Fresh food supply chain full-link traceability and loss early warning method based on block chain

Abstract

The invention discloses a fresh food supply chain full-link traceability and loss early warning method based on a block chain, and the method comprises the steps: 1, carrying out the lightweight collection of full-link data, processing the data, and obtaining a standardized data source, and 2, carrying out the construction of a block chain bottom-layer architecture based on the standardized data source; step 3, establishing a dynamic reputation consensus mechanism, and establishing a trusted environment for data sharing through reputation score driving node collaboration based on a block chain bottom layer architecture; 4, constructing a federated learning network; step 5, risk prediction modeling: performing multi-dimensional risk prediction by utilizing model output of collaborative modeling and combining real-time data; step 6, graded early warning response: based on a risk prediction result, realizing graded automatic response through a grading rule; step 7, dynamic decision optimization; the beneficial effects of the invention are that the dynamic decision optimization function adjusts the strategy of each link of the supply chain according to the early warning result, achieves the intelligent scheduling of logistics, inventory and production, and reduces the loss and cost.

Description

Technical Field

[0001] The present invention relates to the field of fresh produce supply chain early warning, and specifically to a blockchain-based full-link traceability and loss early warning method for fresh produce supply chains. Background Art

[0002] Amid the global wave of digital transformation, supply chain management is facing challenges such as data silos, lack of trust, and delayed decision-making. Traditional supply chains rely on manual data collection and centralized storage, resulting in doubts about data authenticity and inefficient cross-enterprise collaboration, making it difficult to cope with the complex and changing market environment. Recalls caused by data opacity in the food supply chain are increasing annually, and logistics costs due to unreasonable route planning account for a high proportion of losses. At the same time, the maturity of technologies such as the Internet of Things (IoT), blockchain, and artificial intelligence (AI) has provided a possibility for breakthroughs. Based on the above-mentioned technological integration trends, the intelligent supply chain system aims to build a full-link solution from data collection and trusted storage to intelligent decision-making, and promote the upgrade of the supply chain towards automation, intelligence, and trustworthiness. Summary of the Invention

[0003] The purpose of the present invention is to solve the above-mentioned problems, and therefore proposes a blockchain-based full-link traceability and loss warning method for the fresh food supply chain.

[0004] The purpose of the present invention can be achieved through the following technical solution: a blockchain-based full-link traceability and loss warning method for fresh food supply chain, comprising: Step 1: Lightweight data collection across the entire chain, obtaining data from the planting and breeding ends, and processing the data to obtain a standardized data source; Step 2: Build the underlying blockchain architecture based on standardized data sources; Step 3: Establish a dynamic reputation consensus mechanism, relying on the underlying blockchain architecture, drive node collaboration through reputation scoring, and establish a trusted environment for data sharing; Step 4: Federated learning network construction, based on a trusted environment, collaborative modeling; Step 5: Risk prediction modeling, using the model output of collaborative modeling and combining it with real-time data to conduct multi-dimensional risk prediction; Step 6: Grading early warning response: Based on the risk prediction results, graded automated response is achieved through graded rules; Step 7: Dynamic decision optimization, converting hierarchical automated responses into specific execution strategies. At the same time, through reverse tracing, data is accumulated and fed back to step 1 to complete collection optimization.

[0005] Furthermore, the step 1 specifically includes: Deploy sensors at the planting / breeding end to collect biometric and environmental data in real time; Edge computing nodes preprocess data, filter outliers, and compress data; Generate lightweight data fingerprints, perform hash operations on preset key data, and form a data chain pass.

[0006] Furthermore, the step 2 specifically includes: Adopt a decentralized distributed storage solution to shard and store the standardized data source in step 1 on multiple nodes. Ensure data integrity and availability through hash verification and redundant backup to avoid single point failure risks. Implement hierarchical encryption for stored data, support computing operations in the encrypted state, and combine with the federated learning framework to achieve the requirement of data being available but invisible; Leveraging the distributed ledger features of blockchain, a unique hash fingerprint is generated for lightweight data and stored on-chain. Smart contracts are used to automatically verify data sources and operation records, ensuring data immutability and traceability. Build an intelligent cache and hot and cold data tiered storage system, dynamically adjust storage strategies based on data access frequency and importance, and integrate edge computing nodes to achieve local storage and rapid access of data, improving read and write efficiency. Deploy TEE hardware modules on storage nodes to build secure enclaves, provide hardware-level protection for data storage, transmission, and processing, resist malicious attacks and data theft, and form an unalterable data source.

