An intelligent motion control method for stacking platform based on three-dimensional imaging of steel plate blanks

Abstract

An intelligent motion control method for a stacking table based on three-dimensional imaging of steel plate blanks. Three-dimensional data is obtained by scanning steel plates on-site in a steel plant from different angles using multiple laser radars. Based on an AI algorithm, the algorithm compensates for the relatively missing data of each radar using the data scanned by other radars to obtain three-dimensional imaging data of the slabs, stacking tables, and rollers with higher precision. The beneficial effects of the present invention are as follows: the automated control system can greatly reduce the errors caused by manual observation and operation through the naked eye, greatly reduce the number of sensors and inductive switches used, avoid problems caused by poor contact, damage, malfunction of detection components, etc., reduce the probability of equipment downtime for maintenance, and can control the operation of multiple modules at the same time. The operating efficiency is much higher than that of manual operation. The operator only needs to supervise and do some emergency response to emergencies, thereby reducing the labor intensity of the operator and improving production efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of steel plate production, and in particular to an intelligent motion control method for a stacking platform based on three-dimensional imaging of steel plate blanks. Background Art

[0002] After smelting, slabs are conveyed to the plate mill via roller conveyors. Currently, the unstacking, rotation, and conveying processes are performed manually. Operators use their eyes and the images displayed on the monitor via on-site cameras to estimate the relative position and height of the slabs. Based on their experience, they manually raise and lower the hydraulic system of the stacking platform and manually control the pusher to move the steel plates forward and backward, achieving slab position change and conveyance.

[0003] Manual operation will have the following problems:

[0004] 1. Improper operation caused by visual errors may cause collision between the slab and the conveyor platform, causing damage to the conveying system. In serious cases, production needs to be stopped for emergency repairs.

[0005] 2. Due to visual errors, the height of the pallet platform lifting platform is not properly adjusted, causing the steel pushing mechanism to frequently collide with the outer frame of the pallet platform lifting platform when retreating, damaging the conveying system. In serious cases, production needs to be stopped for repairs.

[0006] 3. Improper operation during stacking or destacking may cause the slabs to be unable to be stacked off the line or destacking on the line, interrupting the normal production progress and production process, and thus affecting the smoothness and efficiency of production. Summary of the Invention

[0007] The purpose of the present invention is to address the deficiencies and defects in the existing technology and provide an intelligent motion control method for a stacking platform based on three-dimensional imaging of steel plate blanks. The automated control system can greatly reduce the errors caused by manual observation and operation through the naked eye, greatly reduce the number of sensors and induction switches used, avoid problems caused by poor contact, damage, and failure of detection components, reduce the probability of equipment shutdown and maintenance, and can control the operation of multiple sets of modules at the same time. The operating efficiency is much higher than manual operation. The operator only needs to supervise and do some emergency response to emergencies, thereby reducing the labor intensity of the operator and improving production efficiency.

[0008] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an intelligent motion control method for stacking pallets based on three-dimensional imaging of steel plate blanks, which includes the following specific steps: scanning the steel plates in the steel plant site from different angles by multiple laser radars to obtain three-dimensional data, based on the AI ​​algorithm, according to the relatively missing data of each radar, the algorithm is compensated by the data scanned by other radars to obtain high-precision three-dimensional imaging data of the slabs, stacking pallets and rollers; scene images are generated according to the laser point cloud data, field scene modeling is performed according to the scene images, and the recognition area is reasonably planned; the three-dimensional data formed by the radar scanning will be transmitted to the data center through the network The data is stored in the database, the gaps and misalignment points between the slabs are calibrated according to the stored data, and the model is trained through a deep learning model; the intelligent control system optimizes the PLC parameter adjustment according to the on-site production situation and outputs the optimal decision; the deep learning model can distinguish the edges of each stacked slab, and based on the data scanned by the radar, based on the intelligent control algorithm and PLC optimal decision, guides the stacking platform to automatically rise or fall to the appropriate height according to the size of the slab; after the height is confirmed, the intelligent control system continues to issue instructions, and the pusher automatically pushes the slab onto the roller and moves back and forth to complete the online delivery of the slabs as planned.

