Tracking and positioning adhesive film circulating packaging box data transmission system

Abstract

The invention discloses a traceable and locatable adhesive film circulating packaging box data transmission system, which belongs to the technical field of logistics and specifically comprises a system integration tracking and positioning module, a data processing module, an intelligent prediction module, a circulating management module and a user interaction module. The tracking and positioning module tracks the position of the packaging box in real time, and the data processing module processes and analyzes position data; the intelligent prediction module predicts a logistics demand, optimizes a transportation path and predicts a packaging box maintenance demand based on historical and real-time information by using a deep learning and optimization algorithm, and identifies a risk and deals with the risk in advance; the cycle management module comprehensively manages the life cycle of the packaging box and ensures cyclic utilization; the user interaction module provides a convenient interface, the logistics efficiency is improved, and intelligent management and cyclic utilization of the packaging box are promoted.

Description

Technical Field

[0001] The invention belongs to the technical field of logistics, and in particular is a data transmission system for a plastic film recycling packaging box capable of being tracked and positioned. Background Art

[0002] With the rapid development of e-commerce and logistics industries, the volume of express parcels and logistics transportation has increased dramatically. The unsustainability and environmental pollution problems of traditional disposable packaging materials have become increasingly prominent. A large number of disposable packaging materials are discarded after use, which not only causes a huge waste of resources, but also increases the environmental burden. At the same time, during the logistics and transportation process, the loss, damage and poor management of packaging boxes also occur from time to time, seriously affecting the efficiency of logistics. Therefore, the logistics packaging industry has begun to realize the drawbacks of traditional disposable packaging materials and tried to develop recyclable packaging materials. The initial exploration is to apply RFID and GPS Internet of Things technology to the tracking and positioning of packaging boxes, but due to technical level and cost factors, the application effect is limited.

[0003] For example, the patent with publication number CN104678919A discloses an intelligent monitoring system for packaging boxes, including: a sensor system composed of multiple sensors with different functions, a data acquisition unit, a data transmission unit and a data monitoring unit, wherein the sensor system obtains corresponding temperature signals, humidity signals, pressure difference signals, inclination signals and vibration acceleration signals; the data acquisition unit collects the temperature signals, humidity signals, pressure difference signals, inclination signals and vibration acceleration signals and packages them for transmission to the data transmission unit; the data transmission unit transmits the received data to the data monitoring unit; the data monitoring unit receives the data and stores it as historical data, and also determines whether the data is within a threshold range, and generates an alarm signal when it is not within a preset range, and the alarm signal is transmitted to the data acquisition unit through the data transmission unit. Therefore, this technical solution is convenient for monitoring packaging boxes and occupies little space.

[0004] The above existing technologies all have the following problems: 1) They mainly focus on monitoring the internal environment of the packaging box, such as the measurement of environmental parameters such as temperature and humidity, but lack monitoring of the external environment of the packaging box, especially location tracking and transportation route monitoring. For modern logistics systems that require full-process tracking and positioning of packaging boxes, the functions are relatively simple; 2) Although they can monitor and store data in real time, they lack an intelligent prediction module, and cannot make intelligent predictions on logistics needs, transportation routes, packaging box maintenance needs, etc. based on historical data and real-time information, nor can they identify potential risks and problems and take measures in advance to prevent and respond. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a data transmission system for traceable and positionable recycled film packaging boxes, which integrates five major modules: tracking and positioning, data processing, intelligent prediction, circulation management and user interaction. The tracking and positioning module tracks the location of the packaging box in real time, and the data processing module processes and analyzes the location data. The intelligent prediction module uses deep learning and optimization algorithms to predict logistics demand, optimize transportation routes and predict packaging box maintenance needs based on historical and real-time information, while identifying risks and responding in advance. The circulation management module comprehensively manages the life cycle of the packaging box to ensure recycling. The user interaction module provides a convenient interface, improves logistics efficiency, and promotes intelligent management and recycling of packaging boxes.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A data transmission system for a traceable and positionable film recycling packaging box, comprising: a tracking and positioning module, a data processing module, an intelligent prediction module, a recycling management module, and a user interaction module;

[0008] The tracking and positioning module is used to track and locate the position information of the film recycling packaging box in real time;

[0009] The data processing module is used to process, analyze and store the data collected by the tracking and positioning module;

[0010] The intelligent prediction module is used to make intelligent predictions on logistics demand, transportation routes, and packaging box maintenance demand based on historical data and real-time information using a deep learning algorithm combined with an optimization algorithm, and at the same time, identify potential risks and problems, and take preventive and countermeasure measures in advance;

[0011] The recycling management module is responsible for the entire life cycle management of the film recycling packaging box;

[0012] The user interaction module is used to provide a user interaction interface.