[0007] Furthermore, the step three specifically includes: Relying on the blockchain architecture built in step 2, the decentralized and tamper-proof characteristics of the blockchain are used as the trust foundation. Based on the distributed ledger and encrypted transmission system built in step 2, the behavioral value of the node is quantified through the node reputation scoring system. Specifically: Data quality is assessed based on the integrity and accuracy of the original data stored in the blockchain; Consensus contribution points are calculated based on the verification process log recorded in the blockchain to verify node compliance; The historical performance points and the enterprise transaction data stored on the association chain together build a node reputation scoring system.

[0008] Furthermore, the step 4 is specifically as follows: Each participating node saves a complete log. When a node initiates an online learning request, the blockchain system uses a hash algorithm to verify the integrity of the model version, ensuring that all participants use the same trusted version of the basic model. Dynamically adjust the topology of the federated learning network based on the node activity and historical contribution recorded in the blockchain; The behavior of nodes participating in online learning is rewarded through smart contracts, which encourages enterprises to continuously participate in network construction and model iteration.

[0009] Furthermore, the step five is specifically as follows: Acquire real-time environmental data and biometric data from the planting / breeding end to determine the risk of sudden temperature changes during cold chain transportation; The underlying blockchain architecture is used to input training time series prediction models and analyze historical loss events; The federated learning network constructed in step 4 provides cross-domain collaborative models and risk characteristics.

[0010] Furthermore, the step six is ​​specifically as follows: The hierarchical warning response definition has three levels of warning rules: Based on the risk prediction results of step 5, an automated response is achieved through a three-level warning rule. The three-level warning rule is as follows: a single indicator abnormality in the first level automatically pushes warning information to the responsible party, triggering local equipment adjustments; The secondary warning correlation model triggers the automatic start of the smart contract; When the multi-dimensional risk index of the third-level warning is greater than the threshold, the batch of transactions will be frozen and the insurance claim will be initiated, and a full-link traceability report will be pushed to the regulatory end.

[0011] Compared with the prior art, the present invention has the following beneficial effects: Multi-dimensional risk prediction modeling combines multi-source data to achieve deduction from single-point anomalies to global risks through time series prediction, spatial correlation analysis, and anomaly detection, providing accurate risk warnings. The intelligent graded warning response mechanism automatically triggers different processing processes based on risk levels, realizing automated monitoring-analysis-response and improving risk response efficiency. The dynamic decision-making optimization function adjusts the strategies of each link in the supply chain according to the early warning results, realizes the intelligent scheduling of logistics, inventory and production, and reduces losses and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0013] Figure 1 This is a flowchart of a blockchain-based full-link traceability and loss warning method for the fresh food supply chain of the present invention. DETAILED DESCRIPTION

[0014] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0015] See also Figure 1 As shown in the figure, a blockchain-based full-link traceability and loss warning method for fresh food supply chain includes: Step 1: Lightweight data collection across the entire chain, obtaining data from the planting and breeding ends, and processing the data to obtain a standardized data source; Step 1 specifically includes: Deploy sensors at the planting / breeding end to collect biometric and environmental data in real time; Edge computing nodes preprocess data, filter outliers, and compress data; Generate lightweight data fingerprints, perform hash operations on preset key data, and form a data chain pass.

[0016] In actual use, a matrix of temperature and humidity sensors is deployed in the planting area. Using LoRa wide-area network technology, real-time data transmission is achieved within a 5-kilometer radius of the field. For fruit and vegetable cultivation, PEGS edible sensors are embedded in the fruit to continuously monitor changes in ethylene concentration and provide maturity data for precise harvesting. In livestock farming, each animal is equipped with an RFID electronic ear tag to record biometric data such as diet and activity throughout its growth cycle. Furthermore, a vibration sensor network is deployed in the farm to capture behavioral changes in animal groups and provide early warning of disease transmission risks. A distributed edge computing cluster is built, and the system filters abnormal data using a dynamic threshold algorithm. When a temperature sensor detects an abnormal value of -50°C, it automatically triggers a data verification mechanism for neighboring nodes. For vibration data that exceeds the threshold, wavelet transform downsampling technology is used to optimize the original 100Hz sampling frequency to a 5Hz effective fluctuation range, achieving a data compression ratio of over 80%. The node has a built-in 1TB solid-state drive and adopts a tiered storage strategy for hot and cold data. Important raw data is retained for 30 days and automatically migrated to cloud-based cold storage after expiration. A multi-level data encryption system is established. For key information such as batch ID and quality inspection report, AES-256 symmetric encryption is first performed, and then a unique data fingerprint is generated through the SHA-256 hash algorithm. To ensure that the data cannot be tampered with, each fingerprint is associated with the blockchain timestamp service, and a traceability certificate is generated on the Hyperledger Fabric consortium chain. At the same time, a data fingerprint comparison tool is developed to support users to quickly verify data integrity through hash values, and the error response time is controlled within 200 milliseconds.

