An RFID-based end-to-end tracking management system for container logistics

Abstract

This invention discloses an RFID-based container logistics end-to-end tracking and management system, belonging to the field of logistics management technology. It collects RFID signals under different environments through a signal acquisition module and assesses the degree of environmental interference. Combined with a data transmission monitoring module recording and analyzing latency scenarios, and a throughput analysis module evaluating transmission efficiency under high-load scenarios, it comprehensively understands data transmission performance. A real-time assessment module further manages logistics data according to real-time classification, while a dynamic management module, targeting medium-real-time data, dynamically adjusts the data upload frequency by predicting real-time changes within a fixed future time period. This achieves efficient classification management and resource optimization of container logistics data, effectively improving the transparency, reliability, and decision-making efficiency of the logistics management system, especially in emergency scenarios.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, specifically to an RFID-based container logistics end-to-end tracking management system. Background Technology

[0002] RFID is a technology that transmits and identifies data via radio frequency signals, featuring contactless operation, rapid scanning, and batch reading capabilities. In container logistics, each container can be equipped with a unique RFID tag storing information about the container (such as its serial number, cargo category, and destination). RFID readers are deployed at various nodes in the logistics chain (such as ports, warehouses, and transport vehicles). As containers pass through these nodes, the system automatically scans and records their location, status, and time information. Through integration with a back-end management system, this tracking method can update the dynamic information of containers in real time, achieving transparent management throughout the entire process from shipment to destination. RFID-based tracking management offers advantages such as real-time performance, high efficiency, and accurate data, significantly improving logistics operational efficiency and reducing manual intervention costs. In practical applications, this technology can be widely used in international trade, port operations, and logistics companies, helping them optimize transportation routes, monitor container status (such as temperature and vibration), and prevent cargo loss or damage.

[0003] The existing technology has the following shortcomings:

[0004] In the global logistics chain, the large amount of real-time data generated by RFID systems needs to be uploaded to a centralized back-end management platform. However, due to network conditions, server load, or compatibility issues between different systems, data upload and synchronization may be delayed, especially in maritime or long-distance logistics scenarios. This delay can prevent container logistics management systems from reflecting container status in real time, affecting the efficiency and accuracy of logistics management decisions, especially in emergency scenarios (such as when goods need to be quickly rescheduled). Furthermore, delays in data upload and synchronization in container logistics management systems can directly lead to the inability to update container location information in real time, resulting in the system displaying outdated or incorrect statuses. In such cases, all parties in the logistics chain (such as cargo owners, transportation companies, and ports) cannot accurately grasp the dynamic location of goods, losing transparency regarding transportation progress and container status, potentially triggering a crisis of trust. Summary of the Invention

[0005] The purpose of this invention is to provide an RFID-based container logistics end-to-end tracking and management system to address the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an RFID-based container logistics end-to-end tracking and management system, comprising a signal acquisition module, a signal interference analysis module, a data transmission monitoring module, a throughput analysis module, a real-time evaluation module, and a dynamic management module;

[0007] Signal acquisition module: Test points under different environments are selected at several logistics nodes. The test points include open areas, dense metal areas and high humidity areas. RFID reading and writing devices are used to collect signals from tags at different locations multiple times.

[0008] Signal interference analysis module: Measure the signal strength of the collected RFID tags, compare the signal strength and reading success rate under different environments, determine the correlation between signal strength and reading delay, and evaluate the degree of interference of environmental factors on the signal strength of RFID reading and writing devices;

[0009] Data transmission monitoring module: If the interference level is high, different types of monitoring tools will be deployed at each logistics node to record the entire process of data from reading to uploading. Comparative tests will be conducted in high-latency and low-latency scenarios to measure the transmission delay between each node.

[0010] Throughput Analysis Module: Constructs high-load scenarios and analyzes the fluctuation of the data transmission throughput between each node in the actual transmission process by increasing simulated data traffic, and evaluates the degree of abnormality in data transmission efficiency from RFID reading and writing devices to the back-end management platform.

[0011] Real-time performance evaluation module: Based on the degree of interference of environmental factors on the signal strength of RFID readers and the degree of abnormality in data transmission efficiency from RFID readers to the back-end management platform, evaluate the real-time performance of container logistics data upload and synchronization. Based on the evaluation results, the corresponding container logistics data is divided into high real-time data, medium real-time data and low real-time data, and managed accordingly.

[0012] Dynamic Management Module: For medium-real-time data, it further predicts the real-time changes in the uploading and synchronization of container logistics data within a fixed future time period, and dynamically adjusts the data uploading frequency based on the prediction results.

[0013] Preferably, in the signal interference analysis module, a delay correlation index is generated after analyzing the correlation between signal strength and reading delay. The method for obtaining the delay correlation index is as follows:

[0014] The signal strength X is divided into multiple intervals based on its range, and the read delay Y is divided into multiple intervals based on its time. Each sample data is mapped to the corresponding interval according to its value, forming a discretized dataset. For all sample pairs (xi, yi), the frequency of each interval combination is counted, and then normalized to obtain the probability, as shown in the formula: ;in, Let P(x,y) represent the number of samples where X=x and Y=y. P(x) and P(y) are obtained by summing P(x) with respect to y or x, respectively, and their expressions are as follows: Summing all interval combinations (x, y), the delayed correlation index is calculated as follows: In the formula, This is a delayed correlation index.