[0009] Furthermore, the intelligent control system optimizes PLC parameter adjustment according to on-site conditions and outputs an optimized decision. The optimized decision is specifically as follows: based on the properties of the intelligent control system, analyze and obtain several evaluation factors of the intelligent control system; obtain the control logic system of the intelligent control system, establish a mapping matrix between the PLC parameters of the automated control and several evaluation factors of the intelligent control system; construct an evaluation index model of the intelligent control system based on the evaluation factors, the evaluation index model takes the weight matrix of the evaluation factors and the evaluation factor decision matrix as input, and takes the optimized PLC parameter group of the intelligent control system as output; take several decision values ​​for each PLC parameter according to the set value interval, and combine them into several PLC parameter decision The method comprises the following steps: a) forming a PLC parameter decision group, forming a PLC parameter decision matrix from all PLC parameter decision groups, and transforming the PLC parameter decision matrix into an evaluation factor decision matrix based on a mapping matrix between PLC parameters and several evaluation factors of the intelligent control system; b) adding weight values ​​to several evaluation factors according to the properties of the intelligent control system to obtain a weight matrix of the evaluation factors; c) inputting the weight matrix of the evaluation factors and the evaluation factor decision matrix into the evaluation index model of the intelligent control system to obtain the optimized PLC parameter group; c) performing anti-interference processing on the optimized PLC parameter group to obtain the final PLC parameter group; and d) feeding the final PLC parameter group back to the intelligent control system, which modifies the control parameters according to the final PLC parameter group and controls the intelligent control system to execute the optimized control decision.

[0010] Furthermore, the evaluation index model of the intelligent control system constructed based on the evaluation factors specifically includes: training an evaluation factor scoring model for each evaluation factor respectively, the evaluation factor scoring model takes the numerical value of the evaluation factor as input, and the positive score of the evaluation factor as output; constructing an evaluation factor weight normalization transformation model, the evaluation factor weight normalization transformation model takes the weight matrix of the evaluation factor and the evaluation factor decision matrix as input, and the weight normalization matrix as output; constructing a decision index calculation model, the decision index calculation model takes the weight normalization matrix as input, and the decision index as output.

[0011] Furthermore, the training evaluation factor scoring model specifically includes: adding corresponding positive scores to different evaluation factor values ​​according to historical experience data, and encapsulating the evaluation factor values ​​and positive scores into several groups of training data sets; randomly dividing multiple groups of training data sets into calculation training samples and test samples, the number of training data sets in the training samples accounts for 80%, and the number of training data sets in the test samples accounts for 20%; using the training data sets in the training samples to calculate the training model of the evaluation factor scoring model to obtain multiple preliminary training models; substituting the training data sets in the test samples into the preliminary training model, and screening out the preliminary training model with the highest test fit as the evaluation factor scoring model.

[0012] Furthermore, the step of substituting the training data set in the test sample into the preliminary training model and screening out the preliminary training model with the highest test fit specifically includes: inputting the input variables of all training data sets in the test sample into the preliminary training model to obtain predicted output variables; calculating the regression determination coefficient based on the output variables of all training data sets in the test sample and the predicted output variables; calculating the regression determination coefficient for each preliminary training model; and screening out the preliminary training model with the largest regression determination coefficient.

[0013] Furthermore, the calculation formula of the regression determination coefficient is: Where R 2 is the regression determination coefficient; RSS is the residual sum of squares of the preliminary training model; TSS is the total sum of squares of the preliminary training model.

[0014] Furthermore, the step of inputting the weight matrix of the evaluation factors and the evaluation factor decision matrix into the evaluation index model of the intelligent control system to obtain the optimized PLC parameter group specifically includes: inputting the decision value of each evaluation factor in the evaluation factor decision matrix into the corresponding evaluation factor scoring model, obtaining a number of decision positive scores for each evaluation factor, and combining the several decision positive scores of all evaluation factors into a number of decision positive score groups, and forming the decision positive score matrix A of the decision evaluation factor with the several decision positive score groups; Among them, m is the total number of decision-making positive scoring groups, n is the total number of several evaluation factors of the intelligent control system, x ij The positive score of the jth evaluation factor in the i-th decision positive score group; the weight matrix of the evaluation factors and the positive score matrix D of the decision evaluation factors are input into the evaluation factor weight normalization conversion model, the weight normalization value of each evaluation factor is calculated based on the normalization calculation formula, and the weight normalization values ​​of all evaluation factors are combined into the weight normalization matrix B; Among them, w ij is the weight normalized value of the jth evaluation factor in the i-th decision positive scoring group; based on the weight normalization matrix B, the optimal weight normalization value group S is determined + and the worst weight normalized value set S - ,in,

[0015]

[0016] in, is the maximum value of the weight normalized value of the jth evaluation factor, is the minimum value of the weight normalized value of the jth evaluation factor; the decision index of each decision positive scoring group is calculated according to the index calculation formula; the decision positive scoring group with the largest decision index is screened out, and its corresponding PLC parameter decision group is the optimized PLC parameter group.

