An efficient pipeline packing system based on internet of things

Abstract

The application discloses a kind of high-efficiency flow line packing systems based on Internet of Things.The application, through the collaborative work of sensor module, executor module and controller module, realizes the automatic identification, handling and packing of goods, greatly reduces manual intervention, improves the running speed and efficiency of production line.Artificial intelligence and optimization module can analyze and process data in real time, make decisions quickly, so that the whole system can quickly respond to changes in production demand.Artificial intelligence and optimization module can make more intelligent decisions through deep learning and optimization algorithms, improving the automation level of the system.Model training module can continuously learn, and as time goes on, the system can better adapt to different production environments and needs.Sorting optimization strategy submodule can optimize the process according to production data, reduce unnecessary steps and improve overall efficiency, providing strong support for modern industrial production.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of flow line packaging, and particularly relates to an efficient flow line packaging system based on the Internet of Things. BACKGROUND

[0002] Flow line packaging is an efficient operation mode in modern industrial production, which realizes the automation and continuity of packaging operation by decomposing the packaging process into multiple links and arranging them according to the product flow sequence. Flow line packaging mainly includes product preparation, packaging material supply, product filling, sealing, inspection and handling, etc. Its purpose is to improve packaging efficiency, reduce labor cost and ensure the safety and beauty of products during transportation and storage. Flow line packaging is widely used in the production process of food, medicine, daily chemical, electronics and other industries. With the progress of science and technology, flow line packaging equipment and technology are constantly optimized and upgraded. In short, flow line packaging plays an important role in modern industrial production.

[0003] However, the traditional flow line packaging system often relies on manual operation, which is low in efficiency and prone to errors. At the same time, the system lacks intelligent analysis capability and cannot make quick decisions based on real-time data, nor can it optimize the production process, thereby affecting the overall packaging efficiency. SUMMARY

[0004] The purpose of the present application is to solve the above-mentioned problems and provide an efficient flow line packaging system based on the Internet of Things.

[0005] The technical solution adopted by the present application is as follows: an efficient flow line packaging system based on the Internet of Things, the system comprising: a sensor module, a packaging executor module, a controller module, a communication module, a data processing and storage module, an artificial intelligence and optimization module and a signal processing module;

[0006] The artificial intelligence and optimization module is internally provided with a data processing module, a feature extraction module, a model training module, a sorting optimization strategy submodule and a result output submodule;

[0007] The sensor module is connected with the signal processing module: the original data collected by the sensor module needs to be filtered, amplified and converted by the signal processing module to obtain accurate and stable signals;

[0008] The sensor module is connected with the controller module: the processed signals are sent to the controller module for further data analysis and instruction issuance;

[0009] The sensor module, the packaging executor module and the controller module are connected: the controller module sends control signals to the executor module according to the processed data and the preset program to drive the executor to perform specific packaging operations;

[0010] The controller module receives signals from the sensor module, makes logical judgments and decisions, and sends control instructions to the packaging executor module;

[0011] The controller module is connected with the communication module: data exchange with other systems or cloud platform through the communication module;

[0012] The controller module is connected with the data processing and storage module: the processed data is sent to the data processing and storage module for storage and analysis;

[0013] The communication module is connected with the controller module: responsible for sending instructions from the controller to external systems or receiving instructions from external systems;

[0014] The communication module is connected with the data processing and storage module: realizing remote transmission and backup of data;

[0015] The data processing and storage module is connected with the controller module: receiving data from the controller, storing and preprocessing;

[0016] The data processing and storage module is connected with the artificial intelligence and optimization module: providing data required for training and reasoning for the artificial intelligence module;

[0017] The data processing and storage module is connected with the communication module: realizing remote storage and access of data;

[0018] The artificial intelligence and optimization module is connected with the data processing and storage module: using processed and stored data for model training and optimization;

[0019] The artificial intelligence and optimization module is connected with the controller module: feeding back the optimized control strategy or decision to the controller to guide the operation of the executor;

[0020] The signal processing module is connected with the sensor module: processing the signals collected by the sensor to meet the system requirements.