[0013] Specifically, the tracking and positioning module includes an RFID unit and a path planning unit. The RFID unit is configured with a multi-tag anti-collision strategy, and the multi-tag anti-collision strategy is used to read and write the electronic tag information on the packaging box through a wireless radio frequency method; the path planning unit is configured with a path planning strategy, which is used to plan the optimal transportation route for the packaging box based on real-time traffic data and historical logistics information.

[0014] Specifically, the specific steps of the multi-tag anti-collision strategy include:

[0015] A1: Using the pre-built machine learning prediction model, the initial frame length and preset recognition rounds are intelligently initialized according to historical data and current environmental parameters. When the reader sends a query command to all tags, it indicates that the current frame starts.

[0016] A2: After receiving the query command, each tag randomly selects a time slot to respond and sends its ID in the selected time slot;

[0017] A3: At the same time, the reader listens to the tag's response in each time slot.

[0018] If only one tag responds within the time slot, the reader successfully reads the tag information and marks the time slot as successful;

[0019] If multiple tags respond at the same time in a time slot, a collision will occur and the reader cannot correctly read the information of any tag, and the time slot is marked as a collision;

[0020] If no tag responds in the time slot, the time slot is marked as idle;

[0021] A4: Dynamically adjust the frame length of the next frame through cross-frame optimization strategy according to the collision, idle and resolution conditions of the current frame and previous frames;

[0022] A5: Repeat steps A2 to A4 until the preset recognition round is reached.

[0023] Specifically, the specific steps of the cross-frame optimization strategy in A4 include:

[0024] A4.1: At the end of each frame, the system collects the number of collisions, idle tags, and successfully resolved tags in each frame. At the same time, it captures the time distribution of tag responses, signal strength changes, and environmental interference data, and fuses the data using a head-to-tail splicing method to obtain tag fusion data.

[0025] A4.2: Preprocess and extract features of the label fusion data to obtain label fusion feature data;

[0026] A4.3: Load the pre-trained machine learning prediction model and input the label fusion feature data to obtain the prediction results of the machine learning prediction model;

[0027] A4.4: Based on the prediction results, use a multi-objective optimization algorithm to obtain the optimal frame length;

[0028] A4.5: Establish a real-time feedback mechanism to compare the recognition result of the current frame with the prediction result, and adjust the frame length of the next frame in real time according to the comparison result. At the same time, introduce prediction confidence evaluation and set the confidence threshold h.

[0029] like , then adjust the machine learning prediction model parameters, where y represents the prediction result of the machine learning prediction model for the next frame.

[0030] Specifically, the specific formula of the path planning strategy is:

[0031] in, represents the minimum cost from state s to the target state at time t, t represents the current time, represents the instantaneous cost function at time t, represents the weight of executing decision a, represents the time required to execute decision a to transfer from state s to the next state, It represents the time delay caused by the uncertainty factors encountered when executing decision a from state s to the next state. represents the distance penalty coefficient, Represents the transition from state s to the next state The cumulative distance of Indicates that from the next state Departure, the minimum cost after taking into account the transfer time and additional delay.

[0032] Specifically, the intelligent prediction module includes a scheduling unit, a demand prediction unit and a risk assessment unit; the scheduling unit is used to intelligently schedule the transportation and recycling of packaging boxes based on historical data and real-time information; the demand prediction unit predicts logistics demand through a deep learning model and prepares resources in advance; the risk assessment unit is configured with an improved random forest algorithm to assess the risks that may be encountered in the logistics process.

[0033] Specifically, the specific steps of the improved random forest algorithm include:

[0034] B1: Collect historical risk data of the logistics process, preprocess and extract features of the historical risk data, and obtain historical feature data;

[0035] B2: Select random forest and gradient boosting tree as base models and set the initial parameters of the base models;

[0036] B3: Using the stacking integration strategy, random forest and gradient boosting tree are used as the first-layer model, and the logistic regression model is used as the second-layer meta-model. The output of the first-layer model is used as the feature input of the second-layer meta-model. At the same time, a nonlinear transformation layer is introduced into each tree node in the base model to generate a risk assessment model.