[0017] Furthermore, the normalized calculation formula is: Where, α j is the weight value of the jth evaluation factor, x kj is the positive score of the jth evaluation factor in the kth decision positive score group.

[0018] Furthermore, the indicator calculation formula is: Where C i is the decision indicator of the i-th group of decision-making positive scoring group.

[0019] Furthermore, the anti-interference processing of the optimized PLC parameter group specifically includes: performing anti-interference calculation on each PLC parameter in the optimized PLC parameter group according to the anti-interference formula to obtain the PLC parameter value of each PLC parameter after anti-interference processing; combining the PLC parameter values ​​after anti-interference processing of each PLC parameter to generate a final PLC parameter group; the anti-interference formula is: γ * =D -1 γ+σ, where γ * is the PLC parameter value after anti-interference processing, D -1is the inverse matrix of the stability matrix of the intelligent control system, γ is the PLC parameter value of the optimized PLC parameter group, and σ is the noise of the intelligent control system.

[0020] After adopting the above technical solution, the beneficial effects of the present invention are as follows: the intelligent motion control method of the stacking platform based on three-dimensional imaging of steel plate blanks has at least the following beneficial effects:

[0021] 1. The automated control system assigns tasks to each slab based on the steel plant's production scheduling system. The automated control system plans the movement and motion route of each slab through an intelligent path planning algorithm based on the actual destination and purpose of each slab. The intelligent control program of the intelligent control automation system then controls the corresponding equipment or roller conveyor, so that the slab can automatically reach the corresponding position specified by the scheduling system. This can greatly reduce the errors caused by manual observation and operation, greatly reduce the number of sensors and induction switches used, avoid problems caused by poor contact, damage, and malfunction of detection components, reduce the probability of equipment shutdown and maintenance, and can control the operation of multiple modules at the same time. The operating efficiency is much higher than that of manual operation. The operator only needs to supervise and handle some emergencies, which greatly reduces the labor intensity of the operator and improves production efficiency.

[0022] 2. Determine several evaluation factors of the automation control system based on the properties of the automation control system. Then, combine the control logic system of the automation control system to obtain the internal control matrix between the PLC parameters of the automation control and the evaluation factors. Then, optimize the PLC parameters of several PLC parameter decision groups. This method can achieve automated optimal decision-making for the PLC parameter group, reduce dependence on manual experience, and effectively ensure that the PLC control parameters can achieve the optimal operation of the automation control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 Schematic diagram of the method flow of the present invention;

[0025] Figure 2 Schematic diagram of the flow of the method for constructing an evaluation index model of an intelligent control system based on evaluation factors in the present invention;

[0026] Figure 3Schematic diagram of the method flow for training the evaluation factor scoring model in the present invention;

[0027] Figure 4 Schematic diagram of the process of selecting the preliminary training model with the highest test fit in the present invention. DETAILED DESCRIPTION

[0028] See Figure 1-Figure 4 As shown, the technical solution adopted in this specific embodiment is: it includes the following specific steps:

[0029] S1 uses multiple LiDARs to scan steel plates at the steel mill site from different angles to obtain 3D data. Based on an AI algorithm, the algorithm compensates for the missing data of each LiDAR with the data scanned by other LiDARs to obtain high-precision 3D imaging data of slabs, stacking platforms, and roller tables.

[0030] S2, generates scene images based on laser point cloud data, performs field scene modeling based on the scene images, and reasonably plans the recognition area;

[0031] In step S3, the 3D data generated by radar scanning is transmitted to the data center via the network for storage. The gaps and offset points between the slabs are calibrated based on the stored data, and the model is trained using a deep learning model.

[0032] S4, the intelligent control system optimizes PLC parameter adjustments based on on-site production conditions and outputs the optimal decision;

[0033] S5, a deep learning model can distinguish the edges of each stacked slab. Based on the radar scan data, intelligent control algorithms and PLC optimization decisions, it guides the stacking platform to automatically rise or fall to the appropriate height according to the size of the slab.