[0017] Step 2: Build the underlying blockchain architecture based on standardized data sources; Step 2 specifically includes: Adopt a decentralized distributed storage solution to shard and store the standardized data source in step 1 on multiple nodes. Ensure data integrity and availability through hash verification and redundant backup to avoid single point failure risks. Implement hierarchical encryption for stored data, support computing operations in the encrypted state, and combine with the federated learning framework to achieve the requirement of data being available but invisible; Leveraging the distributed ledger features of blockchain, a unique hash fingerprint is generated for lightweight data and stored on-chain. Smart contracts are used to automatically verify data sources and operation records, ensuring data immutability and traceability. Build an intelligent cache and hot and cold data tiered storage system, dynamically adjust storage strategies based on data access frequency and importance, and integrate edge computing nodes to achieve local storage and rapid access of data, improving read and write efficiency. Deploy TEE hardware modules on storage nodes to build secure enclaves, provide hardware-level protection for data storage, transmission, and processing, resist malicious attacks and data theft, and form an unalterable data source.

[0018] When in use, the lightweight data fingerprint generated in step 1 is received, and the data source and integrity are automatically verified through the smart contract. For example, when the production end submits batch data, the contract compares the data hash with the original fingerprint. If they are consistent, the evidence is automatically uploaded to the chain and the traceability chain is updated. At the same time, zero-knowledge proof technology is used to support nodes to verify the authenticity of data without leaking the data content. Based on the frequency of data access, hot and cold data tiered storage is constructed. Real-time data exceeding the access threshold is stored in the blockchain memory database, and data below the access threshold is migrated to IPFS long-term storage. Combined with the edge computing node caching strategy, data can be called nearby, and the response time of key data queries can be controlled within 500 milliseconds.

[0019] Step 3: Establish a dynamic reputation consensus mechanism, relying on the underlying blockchain architecture, drive node collaboration through reputation scoring, and establish a trusted environment for data sharing; Step three specifically includes: Relying on the blockchain architecture built in step 2, the decentralized and tamper-proof characteristics of the blockchain are used as the trust foundation. Based on the distributed ledger and encrypted transmission system built in step 2, the behavioral value of the node is quantified through the node reputation scoring system. Specifically: Data quality is assessed based on the integrity and accuracy of the original data stored in the blockchain; Consensus contribution points are calculated based on the verification process log recorded in the blockchain to verify node compliance; The historical performance score is associated with the enterprise transaction data stored on the chain, and the three together build a node reputation scoring system; When in use, the dynamic reputation consensus mechanism in step three deeply relies on the blockchain architecture built in step two, and uses the decentralized and tamper-proof characteristics of the blockchain as the trust foundation. On the basis of the distributed ledger and encrypted transmission system constructed in step two, step three quantifies the behavioral value of the node through a three-dimensional reputation scoring system: the data quality score is evaluated based on the integrity and accuracy of the original data stored in the blockchain; the consensus contribution score is based on the verification process log recorded in the blockchain to calculate the compliance of the node; the historical performance score is based on the corporate transaction data stored on the associated chain. The three together build a node reputation scoring system, and dynamic verification node election is carried out through this standard, so that nodes with a reputation that reaches the threshold can obtain verification authority, forming a "reputation-driven verification" collaborative model, which effectively avoids the centralization risk of a few nodes controlling the discourse power in the traditional consensus mechanism. At the same time, in conjunction with the incentive and punishment mechanism, a set of adaptive and traceable data sharing trust environment is established within the blockchain framework to ensure the authenticity and security of supply chain data during the circulation and sharing process.

