l_{N}]$, where $l_{i}$ i is the expected processing time of the $i$-th workpiece, and $N$ represents the total number of sampling points. i ... N ) i represents that the $i$-th workpiece is assigned to a certain equipment (such as "1" represents grinding robot A, "2" represents grinding robot B). Subsequently, generate a processing time allocation vector $L = [l_{1}, l_{2},..., l_{i},..., l_{N}]$, where $l_{i}$ i ... N i is the expected processing time of the $i$-th workpiece, and $N$ represents the total number of sampling points.

[0136] Formula objective:

[0137] ​

[0138] During the iteration process, the scheduling algorithm continuously adjusts Q, R, and L until the objective function F reaches a local or global optimum.

[0139] Step 2.4: Generate the optimized process schedule.

[0140] Integrate the aforementioned scheduling vectors to form the final optimized process schedule.

[0141] First, define the vector S of the final optimized process schedule opt as:

[0142] S opt = [Q, R, L]

[0143] For example, assume that currently 5 workpieces need to be polished. The optimized scheduling plan can be:

[0144] The processing sequence Q = [2, 5, 1, 3, 4], indicating that the second workpiece is processed first, followed by the fifth workpiece, and so on.

[0145] The equipment allocation R = (1, 2, 1, 2, 1), indicating that workpieces 2, 1, and 4 are processed by grinding robot A (number 1), and workpieces 5 and 3 are processed by grinding robot B (number 2).

[0146] The processing time L = [45, 50, 47, 52, 46] seconds, which is adjusted according to the actual processing requirements of the workpieces and the equipment capabilities respectively.

[0147] The final scheduling plan S opt = [Q, R, L] simultaneously meets the production rhythm, equipment load balance, and priority requirements.

[0148] Step 3: Robot path optimization and dynamic adjustment.

[0149] Combined with the MES scheduling results, perform secondary optimization on the offline programming path of the grinding robot to ensure that the path selection takes into account the production rhythm, production line layout, and equipment status. Adopt an adaptive path adjustment algorithm (such as D* dynamic path planning) to perform fine-tuning according to real-time data during the robot execution process, such as avoiding bottlenecks in other processes and adjusting the feed speed.

[0150] Specifically, it includes steps 3.1 to 3.4:

[0151] Step 3.1: Extract the offline programming path and construct the initial path plan.

[0152] According to the optimized process schedule S generated in step 2 opt extract the offline programming path and construct the initial path plan.

[0153] First, obtain the optimized process schedule S from Step 2 opt =[Q, R, L]. According to the position of the workpiece in the processing sequence and the equipment allocation information, extract the initial grinding path corresponding to each workpiece from the offline programming database. Among them, the defined offline path vector is: P0 = [p1, p2,..., p i ,..., p M , P0 represents the initial offline path, and p i is the spatial coordinate of the i-th sampling point in the path, denoted as p i = [x i , y i , z i , where x i , y i , z i respectively represent the positions of the sampling point in the X, Y, and Z axis directions, usually in mm, and M is the total number of path sampling points, determined according to the complexity of the workpiece and the offline programming settings.

[0154] Step 3.2, Real-time data acquisition and path deviation calculation.

[0155] Use real-time sensor data to detect the deviation of the workpiece or the grinding robot during the actual grinding process and calculate the path correction amount.

[0156] During the grinding process, collect real-time data through sensors such as vision, force / torque, and displacement to form a real-time feedback vector: R f = [δ1, δ2,..., δ i ,..., δ M , where δ i represents the deviation vector corresponding to the i-th sampling point of the initial path, denoted as δ i = [Δx i , Δy i , Δz i , and Δx i , Δy i , Δz i respectively represent the real-time deviations of this point in the X, Y, and Z axis directions, in mm. The above deviation values are obtained by comparing the sensor-collected data with the preset path coordinates, and the algorithm parameters are set according to the sensor accuracy.

[0157] At the same time, collect the actual grinding quality data of the workpiece (such as surface roughness) and the equipment working state data, correct them in combination with the offline programming path, and establish an error function E:

[0158]

[0159] In the formula, E represents the average spatial error, which is used to reflect the deviation degree between the grinding path and the actual state.

[0160] Step 3.3, Dynamic Path Adjustment Algorithm and Compensation Calculation.