[0037] B4: Set risk assessment thresholds and online learning mechanisms to enable the risk assessment model to receive new risk data in real time for risk assessment, and automatically update risk assessment model parameters based on the assessment results. At the same time, dynamically adjust the risk assessment threshold based on historical assessment results and actual risk occurrences;

[0038] B5: Regularly evaluate the performance indicators of risk assessment models.

[0039] Specifically, the risk data in B1 includes weather conditions, traffic conditions, cargo types, transportation routes, historical accident records, real-time traffic camera image data and dynamic data of transportation vehicles; the dynamic data of transportation vehicles includes the position, speed and acceleration of transportation vehicles.

[0040] Specifically, the cycle management module includes: a recycling unit, a cleaning unit, a detection unit, and a maintenance prediction unit;

[0041] The recycling unit is responsible for recycling the packaging boxes;

[0042] The cleaning unit is used to clean and disinfect the recycled packaging boxes;

[0043] The detection unit is used to detect the cleaned packaging box;

[0044] The maintenance prediction unit predicts the maintenance requirements of the packaging box based on the data analysis results and arranges a maintenance or replacement plan in advance.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention proposes a data transmission system for traceable and positionable recycled film packaging boxes. The system integrates the Internet of Things, big data, and artificial intelligence advanced technologies to achieve real-time tracking and positioning, intelligent prediction, and full life cycle management of recycled packaging boxes, effectively improving the efficiency and accuracy of logistics management and promoting the recycling and sustainable development of packaging boxes.

[0047] 2. The present invention proposes a data transmission system for traceable and positionable recycled film packaging boxes. By integrating five modules of tracking and positioning, data processing, intelligent prediction, circulation management and user interaction, it realizes all-round intelligent management of packaging boxes. On the one hand, it significantly improves the efficiency and accuracy of logistics transportation. Through real-time tracking and route planning strategies, it ensures that the packaging boxes can be delivered to the destination quickly and accurately, reducing transportation costs and delay risks; on the other hand, through intelligent prediction and risk assessment, the system can predict logistics needs in advance, plan transportation routes and identify potential risks, providing decision-making support for enterprises and enhancing the foresight and adaptability of logistics management; in addition, the system also promotes the recycling of packaging boxes, realizes the full life cycle management of packaging boxes through the circulation management module, and helps promote the development of green logistics. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a structural diagram of a data transmission system for a traceable and positionable film recycling packaging box according to the present invention;

[0049] Figure 2 This is a flow chart of the implementation of a multi-label anti-collision strategy for a data transmission system for a traceable and positionable film recycling packaging box according to the present invention;

[0050] Figure 3 This is a flow chart for realizing a cross-frame optimization strategy of a data transmission system for a traceable and positionable film recycling packaging box according to the present invention;

[0051] Figure 4 This is a flow chart for implementing a path planning unit of a data transmission system for a traceable and positionable adhesive film recycling packaging box according to the present invention. DETAILED DESCRIPTION

[0052] Example 1

[0053] See also Figure 1 , an embodiment of the present invention provides: a data transmission system for a traceable and positionable plastic film recycling packaging box, comprising:

[0054] Tracking and positioning module, data processing module, intelligent prediction module, cycle management module, user interaction module;

[0055] Tracking and positioning module, used to track and locate the location information of plastic film recycling packaging boxes in real time to ensure the accuracy and timeliness of logistics information;

[0056] The data processing module is used to process, analyze and store the data collected by the tracking and positioning module;

[0057] Intelligent prediction module, which uses deep learning algorithms combined with optimization algorithms to make intelligent predictions on logistics demand, transportation routes, and packaging box maintenance needs based on historical data and real-time information. It also identifies potential risks and problems and takes preventive and response measures in advance.

[0058] The recycling management module is responsible for the entire life cycle management of the film recycling packaging box, including recycling, cleaning, testing and maintenance;

[0059] The user interaction module is used to provide a user interaction interface, allowing users to easily query logistics information and submit change requests.

[0060] The tracking and positioning module includes: GPS unit, RFID unit, signal transmission unit, and path planning unit;

[0061] GPS unit, used to obtain the longitude and latitude information of the packaging box through satellite signals to achieve global positioning;

[0062] RFID unit, used to read and write the electronic tag information on the packaging box using wireless radio frequency technology to achieve rapid identification at close range;

[0063] Signal transmission unit, used to transmit the data collected by GPS and RFID to the central processing system to ensure the real-time and accuracy of the information;

[0064] The path planning unit is used to plan the optimal transportation path for the packaging box using the path planning strategy based on real-time traffic data and historical logistics information.