[0021] In a preferred embodiment, the sensor module is internally provided with:

[0022] Cargo identification sensor: including barcode scanner, RFID reader, for reading information on cargo label;

[0023] Position sensor: photoelectric sensor, proximity sensor, for detecting the position of goods on the assembly line;

[0024] Weighing sensor, for measuring the weight of goods;

[0025] Including encoder, for monitoring the running speed of the conveyor belt;

[0026] The sensor module monitors various parameters of the goods in real time through physical contact or non-contact; when the goods pass through the sensor, the sensor converts the physical signal into an electrical signal and further processes it through the signal processing module.

[0027] In a preferred embodiment, the packing executor module comprises:

[0028] Conveyor belt: used to transport goods to the packing area;

[0029] Robot: collaborative robot or Delta robot, used to grab, carry and place goods;

[0030] Packaging equipment sealer, heat shrink machine, used for packaging goods;

[0031] Sorting equipment: slider sorting machine or conveying chain sorting machine, used to sort the packed goods to the designated area;

[0032] The executor module receives instructions from the controller module to drive the corresponding mechanical components to complete the operation; when the goods arrive at the packing position, the robot grabs the goods according to the preset program and places them on the packaging equipment for packaging.

[0033] In a preferred embodiment, the controller module is responsible for processing sensor data and controlling the executor module;

[0034] The controller module receives input signals from the sensor module, processes them according to the preset program logic, and then outputs control signals to the executor module; when the weight sensor detects that the weight of the goods exceeds the preset range, the controller sends a signal to the robot to adjust the grabbing force.

[0035] In a preferred embodiment, the communication module sends and receives data through wired or wireless means; the data of the controller module is uploaded to the cloud server through Ethernet connection, or data exchange is carried out with mobile devices through Bluetooth connection;

[0036] The data processing and storage module has a data processing unit for data cleaning and data fusion operation; the data storage unit of the data processing and storage module includes local database and cloud database; the data processing unit preprocesses the data collected by the sensor, including removing outliers and standardizing data; the processed data is stored in the database for historical data analysis or to provide training data for the artificial intelligence module.

[0037] In a preferred embodiment, the data processing module is responsible for cleaning data, removing invalid or erroneous data, standardizing and normalizing data, and possibly feature fusion and data dimensionality reduction, to prepare for subsequent efficient pipeline data feature extraction;

[0038] The feature extraction module converts the preprocessed pipeline packed data into a form that can be used for machine learning models; including selecting and extracting the most informative features for intrusion detection tasks, which can include CAN message statistics, pipeline packed time series features, and pipeline packed communication patterns.

[0039] In a preferred embodiment, the model training module includes the following steps in the process of using the training model:

[0040] 1 Initialize model parameters: randomly initialize the weights and biases of the pipeline packed data;

[0041] 2 Forward propagation: input the feature parameters of the pipeline packed data, calculate the activation values of the hidden layer and the output layer, and the formula for forward propagation is:

[0042]

[0043] where a j l is the activation value of the jth neuron in the lth layer, w ij l is the weight from the ith neuron in the (l-1)th layer to the jth neuron in the lth layer, b j l is the bias of the jth neuron in the lth layer, and σ is the activation function, and nl-1 is the number of neurons in the (l-1)th layer;

[0044] 3 Calculate the loss function of the pipeline packed data: compare the activation values of the output layer with the actual fault types; the formula for calculating the loss function is:

[0045]

[0046] where L(y,y^) is the loss function, is the independent encoding of the actual fault type of the ith sample, is the activation value of the output layer of the ith sample, mm is the number of samples, and nout is the number of neurons in the output layer;

[0047] 4 Backpropagation: calculate the gradient of the loss function of the pipeline packed data with respect to the weights and biases, and update the weights and biases, and the formula for backpropagation is:

[0048] where, is the gradient of the loss function with respect to the weights w ij l j l is the error of the jth neuron in the lth layer, a j l-1 is the activation value of the jth neuron in the l-1th layer.