[0161] Based on the real-time feedback vector R f and the error function E, the initial path P0 is secondarily optimized using the dynamic path adjustment algorithm to generate a compensation path.

[0162] First, define the path compensation vector ΔP as: ΔP = [δ1, δ2,...δ M , and use the dynamic adjustment formula to calculate the corrected path P * , and its expression is:

[0163] P * = P0 + k × ΔP

[0164] In the formula, P * is the corrected path, P0 is the initial path before correction, k is the compensation factor, and its value range is generally 0.8 to 1.2, which is dynamically adjusted according to the real-time error E. When E is large, k can be appropriately increased (such as increased to 1.1 to 1.2), and when E is small, k can be appropriately decreased (such as decreased to 0.8 to 1.0), and ΔP is the path compensation vector.

[0165] To ensure the path smoothness, the interpolation smoothing algorithm (such as B-spline or Bezier curve) is applied to smooth P * to obtain the final path P opt , and its expression is:

[0166] P opt = Smooth(P * )

[0167] In the formula, Smooth is the smoothing algorithm, which is used to ensure the path continuity and the motion compliance of the grinding robot.

[0168] Step 3.4, Optimal Path Verification and Real-time Update.

[0169] The optimal path P opt generated by the dynamic adjustment is verified in real time and ensured to be continuously updated during the execution of the grinding robot.

[0170] First, use the virtual simulation platform or digital twin system to compare and verify P opt with the real-time device status to ensure that the path will not cause the robot to collide or over-cut. Secondly, define the verification index V, and its expression is:

[0171]

[0172] In the formula, is the index function. If If the preset safety distance is satisfied, then otherwise it is 0. The range of V is between 0 and 1. If V ≥ 0.95, the path is considered safe and reliable. If the verification index V does not meet the requirements, the feedback adjustment compensation factor k is adjusted or ΔP is recalculated until V meets the standard. At the same time, the system continuously and real-time collects new sensor data, and repeats steps 3.2 and 3.3 according to the new deviation information to achieve closed-loop feedback update.

[0173] In this embodiment, the optimal path expression executed by the grinding robot is:

[0174] P exec = P opt

[0175] where P exec = [P exec，1 , P exec，2 ,..., P exec，i ,..., P exec，M , and P exec，i represents the coordinates of the sampling points after real-time verification, which is used to ensure that the grinding robot moves along this path during actual operation.

[0176] Step 4, closed-loop feedback and continuous optimization.

[0177] After the grinding robot executes the task, it transmits the key process parameters (such as grinding force, time, quality inspection results, etc.) back to the MES and IIoT platforms to form a closed-loop feedback of the production process. By analyzing historical data through machine learning, the grinding strategy is further optimized to achieve adaptive iterative optimization.

[0178] Specifically, it includes steps 4.1 to 4.4:

[0179] Step 4.1, real-time data monitoring and feedback collection.

[0180] Real-time monitor the optimal path executed by the robot, and collect its motion state and processing quality data.

[0181] First, based on the optimal path P exec executed by the robot in step 3, real-time collect data, including position data (such as the deviation between the actual path point and the preset path point), motion state data (such as acceleration, force value on the end effector), and processing quality data (such as surface roughness of the workpiece). Define the feedback data vector: F = [f1, f2, f3], where f1 is the average path deviation, f2 is the average grinding force fluctuation (standard deviation), and f3 is the workpiece surface processing quality index.

[0182] Step 4.2, error evaluation and feedback data processing.

[0183] Evaluate the error of the collected feedback data, calculate the deviation between the actual execution path and the ideal path, and quantify other key metrics.

[0184] First, set the ideal path as the optimized path adjusted in step 3, and define the position error vector E p : E p = [e1, e2,..., e i ,..., e M , where

[0185]

[0186] Among them, e i represents the spatial error of the i-th sampling point, with the unit of mm, x opt，i , y opt，i , z opt，i are the preset target coordinate values respectively, and x exec，i , y exec，i , z exec，i are the coordinate values executed by the grinding robot respectively.

[0187] After that, calculate the average path deviation:

[0188]

[0189] At the same time, collect the fluctuations of the grinding force data, calculate its standard deviation f2 and average the surface roughness data f3.

[0190] Step 4.3, Optimization Suggestion Generation Algorithm.