[0015] Preferably, in the throughput analysis module, after analyzing the fluctuation of the data transmission throughput ratio between each node during the actual transmission process, a throughput ratio fluctuation index is generated. The method for obtaining the throughput ratio fluctuation index is as follows:

[0016] Collect the actual throughput utilization rate under high load scenarios Define the throughput utilization curve under ideal conditions. The distance function d(q,c) is used to calculate the difference between the actual value q and the ideal value c, using the Euclidean distance formula: Construct an n×m cumulative distance matrix D to represent the alignment cost of sequences Q and C, where n is the length of the actual sequence Q and m is the length of the ideal sequence C. Recursively calculate D(i,j): D(i,j)=d(qi,cj)+min{D(i−1,j),D(i,j−1),D(i−1,j−1)}; where D(i−1,j) represents the cost of aligning the current point with the previous point in the actual sequence, D(i,j−1) represents the cost of aligning the current point with the previous point in the ideal sequence, and D(i−1,j−1) represents the cost of aligning the current point with the previous point in both sequences simultaneously. The initial condition is: D(0,0)=d(q0,c0); the boundary condition is that the cumulative value of the first row and the first column is the one-way sum on the path. Starting from the lower right corner D(n,m) of the matrix, backtrack along the path with the minimum cumulative cost to find the optimal alignment path w(i); accumulate the cost on the path. The expression is: The throughput utilization rate fluctuation index is obtained by normalizing the calculated cumulative path cost.

[0017] Preferably, in the real-time evaluation module, the real-time performance of container logistics data upload and synchronization is evaluated based on the degree of interference of environmental factors on the signal strength of the RFID reader and the degree of abnormality in the data transmission efficiency from the RFID reader to the back-end management platform.

[0018] The latency correlation index and throughput occupancy ratio fluctuation index are converted into a comprehensive feature vector. This comprehensive feature vector is used as the input to a machine learning model. The machine learning model uses the prediction of the real-time value label of container logistics data upload and synchronization for each set of comprehensive feature vectors as the prediction objective. The training objective is to minimize the sum of prediction errors for all real-time value labels of container logistics data upload and synchronization. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The real-time value of container logistics data upload and synchronization is determined based on the model output. The machine learning model is a multinomial regression model.

[0019] Preferably, the real-time value of the acquired container logistics data upload and synchronization is compared with a gradient standard threshold, which includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold. The real-time value of the container logistics data upload and synchronization is compared with the first standard threshold and the second standard threshold respectively.

[0020] If the real-time value of container logistics data upload and synchronization is greater than the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is high. At this time, a high real-time signal is generated, and the corresponding container logistics data is classified as high real-time data.

[0021] If the real-time performance of container logistics data upload and synchronization is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is in the middle. At this time, a medium real-time signal is generated, and the corresponding container logistics data is divided into medium real-time data.

[0022] If the real-time performance of container logistics data upload and synchronization is less than the first standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is low. At this time, a low real-time performance signal is generated, and the corresponding container logistics data is classified as low real-time data.

[0023] Preferably, in the dynamic management module, for medium-real-time data, the real-time changes in the uploading and synchronization of container logistics data within a fixed future time period are further predicted, and the data uploading frequency is dynamically adjusted based on the prediction results. Specifically:

[0024] For real-time data, which refers to container logistics data uploaded and synchronized within a fixed time period whose real-time value is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, real-time values ​​generated within a future fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold will be collected, and a corresponding dataset will be established. The mean and standard deviation of the dataset will be calculated, and after analysis, the changes in the real-time performance of container logistics data uploaded and synchronized within a future fixed time period will be further predicted based on the analysis results.

[0025] Preferably, if the average real-time value in the dataset is greater than or equal to a reference threshold for the average real-time value, a level-three warning signal is generated, the upload frequency is reduced, the upload cycle is extended, and batch processing is adopted.

[0026] If the mean real-time value is greater than or equal to the reference threshold of the mean real-time value, and the standard deviation of the real-time value is greater than or equal to the reference threshold of the standard deviation of the real-time value, no warning signal will be generated, the upload frequency will be maintained, and monitoring will be strengthened.

[0027] If the average real-time value is less than the reference threshold for the average real-time value, and the standard deviation of the real-time value is greater than or equal to the reference threshold for the standard deviation of the real-time value, a first-level warning signal is generated, the upload frequency is increased, the timeliness of data upload is prioritized, and network conditions are optimized.

[0028] If the average real-time performance value is less than the reference threshold for the average real-time performance value, and the standard deviation of the real-time performance value is less than the reference threshold for the standard deviation of the real-time performance value, a level-two warning signal is generated, the upload frequency is reduced, and resources are concentrated to optimize the transmission link.

[0029] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0030] 1. This invention, through a multi-module collaborative design, significantly improves the real-time performance and stability of the container logistics management system during data upload and synchronization. By employing a signal acquisition module and a signal interference analysis module, it deeply analyzes the RFID signal strength and delay characteristics under different logistics environments, effectively identifying environmental factors affecting data transmission. Combined with a data transmission monitoring module and a throughput analysis module, it accurately measures and evaluates delay and throughput fluctuations during data transmission, providing a scientific basis for optimizing transmission performance. Simultaneously, through a real-time evaluation module, the system can achieve efficient classification and management of logistics data (high real-time, medium real-time, low real-time) and dynamically adjust resource allocation strategies, thereby improving the system's responsiveness to high-priority tasks.