[0049] In a preferred embodiment, the sorting optimization strategy submodule initializes the optimization algorithm according to preset parameters or historical data at system startup; these parameters may include learning rate, regularization coefficient, and iteration number;

[0050] Data reception: The submodule receives input data from the data processing and storage module, which is usually preprocessed and ready for model training or optimization;

[0051] Model training: The optimization strategy submodule trains the model using machine learning algorithms; during the training process, the optimization algorithm adjusts the weights and biases of the model to minimize the loss function;

[0052] Parameter adjustment:

[0053] Learning rate scheduling: Automatically adjust the learning rate according to the training progress or the performance of the validation set to prevent slow convergence or overfitting during training;

[0054] Regularization: Apply L1 or L2 regularization to reduce model complexity and prevent overfitting;

[0055] Dropout: Randomly drop a portion of neurons in the network during training to enhance the generalization ability of the model;

[0056] Performance evaluation: After each iteration, the submodule evaluates the performance of the model, usually using the validation set; if the performance does not improve or decreases, the submodule may adjust the optimization strategy;

[0057] Strategy update: According to the evaluation results, the optimization strategy submodule updates the parameters of the optimization algorithm;

[0058] Output: When the model training reaches the preset iteration number or performance standard, the optimization strategy submodule outputs the final optimized model parameters for use by other parts of the artificial intelligence and optimization module.

[0059] In a preferred embodiment, the operation method of the result output submodule is as follows:

[0060] Receive prediction results: The submodule receives the final prediction results from the artificial intelligence and optimization module, which may be classification labels, regression values, or optimized packaging schemes;​

[0061] Processing prediction results:

[0062] Confidence calculation: for classification problems, use the softmax function to calculate the prediction probability of each class, i.e. confidence;

[0063] Result formatting: convert the prediction results into an easy-to-understand format, and convert the classification labels into corresponding goods names;

[0064] Result verification: before output, the sub-module may verify the results to ensure that they meet the business logic and system requirements;

[0065] Visualization: visualize the results through the monitoring interface, including goods type, packaging scheme, and confidence information;

[0066] Data storage: store the prediction results and related information in the database of the data processing and storage module for historical record query and subsequent analysis;

[0067] Feedback mechanism: according to the prediction results and the actual operation effect, the sub-module may provide feedback to the artificial intelligence and optimization module to further adjust and optimize the model;

[0068] Output instructions: finally, the sub-module outputs the prediction results or optimization scheme in the form of instructions to the controller module, and the controller module controls the packaging executor module to execute specific packaging tasks according to the instructions.

[0069] In a preferred embodiment, the signal processing module first filters and amplifies the analog signals output by the sensors, and then converts the analog signals into digital signals through an analog-to-digital converter (ADC) for further processing by the controller module; the analog signals of the weighing sensor are converted into digital signals to accurately measure the weight of the goods.

[0070] In summary, due to the adoption of the above technical solutions, the present application has the following advantages:

[0071] 1、In the present application, through the cooperative work of the sensor module, the actuator module and the controller module, automatic identification, handling and packaging of goods are realized, greatly reducing manual intervention and improving the running speed and efficiency of the production line. Artificial intelligence and optimization module can analyze and process data in real time, make quick decisions, so that the whole system can quickly respond to changes in production demand. Artificial intelligence and optimization module can make more intelligent decisions through deep learning and optimization algorithms, improving the automation level of the system. Model training module can continuously learn, and as time goes by, the system can better adapt to different production environments and needs. The sorting optimization strategy sub-module can optimize the process according to production data, reduce unnecessary steps and improve overall efficiency. Through data analysis, artificial intelligence and optimization module can predict potential equipment failures and perform maintenance in advance to reduce downtime. The system can automatically adjust the packaging strategy according to different goods characteristics to provide more personalized services. The contribution of artificial intelligence and optimization module to the whole system is that it improves the intelligent level of the system and enhances the adaptive ability of the system, thereby playing a key role in improving production efficiency, reducing error rate, saving cost and improving product quality. By integrating artificial intelligence technology, the system can better adapt to complex and variable production environments and provide strong support for modern industrial production.