[0065] The data processing module includes: a data processing unit, a data analysis unit, a data storage unit, and a privacy protection unit;

[0066] The data processing unit is responsible for cleaning, converting and loading data to ensure data quality and availability;

[0067] Data analysis unit, used to conduct in-depth analysis of the processed data and analyze data development trends;

[0068] A data storage unit, used to store the processed and analyzed data in a database for subsequent query and use;

[0069] The privacy protection unit is responsible for encrypting and processing sensitive data, ensuring the security of data transmission and storage, and protecting user privacy.

[0070] The intelligent prediction module includes: scheduling unit, demand prediction unit, and risk assessment unit;

[0071] The scheduling unit is used to intelligently schedule the transportation and recycling of packaging boxes based on historical data and real-time information;

[0072] Demand forecasting unit, used to predict logistics demand based on machine learning models, prepare resources in advance, and reduce waiting time

[0073] The risk assessment unit is used to use the improved random forest algorithm to evaluate the risks that may be encountered in the logistics process and propose preventive measures.

[0074] The cycle management module includes: recovery unit, cleaning unit, detection unit, and maintenance prediction unit;

[0075] The recycling unit is responsible for the recycling of packaging boxes to ensure that the packaging boxes can be returned to the circulation system in a timely manner;

[0076] A cleaning unit is used to clean and disinfect the recycled packaging boxes to ensure the hygiene and reusability of the packaging boxes;

[0077] The inspection unit is used to inspect the cleaned packaging boxes to ensure that they meet the standards for reuse;

[0078] The maintenance prediction unit predicts the maintenance needs of packaging boxes based on data analysis results and arranges repair or replacement plans in advance.

[0079] Furthermore, the maintenance prediction unit adopts a survival analysis method, and the specific implementation steps of the survival analysis method include:

[0080] C1: Collect historical data of packaging boxes, including but not limited to each use cycle, maintenance records, fault type, maintenance time, and replacement time, ensure the integrity and accuracy of the data, and clean and pre-process the data;

[0081] C2: Defined Events: Events include but are not limited to the failure, repair or reaching the predetermined replacement criteria of the packaging box;

[0082] C3: Define the survival time: the period from when the packaging box is put into use to when the event occurs;

[0083] C4: Select a survival analysis parameter model based on the characteristics of the data and the purpose of the study, and use the collected data to fit the survival analysis parameter model to estimate the parameters of the survival analysis parameter model;

[0084] C5: Use residual analysis to evaluate the goodness of fit of the survival analysis parameter model, wherein residual analysis is a prior art in this field and is not an inventive solution of the present application, and will not be described in detail here;

[0085] C6: Estimate the survival probability of the boxes at different time points based on the fitted survival analysis parameter model, and draw a survival curve to show the survival status of the boxes over time;

[0086] C7: Use the results of survival analysis to predict the reliability or failure rate of packaging boxes in the future. Based on the prediction results, formulate a maintenance or replacement plan to optimize resource allocation.

[0087] It should be noted that the tracking and positioning module, data processing module, intelligent prediction module, cycle management module, and user interaction module are interconnected through data exchange and communication protocols to form an organic whole.

[0088] Furthermore, the tracking and positioning module, the data processing module, the intelligent prediction module, the cycle management module, the user interaction module and the workflow between the modules include:

[0089] The GPS unit in the tracking and positioning module obtains the location information of the packaging box in real time, and sends the location data to the signal transmission unit and the path planning unit. At the same time, when the RFID reader is close to the packaging box, it automatically reads and verifies the tag information, and also sends the read tag information to the signal transmission unit. The signal transmission unit transmits the data collected by the GPS unit and the RFID unit to the data processing module through the wireless network. The path planning unit plans the optimal transportation route according to the GPS data and the preset transportation rules, and sends the planned route information to the scheduling unit;

[0090] The data processing unit receives the data transmitted by the tracking and positioning module, performs preprocessing and cleaning, and sends the processed data to the data analysis unit. The data analysis unit uses statistical analysis methods to analyze the transportation efficiency and usage status of the packaging box, and sends the analysis results to the data storage unit and the intelligent prediction module;

