Remote dynamic monitoring system for import and export commodities based on the Internet of Things

Abstract

This application relates to the field of international trade, and discloses a remote dynamic monitoring system for import and export commodities based on the Internet of Things, including: a perception layer, which is a multimodal sensor network used to collect environmental parameters and location information of import and export commodities; an edge computing layer, which includes sensor nodes, local gateways and regional gateways, and performs data preprocessing, aggregation, feature extraction and distributed collaborative computing; an intelligent prediction and anomaly analysis layer, which predicts environmental parameter trends and detects anomalies through deep learning models; a communication and data transmission layer, which provides event-driven data transmission strategies and dynamic channel selection functions. By deploying a multi-level collaborative computing framework at the edge computing layer, the upload of redundant data is reduced. At the same time, the regional gateway uses a lightweight machine learning model to reduce the processing burden of the central server, and cooperates with the dynamic threshold adjustment mechanism to further improve the accuracy of data analysis and the adaptability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of international trade, and in particular to a remote dynamic monitoring system for imported and exported goods based on the Internet of Things. Background Art

[0002] In the cross-border logistics process, goods may undergo long-term, multi-stage transportation, involving multiple modes of transportation (such as sea, land, and air) and complex environmental conditions (such as temperature changes, vibration, and humidity control). Especially for the transportation of cold chain foods, pharmaceutical products, high-value goods, and dangerous goods, environmental monitoring, status tracking, and abnormal event response are particularly important. In such a scenario, the existing import and export commodity inspection system may not be able to cope with the challenge well.

[0003] First, the current logistics industry relies heavily on simple monitoring equipment, such as temperature and humidity recorders or GPS positioning devices, which usually only collect data in a single dimension and cannot fully reflect the environmental status of the goods during transportation. In addition, most of the collected data is extracted manually after the transportation is completed, which cannot achieve real-time dynamic monitoring and early warning. The existing monitoring system also lacks the ability to intelligently process multimodal data, resulting in insufficient monitoring accuracy and low data processing efficiency.

[0004] Secondly, traditional monitoring systems usually rely on fixed thresholds to trigger alarms, lack the ability to predict the changing trends of environmental parameters, and are difficult to warn of potential risks in a timely manner. For example, in cold chain transportation, if the temperature gradually approaches the dangerous range but does not exceed the threshold, the system may not trigger an alarm, which will eventually lead to damage to the goods. In addition, abnormal data usually requires manual analysis and processing, which is slow to respond and misses the best time for intervention.

[0005] Furthermore, in international logistics scenarios, transportation routes may cover areas with weak network signals such as oceans, deserts, and mountains, resulting in unstable or interrupted transmission of monitoring data. In addition, most existing systems use a fixed-frequency data upload mode, which continuously transmits data regardless of its value, wasting communication bandwidth and equipment power. For high-value or urgent data, the existing system lacks a priority transmission mechanism, resulting in untimely transmission of abnormal information.

[0006] Finally, logistics monitoring of imported and exported goods involves collaboration among multiple parties, including logistics companies, customs supervision departments, and customers. In traditional monitoring systems, data is usually managed independently by logistics companies, with serious data silos and lack of transparency, making it difficult to ensure the credibility and consistency of data. For example, when goods are damaged or abnormal events occur, different parties may have disputes over the source of data. In addition, the existing system lacks an efficient and secure multi-party data sharing platform, resulting in low collaboration efficiency. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention provides a remote dynamic monitoring system for imported and exported goods based on the Internet of Things, which solves the problems of the existing detection system, such as single monitoring data, low processing efficiency, inability to predict and respond to abnormal events in advance, waste of resources and poor stability in data transmission, and lack of transparency and credibility in multi-party collaboration.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a remote dynamic monitoring system for imported and exported goods based on the Internet of Things, comprising:

[0009] The perception layer is a multimodal sensor network used to collect environmental parameters and location information of imported and exported goods;

[0010] The edge computing layer, which includes sensor nodes, local gateways, and regional gateways, performs data preprocessing, aggregation, feature extraction, and distributed collaborative computing;

[0011] Intelligent prediction and anomaly analysis layer, which predicts environmental parameter trends and detects anomalies through deep learning models;

[0012] The communication and data transmission layer provides event-driven data transmission strategies and dynamic channel selection functions;

[0013] The data storage and collaboration layer stores monitoring data based on blockchain technology and manages data relationships through knowledge graphs;

[0014] The security layer uses quantum encryption and zero-knowledge proof technology to ensure the security and integrity of data.