[0034] S6, after the height is confirmed, the intelligent control system continues to issue instructions, and the pusher automatically pushes the slab onto the roller table and moves back and forth to complete the slab on the line as planned. Among them, the intelligent control system optimizes the PLC parameters according to the on-site conditions and outputs the optimal decision. The specific optimal decision is:

[0035] Analyze the properties of the intelligent control system and obtain several evaluation factors of the intelligent control system;

[0036] Obtain the control logic system of the intelligent control system and establish a mapping matrix between the PLC parameters of the automation control and several evaluation factors of the intelligent control system;

[0037] An evaluation index model of an intelligent control system is constructed based on the evaluation factors, wherein the evaluation index model takes a weight matrix of the evaluation factors and an evaluation factor decision matrix as input and takes an optimized PLC parameter group of the intelligent control system as output;

[0038] According to the set value interval, several decision values ​​are taken for each PLC parameter, and the values ​​are combined into several PLC parameter decision groups. All the PLC parameter decision groups are combined into a PLC parameter decision matrix. Based on the mapping matrix between the PLC parameters and several evaluation factors of the intelligent control system, the PLC parameter decision matrix is ​​converted into an evaluation factor decision matrix.

[0039] According to the properties of the intelligent control system, weight values ​​are added to several evaluation factors to obtain the weight matrix of the evaluation factors;

[0040] The weight matrix of the evaluation factors and the decision matrix of the evaluation factors are input into the evaluation index model of the intelligent control system to obtain the optimal PLC parameter group;

[0041] Perform anti-interference processing on the optimized PLC parameter group to obtain the final PLC parameter group;

[0042] The final PLC parameter group is fed back to the intelligent control system, and the intelligent control system modifies the control parameters according to the final PLC parameter group, and controls the intelligent control system to execute the optimized control decision.

[0043] This solution combines the properties of the intelligent control system to determine several evaluation factors of the intelligent control system. Then, combined with the control logic system of the intelligent control system, it obtains the intrinsic control matrix between the PLC parameters of the automated control and the evaluation factors. Then, by optimizing the PLC parameter evaluation of several PLC parameter decision groups, it can achieve the automated optimal decision of the PLC parameter group.

[0044] Among them, when selecting the PLC parameter decision group, those skilled in the art can understand that the smaller the set value interval is, the more the final optimized PLC parameter group is in line with the control requirements of the intelligent control system. However, the smaller the set value interval is, the greater the computing power required for the calculation process. Therefore, in the actual setting process, the staff can set it according to the needs;

[0045] For evaluation factors, the evaluation factors of intelligent control systems with different attributes are different. For example, for intelligent control systems with production and processing attributes, the evaluation factors are production efficiency, product quality, and system stability. The weights of various evaluation factors are also different for different intelligent control systems. Therefore, in this scheme, weight values ​​are added to several evaluation factors according to the attributes of the intelligent control system, so that the calculated optimized PLC parameter group can meet the needs of the intelligent control system.

[0046] The evaluation index model of intelligent control system based on evaluation factors specifically includes:

[0047] Training an evaluation factor scoring model for each evaluation factor, wherein the evaluation factor scoring model takes the numerical value of the evaluation factor as input and takes the positive score of the evaluation factor as output;

[0048] Constructing an evaluation factor weight normalization transformation model, wherein the evaluation factor weight normalization transformation model takes the evaluation factor weight matrix and the evaluation factor decision matrix as input and takes the weight normalization matrix as output;

[0049] A decision indicator calculation model is constructed, wherein the decision indicator calculation model takes a weight normalization matrix as input and takes a decision indicator as output.

[0050] Since the value of each evaluation factor is different during the calculation process, for example, for production efficiency, its evaluation value can be production time, output, etc., therefore, in this scheme, an evaluation factor scoring model is trained for each evaluation factor, and a regression calculation is performed on the positive score of the evaluation factor. The larger the positive score value, the better the evaluation factor evaluation.

[0051] The training evaluation factor scoring model specifically includes:

[0052] According to historical experience data, different evaluation factor values ​​are assigned corresponding positive scores, and the evaluation factor values ​​and positive scores are packaged into several sets of training data sets;

[0053] The plurality of training data sets are randomly divided into calculation training samples and test samples, wherein the number of the training data sets in the training samples accounts for 80% and the number of the training data sets in the test samples accounts for 20%;

[0054] Utilizing the training data set in the training sample to perform training model calculations for the evaluation factor scoring model, and obtaining multiple preliminary training models;

[0055] The training data set in the test sample is substituted into the preliminary training model, and the preliminary training model with the highest test fit is selected as the evaluation factor scoring model.