[0191] Generate targeted optimization suggestions based on the feedback data F to adjust the offline path compensation factor and feed parameters of the grinding robot.

[0192] First, define the optimization suggestion vector: G = [g1, g2, g3], where g1 represents the change in the compensation factor to be adjusted, g2 represents the adjustment of the feed speed to be modified, and g3 represents the recommended recalibration time or trigger condition.

[0193] Adjust the compensation factor k according to the magnitude of the feedback data f1. For example, let the current compensation factor be k old (which has the same purpose as the adjustment factor k, the difference is that k old refers to the adjustment compensation factor at the current moment, used to compensate and correct the current offline path according to the current adjustment compensation factor k old ), and the formula can be used:

[0194]

[0195] Among them, Δk is the change amount of the recommended compensation factor, and its usual range is 0.05 - 0.1.

[0196] For the feed speed, if the deviation of the grinding quality f3 (such as higher than the target by 1.0 μm) or the excessive force fluctuation f2 (such as exceeding 3.0 N) is found in the feedback data, it is recommended to reduce the feed speed v, and its formula is:

[0197]

[0198] Among them, the recommended value of Δv is 0.01 - 0.05 m / s.

[0199] According to the system response situation, it is recommended to set the re - calibration trigger condition g3, such as:

[0200]

[0201] Step 4.4, Output and callback of feedback data and optimization suggestions.

[0202] Transmit the generated feedback data and optimization suggestions to the control system to achieve closed - loop feedback and guide the subsequent dynamic adjustment of the path.

[0203] The specific application scenarios of the present invention are as follows:

[0204] Case 1: Path optimization for high - rhythm production of small - sized Class A workpieces

[0205] 1. Specific parameters of the MES scheduling result:

[0206] Production beat time: 45 seconds per workpiece (each workpiece is required to complete grinding within 45 seconds).

[0207] Workpiece priority: Class A workpieces take precedence over Class B workpieces. The current order is mainly composed of Class A workpieces and needs to be completed first.

[0208] Equipment load situation: The grinding robot A on the production line performs the grinding task, and the current load rate is about 85%, approaching its full - load state.

[0209] Other parameters: Class A workpieces are small - sized aluminum alloy parts with medium surface finish requirements; the current order of Class B workpieces is small and has little impact on production capacity.

[0210] 2. Original offline programming path:

[0211] Grinding sequence: The robot A grinds each surface and edge of the workpiece in a fixed order. For example, it grinds the outer surface first and then processes all edges.

[0212] Feed speed: The initial programming sets the feed speed to 0.2 m / s to ensure surface quality, but does not fully consider the 45 - second beat requirement.

[0213] Tool change timing: The robot uses two grinding tools (coarse grinding wheel and polishing wheel). The original path is to change the tool once after all surfaces are coarsely ground and then perform polishing, that is, fine polishing starts only after the coarse grinding of each workpiece is completed.

[0214] Path characteristics: This initial path is not optimized for a 45-second cycle. The actual single-piece grinding time is about 50 seconds, with redundant idle travel and waiting time.

[0215] 3. Secondary optimized path based on MES scheduling:

[0216] Optimized path sequence: Combining the cycle requirements provided by MES, re-plan the grinding sequence to reduce the time for the robot to move back and forth between different surfaces. For example, adjust it to continuously grind adjacent surfaces first and then move to the next area, avoiding frequent jumping movements, thereby reducing the idling time.

[0217] Shorten the tool change time: In order to complete grinding within 45 seconds, the tool change strategy is optimized. Before the end of coarse grinding, the robot moves closer to the tool change device in advance and completes the tool change at the fastest speed, shortening the time required for tool change. When necessary, sectional machining can also be adopted: divide the workpiece into sections, and immediately perform polishing on the section after the coarse grinding of the section is completed. Cycle in this way for each section to avoid long and concentrated tool change pauses.

[0218] Adjust the feed speed: On the premise of ensuring quality, increase the feed speed to 0.25 m / s to speed up the grinding process. For areas with lower surface requirements, further increase the speed to save time, while maintain an appropriate speed for areas with high requirements to balance quality and efficiency.

[0219] Optimization effect: Through the above optimizations, the single-piece grinding time of robot A is compressed to about 45 seconds, synchronized with the production cycle. Class A workpieces are still processed in priority order, and Class B workpieces are inserted when there is idle time, without affecting the timely completion of Class A workpieces.