[0015] Preferably, the perception layer adjusts the sampling interval of the multimodal sensor through a dynamic sampling rate control mechanism, and the dynamic sampling rate is determined by the sampling risk R risk and power consumption constraint E sensor Determine, the formula is as follows:

[0016] R t =R base +β1R risk -β2E sensor

[0017] Among them, R t is the dynamic sampling rate of the sensor at time t, R base is the basic sampling rate, β1, β2 are adjustment parameters.

[0018] Preferably, the edge computing layer uses multi-resolution wavelet transform to reduce noise on sensor data, and its signal decomposition formula is as follows:

[0019]

[0020] Among them, cj,k is the wavelet decomposition coefficient, representing the signal strength at different resolutions, X(t) is the original sensor signal at time t, ψ j,k (t) is the wavelet basis function.

[0021] Preferably, the edge gateway of the edge computing layer performs dimension reduction on the multimodal data through principal component analysis, and the data matrix after dimension reduction is expressed as:

[0022] X reduced =W·X

[0023] Among them, X reduced is the data matrix after dimensionality reduction, retaining the main feature information, W is the dimensionality reduction matrix, and X is the original data matrix.

[0024] Preferably, the intelligent prediction and abnormal analysis layer includes:

[0025] The bidirectional LSTM model is used to model time series data, and its hidden state update formula is:

[0026] h t =LSTM(X t ,h t-1 )

[0027] Among them, h t is the hidden state at time t, which contains the memory and context information of the time series data, X t is the input data at time t, including environmental parameters collected by multimodal sensors, h t-1 is the hidden state of the previous time step, passing historical information to the current moment;

[0028] Convolutional neural networks are used to extract spatial features of environmental parameters.

[0029] Preferably, the communication and data transmission layer schedules data transmission through an event-driven priority strategy, and the priority is calculated by the following formula:

[0030]

[0031] Among them, U(E) is the urgency of the event, I(D) is the importance of the data, and γ1 and γ2 are weight coefficients.

[0032] Preferably, the data storage and collaboration layer adopts a technology combining blockchain and knowledge graph, and the data storage structure of the blockchain is calculated by the following formula:

[0033] H block =Hash(H prev +D current +T timestamp )

[0034] Among them, H block is the hash value of the current block, which is used to uniquely identify the content of the block. prev is the hash value of the previous block, which is used to establish the chain structure of blocks in the blockchain. current is the monitoring data stored in the current block, T timestamp It is the timestamp of the current block, recording the time when the block was generated. Hash is a hash function used to generate a unique summary value of a fixed length.

[0035] Preferably, the security layer is based on an encryption module of quantum key distribution, and its key generation formula is:

[0036] K=H(A⊕B)

[0037] Among them, K is the generated encryption key used to protect the transmitted data, A and B are the quantum state measurement results, generated by the communicating parties respectively, H is the hash function used to map the result of A⊕B to the key, and ⊕ is the bitwise exclusive, which is used to randomize the measurement results;

[0038] The security layer is based on a data integrity verification module based on zero-knowledge proof, and the verification formula is:

[0039] P(g x mod p)=y without revealingx.

[0040] Among them, P is the verification function used to prove the integrity of the data, g is the basic generator used to calculate the modular exponentiation, x is the private data known only by the prover, p is a large prime number used as the base of the modular operation, and y is the public value used to verify whether x is correct.