[0031] 2. This invention significantly reduces data upload latency, improves collaboration efficiency between logistics nodes, and provides real-time, accurate cargo dynamic information to all parties in the logistics chain (such as cargo owners, transportation companies, and ports). It enhances the transparency of the logistics system and reduces trust risks and management decision-making errors caused by information lag. Furthermore, through the prediction and frequency adjustment of real-time data via the dynamic management module, it optimizes system resource utilization and balances network load and transmission performance, demonstrating excellent performance, particularly in emergency and high-load scenarios. The application of this invention not only achieves real-time tracking and management of the entire container logistics process but also provides reliable support for the digital and intelligent development of the logistics chain. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0033] Figure 1 This is a system module diagram of the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] For examples, please refer to Figure 1 As shown in this embodiment, a container logistics end-to-end tracking and management system based on RFID includes a signal acquisition module, a signal interference analysis module, a data transmission monitoring module, a throughput analysis module, a real-time evaluation module, and a dynamic management module.

[0036] Signal acquisition module: Test points under different environments are selected at several logistics nodes. The test points include open areas, dense metal areas and high humidity areas. RFID reading and writing devices are used to collect signals from tags at different locations multiple times.

[0037] Signal interference analysis module: Measure the signal strength of the collected RFID tags, compare the signal strength and reading success rate under different environments, determine the correlation between signal strength and reading delay, and evaluate the degree of interference of environmental factors on the signal strength of RFID reading and writing devices;

[0038] Data transmission monitoring module: If the interference level is high, different types of monitoring tools will be deployed at each logistics node to record the entire process of data from reading to uploading. Comparative tests will be conducted in high-latency and low-latency scenarios to measure the transmission delay between each node.

[0039] Throughput Analysis Module: Constructs high-load scenarios and analyzes the fluctuation of the data transmission throughput between each node in the actual transmission process by increasing simulated data traffic, and evaluates the degree of abnormality in data transmission efficiency from RFID reading and writing devices to the back-end management platform.

[0040] Real-time performance evaluation module: Based on the degree of interference of environmental factors on the signal strength of RFID readers and the degree of abnormality in data transmission efficiency from RFID readers to the back-end management platform, evaluate the real-time performance of container logistics data upload and synchronization. Based on the evaluation results, the corresponding container logistics data is divided into high real-time data, medium real-time data and low real-time data, and managed accordingly.

[0041] Dynamic Management Module: For medium-real-time data, it further predicts the real-time changes in the uploading and synchronization of container logistics data within a fixed future time period, and dynamically adjusts the data uploading frequency based on the prediction results.

[0042] In the signal acquisition module, test points are set up in representative logistics environments to cover typical RFID application scenarios. Open areas: Few obstacles, good signal propagation conditions. Test locations: Open areas of port storage yards and logistics parks. Signal strength and read success rate are tested under optimal conditions as a performance benchmark. Dense metal areas: Dense metal objects may cause signal reflection and shielding. Test locations: Warehouses storing large quantities of metal goods and transport vehicles carrying metal products. The impact of metal interference on RFID signals is evaluated. High humidity areas: Higher water content may absorb or attenuate certain frequency bands of radio frequency signals. Test locations: Cold storage, coastal logistics scenarios, and rainy operating environments. The impact of humidity on signal propagation and tag reading is tested.

[0043] Deploy RFID readers at each test point to ensure coverage matches actual logistics operations. Select UHF RFID readers suitable for logistics scenarios, typically in the 860-960MHz band. Tags should support multi-environment applications (e.g., anti-metal tags, waterproof tags). Install readers according to typical usage methods (e.g., fixed installation at entrances, or handheld for scanning). Ensure clear line-of-sight between the reader and the tag (avoid obstructions) and record the distance between the tag and the reader. Adjust the reader's transmit power (e.g., 0.1W to 2W) and antenna angle to optimize signal reception range. Configure the tag's data storage fields, including unique ID, cargo type, etc.

[0044] Perform multiple signal acquisitions on tags at different locations, recording signal strength and read results. Deploy several tags at the test points, ensuring the tags are in easily readable locations (directly facing the antenna or close to the reader), or partially obscured or mounted on irregular surfaces (such as metal). Scan each tag at least 20 times using the reader, recording data including: Signal Strength Indicator (RSSI): the power of the signal returned by the tag; Read Success Rate: the percentage of successful reads out of total attempts; Read Latency: the time taken from initiating a read to receiving tag data. Record environmental parameters (such as metal object density, humidity, and temperature) at each test point and correlate them with the signal data for analysis. Store the acquisition results locally and upload them to a central management platform for centralized analysis.

[0045] Signal interference analysis module: Measure the signal strength of the collected RFID tags, compare the signal strength and reading success rate under different environments, determine the correlation between signal strength and reading delay, and evaluate the degree of interference of environmental factors on the signal strength of RFID reading and writing devices.

