Supply chain platform data management system based on intelligent Internet of Things

Abstract

The invention relates to the technical field of intelligent Internet of Things supply chain management, in particular to an intelligent Internet of Things-based supply chain platform data management system, which comprises a data acquisition module, a contract execution module, a node control module, a credit evaluation module and a data asset module. The data acquisition module acquires material physical state information to generate a supply chain panoramic data stream; the contract execution module extracts key performance data and automatically verifies the key performance data according to intelligent contract logic, and generates a credible contract execution signal if a condition is met; the node control module triggers logistics and storage nodes to automatically control and synchronously update resource state records according to the signals; the credit evaluation module extracts features to generate a financial evaluation vector and calculates a dynamic credible score; and the data asset module is packaged into a standardized asset packet to synchronously authorize the alliance node. The system can realize the technical effects of contract automatic execution, node linkage control, credit dynamic evaluation and alliance collaborative decision.

Description

Technical Field

[0001] This invention relates to the field of smart IoT supply chain management technology, and in particular to a supply chain platform data management system based on smart IoT. Background Technology

[0002] In traditional supply chain operations, IoT devices collect physical status information of materials to form a basic data stream. Contract-related performance data is often manually screened, and manual verification is required to ensure compliance with contract terms. Control operations at logistics and warehousing nodes rely on manual instructions, and node resource status records are manually entered and updated. Credit assessments in supply chain finance often use static historical data without analyzing real-time node resource status. Data from participating nodes is stored in a fragmented manner, lacking standardized data assets, and data synchronization between alliance nodes relies on manual transmission and format conversion. This manual contract verification and node control model is prone to data screening biases, lacks reliable verification results, and fails to establish real-time linkage between node control actions and contract execution status. Static credit assessments cannot adapt to the constantly changing performance status of the supply chain, and fragmented data cannot provide a unified data carrier for collaborative decision-making among alliance nodes.

[0003] The industry is unable to extract key performance data of specific contracts from the supply chain panoramic data stream and automatically verify and generate trusted execution event signals based on smart contract logic to trigger automated node control. It is also unable to perform deep feature extraction on updated resource status records to generate financial assessment feature vectors, calculate dynamic trust scores, and package them into standardized data assets to synchronize to authorized alliance nodes. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a supply chain platform data management system based on the Internet of Things.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a supply chain platform data management system based on the Internet of Things, comprising: The data acquisition module collects real-time physical status information of physical materials in the supply chain and generates a panoramic data stream of the supply chain. The contract execution module extracts key performance data related to specific supply chain contracts from the supply chain panoramic data stream, and performs automatic verification on the key performance data according to the preset smart contract logic. When the key performance data meets the triggering conditions of the smart contract logic, a reliable contract execution event signal is automatically generated. The node control module, based on the trusted contract execution event signal, triggers automated control commands for preset logistics nodes or warehousing nodes in the supply chain, and synchronously updates the resource status records of the corresponding nodes. The credit assessment module performs deep feature extraction on the updated resource status records to generate data feature vectors for supply chain finance assessment. The data feature vectors are compared and analyzed with the historical performance behavior feature database to calculate the dynamic credibility score of the current supply chain link. The data asset module packages the dynamic trust score and the associated key performance data to generate a standardized supply chain data asset package, and synchronizes the supply chain data asset package to authorized supply chain alliance nodes through a preset secure channel to drive collaborative decision-making within the alliance.

[0006] As a further aspect of the present invention, the real-time acquisition of physical state information of physical materials in the supply chain to generate a panoramic supply chain data stream includes: The physical status information of physical materials in the supply chain is collected in real time through IoT sensing devices, and the original material status stream is generated. It receives digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces and generates structured business information flow. Standardized fusion processing is performed on the raw material state flow and the structured business information flow to generate a unified supply chain panoramic data flow with spatiotemporal labels; The step of collecting physical state information of physical materials in the supply chain in real time through IoT sensing devices and generating the original material state flow specifically includes: In the warehousing process of the supply chain, RFID readers and gravity sensors are deployed on warehouse shelves, pallets and key aisles to capture inbound and outbound events of goods, inventory location and real-time weight changes, and generate warehousing perception information. In the transportation link of the supply chain, temperature and humidity sensors, vibration sensors and positioning modules are deployed inside the transportation vehicle to capture environmental parameters, transportation vibration trajectory and real-time geographical location of goods during the journey, and generate in-transit perception information. In the production or sorting process of the supply chain, visual recognition cameras and proximity sensors are deployed at key nodes of the production line to capture the flow status of materials, work progress and abnormal stop events, and generate work perception information. The warehouse sensing information, in-transit sensing information, and operation sensing information are encapsulated according to a unified Internet of Things data protocol, and each encapsulated data is added with a data acquisition device identifier, data acquisition timestamp, and location code to form an initial sensing data packet sequence. The initial sensing data packet sequence is subjected to integrity verification and timestamp sorting to form a time-sequential raw material state flow.