[0091] The dispatch unit of the intelligent prediction module dispatches the transportation of packaging boxes according to the results of the route planning unit and the real-time transportation demand, and then sends the dispatch instructions to the transportation vehicles or operators and receives feedback information. The demand forecasting unit predicts future logistics demand based on historical data and real-time information, and sends the forecast results to the dispatch unit and the cycle management module to support decision-making. The risk assessment unit identifies potential risks and problems in the transportation process, takes preventive and response measures in advance, and sends the risk assessment results to the dispatch unit and the user interaction module;

[0092] The recycling unit of the circulation management module classifies and recycles the packaging boxes according to the usage status and degree of damage, and sends the recycling information to the cleaning unit and the maintenance prediction unit. After receiving the information sent by the recycling unit, the cleaning unit cleans and disinfects the recycled packaging boxes, and sends the cleaning results to the detection unit. The detection unit detects the cleaned packaging boxes, evaluates their performance and safety, and sends the detection results to the maintenance prediction unit and the user interaction module. After receiving the information, the maintenance prediction unit predicts the maintenance needs of the packaging boxes based on the detection results and historical data, and sends the maintenance prediction results to the scheduling unit and the user interaction module.

[0093] Example 2

[0094] See also Figure 2 and Figure 3 In this embodiment, a multi-tag anti-collision strategy is adopted in the RFID unit. The specific steps of the multi-tag anti-collision strategy include:

[0095] A1: Using the pre-built machine learning prediction model, the initial frame length and preset recognition rounds are intelligently initialized according to historical data and current environmental parameters. When the reader sends a query command to all tags, it indicates that the current frame starts.

[0096] Furthermore, the specific steps of A1 include:

[0097] A1.1: Collect historical data, including frame length, recognition rounds, and corresponding recognition efficiency or performance indicators under different environmental parameters. Different environmental parameters include signal strength, environmental interference level, and tag distribution density.

[0098] A1.2: Extract frame length and recognition round features from historical data, such as current environment parameters, number of tags, historical frame length, and recognition efficiency, and perform feature screening to obtain filtered historical data;

[0099] A1.3: Use the filtered historical data to train the regression model, and optimize the prediction performance of the regression model by adjusting the regression model parameters and training strategy. The specific process of training the regression model using known data is the prior art content in this field, and is not the inventive solution of this application, and will not be described in detail here;

[0100] A1.4: During system operation, the current environmental parameters and tag status are monitored in real time, where the tag status includes the number of tags and distribution density;

[0101] A1.5: Input the real-time environmental parameters into the regression model to obtain the predicted optimal frame length and recognition rounds. Based on the prediction results, intelligently initialize the initial frame length and preset recognition rounds of the reader to optimize recognition efficiency and system performance.

[0102] A2: After receiving the query command, each tag randomly selects a time slot to respond and sends its ID in the selected time slot;

[0103] Furthermore, the specific steps of A2 include:

[0104] A2.1: The reader broadcasts a query command to all tags, starting an identification cycle;

[0105] A2.2: After receiving the query command, each tag is ready to participate in the identification process;

[0106] A2.3: Each tag generates a random number internally, which determines in which time slot the tag will respond;

[0107] A2.4: The tag waits in the selected time slot until its time slot is reached;

[0108] A2.5: When the time slot is reached, the tag sends its ID or other identification information to the reader.

[0109] A3: At the same time, the reader listens to the tag's response in each time slot.

[0110] If only one tag responds within the time slot, the reader successfully reads the tag information and marks the time slot as successful;

[0111] If multiple tags respond at the same time in a time slot, a collision will occur and the reader cannot correctly read the information of any tag, and the time slot is marked as a collision;

[0112] If no tag responds in the time slot, the time slot is marked as idle;

[0113] A4: Dynamically adjust the frame length of the next frame through cross-frame optimization strategy according to the collision, idle and resolution conditions of the current frame and previous frames;

[0114] Furthermore, the goal of the cross-frame optimization strategy is to increase the frame length to disperse the responses of tags if the collision rate is high, and to reduce the frame length to improve the recognition efficiency if the collision rate is low and there are more idle time slots.

[0115] A5: Repeat steps A2 to A4 until the preset recognition round is reached.

[0116] The specific steps of the cross-frame optimization strategy in A4 include:

[0117] A4.1: At the end of each frame, the system collects the number of collisions, idle tags, and successfully resolved tags in each frame. At the same time, it captures the time distribution of tag responses, signal strength changes, and environmental interference data, and fuses the data using a head-to-tail splicing method to obtain tag fusion data.