[0046] At each test point (e.g., open area, dense metal area, high humidity area), use an RFID reader to read the designated tag multiple times. Record the received signal strength (RSSI, Received Signal Strength Indicator) returned for each read. RSSI is usually expressed as a negative value in decibels per milliwatt (dBm); the closer the value is to 0, the stronger the signal. Calculate the read success rate: At each test point, read the same tag multiple times and record the ratio of successful reads to the total number of attempts. For example: Analyze the relationship between signal strength and read success rate: Compare signal strength and read success rate under different environments, and observe whether the following trends exist: High signal strength (e.g., -30dBm to -50dBm): Usually corresponds to a higher read success rate. Low signal strength (e.g., below -70dBm): May lead to read failures or a significant decrease in the success rate.

[0047] Measure the time (in milliseconds, ms) from initiating a read request to successfully reading data. Record the latency for each read at each test point and take the average as the representative value for that point. Compare the trend between signal strength and read latency: High signal strength: Generally lower latency, rapid reading. High signal strength may lead to multiple read attempts or communication errors, increasing latency. Determine whether a decrease in signal strength directly causes an increase in read latency, thus revealing the specific impact of the environment on read performance.

[0048] After analyzing the correlation between signal strength and reading delay, a delay correlation index is generated to assess the degree of interference from environmental factors on the signal strength of RFID readers. The method for obtaining the delay correlation index is as follows:

[0049] The signal strength X is divided into multiple intervals based on its range (e.g., -90dBm to -30dBm divided into 10 intervals). The read latency Y is divided into multiple intervals based on time (e.g., 0ms to 200ms divided into 10 intervals). Each sample data is mapped to the corresponding interval according to its value, forming a discretized dataset.

[0050] For all sample pairs (xi, yi), count the frequency of each interval combination, then normalize to obtain the probability, as shown in the formula: ;in, Let P(x,y) represent the number of samples where X=x and Y=y. P(x) and P(y) are obtained by summing P(x) with respect to y or x, respectively, and their expressions are as follows: Summing over all possible interval combinations (x, y), the delayed correlation index is calculated as follows: In the formula, This is a delayed correlation index.

[0051] A high latency correlation index indicates a strong dependence between RFID signal strength and read latency, meaning that fluctuations in signal strength significantly impact read latency. This typically implies that environmental factors (such as metallic interference, high humidity, and obstructions) significantly interfere with signal propagation, leading to reduced or unstable signal strength and consequently, a substantial increase in read latency. For instance, in areas with dense metal concentrations or high humidity environments, signal strength attenuation is significant, requiring the system to retries multiple times to successfully read data, thus increasing latency.

[0052] A low latency correlation index indicates a weak relationship between signal strength and read latency, meaning the read latency is not significantly affected by fluctuations in signal strength. This may suggest that the signal propagation of the RFID reader / writer is relatively stable in the current environment, with low interference from environmental factors. For example, in open areas or unobstructed logistics nodes, signal strength remains high, data can be read quickly, and latency remains relatively stable and is not significantly affected by signal strength.

[0053] Data transmission monitoring module: If the interference level is high, different types of monitoring tools will be deployed at each logistics node to record the entire process of data from reading to uploading. Comparative tests will be conducted in high-latency and low-latency scenarios to measure the transmission delay between each node.

[0054] Focus on monitoring key logistics nodes with severe signal interference (such as areas with dense metal deposits, high humidity areas, or complex transportation routes). Simultaneously, select nodes with relatively stable signals (such as open areas) as comparison samples. Deploy appropriate monitoring equipment to cover the entire data flow from RFID readers to the back-end management platform: Reader performance monitor: records tag read success rate, signal strength, and read latency. Edge computing device: monitors local cache and data processing time. Network traffic analysis tool: monitors transmission latency and throughput during data upload. Tool deployment methods: Fixed monitoring: installed in fixed locations at ports or warehouses for long-term monitoring of node performance. Mobile monitoring: equipped with handheld RFID devices or vehicle-mounted edge computing devices for monitoring dynamic data during transportation.

[0055] Record the complete data transmission process and categorize and store test results under high-latency and low-latency scenarios. When reading data from RFID tags, record parameters such as signal strength, read success rate, and read latency. Record the data upload time from the edge device to the backend management platform, including network transmission latency and processing time. Scenarios with strong signal interference, unstable network, and latency exceeding predetermined thresholds are included. Scenarios with stable signal, smooth network, and latency within the normal range are also included.

[0056] Test data transmission performance under scenarios with varying levels of interference and analyze key factors affecting latency. Design comparative tests: Run data transmission tests in both high-latency scenarios (e.g., densely metal areas) and low-latency scenarios (e.g., open areas). Test the latency distribution of each node in the transmission path, including: read latency from tag to reader; upload latency from edge device to backend. Adjust environmental variables (e.g., add interference sources, change signal frequency) to simulate different levels of interference. Test changes in network status (e.g., compare high-traffic congestion with idle periods). Compare data under high-latency and low-latency scenarios: Is signal strength significantly reduced? Is data upload latency significantly increased? Are there any abnormal retransmissions or packet loss in high-latency scenarios?

[0057] Generate a latency analysis report to identify latency bottlenecks and provide optimization directions. Calculate total latency: Total latency = Read latency + Data upload latency. Calculate the mean and standard deviation for all test samples to assess latency distribution. Based on the latency data, analyze whether the problem is mainly concentrated in the read or upload phase: If read latency is high, optimize the signal transmission performance of the reader and tags. If upload latency is high, optimize network bandwidth or edge computing performance. Provide a latency distribution chart (comparison of high and low latency scenarios). Assess the impact of environmental factors on latency (e.g., the attenuation rate of signal due to metallic interference or humidity). Propose optimization suggestions (e.g., improve equipment layout, introduce anti-interference tags).