[0041] Preferably, the data storage and collaboration layer further includes a monitoring interface that supports multiple languages, and users can query product status and abnormal events in real time through a visual dashboard, and set monitoring thresholds through personalized rules.

[0042] Preferably, the system supports a federated learning framework that aggregates the parameters of the regional gateway model through the following formula:

[0043]

[0044] Among them, w t is the model parameter of the kth region, n k is the sample size.

[0045] The present invention provides a remote dynamic monitoring system for import and export commodities based on the Internet of Things. It has the following beneficial effects:

[0046] 1. The present invention realizes distributed data processing among sensor nodes, local gateways and regional gateways by deploying a multi-level collaborative computing framework at the edge computing layer. The edge nodes undertake data preprocessing and feature extraction tasks, reducing the uploading of redundant data. At the same time, the regional gateway uses a lightweight machine learning model to perform intelligent analysis and real-time anomaly detection on the monitoring data, reducing the processing burden of the central server. Through the dynamic threshold adjustment mechanism, the system can flexibly adapt to different transportation conditions and commodity types, further improving the accuracy of data analysis and the adaptability of the system.

[0047] 2. The present invention adopts a dynamic data transmission strategy based on event priority, which only uploads high-priority data when a specific abnormal event is triggered, and delays the transmission or uploads normal data in batches. The transmission priority is calculated by the urgency and importance of the data, thereby avoiding the waste of communication resources. In addition, the system has dynamic channel switching capabilities and can flexibly choose between various communication technologies according to network coverage, transmission reliability and energy consumption, ensuring the stability and efficiency of data transmission in complex scenarios. This strategy significantly reduces the energy consumption and communication costs of data transmission and extends the operating time of the equipment.

[0048] 3. The present invention significantly improves the system's anomaly detection capability through cross-modal data fusion technology and deep learning models. The system uses an attention mechanism to fuse data streams from different sensors, extracts multimodal features, combines a bidirectional LSTM network to model time dependencies, and a convolutional neural network to capture spatial features, thereby achieving environmental parameter trend prediction and abnormal event classification. Through a risk grading mechanism, the system can prioritize abnormal events and quickly issue warnings, thereby helping users take countermeasures in advance and reduce possible risks and losses during transportation.

[0049] 4. The present invention records monitoring data through blockchain technology to ensure the integrity and non-tamperability of the data. At the same time, it automatically triggers abnormal event alarms through smart contracts to reduce the delay caused by manual operations. Combined with knowledge graph technology, the system models commodities, environmental parameters and transportation events as semantic relationships, providing users with a transparent and easy-to-query data management platform. By providing a standardized API interface, logistics companies, customs and customers can efficiently share monitoring data, thereby significantly improving the efficiency of cross-organizational collaboration. Based on the deep integration of blockchain and knowledge graph, this system has shown strong advantages in data sharing, collaborative transparency and decision support. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a system logic diagram of the present invention;

[0051] Figure 2 It is a schematic diagram of the system layer distribution of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] Please see attached Figure 1 -Attached Figure 2 The embodiment of the present invention provides a remote dynamic monitoring system for imported and exported goods based on the Internet of Things, including:

[0054] The perception layer is a multimodal sensor network used to collect environmental parameters and location information of imported and exported goods;

[0055] The edge computing layer, which includes sensor nodes, local gateways, and regional gateways, performs data preprocessing, aggregation, feature extraction, and distributed collaborative computing;

[0056] Intelligent prediction and anomaly analysis layer, which predicts environmental parameter trends and detects anomalies through deep learning models;

[0057] The communication and data transmission layer provides event-driven data transmission strategies and dynamic channel selection functions;

[0058] The data storage and collaboration layer stores monitoring data based on blockchain technology and manages data relationships through knowledge graphs;

[0059] The security layer uses quantum encryption and zero-knowledge proof technology to ensure data security and integrity;

[0060] The system supports the federated learning framework and aggregates the parameters of the regional gateway model through the following formula:

[0061]

[0062] Among them, w t is the model parameter of the kth region, n k is the sample size.