[0007] As a further aspect of the present invention, the step of receiving digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces and generating a structured business information flow includes: Define a set of standardized supply chain business data interfaces, which cover core business documents such as purchase orders, sales orders, delivery notes, warehouse receipt notes, invoices, and settlement notes; Through message queues or remote procedure calls, business data messages conforming to the interface specifications are received asynchronously from the supplier's business resource planning system, the distributor's customer relationship management system, and the carrier's transportation management system. The received business data messages are parsed and validated, and the business entities, business relationships, business events and key attribute values ​​are extracted. The unstructured message content is then transformed into an internally unified data model instance. Each data model instance is appended with its source system identifier, receiving timestamp, and associated business order number to form a standard structured business record; All received structured business records are sorted and indexed according to their business occurrence time or logical sequence to form a structured business information flow.

[0008] As a further aspect of the present invention, the step of performing standardized fusion processing on the original material state flow and the structured business information flow to generate a unified supply chain panoramic data flow with spatiotemporal labels includes: Establish material identification mapping rules to associate and map the specific material identifications collected by IoT devices in the original material status flow with the material codes recorded in business documents in the structured business information flow. Establish a spatiotemporal coordinate system, unify the equipment location codes in the raw material state flow and the logistics node codes in the structured business information flow, and map the equipment location codes in the raw material state flow and the logistics node codes in the structured business information flow to the same set of supply chain network coordinates; Based on the material identification mapping relationship and the unified supply chain network coordinates, the time window alignment and spatial location matching are performed on the perception data packet sequence in the original material state flow and the business records in the structured business information flow. The successfully matched perception data packet sequence is concatenated with the business record to generate a fusion record, which contains both the physical state data of the material and the associated business information data. Each fusion record is labeled with its corresponding material identifier, fusion time point, and specific location in the supply chain network coordinates, forming a unified, spatiotemporally labeled panoramic supply chain data stream.

[0009] As a further aspect of the present invention, the step of extracting key performance data related to specific supply chain contracts from the supply chain panoramic data stream and automatically verifying the key performance data according to preset smart contract logic includes: Analyze the digital terms of supply chain contracts, identify key performance events, performance conditions, performance standards and related constraint parameters stipulated in the terms, and convert them into computable state machine logic; From the unified, spatiotemporally labeled supply chain panoramic data stream, filter out the fused records related to the material identification, participants, and logistics routes involved in the supply chain contract; Based on the triggering events defined by the state machine logic, specific performance observation data is further extracted from the selected fusion records. The performance observation data includes actual delivery time, actual arrival location, actual environmental indicators, and actual quantity specifications. The extracted performance observation data is automatically compared with the performance standards and constraint parameters agreed in the state machine logic to determine whether the performance observation data meets the agreed condition range. The detailed process and results of the comparison are recorded as an immutable verification log, which serves as proof of the fulfillment of the key performance data.

[0010] As a further aspect of the present invention, the step of automatically generating a reliable contract execution event signal when the key performance data meets the triggering conditions of the smart contract logic includes: In the smart contract logic, multiple execution stage nodes are preset, and a set of performance conditions to be satisfied is associated with each execution stage node; The output of the verification log is monitored in real time. When the verification log confirms that the key performance data has reached or exceeded the performance condition set associated with a certain execution stage node, the triggering condition is determined to be met. Generate a standard format contract execution event signal, which includes at least a unique contract identifier, the execution stage node number that was triggered, the trigger timestamp, and a credential hash value pointing to the corresponding verification log. The generated contract execution event signal is published to the distributed event bus, and the contract execution event signal and its associated verification log credential hash value are stored in a distributed storage with tamper-proof characteristics.