[0058] Throughput Analysis Module: Constructs high-load scenarios and analyzes the fluctuation of the data transmission throughput between each node in the actual transmission process by increasing simulated data traffic, and evaluates the degree of abnormality in data transmission efficiency from RFID reading and writing devices to the back-end management platform.

[0059] Simulate scenarios of surging data traffic in the logistics chain to test the maximum throughput and stability of the transmission system. Simulate peak periods (such as concentrated loading and unloading of goods at ports or peak inbound and outbound times at warehouses). Set up multiple RFID devices to simultaneously read and upload data, increasing the data volume. Use data generation tools or scripts to simulate large-scale RFID tag readings. Each data entry includes tag ID, location, timestamp, and status information. Set the data frequency to 2 times, 5 times, or even higher than normal traffic. Gradually increase the data traffic in the simulated environment and observe changes in system throughput. Test traffic performance under different network conditions (such as bandwidth limitations and network congestion).

[0060] Measure the data throughput between nodes and analyze its performance under high load conditions. Define throughput: Throughput: The amount of data successfully transmitted per unit time (unit: Mbps or pieces / second). Formula: Throughput = Total amount of data successfully transmitted / Time; record the amount of data successfully read per second by the RFID device. Measure the data transmission speed from the reading device to the edge device. Measure the upload throughput from the edge device to the backend. Monitor the data throughput utilization rate under high load conditions, recording the instantaneous peak, average, and minimum throughput values. Capture throughput fluctuations and calculate the fluctuation amplitude.

[0061] The throughput utilization ratio of each node is defined as the ratio of actual throughput to maximum theoretical throughput: Utilization ratio = Actual throughput / Maximum throughput; compare the throughput utilization ratio under high load and normal load conditions. Calculate the throughput fluctuation range: Fluctuation range = (Peak throughput - Minimum throughput) / Average throughput; when the utilization ratio is close to 100% or the fluctuation range is too high, it may mean that there is a transmission bottleneck or anomaly in the node.

[0062] After analyzing the fluctuation of the data transmission throughput ratio between each node during the actual transmission process, a throughput ratio fluctuation index is generated to assess the degree of abnormality in data transmission efficiency from the RFID reader / writer to the back-end management platform. The method for obtaining the throughput ratio fluctuation index is as follows:

[0063] Actual throughput utilization time series Q: Collects the actual throughput utilization rate under high load scenarios. Recorded at fixed time intervals (e.g., per second or per minute). Ideal throughput utilization time series C: Defines the throughput utilization curve under ideal conditions. It is usually a stable constant value or an expected trend curve.

[0064] The distance function d(q,c) is used to calculate the difference between the actual value q and the ideal value c, using the Euclidean distance formula: Construct an n×m cumulative distance matrix D to represent the alignment cost of sequences Q and C, where n is the length of the actual sequence Q and m is the length of the ideal sequence C. Recursively calculate D(i,j): D(i,j)=d(qi,cj)+min{D(i−1,j),D(i,j−1),D(i−1,j−1)}; D(i−1,j) represents the cost of aligning the current point with the previous point in the actual sequence, D(i,j−1) represents the cost of aligning the current point with the previous point in the ideal sequence, and D(i−1,j−1) represents the cost of aligning the current point with the previous point in both sequences simultaneously. The initial condition is: D(0,0)=d(q0,c0); the boundary condition is that the cumulative value of the first row and the first column is the one-way sum on the path. Starting from the lower right corner D(n,m) of the matrix, backtrack along the path with the minimum cumulative cost to find the optimal alignment path w(i); accumulate the cost on the path. The expression is: The throughput utilization rate fluctuation index is obtained by normalizing the calculated cumulative path cost.

[0065] A high throughput fluctuation index indicates a significant difference between the actual throughput and the ideal state, suggesting obvious fluctuations or instability during transmission. This typically reflects an abnormally high level of data transmission efficiency from RFID readers to the back-end management platform, potentially caused by network congestion, insufficient node device performance, or improper data traffic management. A high fluctuation index means that the system's transmission efficiency is significantly affected, potentially leading to data delays, packet loss, or insufficient throughput, severely impacting the real-time performance and reliability of the logistics system.

[0066] When the throughput utilization rate fluctuation index is low, it indicates that the actual throughput is close to the ideal state, the data transmission process is relatively stable, and the fluctuation range is small. In this case, the data transmission efficiency from the RFID reader to the back-end management platform is not abnormally high, and the transmission link can maintain good performance under high load scenarios. The low fluctuation index reflects that the system has strong anti-interference capabilities and resource allocation capabilities, and can effectively meet the large-scale data transmission needs between logistics nodes.

[0067] Real-time performance evaluation module: Based on the degree of interference of environmental factors on the signal strength of RFID readers and the degree of abnormality in data transmission efficiency from RFID readers to the back-end management platform, evaluate the real-time performance of container logistics data upload and synchronization. Based on the evaluation results, the corresponding container logistics data is divided into high real-time data, medium real-time data and low real-time data, and managed accordingly.