[0063] Specifically, data is collected from multimodal sensors in the perception layer, and then denoised, aggregated, and feature extracted in the edge computing layer, which reduces redundant data and improves computing efficiency. The data after preliminary processing is transmitted to the intelligent prediction and anomaly analysis layer, and the deep learning model is used to predict environmental parameter trends and identify anomaly risks. High-priority data is uploaded to the cloud first through the communication and data transmission layer using an event-driven strategy, and the data is stored in the blockchain to ensure integrity and transparency. For example, during cold chain transportation, after the perception layer collects temperature anomaly information, the edge computing layer uses the graph neural network to determine the scope of anomaly impact, and the intelligent analysis layer predicts the temperature change trend, which ultimately triggers the priority transmission mechanism of the communication layer and uploads high-risk data to the cloud. The data storage and collaboration layer records abnormal events and helps users identify potential risk factors through knowledge graphs and historical data association analysis. The entire process is encrypted and protected by the security layer to ensure data transmission security and privacy protection.

[0064] The perception layer adjusts the sampling interval of the multimodal sensor through a dynamic sampling rate control mechanism. The dynamic sampling rate is determined by the sampling risk R risk and power consumption constraint E sensor Determine, the formula is as follows:

[0065] R t =R base +β1R risk -β2E sensor

[0066] Among them, R t is the dynamic sampling rate of the sensor at time t, R base is the basic sampling rate, β1 and β2 are adjustment parameters;

[0067] The edge computing layer uses multi-resolution wavelet transform to reduce the noise of sensor data. The signal decomposition formula is as follows:

[0068]

[0069] Among them, c j,k is the wavelet decomposition coefficient, representing the signal strength at different resolutions, X(t) is the original sensor signal at time t, ψ j,k (t) is the wavelet basis function;

[0070] The edge gateway of the edge computing layer reduces the dimension of multimodal data through principal component analysis. The data matrix after dimension reduction is expressed as:

[0071] X reduced =W·X

[0072] Among them, X reducedis the data matrix after dimensionality reduction, retaining the main feature information, W is the dimensionality reduction matrix, and X is the original data matrix.

[0073] Specifically, the perception layer consists of a multimodal sensor network deployed in the import and export commodity transportation environment, which is used to collect dynamic data of commodities and their surroundings in real time, including environmental parameters such as temperature, humidity, vibration, light intensity, gas concentration, as well as GPS positioning information and the operating status of the sensor equipment itself (such as power and communication status). This layer dynamically adjusts the sampling frequency through a dynamic sampling rate control formula to optimize energy consumption and sampling efficiency. In addition, it also uses signal noise reduction and data compression algorithms (such as wavelet transform and differential coding) to reduce redundant data and ensure the accuracy and real-time performance of sensor data.

[0074] The edge computing layer includes a three-layer collaborative computing architecture of sensor nodes, local gateways, and regional gateways, which is mainly responsible for preliminary data processing, feature extraction, aggregation and dimensionality reduction, and distributed collaborative computing. The sensor node layer uses a multi-resolution algorithm to reduce the noise of local data, the local gateway layer uses principal component analysis to reduce the dimensionality of multimodal data, and the regional gateway layer uses graph neural networks to model the correlation features between sensor nodes. In addition, this layer implements global model updates across regional gateways through a federated learning framework, effectively solving data privacy issues while reducing the computing pressure of centralized processing.