[0011] As a further aspect of the present invention, the step of triggering automated control commands for preset logistics nodes or warehousing nodes in the supply chain based on the trusted contract execution event signal, and synchronously updating the resource status records of the corresponding nodes, includes: Maintain a predefined control rule base, where each rule defines the mapping relationship between a specific contract execution event signal and one or more automated control instructions; Listen to the distributed event bus and capture published contract execution event signals; Using captured contract execution event signals as input, the system performs pattern matching in the control rule base to find and trigger matching control rules. Execute the matching control rules to generate specific automated control instructions with operational semantics, including "open the designated warehouse door", "allow outbound", "sort to a specific channel", and "trigger payment application". The generated automated control commands are sent to designated logistics equipment or warehouse management subsystems for execution via the control network. After the instruction is issued and executed, the resource status records of the corresponding logistics node or warehousing node stored in the supply chain platform database are automatically modified according to the semantic logic of the instruction. The resource status records include inventory quantity, warehouse occupancy status, and equipment working status.

[0012] As a further aspect of the present invention, the step of performing deep feature extraction on the updated resource status record to generate a data feature vector for supply chain finance assessment includes: From the updated resource status records, extract multi-dimensional time-series data related to the target supply chain financing entity over a continuous period of time, including inventory turnover sequence, order delivery on-time rate sequence, and in-transit goods value change sequence; The multidimensional time-series data is subjected to feature engineering processing to calculate statistical features, trend features, stability features, and volatility features, generating a primary feature set; By combining the historical records of the contract execution event signals, relevant historical performance quality characteristics are extracted from the verification log, including the number of defaults, performance delay distribution, and accuracy of condition fulfillment. The primary feature set is fused with the historical performance quality features and input into a pre-trained deep feature encoding network; The deep feature encoding network performs dimensionality reduction and abstraction on the fused features through multi-layer nonlinear transformation, and finally outputs a fixed-dimensional, dense data feature vector, which is used to characterize the operational health and risk status of the supply chain links.

[0013] As a further aspect of the present invention, the step of comparing and analyzing the data feature vector with the historical performance behavior feature database to calculate the dynamic reliability score of the current supply chain link includes: Maintain a historical performance behavior feature database, which stores data feature vectors generated by different supply chain participants in different business scenarios and their corresponding actual performance outcome labels. Calculate the similarity between the currently generated data feature vector and each historical data feature vector in the historical performance behavior feature database; Select the feature vectors of several historical data with the highest similarity as the nearest neighbor sample set; Analyze the actual performance outcome label corresponding to each sample in the nearest neighbor sample set. The actual performance outcome label includes "full performance", "partial breach", and "serious breach". Based on the similarity weight of neighboring samples, a weighted vote is performed on the performance outcome labels to calculate the predicted probability of various performance outcomes occurring in the current supply chain link in the future. According to the preset scoring model, the predicted probability is mapped to a quantified score, which is the dynamic reliability score of the current supply chain link.

[0014] As a further aspect of the present invention, the dynamic trust score and the associated key performance data are packaged to generate a standardized supply chain data asset package, and the supply chain data asset package is synchronized to authorized supply chain alliance nodes through a preset secure channel to drive collaborative decision-making within the alliance, including the following steps: Create a standard data asset package structure template, which includes three parts: data packet header, body data, and data signature; The header of the data packet contains the unique identifier of the asset package, the generation time, the data validity period, the identifier of the supply chain link to which it belongs, and the list of target receiving alliance nodes. The main data encapsulates the dynamic trust score, the key performance data on which the dynamic trust score is calculated, and index information for the relevant verification logs that support the key performance data. The hash values ​​of the data packet header and the main data are signed using the private key of the data generator. The signature result is then filled into the data signature part to form a complete supply chain data asset package. Based on the target receiving alliance node list in the data packet header, the supply chain data asset package is distributed to each authorized supply chain alliance node through a peer-to-peer encrypted communication link or an alliance permissioned blockchain network. After receiving the supply chain data asset package, the alliance node can use the trusted data encapsulated therein to drive collaborative decision-making within the node. The collaborative decision-making includes shared adjustments to demand forecasts, joint optimization of safety stock, and automatic allocation of inventory across enterprises.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Key performance data related to specific supply chain contracts is extracted from the overall supply chain data stream. This data is then automatically verified according to pre-defined smart contract logic. When the key performance data meets the triggering conditions of the smart contract logic, a reliable contract execution event signal is automatically generated. Based on this reliable signal, automated control instructions are triggered at pre-defined logistics or warehousing nodes, simultaneously updating the resource status records of the corresponding nodes. The extraction and verification of performance data are entirely automated, strictly adhering to the pre-defined smart contract logic. The generated contract execution event signal possesses objective and reliable attributes. Automated control instructions are directly triggered by the reliable signal, establishing a real-time binding between node control actions and contract execution status. Updates to resource status records are synchronized with the execution of control instructions, ensuring a real-time correspondence between node resource status and contract performance status.