[0072] 2、In the present application, the automated packaging process reduces the dependence on manpower, especially in high-intensity and repetitive labor, saving labor costs. By precisely controlling the packaging process, the risk of material waste and product damage is reduced, thereby reducing costs. Artificial intelligence and optimization module ensure the consistency of packaging quality of each package, improving the standardization of products. BRIEF DESCRIPTION OF DRAWINGS

[0073] Fig. 1 is the overall system block diagram of the present application;

[0074] Fig. 2 is the artificial intelligence and optimization module in the present application. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0076] Reference Figs. 1-2 ,

[0077] Example:

[0078] An efficient pipeline packaging system based on Internet of Things, the system comprises: a sensor module, a packaging executor module, a controller module, a communication module, a data processing and storage module, an artificial intelligence and optimization module, and a signal processing module.

[0079] The artificial intelligence and optimization module is internally provided with a data processing module, a feature extraction module, a model training module, a sorting optimization strategy submodule, and a result output submodule.

[0080] The sensor module is connected with the signal processing module: the raw data collected by the sensor module needs to be filtered, amplified, converted, etc. through the signal processing module to obtain accurate and stable signals.

[0081] The sensor module is connected with the controller module: the processed signals are sent to the controller module for further data analysis and instruction issuance.

[0082] The sensor module is connected with the controller module: the controller module sends control signals to the executor module according to the processed data and the preset program to drive the executor to perform specific packaging operations.

[0083] The controller module receives signals from the sensor module, makes logical judgments and decisions, and sends control instructions to the packaging executor module.

[0084] The controller module is connected with the communication module: data exchange is carried out with other systems or cloud platforms through the communication module.

[0085] The controller module is connected with the data processing and storage module: the processed data is sent to the data processing and storage module for storage and analysis.

[0086] The communication module is connected with the controller module: responsible for sending instructions of the controller to external systems or receiving instructions from external systems.

[0087] The communication module is connected with the data processing and storage module: realizing remote transmission and backup of data.

[0088] The data processing and storage module is connected with the controller module: receiving data from the controller, storing and preprocessing.

[0089] The data processing and storage module is connected with the artificial intelligence and optimization module: providing data required for training and reasoning of the artificial intelligence module.

[0090] The data processing and storage module is connected with the communication module: realizing remote storage and access of data.

[0091] The artificial intelligence and optimization module is connected with the data processing and storage module: using the processed and stored data for model training and optimization.

[0092] Artificial intelligence and optimization module is connected with controller module: feedback the optimized control strategy or decision to the controller to guide the operation of the actuator.

[0093] Signal processing module is connected with sensor module: process the signals collected by sensors to meet the system requirements.

[0094] The sensor module is internally provided with:

[0095] Cargo identification sensor: such as barcode scanner, RFID reader, etc., used to read the information on the cargo label.

[0096] Position sensor: photoelectric sensor, proximity sensor, used to detect the position of the goods on the assembly line.

[0097] Weighing sensor, used to measure the weight of the goods.

[0098] Such as encoder, used to monitor the running speed of the conveyor belt;

[0099] The sensor module monitors various parameters of the goods in real time through physical contact or non-contact methods. When the goods pass through the sensor, the sensor converts the physical signal into an electrical signal and further processes it through the signal processing module. The barcode scanner emits a light beam when the goods pass through, reads the barcode information and converts it into electronic data.

[0100] The packaging actuator module includes:

[0101] Conveyor belt: used to transport goods to the packaging area.

[0102] Robot: collaborative robot or Delta robot, used to grab, carry and place goods.

[0103] Packaging equipment: sealing machine, heat shrink machine, etc., used to package goods.

[0104] Sorting equipment: slider sorting machine or conveyor chain sorting machine, used to sort the packaged goods to the designated area.