[0056] Substituting the training data set in the test sample into the preliminary training model and selecting the preliminary training model with the highest test fit specifically includes:

[0057] Input all the input variables of the training data set in the test sample into the preliminary training model to obtain the predicted output variables;

[0058] Calculate the regression coefficient of determination based on the output variables of all training datasets and the predicted output variables in the test sample;

[0059] Calculate the regression coefficient of determination for each preliminary training model;

[0060] The preliminary training model with the largest regression coefficient of determination is selected, wherein the calculation formula of the regression coefficient of determination is:

[0061]

[0062] Where R 2 is the regression determination coefficient; RSS is the residual sum of squares of the preliminary training model; TSS is the total sum of squares of the preliminary training model.

[0063] The step of inputting the weight matrix of the evaluation factors and the decision matrix of the evaluation factors into the evaluation index model of the intelligent control system to obtain the optimized PLC parameter group specifically includes:

[0064] Input the decision value of each evaluation factor in the evaluation factor decision matrix into the corresponding evaluation factor scoring model to obtain a number of decision positive scores for each evaluation factor, and combine the several decision positive scores of all evaluation factors into a number of decision positive score groups, and then combine the several decision positive score groups into a decision evaluation factor positive score matrix A;

[0065]

[0066] Among them, m is the total number of decision-making positive scoring groups, n is the total number of several evaluation factors of the intelligent control system, x ij is the positive score of the jth evaluation factor in the i-th decision positive score group;

[0067] The weight matrix of the evaluation factors and the positive score matrix D of the decision evaluation factors are input into the evaluation factor weight normalization transformation model. The normalized weight value of each evaluation factor is calculated based on the normalization calculation formula, and the normalized weight values ​​of all evaluation factors are combined into the weight normalization matrix B.

[0068]

[0069] Among them, w ij is the normalized value of the weight of the jth evaluation factor in the i-th decision positive scoring group;

[0070] Determine the optimal weight normalization value group S based on the weight normalization matrix B + and the worst weight normalized value set S - ,in,

[0071]

[0072] in, is the maximum value of the weight normalized value of the jth evaluation factor, is the minimum value of the weight normalized value of the j-th evaluation factor;

[0073] Calculate the decision index of each decision-making positive scoring group according to the index calculation formula;

[0074] The decision-making positive score group with the largest decision index is screened out, and its corresponding PLC parameter decision group is the optimized PLC parameter group.

[0075] The normalized calculation formula is:

[0076]

[0077] Where, α j is the weight value of the jth evaluation factor, x kj is the positive score of the jth evaluation factor in the kth decision positive score group

[0078] The indicator calculation formula is:

[0079]

[0080] Where C i is the decision indicator of the i-th group of decision-making positive scoring group.

[0081] The anti-interference processing for the optimized PLC parameter group specifically includes:

[0082] Perform anti-interference calculation on each PLC parameter in the optimized PLC parameter group according to the anti-interference formula to obtain the PLC parameter value after anti-interference processing of each PLC parameter;

[0083] The PLC parameter values ​​after anti-interference processing of each PLC parameter are combined to generate the final PLC parameter group;

[0084] The anti-interference formula is:

[0085] γ * =D -1 γ+σ

[0086] Where, γ * is the PLC parameter value after anti-interference processing, D -1 is the inverse matrix of the stability matrix of the intelligent control system, γ is the PLC parameter value of the optimized PLC parameter group, and σ is the noise of the intelligent control system.

[0087] By performing anti-interference processing on the optimized PLC parameter group, the final PLC parameter group obtained can effectively isolate external and internal noise interference, reduce parameter fluctuations during control system operation, and effectively ensure the stability of the intelligent control system during execution.

[0088] The automated control system assigns tasks to each slab based on the steel plant's production scheduling system. Based on the actual destination and purpose of each slab, the automated control system uses an intelligent path planning algorithm to plan the movements and routes of each slab. The intelligent control program of the intelligent control automation system then controls the corresponding equipment or rollers, allowing the slabs to automatically reach the corresponding positions specified by the scheduling system. This can greatly reduce errors caused by manual visual observation and operation, greatly reduce the number of sensors and inductive switches used, avoid problems caused by poor contact, damage, or malfunction of detection components, reduce the probability of equipment downtime for maintenance, and simultaneously control the operation of multiple modules. The operating efficiency is much higher than that of manual operation. Operators only need to supervise and handle emergencies, which greatly reduces the operator's labor intensity and improves production efficiency.