[0068] The latency correlation index and throughput occupancy ratio fluctuation index are converted into a comprehensive feature vector. This comprehensive feature vector is used as the input to a machine learning model. The machine learning model uses the prediction of the real-time value label of container logistics data upload and synchronization for each set of comprehensive feature vectors as the prediction objective. The training objective is to minimize the sum of prediction errors for all real-time value labels of container logistics data upload and synchronization. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The real-time value of container logistics data upload and synchronization is determined based on the model output. The machine learning model is a multinomial regression model.

[0069] The method for obtaining the real-time value of container logistics data upload and synchronization is as follows: Obtain the corresponding function expression from the training data of the comprehensive feature vector of the trained machine learning model. In the formula, It is the output function of the model. GHM is the latency correlation index, and GHM is the throughput utilization rate fluctuation index. This refers to the real-time value for uploading and synchronizing container logistics data.

[0070] The real-time value of the uploaded and synchronized container logistics data is compared with the gradient standard threshold, which includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold. The real-time value of the uploaded and synchronized container logistics data is compared with the first standard threshold and the second standard threshold respectively.

[0071] If the real-time value of container logistics data upload and synchronization is greater than the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is high. At this time, a high real-time signal is generated, and the corresponding container logistics data is classified as high real-time data.

[0072] If the real-time performance of container logistics data upload and synchronization is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is in the middle. At this time, a medium real-time signal is generated, and the corresponding container logistics data is divided into medium real-time data.

[0073] If the real-time performance of container logistics data upload and synchronization is less than the first standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is low. At this time, a low real-time performance signal is generated, and the corresponding container logistics data is classified as low real-time data.

[0074] Highly real-time data is typically associated with mission-critical or emergency situations, requiring immediate upload and processing. Delays in this type of data can prevent systems from responding promptly, thus impacting critical aspects of logistics operations.

[0075] High-real-time data is processed through a priority queue and immediately uploaded to the backend management platform. Low-latency transmission protocols (such as MQTT and HTTPS) or dedicated transmission channels are used to ensure timely transmission. When high-real-time data is detected (e.g., device malfunction, path deviation, excessive latency), an alarm signal is triggered, notifying relevant personnel or the system. Alarms are sent to users through multiple channels (SMS, email, app push). Edge computing devices or network resources are prioritized for high-real-time data to ensure that it is not consumed by low-priority tasks in terms of bandwidth or computing resources. Low-real-time tasks are paused when necessary to ensure the integrity and timeliness of high-real-time data.

[0076] Mid-real-time data has relatively moderate timeliness requirements and is mostly used for periodic uploading and analysis. Its processing can be slightly delayed, but it needs to be completed within a specified time frame to support logistics planning and resource allocation. Mid-real-time data is uploaded to the backend management platform in batches. Upload tasks are triggered periodically at set time intervals (e.g., every 10 minutes or hour) to avoid continuous occupation of network resources. The caching function of edge devices is used to temporarily store mid-real-time data, uploading it when network conditions stabilize or system load decreases. This prevents mid-real-time data transmission failures or the occupation of high-real-time channels during network congestion. Before uploading, mid-real-time data undergoes deduplication, cleaning, and aggregation to reduce redundant data and improve transmission efficiency. For example, continuous temperature data from multiple sensors can be uploaded using average or trend data instead of raw data.

[0077] Low real-time data is typically auxiliary or historical records of the logistics process and has no direct impact on the system's immediate response. The uploading and processing of this type of data can be significantly delayed, making it suitable for off-peak hours or background tasks.

[0078] Upload low-real-time data during off-peak hours (such as at night or when network traffic is low) to avoid consuming system resources. Use background tasks for batch processing to reduce the burden on the main system. Store low-real-time data on edge devices or local servers and periodically archive and upload it to the cloud. Long-term upload intervals (such as daily or weekly) can be set. Highly compress low-real-time data to retain core information and reduce upload bandwidth requirements. For example, simplify hourly records of environmental data into daily summaries. Use it for historical trend analysis, performance evaluation, and predictive maintenance, which do not require real-time processing. Combine it with big data analytics tools to mine long-term accumulated data to support future decision optimization.

[0079] Dynamic Management Module: For medium-real-time data, it further predicts the real-time changes in the uploading and synchronization of container logistics data within a fixed future time period, and dynamically adjusts the data uploading frequency based on the prediction results.

[0080] For real-time data, which refers to container logistics data uploaded and synchronized within a fixed time period whose real-time value is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, real-time values ​​generated within a future fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold will be collected, and a corresponding dataset will be established. The mean and standard deviation of the dataset will be calculated, and after analysis, the changes in the real-time performance of container logistics data uploaded and synchronized within a future fixed time period will be further predicted based on the analysis results.

[0081] If the mean real-time performance value within the dataset is greater than or equal to a reference threshold for the mean real-time performance value, and the standard deviation of the real-time performance value is less than a reference threshold for the standard deviation of the real-time performance value, it indicates high real-time performance, suggesting relatively stable data upload and synchronization performance. A small standard deviation indicates low fluctuation in real-time performance values ​​and strong consistency in data characteristics. In this case, a level-three warning signal is generated, appropriately reducing the upload frequency to alleviate system load while ensuring effective data upload. The upload cycle can be extended, and batch upload methods can be used to process data.