[0105] The actuator module receives instructions from the controller module to drive the corresponding mechanical components to complete the operation. When the goods arrive at the packaging position, the robot grabs the goods according to the preset program and places them on the packaging equipment for packaging.

[0106] The controller module is responsible for processing sensor data and controlling the actuator module.

[0107] The controller module receives input signals from the sensor module, processes them according to pre-set program logic, and outputs control signals to the actuator module. When the weight sensor detects that the weight of the goods exceeds the pre-set range, the controller sends a signal to the robot to adjust the gripping force

[0108] The communication module sends and receives data through wired or wireless means. The data from the controller module is uploaded to the cloud server through an Ethernet connection, or exchanged with mobile devices through a Bluetooth connection;

[0109] The data processing and storage module includes a data processing unit for data cleaning, data fusion, etc. The data storage unit includes a local database and a cloud database. The data processing unit preprocesses the data collected by the sensor, such as removing outliers and standardizing data. The processed data is stored in the database for historical data analysis or to provide training data for the artificial intelligence module.

[0110] The data processing module is responsible for cleaning data, removing invalid or erroneous data, standardizing and normalizing data, and possibly feature fusion and data dimensionality reduction, to prepare for subsequent efficient pipeline data feature extraction;

[0111] The feature extraction module converts the preprocessed pipeline packed data into a form that can be used for machine learning models; including selecting and extracting features that are most informative for intrusion detection tasks, which can include CAN message statistics, pipeline packing time series features, and pipeline packing communication patterns.

[0112] The model training module includes the following steps in using the training model:

[0113] 1. Initialize model parameters: randomly initialize the weights and biases of the pipeline packed data;

[0114] 2. Forward propagation: input the feature parameters of the pipeline packed data, calculate the activation values of the hidden layer and the output layer, and the formula for forward propagation is:

[0115] where a j l is the activation value of the jth neuron in the lth layer, w ij l is the weight from the ith neuron in the l-1th layer to the jth neuron in the lth layer, b j l is the bias of the jth neuron in the lth layer, σ is the activation function, and n l-1 is the number of neurons in the l-1th layer.

[0116] 3. Compute the loss function for the packed data pipeline: compare the activation values of the output layer with the actual fault types and calculate the difference; the formula for calculating the loss function is:

[0117] where L(y, y^) is the loss function, yj(i) is the independent code of the actual fault type of the i-th sample, y^ji is the activation value of the output layer of the i-th sample, mm is the number of samples, and nout is the number of neurons in the output layer;

[0118] 4. Backpropagation: calculate the gradient of the loss function for the packed data pipeline with respect to the weights and biases, and update the weights and biases; the formula for backpropagation is:

[0119] where is the gradient of the loss function with respect to the weight w ij l ,

[0120] δ j l is the error of the j-th neuron in the l-th layer, a j l-1 is the activation value of the j-th neuron in the l-1-th layer.

[0121] The sorting optimization strategy submodule initializes the optimization algorithm according to the preset parameters or historical data when the system starts. These parameters may include learning rate, regularization coefficient, number of iterations, etc.

[0122] Data reception: the submodule receives input data from the data processing and storage module, which is usually pre-processed and ready for model training or optimization.

[0123] Model training: the optimization strategy submodule uses machine learning algorithms to train the model. During the training process, the optimization algorithm adjusts the weights and biases of the model to minimize the loss function.

[0124] Parameter adjustment:

[0125] Learning rate scheduling: automatically adjust the learning rate according to the training progress or the performance of the validation set to prevent slow convergence or overfitting during the training process.

[0126] Regularization: apply L1 or L2 regularization to reduce model complexity and prevent overfitting.

[0127] Dropout: randomly drop a portion of neurons in the network during training to enhance the generalization ability of the model.

[0128] Performance Evaluation: After each iteration, the sub-module evaluates the model's performance, typically using a validation set. If performance does not improve or declines, the sub-module may adjust the optimization strategy.