[0020] Step 4: Federated learning network construction, based on a trusted environment, collaborative modeling; Step 4 is as follows: Each participating node saves a complete log. When a node initiates an online learning request, the blockchain system uses a hash algorithm to verify the integrity of the model version, ensuring that all participants use the same trusted version of the basic model. Dynamically adjust the topology of the federated learning network based on the node activity and historical contribution recorded in the blockchain; The behavior of nodes participating in online learning is rewarded through smart contracts, which encourages enterprises to continuously participate in network construction and model iteration; When in use, based on the blockchain trusted consensus environment built in step three, cross-enterprise data can be made "available but invisible", providing a cross-domain collaborative model for the intelligent prediction in step five. Through the above-mentioned network learning network construction method, data privacy and security are guaranteed, and the deep integration and efficient utilization of multi-party data value are achieved. For example, using the federated learning framework, each enterprise trains the model locally and only uploads the updated values ​​of the model parameters, completing collaborative modeling under the premise of protecting data privacy, thereby providing a cross-domain collaborative model for risk prediction.

[0021] Step 5: Risk prediction modeling, using the model output of collaborative modeling and combining it with real-time data to conduct multi-dimensional risk prediction; Step 5 is as follows: Acquire real-time environmental data and biometric data from the planting / breeding end to determine the risk of sudden temperature changes during cold chain transportation; The underlying blockchain architecture is used to input training time series prediction models and analyze historical loss events; The federated learning network constructed in step 4 provides a cross-domain collaborative model and risk characteristics; During use, based on step one, real-time environmental data (such as temperature, humidity, vibration value) and biometric data (ethylene concentration, microbial indicators) of the planting end / breeding end are obtained. These real-time edge data serve as the input part of the input layer of step five, providing dynamic perception capabilities for risk prediction. For example, the temperature and humidity data collected by the edge node are directly used for anomaly detection to judge the risk of sudden temperature changes during cold chain transportation. The underlying blockchain architecture built based on step two provides tamper-proof evidence data for step five. These data are used to train time series prediction models (such as LSTM to predict quality decay trends) and analyze historical loss events to optimize risk prediction strategies. The dynamic reputation consensus mechanism established based on step three ensures the credibility of the data used in step five, the three-dimensional reputation scoring system evaluates the data quality, and the federated learning network built based on step four provides step five with a cross-domain collaborative model and risk characteristics.

[0022] Step 6: Grading early warning response: Based on the risk prediction results, graded automated response is achieved through graded rules; Step six is ​​as follows: The hierarchical warning response definition has three levels of warning rules: Based on the risk prediction results of step 5, an automated response is achieved through a three-level warning rule. The three-level warning rule is as follows: Level 1 alerts a single indicator anomaly (such as the temperature exceeding the threshold of ±2°C), automatically pushes warning information to the responsible party, and triggers local equipment adjustments (such as increasing the temperature of the warehouse air conditioner); When the secondary warning correlation model is triggered (e.g., the ethylene concentration of a batch of apples exceeds the standard and the adjacent bananas have respiratory climacteric characteristics), the smart contract is automatically activated (e.g., generating a dispatch instruction for an isolated storage location); When the multi-dimensional risk index of the third-level warning exceeds the threshold, the batch of transactions will be frozen and insurance claims will be initiated. At the same time, a full-link traceability report will be pushed to the regulatory end; Step 7: Dynamic decision optimization: Convert hierarchical automated responses into specific execution strategies. At the same time, through reverse tracing, accumulated data is fed back to step 1 to complete data collection optimization. Step seven is as follows: On the logistics side: the route is dynamically adjusted according to the early warning results, and the GIS system synchronizes the optimal route in real time; on the inventory side: terminal demand is predicted through learning algorithms, and expiring products within one month of their shelf life are automatically allocated to promotion channels; historical loss data is fed back to the planting / breeding links to optimize planting plans. When a loss event occurs, the system automatically backtracks the blockchain data and generates a four-dimensional traceability report to verify the compliance of transportation temperature and humidity, trace warehouse in and out records, verify the authenticity of the product source through DNA fingerprint technology, and associate the historical credit score of the responsible party as an important basis for determining responsibility.