[0220] Case 2: Multi-process grinding optimization of Class A precision parts and Class B structural parts

[0221] 1. Specific parameters of MES scheduling results:

[0222] Production cycle time: 90 seconds per workpiece (the planned time for each workpiece to complete grinding is about 90 seconds, including two processes of coarse grinding and fine grinding).

[0223] Workpiece priority: Class A workpieces (such as appearance parts, requiring high surface finish) take precedence over Class B workpieces (such as internal structural parts, with lower requirements for surface finish). MES issues instructions to first meet the delivery deadlines of Class A workpieces.

[0224] Equipment load situation: Grinding robot A is responsible for processing all workpieces at this station. The current load rate is about 75%, and it needs to take into account the tasks of both Class A and Class B workpieces.

[0225] Other parameters: Class A workpieces require both rough grinding and fine grinding to meet surface requirements; Class B workpieces only require rough grinding to remove burrs (fine grinding is not necessary), which affects the grinding path and tool change requirements.

[0226] 2. Original offline programming path:

[0227] Grinding sequence: Original path In order to improve equipment utilization, a process merging strategy is adopted. For example, the rough grinding of multiple workpieces that arrive in succession is completed first, and then the tool is changed for fine grinding. Specifically, robot A first completes the rough grinding of a batch of workpieces (including A and B) one by one, and then replaces the fine grinding tool and completes the fine grinding one by one.

[0228] Feed speed: The feed speed is set to 0.3m / s in the rough grinding stage and 0.15m / s in the fine grinding stage to ensure high surface quality for Class A workpieces. This speed setting is consistent for all workpieces and is not differentiated according to workpiece priority.

[0229] Tool change timing: The original programming is to change the tool once for fine grinding after all rough grinding tasks are completed, so that only one tool change occurs for each batch of workpieces. However, this means that after the rough grinding of a certain type A workpiece is completed, it is necessary to wait until all the workpieces in the batch are finished with rough grinding and the tool is changed before fine grinding can be carried out, which prolongs the overall processing time of the type A workpieces.

[0230] Problem: Although batch processing reduces the frequency of tool changes, the fine grinding of Class A workpieces is delayed and cannot be completed as soon as possible, which violates its priority delivery requirements.

[0231] 3. Secondary optimization path based on MES scheduling:

[0232] Optimize the path sequence: According to the priority requirements provided by MES, adjust the grinding sequence to complete all the processes of Class A workpieces first. Specifically, for Class A workpieces, instead of waiting for all workpieces to be rough-ground, change the tool immediately after the rough grinding of the workpiece to perform fine grinding, to ensure that Class A workpieces are completed as soon as possible. The fine grinding of Class B workpieces can be postponed and processed uniformly, or only rough grinding can be performed and temporarily stored, and then fine grinding can be performed when idle.

[0233] Shorten tool change time: In view of the increase in tool changes for single-piece Class A workpieces, the tool change time is reduced by optimizing the program. For example, at the end of the rough grinding of Class A workpieces, the robot slows down in advance and approaches the tool change position, switching to the fine grinding tool at the fastest speed. For multiple consecutive Class A workpieces, try to complete the rough grinding in sequence and then fine grind them one by one immediately, quickly change the tool between each piece, and use the overlapping time of the robot movement to shorten the pause.

[0234] Adjust the feed rate: Adopt different feed rates for workpieces with different priorities. For type A workpieces, appropriately increase the speed during rough grinding and finish grinding (e.g., increase to 0.35 m / s during rough grinding) to speed up processing and shorten the total time while ensuring quality; while for type B workpieces, due to their lower priority, slightly reduce the rough grinding speed to 0.25 m / s, thereby extending their processing time and giving up time to prioritize the completion of type A workpieces.

[0235] Optimization effect: Through the above adjustments, type A workpieces can be completed independently as soon as possible without waiting for the batch process. The total processing time per piece is shortened from the original approximately 100 seconds to nearly 90 seconds, meeting the priority delivery requirements. At the same time, although the finish grinding time of type B workpieces is slightly postponed or processed at a slower speed, it is still within the allowable range of the beat and does not affect the overall compliance of the production rhythm.