[0129] Strategy Update: Based on the evaluation results, the optimization strategy sub-module updates the parameters of the optimization algorithm, such as changing the learning rate, adjusting the regularization coefficient, etc.

[0130] Output: When the model training reaches the pre-set number of iterations or performance standards, the optimization strategy sub-module outputs the final optimized model parameters for use by other parts of the artificial intelligence and optimization module.

[0131] The running method of the result output sub-module is as follows:

[0132] Receive Prediction Results: The sub-module receives the final prediction results from the artificial intelligence and optimization module, which may be classification labels, regression values, or optimized packaging schemes.

[0133] Process Prediction Results:

[0134] Confidence Calculation: For classification problems, use the softmax function to calculate the predicted probability of each class, i.e., the confidence.

[0135] Result Formatting: Convert the prediction results into an easy-to-understand format, such as converting classification labels into corresponding cargo names.

[0136] Result Verification: Before output, the sub-module may verify the results to ensure they meet business logic and system requirements.

[0137] Visualization: Visualize the results through the monitoring interface, including cargo types, packaging schemes, confidence, etc.

[0138] Data Storage: Store prediction results and related information in the database of the data processing and storage module for historical record queries and subsequent analysis.

[0139] Feedback Mechanism: Based on the prediction results and actual operation effects, the sub-module may provide feedback to the artificial intelligence and optimization module for further adjustment and optimization of the model.

[0140] Output Instructions: Finally, the sub-module outputs the prediction results or optimization scheme in the form of instructions to the controller module, which controls the packaging executor module to execute specific packaging tasks.

[0141] The signal processing module first filters and amplifies the analog signals output by the sensors, and then converts the analog signals into digital signals through an analog-to-digital converter (ADC) for further processing by the controller module. The analog signals of the weighing sensors are converted into digital signals to accurately measure the weight of the goods.

[0142] In the present application, through the cooperative work of the sensor module, the actuator module and the controller module, automatic identification, handling and packaging of goods are realized, greatly reducing manual intervention and improving the running speed and efficiency of the production line. The artificial intelligence and optimization module can analyze and process data in real time and make quick decisions, so that the entire system can quickly respond to changes in production demand. The artificial intelligence and optimization module can make more intelligent decisions through deep learning and optimization algorithms, improving the automation level of the system. The model training module can continuously learn, and as time goes on, the system can better adapt to different production environments and needs. The sorting optimization strategy sub-module can optimize the process according to production data, reduce unnecessary steps, and improve overall efficiency. Through analysis of data, the artificial intelligence and optimization module can predict potential equipment failures and perform maintenance in advance to reduce downtime. The system can automatically adjust the packaging strategy according to different characteristics of the goods, providing more personalized services. The contribution of the artificial intelligence and optimization module to the entire system lies in its improvement of the intelligent level of the system and enhancement of the adaptive ability of the system, thereby playing a key role in improving production efficiency, reducing error rate, saving cost and improving product quality. By integrating artificial intelligence technology, the system can better adapt to complex and variable production environments, providing strong support for modern industrial production.

[0143] In the present application, the automated packaging process reduces the dependence on manpower, especially in high-intensity and repetitive labor, saving labor costs. By precisely controlling the packaging process, the risk of material waste and product damage is reduced, thereby reducing costs. The artificial intelligence and optimization module ensures consistent packaging quality for each package, improving the standardization of products.