[0089] Combined with the properties of the automation control system, several evaluation factors of the automation control system are determined. Then, combined with the control logic system of the automation control system, the internal control matrix between the PLC parameters of the automation control and the evaluation factors is obtained. Then, by optimizing the PLC parameter evaluation of several PLC parameter decision groups, this method can achieve automated optimal decision-making for the PLC parameter group, reduce the degree of dependence on manual experience, and effectively ensure that the PLC control parameters can achieve the optimal operation of the automation control system.

[0090] The above description is only used to illustrate the technical solution of the present invention and is not intended to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in this field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. An intelligent motion control method for a stacking platform based on three-dimensional imaging of steel plate blanks, characterized by: It includes the following specific steps: S1 uses multiple LiDARs to scan steel plates at the steel mill site from different angles to obtain 3D data. Based on an AI algorithm, the algorithm compensates for the missing data of each radar with the data scanned by other radars to obtain 3D imaging data of slabs, stacks, and rollers. S2, generates scene images based on laser point cloud data, performs field scene modeling based on the scene images, and reasonably plans the recognition area; In step S3, the 3D data generated by radar scanning is transmitted to the data center via the network for storage. The gaps and offset points between the slabs are calibrated based on the stored data, and the model is trained using a deep learning model. S4, the intelligent control system optimizes PLC parameter adjustments based on on-site production conditions and outputs the optimal decision; S5, a deep learning model can distinguish the edges of each stacked slab. Based on the radar scan data, intelligent control algorithms and PLC optimization decisions, it guides the stacking platform to automatically rise or fall to the appropriate height according to the size of the slab. S6, after the height is confirmed, the intelligent control system continues to issue instructions, the pusher automatically pushes the slab onto the roller table and moves back and forth to complete the slab on-line as planned. The intelligent control system optimizes the PLC parameters according to the on-site conditions and outputs the optimal decision. The specific optimal decision is: Analyze the properties of the intelligent control system and obtain several evaluation factors of the intelligent control system; Obtain the control logic system of the intelligent control system and establish a mapping matrix between the PLC parameters of the automation control and several evaluation factors of the intelligent control system; An evaluation index model of an intelligent control system is constructed based on the evaluation factors, wherein the evaluation index model takes a weight matrix of the evaluation factors and an evaluation factor decision matrix as input and takes an optimized PLC parameter group of the intelligent control system as output; According to the set value interval, several decision values ​​are taken for each PLC parameter, and the values ​​are combined into several PLC parameter decision groups. All the PLC parameter decision groups are combined into a PLC parameter decision matrix. Based on the mapping matrix between the PLC parameters and several evaluation factors of the intelligent control system, the PLC parameter decision matrix is ​​converted into an evaluation factor decision matrix. According to the properties of the intelligent control system, weight values ​​are added to several evaluation factors to obtain the weight matrix of the evaluation factors; The weight matrix of the evaluation factors and the decision matrix of the evaluation factors are input into the evaluation index model of the intelligent control system to obtain the optimal PLC parameter group; Perform anti-interference processing on the optimized PLC parameter group to obtain the final PLC parameter group; The final PLC parameter group is fed back to the intelligent control system, and the intelligent control system modifies the control parameters according to the final PLC parameter group, and controls the intelligent control system to execute the optimized control decision.

2. The method for intelligent motion control of a pallet stack based on three-dimensional imaging of steel plate blanks according to claim 1, characterized in that: The evaluation index model for constructing an intelligent control system based on evaluation factors specifically includes: Training an evaluation factor scoring model for each evaluation factor, wherein the evaluation factor scoring model takes the numerical value of the evaluation factor as input and takes the positive score of the evaluation factor as output; Constructing an evaluation factor weight normalization transformation model, wherein the evaluation factor weight normalization transformation model takes the evaluation factor weight matrix and the evaluation factor decision matrix as input and takes the weight normalization matrix as output; A decision indicator calculation model is constructed, wherein the decision indicator calculation model takes a weight normalization matrix as input and takes a decision indicator as output.