[0118] Furthermore, the data collection steps include:

[0119] Collision detection: The system determines the collision between tags by detecting whether the received signals overlap or conflict;

[0120] Idle detection: In each time frame, if no tag responds, it is recorded as idle state;

[0121] Successful parsing: records the number of tags that were successfully parsed, i.e., those that were correctly identified without collision;

[0122] Time distribution: record the time point of each tag response to analyze the time distribution of the response;

[0123] Signal Strength: Captures signal strength data for each tag to assess signal quality and distance from the tag to the reader;

[0124] Environmental interference: Monitor and record interference factors in the environment, such as wireless signal sources and obstacles.

[0125] Data collection loop:

[0126] 1) Set the length of the time frame;

[0127] 2) In each time frame, perform the data collection step;

[0128] 3) Check whether there is a new tag response and record the relevant data;

[0129] 4) Check whether the time frame has ended. If not, return to continue collecting data; if it has ended, perform data fusion.

[0130] A4.2: Preprocessing and feature extraction are performed on the label fusion data to obtain label fusion feature data. In the present invention, the feature extraction adopts the principal component analysis method, and the calculation process of the principal component analysis method is the prior art content in this field, which is not the inventive solution of the present application and is not described in detail here;

[0131] A4.3: Load the pre-trained machine learning prediction model and input the label fusion feature data to obtain the prediction result of the machine learning prediction model. The training process of the machine learning prediction model is the prior art content in this field and is not the inventive solution of this application, so it will not be described in detail here.

[0132] A4.4: Based on the prediction results, a multi-objective optimization algorithm is used to obtain the optimal frame length, wherein the multi-objective optimization algorithm adopts the NSGA-II algorithm, and the calculation formula of the NSGA-II algorithm is the prior art content in this field, which is not the inventive solution of the present application and is not described in detail here;

[0133] A4.5: Establish a real-time feedback mechanism to compare the recognition result of the current frame with the prediction result, and adjust the frame length of the next frame in real time according to the comparison result. At the same time, introduce prediction confidence evaluation and set the confidence threshold h.

[0134] like , then adjust the machine learning prediction model parameters, where y represents the prediction result of the machine learning prediction model for the next frame.

[0135] In the present invention, the multi-tag anti-collision algorithm plays a vital role in the RFID unit, and its role and advantages are mainly reflected in the following aspects:

[0136] (1) In a traditional RFID system, when multiple tags enter the recognition range of the reader at the same time, they may respond to the reader's query command at the same time, resulting in signal interference, i.e., "collision". Such collision will significantly reduce the reading efficiency of the system. In the present invention, the anti-collision algorithm enables the tags to communicate with the reader in an orderly and one-by-one manner through set rules and strategies, thereby avoiding the occurrence of collisions and improving the reading efficiency of the system.

[0137] (2) In the absence of an anti-collision algorithm, the simultaneous response of multiple tags may cause the reader to be unable to accurately identify the information of each tag, and may even misread the information of multiple tags as the information of the same tag. In the present invention, the anti-collision algorithm can ensure that each tag is independently and accurately identified, thereby improving the recognition accuracy of the system, and has the advantages of wide applicability, strong flexibility and good scalability.

[0138] The specific formula of the path planning strategy is:

[0139] in, represents the minimum cost from state s to the target state at time t, t represents the current time, represents the instantaneous cost function at time t, represents the weight of executing decision a, represents the time required to execute decision a to transfer from state s to the next state, It represents the time delay caused by the uncertainty factors encountered when executing decision a from state s to the next state. represents the distance penalty coefficient, Represents the transition from state s to the next state The cumulative distance of Indicates that from the next state Departure, the minimum cost after taking into account the transfer time and additional delay.

[0140] Among them, state s represents the multi-dimensional information of geographical location, time, remaining distance, and vehicle status, and decision a represents the selection of different roads, speeds, and costs. Including driving time, fuel consumption, fines and other factors; in the present invention, time, distance, immediate cost, uncertainty and decision weight are innovatively combined to make the path planning more in line with the actual situation and user needs.

[0141] Example 3

[0142] See also Figure 4 The specific steps of the improved random forest algorithm in this embodiment include:

[0143] B1: Collect historical risk data of the logistics process, preprocess and extract features of the historical risk data, and obtain historical feature data;

[0144] B2: Select random forest and gradient boosting tree as base models, and set the initial parameters of the base models, including the number of trees, the maximum depth of the tree, the minimum number of samples required to split a node, and the learning rate;

[0145] B3: Using the stacking integration strategy, random forest and gradient boosting tree are used as the first-layer model, and the logistic regression model is used as the second-layer meta-model. The output of the first-layer model is used as the feature input of the second-layer meta-model. At the same time, a nonlinear transformation layer is introduced into each tree node in the base model to generate a risk assessment model.