[0236] Case 3: Load balancing optimization for collaborative grinding of two robots

[0237] 1. Specific parameters of the MES scheduling result:

[0238] Production beat time: 60 seconds per workpiece (the production line beat is 60 seconds, and one workpiece needs to be ground per minute).

[0239] Workpiece priority: Type A workpieces take precedence over type B workpieces. In the current order, the number of type A workpieces accounts for about 70%, and type B accounts for about 30%. Type A orders are urgent and need to be processed quickly.

[0240] Equipment load situation: The current load rate of grinding robot A is 85%, while the load rate of robot B is 60%. Robot A has a heavy task and has become a bottleneck, while robot B still has spare production capacity.

[0241] Other parameters: The two robots belong to parallel workstations and originally processed fixed types of workpieces respectively (robot A processes type A workpieces, and robot B processes type B workpieces). The process flow and grinding tools are slightly different but generally similar.

[0242] 2. Original offline programming path:

[0243] Grinding sequence: Robots A and B grind independently in a fixed order according to their respective workpiece types. Robot A continuously processes the queue of type A workpieces, and robot B processes the queue of type B workpieces. As a result, robot A often operates continuously without idle time, while robot B will wait for the next batch after completing the current batch of type B workpieces, resulting in idle time.

[0244] Feed rate: Both robots operate according to standard parameters, and A and B use the same default feed rate (e.g., 0.3 m / s) without adjustment according to their respective load conditions. Robot A is approaching its beat limit due to long-term high-load operation.

[0245] Tool change timing: Each robot changes tools according to its respective process requirements. Assume that type A workpieces need to change tools once (rough grinding → finish grinding), and type B workpieces use a single grinding tool and do not require frequent tool changes. The original programming did not consider the task connection and tool change coordination between the two robots.

[0246] Existing problems: Under this setting, robot A is busy processing a large number of type A workpieces, and robot B is often idle. Although the overall production line barely meets the 60-second beat, there is a backlog in the queue of type A workpieces, posing a risk of delay, and the equipment utilization rate is uneven.

[0247] 3. Secondary optimization path based on MES scheduling:

[0248] Optimize task allocation: According to the MES scheduling results, reallocate some type A workpieces to be processed by robot B to achieve load balance. Adjust the offline programming so that robot B can execute the grinding path of type A workpieces. For example, modify the program of robot B to add the grinding sequence and actions of type A workpieces, so that it can take over the processing of type A workpieces when there is no type B task. In this way, the two robots can process type A workpieces in parallel (grinding different workpieces respectively) and complete more type A workpieces per unit time.

[0249] Shorten tool change and waiting time: Since robot B starts to undertake type A workpieces, it also needs to change tools for rough grinding and finish grinding. Optimize the program to arrange for robot B to prepare for tool switching in advance during idle waiting to ensure that there is no additional tool change time when taking over type A workpieces. For robot A, if the secondary surfaces or steps of some type A workpieces can be done by robot B, then robot A can skip the corresponding operations and tool changes and directly transfer to the processing of the next workpiece, reducing waiting time.

[0250] Adjust the feed speed: After load balancing, the pressure on robot A is reduced, and it can grind at a stable speed to ensure quality; when robot B undertakes more tasks, the feed speed can be slightly increased within the allowable range (for example, from 0.3 m / s to 0.35 m / s) to shorten the processing time of each type A workpiece for it and help the overall production meet the beat requirements.

[0251] Optimization effect: After secondary optimization, the two robots work together: the load of robot A drops to about 70%, and that of robot B increases to about 75%, and the production capacity utilization is more balanced. The grinding path of each workpiece is rearranged according to the scheduling requirements to ensure that the processing of type A workpieces is given priority within the 60-second beat. Type B workpieces are completed by robot B during the gaps without affecting the priority production of type A workpieces. The overall production efficiency is improved, and the on-time delivery rate of key workpieces is increased.

[0252] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The meaning of "a plurality" is two or more, unless specifically and explicitly defined otherwise.