3. The intelligent motion control method for stacking pallets based on three-dimensional imaging of steel plate blanks according to claim 2, characterized in that: The training evaluation factor scoring model specifically includes: According to historical experience data, different evaluation factor values ​​are assigned corresponding positive scores, and the evaluation factor values ​​and positive scores are packaged into several sets of training data sets; The plurality of training data sets are randomly divided into calculation training samples and test samples, wherein the number of the training data sets in the training samples accounts for 80% and the number of the training data sets in the test samples accounts for 20%; Utilizing the training data set in the training sample to perform training model calculations for the evaluation factor scoring model, and obtaining multiple preliminary training models; The training data set in the test sample is substituted into the preliminary training model, and the preliminary training model with the highest test fit is selected as the evaluation factor scoring model.

4. The method for intelligent motion control of a pallet stack based on three-dimensional imaging of steel plate blanks according to claim 3, characterized in that: Substituting the training data set in the test sample into the preliminary training model and selecting the preliminary training model with the highest test fit specifically includes: Input all the input variables of the training data set in the test sample into the preliminary training model to obtain the predicted output variables; Calculate the regression coefficient of determination based on the output variables of all training datasets and the predicted output variables in the test sample; Calculate the regression coefficient of determination for each preliminary training model; Screen out the preliminary training model with the largest regression determination coefficient.

5. The method for intelligent motion control of a pallet stacking platform based on three-dimensional imaging of steel plate blanks according to claim 4, characterized in that: The calculation formula of the regression determination coefficient is: Where R 2 is the regression determination coefficient; RSS is the residual sum of squares of the preliminary training model; TSS is the total sum of squares of the preliminary training model.

6. The method for intelligent motion control of a pallet stacking platform based on three-dimensional imaging of steel plate blanks according to claim 1, characterized in that: The step of inputting the weight matrix of the evaluation factors and the decision matrix of the evaluation factors into the evaluation index model of the intelligent control system to obtain the optimized PLC parameter group specifically includes: Input the decision value of each evaluation factor in the evaluation factor decision matrix into the corresponding evaluation factor scoring model to obtain a number of decision positive scores for each evaluation factor, and combine the several decision positive scores of all evaluation factors into a number of decision positive score groups, and then combine the several decision positive score groups into a decision evaluation factor positive score matrix A; Among them, m is the total number of decision-making positive scoring groups, n is the total number of several evaluation factors of the intelligent control system, x ij is the positive score of the jth evaluation factor in the i-th decision positive score group; The weight matrix of the evaluation factors and the positive score matrix D of the decision evaluation factors are input into the evaluation factor weight normalization transformation model. The normalized weight value of each evaluation factor is calculated based on the normalization calculation formula, and the normalized weight values ​​of all evaluation factors are combined into the weight normalization matrix B. Among them, w ij is the normalized value of the weight of the jth evaluation factor in the i-th decision positive scoring group; Determine the optimal weight normalization value group S based on the weight normalization matrix B + and the worst weight normalized value set S - ,in, in, is the maximum value of the weight normalized value of the jth evaluation factor, is the minimum value of the weight normalized value of the j-th evaluation factor; Calculate the decision index of each decision-making positive scoring group according to the index calculation formula; The decision-making positive score group with the largest decision index is screened out, and its corresponding PLC parameter decision group is the optimized PLC parameter group.

7. The method for intelligent motion control of a pallet stacking platform based on three-dimensional imaging of steel plate blanks according to claim 6, characterized in that: The normalized calculation formula is: Where, α j is the weight value of the jth evaluation factor, x kj is the positive score of the jth evaluation factor in the kth decision positive score group.

8. The method for intelligent motion control of a pallet stacking platform based on three-dimensional imaging of steel plate blanks according to claim 6, characterized in that: The indicator calculation formula is: Where C i is the decision indicator of the i-th group of decision-making positive scoring group.

9. The method for intelligent motion control of a pallet stacking platform based on three-dimensional imaging of steel plate blanks according to claim 1, characterized in that: The anti-interference processing for the optimized PLC parameter group specifically includes: Perform anti-interference calculation on each PLC parameter in the optimized PLC parameter group according to the anti-interference formula to obtain the PLC parameter value after anti-interference processing of each PLC parameter; The PLC parameter values ​​after anti-interference processing of each PLC parameter are combined to generate the final PLC parameter group; The anti-interference formula is: c * =D -1 c+s Where, γ * is the PLC parameter value after anti-interference processing, D -1 is the inverse matrix of the stability matrix of the intelligent control system, γ is the PLC parameter value of the optimized PLC parameter group, and σ is the noise of the intelligent control system.

Images