[0082] If the mean real-time performance value is greater than or equal to the reference threshold for the mean real-time performance value, and the standard deviation of the real-time performance value is greater than or equal to the reference threshold for the standard deviation of the real-time performance value, it indicates that the real-time performance value is high and the overall data upload performance is good, but there are large fluctuations and instability. A high standard deviation indicates that the data may be inconsistent in some time periods, and the cause of the fluctuation needs to be investigated. In this case, no warning signal is generated, the current upload frequency is maintained, and the existing upload frequency is maintained to avoid delays or data loss caused by data fluctuations. Enhanced monitoring is implemented to analyze the sources of fluctuations (such as network congestion, insufficient node performance) and optimization is carried out.

[0083] If the mean real-time performance value is less than the reference threshold for the mean real-time performance value, and the standard deviation of the real-time performance value is greater than or equal to the reference threshold for the standard deviation of the real-time performance value, the real-time performance value is low, indicating a significant problem with data upload performance. A high standard deviation and large fluctuations may be affected by environmental factors or system load. In this case, a level one warning signal is generated to increase the upload frequency and optimize network conditions: increase the upload frequency to prioritize ensuring the timeliness of data uploads. Optimization measures are taken for the sources of fluctuation (such as network congestion or insufficient hardware performance) to improve data transmission efficiency.

[0084] If the average real-time performance value is less than the reference threshold for the average real-time performance value, and the standard deviation of the real-time performance value is less than the reference threshold for the standard deviation of the real-time performance value, it indicates low real-time performance and poor data upload performance, but with small fluctuations. The problem manifests as overall poor upload performance rather than random anomalies. A low standard deviation indicates that the problem is persistent. At this point, a secondary warning signal is generated to further reduce the upload frequency and focus on optimizing transmission performance: while reducing the upload frequency, the transmission link is optimized (e.g., increasing bandwidth, improving hardware performance). This concentrates resources to solve long-term data transmission performance problems.

[0085] It should be noted that Level 1 warning signals are more important than Level 2 warning signals, and Level 2 warning signals are more important than Level 3 warning signals. Relevant personnel can take corresponding management measures according to the different warning signal levels.

[0086] In this embodiment, firstly, the signal acquisition module collects RFID signals at logistics nodes in different environments (such as open areas, densely packed metal areas, and high-humidity areas) and measures the signal strength and read success rate. Next, the signal interference analysis module evaluates the correlation between signal strength and read latency, as well as the degree of environmental interference. If the signal interference is high, the data transmission monitoring module deploys monitoring tools to record the entire process from data reading to uploading, conducting transmission latency tests for both high-latency and low-latency scenarios. In the throughput analysis module, the fluctuation of throughput occupancy during data transmission is analyzed by simulating high-load scenarios to assess the degree of RFID data transmission efficiency anomalies. The real-time performance evaluation module integrates the above results and classifies the data's real-time performance into high-real-time, medium-real-time, and low-real-time data. Finally, the dynamic management module dynamically adjusts the data upload frequency for medium-real-time data by predicting real-time changes within a fixed time period, ensuring an optimal balance between system resource utilization and transmission efficiency.

[0087] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0088] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0089] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An RFID-based container logistics end-to-end tracking management system, characterized in that: It includes a signal acquisition module, a signal interference analysis module, a data transmission monitoring module, a throughput analysis module, a real-time evaluation module, and a dynamic management module; Signal acquisition module: Test points under different environments are selected at several logistics nodes. The test points include open areas, dense metal areas and high humidity areas. RFID reading and writing devices are used to collect signals from tags at different locations multiple times. Signal interference analysis module: Measure the signal strength of the collected RFID tags, compare the signal strength and reading success rate under different environments, determine the correlation between signal strength and reading delay, and evaluate the degree of interference of environmental factors on the signal strength of RFID reading and writing devices; Data transmission monitoring module: If the interference level is high, different types of monitoring tools will be deployed at each logistics node to record the entire process of data from reading to uploading. Comparative tests will be conducted in high-latency and low-latency scenarios to measure the transmission delay between each node. Throughput Analysis Module: Constructs high-load scenarios and analyzes the fluctuation of the data transmission throughput between each node in the actual transmission process by increasing simulated data traffic, and evaluates the degree of abnormality in data transmission efficiency from RFID reading and writing devices to the back-end management platform. Real-time performance evaluation module: Based on the degree of interference of environmental factors on the signal strength of RFID readers and the degree of abnormality in data transmission efficiency from RFID readers to the back-end management platform, evaluate the real-time performance of container logistics data upload and synchronization. Based on the evaluation results, the corresponding container logistics data is divided into high real-time data, medium real-time data and low real-time data, and managed accordingly. Dynamic Management Module: For medium-real-time data, it further predicts the real-time changes in the uploading and synchronization of container logistics data within a fixed future time period, and dynamically adjusts the data uploading frequency based on the prediction results.

2. The RFID-based container logistics end-to-end tracking management system according to claim 1, characterized in that: In the signal interference analysis module, a delay correlation index is generated after analyzing the correlation between signal strength and reading delay. The method for obtaining the delay correlation index is as follows: The signal strength X is divided into multiple intervals based on its range, and the read delay Y is divided into multiple intervals based on its time. Each sample data is mapped to the corresponding interval according to its value, forming a discretized dataset. For all sample pairs (xi, yi), the frequency of each interval combination is counted, and then normalized to obtain the probability, as shown in the formula: ;in, Let P(x,y) represent the number of samples where X=x and Y=y. P(x) and P(y) are obtained by summing P(x) with respect to y or x, respectively, and their expressions are as follows: Summing all interval combinations (x, y), the delayed correlation index is calculated as follows: In the formula, This is a delayed correlation index.