[0253] In the present invention, unless otherwise clearly defined and limited, terms such as "installed", "connected", "coupled", "fixed", etc. shall be construed in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0254] In the present invention, unless otherwise clearly defined and limited, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "beneath" and "underneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

[0255] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0256] Any process or method description represented in a flowchart or otherwise described herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where functions may be performed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0257] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered a sequenced list of executable instructions for implementing a logical function, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0258] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0259] Those of ordinary skill in the art can understand that all or part of the steps carried out in implementing the above-described embodiment methods can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0260] In addition, in each of the embodiments of the present invention, the functional units can be integrated in a processing module, or each unit can exist physically alone, or two or more units can be integrated in a module. The above-integrated module can be implemented in the form of hardware or in the form of a software functional module. When the above-integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0261] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A robot grinding path optimization method based on the linkage of industrial Internet of Things and MES, characterized in that, The method includes: Deploying a multi-sensor network through the Industrial Internet of Things (IIoT) to obtain real-time operation data of production line equipment, and using edge computing technology to preprocess the real-time operation data to construct a structured data stream; Using reinforcement learning or heuristic algorithms based on the structured data stream from the IIoT, and combining order requirements, process requirements, and historical operation data, to calculate the optimal task allocation plan for the grinding process and upstream and downstream processes; Invoking an adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the optimal task allocation plan to generate an optimal path; Performing real-time monitoring on the optimal path executed by the grinding robot, and transmitting the real-time monitoring data back to the MES system and the IIoT platform, and analyzing the historical operation data through machine learning to perform adaptive iterative optimization on the optimal path; Wherein, the production line equipment includes a grinding robot, a stamping machine, and inspection equipment, the heuristic algorithms include a genetic algorithm and a particle swarm optimization algorithm, and the real-time monitoring data includes the position data, motion state data, and workpiece processing quality data of the grinding robot.

2. The robot grinding path optimization method based on the linkage between industrial Internet of Things and MES according to claim 1, wherein The deploying a multi-sensor network through the IIoT to obtain real-time operation data of production line equipment, and using edge computing technology to preprocess the real-time operation data to construct a structured data stream includes: Collecting raw operation data from the production line equipment, and using an edge computing gateway to send the raw operation data to a preprocessing unit to preprocess the raw operation data through a filtering function to obtain digitized preprocessed data; Fusing the digitized preprocessed data of different devices, adding a unified time stamp according to the data collection time to form a time-series fused data vector, and constructing the structured data stream based on the fused data vector; Wherein, the raw operation data includes the motor temperature, grinding force, and end position of the grinding robot, and also includes the stamping pressure and die temperature of the stamping machine and the optical and dimensional sensor data of the inspection equipment.

3. The robot grinding path optimization method based on the linkage between industrial Internet of Things and MES according to claim 2, characterized in that, The using reinforcement learning or heuristic algorithms based on the structured data stream from the IIoT, and combining order requirements, process requirements, and historical operation data, to calculate the optimal task allocation plan for the grinding process and upstream and downstream processes includes: Parsing the operation status, processing progress, and beat information of each device from the structured data stream, and parsing the current task queue and the initial process requirements of each process through the MES system according to the operation status, processing progress, and beat information of each device to generate a task demand vector; Defining the production scheduling objective as minimizing the overall processing time and waiting time, introducing equipment load balance constraints at the same time, constructing a production scheduling optimization model, and using the task demand vector as the input of the production scheduling optimization model to output the production scheduling objective function and the task quantity constraints of each device; Among them, the operating status includes the operating status of the grinding robot, the stamping machine, and the inspection equipment. The operating status of the grinding robot includes temperature, grinding force, and displacement. The operating status of the stamping machine includes stamping pressure and die temperature. The operating status of the inspection equipment includes dimensional deviation. The processing progress is the percentage of the processing progress of the current workpiece. The cycle time information is the time required for the grinding robot to process a single piece. The task requirement vector is jointly composed of the processing progress percentages of the current processes, the process standards required to be completed for each process, and the real-time operating data obtained by each device.

4. The robot grinding path optimization method based on the linkage of industrial Internet of Things and MES according to claim 3, characterized in that, The step of using reinforcement learning or heuristic algorithm based on the structured data stream from the industrial Internet of Things, and combining order requirements, process requirements, and historical operating data to calculate the optimal task allocation plan for the grinding process and upstream and downstream processes further includes: Using the heuristic algorithm to construct a scheduling vector, and simultaneously generating an equipment allocation vector and a processing time allocation vector to solve the production scheduling objective function and obtain a local or global optimal solution of the production scheduling objective function; Obtaining the finally optimized process scheduling vector defined based on the local or global optimal solution of the production scheduling objective function; Among them, the scheduling vector is composed of the processing sequence numbers of each workpiece, the equipment allocation vector is composed of the corresponding equipment allocated to each workpiece, the processing time allocation vector is composed of the expected processing time of each workpiece, and the process scheduling vector is jointly composed of the scheduling vector, the equipment allocation vector, and the processing time allocation vector.