[0075] The intelligent prediction and anomaly analysis layer includes:

[0076] The bidirectional LSTM model is used to model time series data, and its hidden state update formula is:

[0077] h t =LSTM(X t ,h t-1 )

[0078] Among them, h t is the hidden state at time t, which contains the memory and context information of the time series data, X t is the input data at time t, including environmental parameters collected by multimodal sensors, h t-1 is the hidden state of the previous time step, passing historical information to the current moment;

[0079] Use convolutional neural networks to extract spatial features of environmental parameters;

[0080] The communication and data transmission layer schedules data transmission through an event-driven priority strategy. The priority is calculated by the following formula:

[0081]

[0082] Among them, U(E) is the urgency of the event, I(D) is the importance of the data, and γ1 and γ2 are weight coefficients;

[0083] The data storage and collaboration layer uses a technology that combines blockchain and knowledge graph. The data storage structure of the blockchain is calculated using the following formula:

[0084] H block =Hash(H prev +D current +T timestamp )

[0085] Among them, H block is the hash value of the current block, which is used to uniquely identify the content of the block. prev is the hash value of the previous block, which is used to establish the chain structure of blocks in the blockchain. current is the monitoring data stored in the current block, T timestamp It is the timestamp of the current block, recording the time when the block was generated. Hash is a hash function used to generate a unique summary value of a fixed length.

[0086] The security layer is based on the encryption module of quantum key distribution, and its key generation formula is:

[0087] K=H(A⊕B)

[0088] Among them, K is the generated encryption key used to protect the transmitted data, A and B are the quantum state measurement results, generated by the communicating parties respectively, H is the hash function used to map the result of A⊕B to the key, and ⊕ is the bitwise exclusive, which is used to randomize the measurement results;

[0089] The security layer is based on the data integrity verification module of zero-knowledge proof. The verification formula is:

[0090] P(g x mod p)=y without revealing ingx.

[0091] Among them, P is the verification function used to prove the integrity of the data, g is the basic generator used to calculate the modular power, x is the private data known only by the prover, p is a large prime number used as the base of the modular operation, and y is the public value used to verify whether x is correct;

[0092] The data storage and collaboration layer further includes a monitoring interface with multi-language support. Users can query product status and abnormal events in real time through a visual dashboard, and set monitoring thresholds through personalized rules.

[0093] Specifically, the intelligent prediction and anomaly analysis layer is the core analysis module of the system, which uses a multi-task learning framework to simultaneously complete environmental parameter trend prediction and anomaly type classification. The time series and spatial features of multimodal data are fused through the attention mechanism, the bidirectional LSTM network is used to capture the temporal dependency, and the cross-modal spatial features are extracted through the convolutional network. In addition, this layer performs risk classification on the results output by the model, and explains the causes of abnormal events through methods such as SHAP values, thereby providing users with intuitive and transparent risk warnings.

[0094] The communication and data transmission layer mainly optimizes data transmission efficiency and reliability through event-driven data transmission strategies and dynamic channel switching technology. The transmission priority is calculated based on the urgency of the event and the importance of the data, and the optimal channel is selected based on the current network conditions (such as the bandwidth and power consumption of NB-IoT, LoRa, and 5G). In addition, to solve the problem of signal interruption in complex transportation scenarios, this layer designs breakpoint resumption and caching mechanisms to ensure data integrity and reliability.

[0095] The data storage and collaboration layer uses blockchain technology to record commodity monitoring data to ensure the immutability and transparency of the data. At the same time, the relationship between commodities, environmental parameters and transportation events is modeled through the knowledge graph to form a semantic data management framework. Each block is connected through a hash function to generate a data chain. The knowledge graph describes the relationship between data through triples, provides standardized API interfaces for multiple users (such as customs, logistics parties and customers), and supports data query and sharing.

[0096] The security layer protects the security of data during transmission and storage through quantum encryption and zero-knowledge proof technology. Quantum key distribution is used to generate highly secure keys, combined with symmetric and asymmetric encryption algorithms to achieve data transmission encryption. Zero-knowledge proof is used to verify data integrity and the credibility of transmission while protecting user privacy. In addition, this layer also integrates hierarchical permission management to ensure that different users can only access data within their authorized scope.