[0023] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A blockchain-based full-link traceability and loss warning method for fresh food supply chain, characterized by: include: Step 1: Lightweight data collection across the entire chain, obtaining data from the planting and breeding ends, and processing the data to obtain a standardized data source; Step 2: Build the underlying blockchain architecture based on standardized data sources; Step 3: Establish a dynamic reputation consensus mechanism, relying on the underlying blockchain architecture, drive node collaboration through reputation scoring, and establish a trusted environment for data sharing; Step 4: Federated learning network construction, based on a trusted environment, collaborative modeling; Step 5: Risk prediction modeling, using the model output of collaborative modeling and combining it with real-time data to conduct multi-dimensional risk prediction; Step 6: Grading early warning response: Based on the risk prediction results, graded automated response is achieved through graded rules; Step 7: Dynamic decision optimization, converting hierarchical automated responses into specific execution strategies. At the same time, through reverse tracing, data is accumulated and fed back to step 1 to complete collection optimization.

2. A blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 1, characterized in that: The step 1 specifically includes: Deploy sensors at the planting / breeding end to collect biometric and environmental data in real time; Edge computing nodes preprocess data, filter outliers, and compress data; Generate lightweight data fingerprints, perform hash operations on preset key data, and form a data chain pass.

3. The blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 1 is characterized in that: The second step specifically includes: Adopt a decentralized distributed storage solution to shard and store the standardized data source in step 1 on multiple nodes. Ensure data integrity and availability through hash verification and redundant backup to avoid single point failure risks. Implement hierarchical encryption for stored data, support computing operations in the encrypted state, and combine with the federated learning framework to achieve the requirement of data being available but invisible; Leveraging the distributed ledger features of blockchain, a unique hash fingerprint is generated for lightweight data and stored on-chain. Smart contracts are used to automatically verify data sources and operation records, ensuring data immutability and traceability. Build an intelligent cache and hot and cold data tiered storage system, dynamically adjust storage strategies based on data access frequency and importance, and integrate edge computing nodes to achieve local storage and rapid access of data, improving read and write efficiency. Deploy TEE hardware modules on storage nodes to build secure enclaves, provide hardware-level protection for data storage, transmission, and processing, resist malicious attacks and data theft, and form an unalterable data source.

4. A blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 3, characterized in that: The step three specifically includes: Relying on the blockchain architecture built in step 2, the decentralized and tamper-proof characteristics of the blockchain are used as the trust foundation. Based on the distributed ledger and encrypted transmission system built in step 2, the behavioral value of the node is quantified through the node reputation scoring system. Specifically: Data quality is assessed based on the integrity and accuracy of the original data stored in the blockchain; Consensus contribution points are calculated based on the verification process log recorded in the blockchain to verify node compliance; The historical performance points and the enterprise transaction data stored on the association chain together build a node reputation scoring system.

5. The blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 4 is characterized in that: The step 4 is specifically as follows: Each participating node saves a complete log. When a node initiates an online learning request, the blockchain system uses a hash algorithm to verify the integrity of the model version, ensuring that all participants use the same trusted version of the basic model. Dynamically adjust the topology of the federated learning network based on the node activity and historical contribution recorded in the blockchain; The behavior of nodes participating in online learning is rewarded through smart contracts, which encourages enterprises to continuously participate in network construction and model iteration.

6. A blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 5, characterized in that: The step five is specifically as follows: Acquire real-time environmental data and biometric data from the planting / breeding end to determine the risk of sudden temperature changes during cold chain transportation; The underlying blockchain architecture is used to input training time series prediction models and analyze historical loss events; The federated learning network constructed in step 4 provides cross-domain collaborative models and risk characteristics.

7. A blockchain-based full-link traceability and loss warning method for fresh food supply chain according to claim 6, characterized in that: The step six is ​​specifically as follows: The hierarchical warning response definition has three levels of warning rules: Based on the risk prediction results of step 5, an automated response is achieved through a three-level warning rule. The three-level warning rule is as follows: a single indicator abnormality in the first level automatically pushes warning information to the responsible party, triggering local equipment adjustments; The secondary warning correlation model triggers the automatic start of the smart contract; When the multi-dimensional risk index of the third-level warning is greater than the threshold, the batch of transactions will be frozen and the insurance claim will be initiated, and a full-link traceability report will be pushed to the regulatory end.

Images