5. The robot grinding path optimization method based on the linkage between industrial Internet of Things and MES according to claim 4, characterized in that, The step of calling the adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the optimal task allocation plan by the grinding robot to generate an optimal path includes: According to the position of the workpiece in the processing sequence and the equipment allocation information in the process scheduling vector, extracting the initial grinding path corresponding to each workpiece from the offline programming database to obtain the defined offline path vector; During the grinding process, collecting real-time feedback data through multiple sensors to form a real-time feedback vector, and correcting it in combination with the offline programming path according to the actual grinding quality data of the workpiece and the actual working state data of the equipment, and establishing an error function; Among them, the offline path vector is composed of the spatial coordinates of each sampling point in the initial grinding path, and the expression of the error function is: Where, E represents the average spatial error, which is used to reflect the deviation degree between the grinding path and the actual state, M is the total number of path sampling points, and Δx i , Δy i , Δz i are respectively the real-time deviations of the i-th sampling point in the path in the X, Y, and Z axis directions.

6. The robot grinding path optimization method based on the linkage of industrial Internet of Things and MES according to claim 5, wherein The step of calling the adaptive path fine-tuning algorithm to fine-tune the offline programming path of the grinding robot during the execution of the optimal task allocation plan by the grinding robot to generate an optimal path further includes: Based on the real-time feedback vector and the average spatial error, obtaining the defined path compensation vector composed of the deviation vectors of each sampling point in the initial grinding path, and calling the dynamic adjustment formula to calculate the corrected path, and using the interpolation smoothing algorithm to smooth the corrected path to obtain the adjusted optimal path vector; Comparing and verifying the optimal path vector with the actual equipment status through a virtual simulation platform or a digital twin system, and generating the optimal path executed by the grinding robot in combination with the defined verification index; Among them, the dynamic adjustment formula is: P * = P0 + k × ΔP where P * is the corrected path, P0 is the initial path before correction, k is the compensation factor, and ΔP is the path compensation vector.

7. The robot grinding path optimization method based on the linkage between industrial Internet of Things and MES according to claim 6, characterized in that, Performing real-time monitoring on the optimal path executed by the grinding robot, and transmitting the real-time monitoring data back to the MES system and the industrial Internet of Things platform, and analyzing the historical operation data through machine learning to perform adaptive iterative optimization on the optimal path, including: Obtaining the position data, motion state data, and workpiece processing quality data of the grinding robot when executing the optimal path through real-time monitoring of the optimal path executed by the grinding robot, so as to output a feedback data vector defined by the average path deviation, the standard deviation of the average grinding force, and the workpiece surface processing quality index; Defining a position error vector with the optimal path executed by the grinding robot as the ideal path, calculating the path deviation between the actual execution path and the ideal path, so as to perform error evaluation on the feedback data vector.

8. The robot grinding path optimization method based on the linkage between industrial Internet of Things and MES according to claim 7, characterized in that Performing real-time monitoring on the optimal path executed by the grinding robot, and transmitting the real-time monitoring data back to the MES system and the industrial Internet of Things platform, and analyzing the historical operation data through machine learning to perform adaptive iterative optimization on the optimal path, further including: Adjusting the compensation factor in the dynamic adjustment formula according to the path deviation between the actual execution path and the ideal path, and obtaining the recommended re-calibration trigger condition according to the system response situation, outputting an optimization recommendation vector, and at the same time transmitting the feedback data vector and the optimization recommendation vector back to the MES system and the industrial Internet of Things platform; Among them, the optimization recommendation vector is jointly composed of the change amount of the compensation factor to be adjusted, the feed speed adjustment to be modified, and the recommended re-calibration time or trigger condition.

9. A terminal, including a processor and a storage medium; characterized in that: The storage medium is used for storing instructions; The processor is used to operate according to the instructions to execute the steps of the method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it realizes the steps of the method according to any one of claims 1-8.

Images