[0097] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The remote dynamic monitoring system for import and export commodities based on the Internet of Things is characterized by: include: The perception layer is a multimodal sensor network used to collect environmental parameters and location information of imported and exported goods; The edge computing layer, which includes sensor nodes, local gateways, and regional gateways, performs data preprocessing, aggregation, feature extraction, and distributed collaborative computing; Intelligent prediction and anomaly analysis layer, which predicts environmental parameter trends and detects anomalies through deep learning models; The communication and data transmission layer provides event-driven data transmission strategies and dynamic channel selection functions; The data storage and collaboration layer stores monitoring data based on blockchain technology and manages data relationships through knowledge graphs; The security layer uses quantum encryption and zero-knowledge proof technology to ensure the security and integrity of data.

2. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The perception layer adjusts the sampling interval of the multimodal sensor through a dynamic sampling rate control mechanism, and the dynamic sampling rate is determined by the sampling risk R risk and power consumption constraint E sensor Determine, the formula is as follows: R t =R base +β1R risk -β2E sensor Among them, R t is the dynamic sampling rate of the sensor at time t, R base is the basic sampling rate, β1, β2 are adjustment parameters.

3. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The edge computing layer uses multi-resolution wavelet transform to reduce the noise of sensor data, and its signal decomposition formula is as follows: Among them, c j,k is the wavelet decomposition coefficient, representing the signal strength at different resolutions, X(t) is the original sensor signal at time t, ψ j,k (t) is the wavelet basis function.

4. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The edge gateway of the edge computing layer reduces the dimension of multimodal data through principal component analysis. The data matrix after dimension reduction is expressed as: X reduced =W·X Among them, X reduced is the data matrix after dimensionality reduction, retaining the main feature information, W is the dimensionality reduction matrix, and X is the original data matrix.

5. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The intelligent prediction and abnormal analysis layer includes: The bidirectional LSTM model is used to model time series data, and its hidden state update formula is: h t =LSTM(X t ,h t-1 ) Among them, h t is the hidden state at time t, which contains the memory and context information of the time series data, X t is the input data at time t, including the environmental parameters collected by multimodal sensors, h t-1 is the hidden state of the previous time step, passing historical information to the current moment; Convolutional neural networks are used to extract spatial features of environmental parameters.

6. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The communication and data transmission layer schedules data transmission through an event-driven priority strategy, and the priority is calculated by the following formula: Among them, U(E) is the urgency of the event, I(D) is the importance of the data, and γ1 and γ2 are weight coefficients.

7. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The data storage and collaboration layer adopts the technology of combining blockchain and knowledge graph. The data storage structure of blockchain is calculated by the following formula: H block =Hash(H prev +D current +T timestamp ) Among them, H block is the hash value of the current block, which is used to uniquely identify the content of the block. prev is the hash value of the previous block, which is used to establish the chain structure of blocks in the blockchain. current is the monitoring data stored in the current block, T timestamp It is the timestamp of the current block, recording the time when the block was generated. Hash is a hash function used to generate a unique summary value of a fixed length.

8. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The security layer is based on the encryption module of quantum key distribution, and its key generation formula is: K=H(A⊕B) Among them, K is the generated encryption key used to protect the transmitted data, A and B are the quantum state measurement results, generated by the communicating parties respectively, H is the hash function used to map the result of A⊕B to the key, and ⊕ is the bitwise exclusive, which is used to randomize the measurement results; The security layer is based on a data integrity verification module based on zero-knowledge proof, and the verification formula is: P(g x mod p)=y without revealingx. Among them, P is the verification function used to prove the integrity of the data, g is the basic generator used to calculate the modular exponentiation, x is the private data known only by the prover, p is a large prime number used as the base of the modular operation, and y is the public value used to verify whether x is correct.

9. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The data storage and collaboration layer further includes a monitoring interface that supports multiple languages. Users can query product status and abnormal events in real time through a visual dashboard and set monitoring thresholds through personalized rules.

10. The remote dynamic monitoring system for import and export commodities based on the Internet of Things according to claim 1 is characterized in that: The system supports the federated learning framework and aggregates the parameters of the regional gateway model through the following formula: Among them, w t is the model parameter of the kth region, n k is the sample size.

Images