[0146] Furthermore, the specific steps of B3 include:

[0147] B3.1: Obtain historical feature data and perform preprocessing;

[0148] B3.2: Training the first layer base model:

[0149] (1) Use the preprocessed historical feature data to train the random forest model and optimize the random forest model by adjusting the number of trees, maximum depth, and minimum number of samples required for splitting;

[0150] (2) Use the preprocessed historical feature data to train the gradient boosting tree model, and adjust the learning rate, number of trees, and tree depth parameters to optimize the gradient boosting tree model;

[0151] (3) Introducing nonlinear transformation layer:

[0152] Apply nonlinear transformations to the output or internal nodes of each base model, mainly by performing nonlinear transformations at the feature level and then inputting the transformed features into the base model, or by performing nonlinear combinations of the outputs of the base models after they have been output;

[0153] B3.3: Generate the input of the second layer model: the output results of the two base models are used as new features;

[0154] B3.4: Training the second layer meta-model:

[0155] Use logistic regression as a meta-model, take the output of the first-layer model as input features, train the logistic regression model, optimize the classification performance, and obtain a risk assessment model;

[0156] B3.5: Use accuracy on the test set to evaluate the performance of the risk assessment model, and deploy the trained risk assessment model to the production environment for risk assessment.

[0157] B4: Set risk assessment thresholds and online learning mechanisms to enable the risk assessment model to receive new risk data in real time for risk assessment, and automatically update risk assessment model parameters based on the assessment results. At the same time, dynamically adjust the risk assessment threshold based on historical assessment results and actual risk occurrences;

[0158] Furthermore, the specific steps of B4 include:

[0159] B4.1: Develop an interface to enable the risk assessment model to receive new risk data from the logistics system in real time and conduct risk assessment immediately;

[0160] B4.2: Using an online learning algorithm, according to the evaluation results of new risk data and the actual risk occurrence, the parameters of the model are automatically adjusted to adapt the model to the changing environment. The online learning algorithm in the present invention adopts a gradient descent algorithm, which is a prior art in the field and is not an inventive solution of the present application, and will not be described in detail here;

[0161] B4.3: Based on the comparison between historical assessment results and actual risk occurrence, dynamically adjust the risk assessment threshold to ensure the accuracy and timeliness of the assessment results. For example, when the actual risk occurrence frequency is higher than expected, the threshold can be lowered to increase risk sensitivity.

[0162] B5: Regularly evaluate the performance indicators of risk assessment models.

[0163] The risk data in B1 include weather conditions, traffic conditions, cargo types, transportation routes, historical accident records, real-time traffic camera image data and dynamic data of transportation vehicles; the dynamic data of transportation vehicles include the position, speed and acceleration of transportation vehicles.

[0164] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0165] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0166] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

Claims

1. A data transmission system for tracing and locating plastic film recycling packaging boxes, characterized in that: The system includes: tracking and positioning module, data processing module, intelligent prediction module, cycle management module, and user interaction module; The tracking and positioning module is used to track and locate the position information of the film recycling packaging box in real time; The data processing module is used to process, analyze and store the data collected by the tracking and positioning module; The intelligent prediction module is used to make intelligent predictions on logistics demand, transportation routes, and packaging box maintenance demand based on historical data and real-time information using a deep learning algorithm combined with an optimization algorithm, and at the same time, identify potential risks and problems, and take preventive and countermeasure measures in advance; The recycling management module is responsible for the entire life cycle management of the film recycling packaging box; The user interaction module is used to provide a user interaction interface.

2. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 1, characterized in that: The tracking and positioning module includes an RFID unit and a path planning unit. The RFID unit is configured with a multi-tag anti-collision strategy, and the multi-tag anti-collision strategy is used to read and write the electronic tag information on the packaging box through a wireless radio frequency method; the path planning unit is configured with a path planning strategy, which is used to plan the optimal transportation route for the packaging box based on real-time traffic data and historical logistics information.

3. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 2, characterized in that: The specific steps of the multi-tag anti-collision strategy include: A1: Using the pre-built machine learning prediction model, the initial frame length and preset recognition rounds are intelligently initialized according to historical data and current environmental parameters. When the reader sends a query command to all tags, it indicates that the current frame starts. A2: After receiving the query command, each tag randomly selects a time slot to respond and sends its ID in the selected time slot; A3: At the same time, the reader listens to the tag's response in each time slot. If only one tag responds within the time slot, the reader successfully reads the tag information and marks the time slot as successful; If multiple tags respond at the same time in a time slot, a collision will occur and the reader cannot correctly read the information of any tag, and the time slot is marked as a collision; If no tag responds in the time slot, the time slot is marked as idle; A4: Dynamically adjust the frame length of the next frame through cross-frame optimization strategy according to the collision, idle and resolution conditions of the current frame and the previous M frames; A5: Repeat steps A2 to A4 until the preset recognition round is reached.

4. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 3, characterized in that: The specific steps of the cross-frame optimization strategy in A4 include: A4.1: At the end of each frame, the system collects the number of collisions, idle tags, and successfully resolved tags in each frame. At the same time, it captures the time distribution of tag responses, signal strength changes, and environmental interference data, and fuses the data using a head-to-tail splicing method to obtain tag fusion data. A4.2: Preprocess and extract features of the label fusion data to obtain label fusion feature data; A4.3: Load the pre-trained machine learning prediction model and input the label fusion feature data to obtain the prediction results of the machine learning prediction model; A4.4: Based on the prediction results, use a multi-objective optimization algorithm to obtain the optimal frame length; A4.5: Establish a real-time feedback mechanism to compare the recognition result of the current frame with the prediction result, and adjust the frame length of the next frame in real time according to the comparison result. At the same time, introduce prediction confidence evaluation and set the confidence threshold h. like , then adjust the machine learning prediction model parameters, where y represents the prediction result of the machine learning prediction model for the next frame.

5. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 4, characterized in that: The specific formula of the path planning strategy is: in, represents the minimum cost from state s to the target state at time t, t represents the current time, represents the instantaneous cost function at time t, represents the weight of executing decision a, represents the time required to execute decision a to transfer from state s to the next state, It represents the time delay caused by the uncertainty factors encountered when executing decision a from state s to the next state. represents the distance penalty coefficient, Represents the transition from state s to the next state The cumulative distance of Indicates that from the next state Departure, the minimum cost after taking into account the transfer time and additional delay.

6. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 5, characterized in that: The intelligent prediction module includes a scheduling unit, a demand prediction unit and a risk assessment unit; the scheduling unit is used to intelligently schedule the transportation and recycling of packaging boxes based on historical data and real-time information; the demand prediction unit predicts logistics demand through a deep learning model and prepares resources in advance; the risk assessment unit is configured with an improved random forest algorithm to assess the risks that may be encountered in the logistics process.

7. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 6, characterized in that: The specific steps of the improved random forest algorithm include: B1: Collect historical risk data of the logistics process, preprocess and extract features of the historical risk data, and obtain historical feature data; B2: Select random forest and gradient boosting tree as base models and set the initial parameters of the base models; B3: Using the stacking integration strategy, random forest and gradient boosting tree are used as the first-layer model, and the logistic regression model is used as the second-layer meta-model. The output of the first-layer model is used as the feature input of the second-layer meta-model. At the same time, a nonlinear transformation layer is introduced into each tree node in the base model to generate a risk assessment model. B4: Set risk assessment thresholds and online learning mechanisms to enable the risk assessment model to receive new risk data in real time for risk assessment, and automatically update risk assessment model parameters based on the assessment results. At the same time, dynamically adjust the risk assessment threshold based on historical assessment results and actual risk occurrences; B5: Regularly evaluate the performance indicators of risk assessment models.

8. The data transmission system for tracing and locating plastic film recycling packaging boxes according to claim 7, characterized in that: The risk data in B1 include weather conditions, traffic conditions, cargo types, transportation routes, historical accident records, real-time traffic camera image data and dynamic data of transportation vehicles; the dynamic data of transportation vehicles include the position, speed and acceleration of transportation vehicles.

9. A data transmission system for tracing and locating plastic film recycling packaging boxes as claimed in claim 8, characterized in that: The cycle management module includes: a recycling unit, a cleaning unit, a detection unit, and a maintenance prediction unit; The recycling unit is responsible for recycling the packaging boxes; The cleaning unit is used to clean and disinfect the recycled packaging boxes; The detection unit is used to detect the cleaned packaging box; The maintenance prediction unit predicts the maintenance requirements of the packaging box based on the data analysis results and arranges a maintenance or replacement plan in advance.

Images