3. The RFID-based container logistics end-to-end tracking management system according to claim 2, characterized in that: In the throughput analysis module, after analyzing the fluctuation of the data transmission throughput ratio between each node during the actual transmission process, a throughput ratio fluctuation index is generated. The method for obtaining the throughput ratio fluctuation index is as follows: Collect the actual throughput utilization rate under high load scenarios and construct the actual sequence. Define the throughput utilization ratio curve under ideal conditions and construct the ideal sequence. The distance function d(q,c) is used to calculate the difference between the actual value q and the ideal value c, using the Euclidean distance formula: ; Construct an n×m cumulative distance matrix D to represent the alignment cost of sequences Q and C, where n is the length of the actual sequence Q and m is the length of the ideal sequence C. Recursively calculate D(i,j): D(i,j)=d(qi,cj)+min{D(i−1,j),D(i,j−1),D(i−1,j−1)}; where D(i−1,j) represents the cost of aligning the current point with the previous point in the actual sequence, D(i,j−1) represents the cost of aligning the current point with the previous point in the ideal sequence, and D(i−1,j−1) represents the cost of aligning the current point with the previous point in both sequences simultaneously. The initial condition is: D(0,0)=d(q0,c0); the boundary condition is that the cumulative value of the first row and the first column is the one-way sum on the path. Starting from the lower right corner D(n,m) of the matrix, backtrack along the path with the minimum cumulative cost to find the optimal alignment path w. The cost of accumulating paths The expression is: The throughput utilization rate fluctuation index is obtained by normalizing the calculated cumulative path cost.

4. The RFID-based container logistics end-to-end tracking management system according to claim 3, characterized in that: In the real-time performance evaluation module, the real-time performance of container logistics data upload and synchronization is evaluated based on the degree of interference of environmental factors on the signal strength of RFID readers and the degree of abnormality in the data transmission efficiency from RFID readers to the back-end management platform. The latency correlation index and throughput occupancy ratio fluctuation index are converted into a comprehensive feature vector. This comprehensive feature vector is used as the input to a machine learning model. The machine learning model uses the prediction of the real-time value label of container logistics data upload and synchronization for each set of comprehensive feature vectors as the prediction objective. The training objective is to minimize the sum of prediction errors for all real-time value labels of container logistics data upload and synchronization. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The real-time value of container logistics data upload and synchronization is determined based on the model output. The machine learning model is a multinomial regression model.

5. The RFID-based container logistics end-to-end tracking management system according to claim 4, characterized in that: The real-time value of the uploaded and synchronized container logistics data is compared with the gradient standard threshold, which includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold. The real-time value of the uploaded and synchronized container logistics data is compared with the first standard threshold and the second standard threshold respectively. If the real-time value of container logistics data upload and synchronization is greater than the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is high. At this time, a high real-time signal is generated, and the corresponding container logistics data is classified as high real-time data. If the real-time performance of container logistics data upload and synchronization is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is in the middle. At this time, a medium real-time signal is generated, and the corresponding container logistics data is divided into medium real-time data. If the real-time performance of container logistics data upload and synchronization is less than the first standard threshold, it indicates that the real-time performance of container logistics data upload and synchronization is low. At this time, a low real-time performance signal is generated, and the corresponding container logistics data is classified as low real-time data.

6. The RFID-based container logistics end-to-end tracking management system according to claim 1, characterized in that: In the dynamic management module, for medium-real-time data, the real-time changes in container logistics data upload and synchronization within a fixed future time period are further predicted, and the data upload frequency is dynamically adjusted based on the prediction results. Specifically: For real-time data, which refers to container logistics data uploaded and synchronized within a fixed time period whose real-time value is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, real-time values ​​generated within a future fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold will be collected, and a corresponding dataset will be established. The mean and standard deviation of the dataset will be calculated, and after analysis, the changes in the real-time performance of container logistics data uploaded and synchronized within a future fixed time period will be further predicted based on the analysis results.

7. The RFID-based container logistics end-to-end tracking management system according to claim 6, characterized in that: If the average real-time value in the dataset is greater than or equal to the reference threshold for the average real-time value, a level three warning signal is generated, the upload frequency is reduced, the upload cycle is extended, and batch processing is adopted. If the mean real-time value is greater than or equal to the reference threshold of the mean real-time value, and the standard deviation of the real-time value is greater than or equal to the reference threshold of the standard deviation of the real-time value, no warning signal will be generated, the upload frequency will be maintained, and monitoring will be strengthened. If the average real-time value is less than the reference threshold for the average real-time value, and the standard deviation of the real-time value is greater than or equal to the reference threshold for the standard deviation of the real-time value, a first-level warning signal is generated, the upload frequency is increased, the timeliness of data upload is prioritized, and network conditions are optimized. If the average real-time performance value is less than the reference threshold for the average real-time performance value, and the standard deviation of the real-time performance value is less than the reference threshold for the standard deviation of the real-time performance value, a level-two warning signal is generated, the upload frequency is reduced, and resources are concentrated to optimize the transmission link.

Images