[0144] It is to be noted that, in the present text, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", "has", "having", "includes", "including", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or even inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a", "has... a", "includes... a", or "has... a" does not, without more constraints, preclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0145] The above examples are merely intended to illustrate the technical solutions of the present application, but not to limit the same; even though the present application has been described in detail with reference to the foregoing examples, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing examples can be modified, or some technical features thereof can be replaced by equivalents; and such modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An efficient pipeline packing system based on Internet of Things, characterized in that: The system comprises a sensor module, a packaging executor module, a controller module, a communication module, a data processing and storage module, an artificial intelligence and optimization module, and a signal processing module; The artificial intelligence and optimization module is internally provided with a data processing module, a feature extraction module, a model training module, a sorting optimization strategy submodule, and a result output submodule; The sensor module is connected with the signal processing module: the original data collected by the sensor module needs to be filtered, amplified, and converted by the signal processing module to obtain accurate and stable signals; The sensor module is connected with the controller module: the processed signals are sent to the controller module for further data analysis and instruction issuance; The sensor module and the packaging executor module are connected with the controller module: the controller module sends control signals to the packaging executor module according to the processed data and the preset program to drive the executor to perform specific packaging operations; The controller module receives signals from the sensor module, makes logical judgments and decisions, and sends control instructions to the packaging executor module; The controller module is connected with the communication module: data exchange is performed with other systems through the communication module; The controller module is connected with the data processing and storage module: the processed data is sent to the data processing and storage module for storage and analysis; The communication module is connected with the controller module: responsible for sending instructions of the controller to external systems or receiving instructions from external systems; The communication module is connected with the data processing and storage module: remote transmission and backup of data are realized; The data processing and storage module is connected with the controller module: receives data from the controller for storage and preprocessing; The data processing and storage module is connected with the artificial intelligence and optimization module: provides data required for training and reasoning for the artificial intelligence module; The data processing and storage module is connected with the communication module: remote storage and access of data are realized; The artificial intelligence and optimization module is connected with the data processing and storage module: uses the processed and stored data for model training and optimization; The artificial intelligence and optimization module is connected with the controller module: feeds back the optimized control strategy or decision to the controller to guide the operation of the executor; The signal processing module is connected with the sensor module: processes the signals collected by the sensor module to meet the system requirements, and the model training module includes the following steps in the process of using the training model:

1. Initialize model parameters: randomly initialize the weights and biases of the pipeline packaging data; 2. Forward propagation: input the feature parameters of the pipeline packaging data, calculate the activation values of the hidden layer and the output layer, and the formula for forward propagation is: where a j l is the activation value of the jth neuron of the lth layer, w ij l is the weight from the ith neuron of the l-1th layer to the jth neuron of the lth layer, b j l is the bias of the jth neuron of the lth layer, σ is an activation function, n l-1 is the number of neurons of the l-1th layer; 3. Calculate the loss function of the pipeline packaging data: compare the difference between the activation values of the output layer and the actual fault type; the formula for calculating the loss function is: Where L(y,y^) is the loss function, yj(i) is the independent code of the actual fault type of the i th sample, y^ji is the activation value of the output layer of the i th sample, mm is the number of samples, and nout is the number of neurons in the output layer; 4 Backpropagation: calculate the gradient of the loss function of the data pipeline packed data on weights and biases, and update the weights and biases, the formula of backpropagation is: wherein, is the gradient of the loss function with respect to the weights w ij l , delta j l is the error of the jth neuron of the lth layer, a j l-1 is the activation value of the jth neuron of the l-1th layer.

2. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The internal of the sensor module is provided with: Cargo identification sensor: including barcode scanner, RFID reader, for reading the information on the cargo label; Position sensor: photoelectric sensor, proximity sensor, for detecting the position of the goods on the pipeline; Weighing sensor, for measuring the weight of the goods; Encoder, for monitoring the running speed of the conveyor belt; The sensor module monitors various parameters of the goods in real time through physical contact or non-contact mode; when the goods pass through the sensor module, the sensor module converts the physical signal into an electrical signal and further processes it through the signal processing module.

3. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The packing executor module includes: Conveyor belt: for conveying goods to the packing area; Robot: collaborative robot or Delta robot, for grabbing, carrying and placing goods; Packaging equipment sealing machine, heat shrinkage machine, for packaging goods; Sorting equipment: slider sorting machine or conveying chain sorting machine, for sorting the packed goods to the designated area; The packing executor module receives instructions from the controller module to drive the corresponding mechanical components to complete the operation; when the goods arrive at the packing position, the robot grabs the goods according to the preset program and places them on the packaging equipment for packaging.

4. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The controller module is responsible for processing sensor data and controlling the packing executor module; The controller module receives input signals from the sensor module, processes them according to the preset program logic, and then outputs control signals to the packing executor module; when the weighing sensor detects that the weight of the goods exceeds the preset range, the controller sends a signal to the robot to adjust the grabbing force.

5. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The communication module sends and receives data through wired or wireless methods; the data of the controller module is uploaded to the cloud server through Ethernet connection, or data exchange is carried out with mobile devices through Bluetooth connection; The data processing and storage module's data processing unit: performs data cleaning and data fusion operations; the data storage unit of the data processing and storage module: includes local database and cloud database; the data processing unit preprocesses the data collected by the sensor module, including removing outliers and standardizing data; the processed data is stored in the database for historical data analysis or to provide training data for the artificial intelligence module.

6. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The data processing module is responsible for cleaning data, removing invalid or incorrect data, standardizing and normalizing data, and feature fusion and data dimensionality reduction, to prepare for the subsequent efficient pipeline packing data feature extraction; The feature extraction module converts the preprocessed pipeline packing data into a form that can be used for machine learning models; Including selecting and extracting the most informative features for intrusion detection tasks, these features include CAN message statistics, pipeline packing time series features, and pipeline packing communication patterns.

7. The high efficient pipelined packing system based on internet of things of claim 1, wherein: The sorting optimization strategy submodule initializes the optimization algorithm based on preset parameters or historical data when the system starts. These parameters include learning rate, regularization coefficient, and iteration number. Data reception: The submodule receives input data from the data processing and storage module. These data are usually pre-processed and ready for model training or optimization. Model training: The optimization strategy submodule trains the model using machine learning algorithms. During the training process, the optimization algorithm adjusts the model's weights and biases to minimize the loss function. Parameter adjustment: Learning rate scheduling: Automatically adjust the learning rate based on training progress or validation set performance to prevent slow convergence or overfitting during training. Regularization: Apply L1 or L2 regularization to reduce model complexity and prevent overfitting. Dropout: Randomly drop a portion of neurons in the network during training to enhance the model's generalization ability. Performance evaluation: After each iteration, the submodule evaluates the model's performance, usually using the validation set. If the performance does not improve or decreases, the submodule adjusts the optimization strategy. Strategy update: Based on the evaluation results, the optimization strategy submodule updates the parameters of the optimization algorithm. Output: When the model training reaches the preset iteration number or performance standard, the optimization strategy submodule outputs the final optimized model parameters for use by other parts of the artificial intelligence and optimization module.

8. The high efficient pipelined packing system based on Internet of Things of claim 1, wherein: The running method of the result output submodule is as follows: Receive prediction results: The submodule receives the final prediction results from the artificial intelligence and optimization module. These results are classification labels, regression values, or optimized packaging schemes. Process prediction results: Confidence calculation: For classification problems, use the softmax function to calculate the predicted probability of each class, i.e., confidence. Result formatting: Convert the prediction results into an easy-to-understand format, converting classification labels into corresponding cargo names. Result verification: Before output, the submodule verifies the results to ensure they meet business logic and system requirements. Visualization: Visualize the results through the monitoring interface, including cargo type, packaging scheme, and confidence information. Data storage: Store prediction results and related information in the database of the data processing and storage module for historical record queries and subsequent analysis. Feedback mechanism: Based on the prediction results and actual operation effects, the submodule provides feedback to the artificial intelligence and optimization module for further model adjustment and optimization. Output instructions: Finally, the submodule outputs the prediction results or optimization scheme in the form of instructions to the controller module, which controls the packaging executor module to perform specific packaging tasks.

9. The high efficient pipelined packing system based on internet of things of claim 1, wherein: The signal processing module first filters and amplifies the analog signals output by the sensor module, then converts the analog signals to digital signals through an analog-to-digital converter (ADC) for further processing by the controller module. The analog signals of the weighing sensor are converted to digital signals to accurately measure the weight of the cargo.

Images