[0016] Deep feature extraction is performed on the updated resource status records to generate data feature vectors for supply chain finance assessment. These feature vectors are then compared and analyzed against a historical performance behavior feature database to calculate a dynamic credibility score for the current supply chain stage. This dynamic credibility score and associated key performance data are packaged into a standardized supply chain data asset package, which is then synchronized to authorized supply chain alliance nodes via a pre-defined secure channel. Deep feature extraction creates a dedicated data dimension adapted to supply chain finance assessment. The comparison and analysis of the feature vectors with the historical performance database directly outputs a dynamic credibility score, which is adjusted in real time according to resource status. The standardized data asset package integrates performance data and the credibility score into a unified data carrier. The pre-defined secure channel ensures the compliance of data synchronization, and authorized alliance nodes can directly conduct collaborative decision-making actions based on the standardized asset package. Attached Figure Description

[0017] Figure 1 This is a sequence diagram of the supply chain platform data management system based on the smart Internet of Things as described in this invention; Figure 2 A flowchart for receiving digital business information and generating a structured business information flow; Figure 3 A trend chart illustrating the co-evolution of supply chain dynamic credit scoring and on-time delivery rate; Figure 4 Heatmap of the changes in the status of core resources at each stage of the supply chain; Figure 5 Monthly trend chart for dynamic credibility score of the supply chain. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 This invention provides a supply chain platform data management system based on the Internet of Things (IoT), the system comprising: The data acquisition module is responsible for collecting real-time physical status information of physical materials in the supply chain, generating a panoramic supply chain data stream. The contract execution module extracts key performance data related to specific supply chain contracts from this panoramic data stream and automatically verifies the key performance data according to preset smart contract logic. When the key performance data meets the triggering conditions of the smart contract logic, a reliable contract execution event signal is automatically generated. The node control module, based on the reliable contract execution event signal, triggers automated control commands for preset logistics or warehousing nodes in the supply chain and synchronously updates the resource status records of the corresponding nodes. The credit assessment module performs deep feature extraction on the updated resource status records, generating a data feature vector for supply chain finance assessment. This data feature vector is compared and analyzed with a historical performance behavior feature database to calculate the dynamic reliability score of the current supply chain link. The data asset module packages the dynamic reliability score and associated key performance data to generate a standardized supply chain data asset package, and synchronizes the supply chain data asset package to authorized supply chain alliance nodes through a preset secure channel to drive collaborative decision-making within the alliance.

[0021] In one embodiment of the present invention, the physical state information of physical materials in the supply chain is collected in real time by IoT sensing devices, and a raw material state stream is generated. Digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces is received, and a structured business information stream is generated. Standardized fusion processing is performed on the raw material state stream and the structured business information stream to generate a unified, spatiotemporally labeled panoramic data stream of the supply chain.

[0022] The specific steps for collecting real-time physical state information of physical materials in the supply chain through IoT sensing devices and generating raw material state flow include: In the warehousing stage of the supply chain, RFID readers and gravity sensors are deployed on warehouse shelves, pallets, and key passages to capture inbound and outbound events, inventory locations, and real-time weight changes, generating warehousing sensing information. In the transportation stage of the supply chain, temperature and humidity sensors, vibration sensors, and positioning modules are deployed inside transport vehicles to capture environmental parameters, transportation vibration trajectories, and real-time geographical locations of goods during transit, generating in-transit sensing information. In the production or sorting stage of the supply chain, visual recognition cameras and proximity sensors are deployed at key nodes of the production line to capture the flow status of materials, work progress, and abnormal stop events, generating work sensing information. The warehousing sensing information, in-transit sensing information, and work sensing information are encapsulated according to a unified IoT data protocol, and each encapsulated data is added with a data acquisition device identifier, data acquisition timestamp, and location code to form an initial sensing data packet sequence. The initial sensing data packet sequence is then subjected to integrity verification and timestamp sorting to form a time-sequential raw material state flow.

[0023] In practical implementation, the data acquisition module of the supply chain platform data management system based on the Internet of Things (IoT) generates a panoramic supply chain data stream. Its implementation involves the collection, processing, and fusion of multi-source data. Specifically, IoT sensing devices deployed at various stages of the supply chain collect real-time physical state information of physical materials. The raw information collected by these devices is aggregated and processed to form a continuous raw material state stream. Furthermore, business systems at upstream and downstream nodes of the supply chain receive digitized business information through predefined open interfaces. This information is parsed and structured to form an ordered structured business information stream. Finally, a standardized fusion processing flow is executed on the raw material state stream and the structured business information stream, ultimately outputting a unified panoramic supply chain data stream where each data unit has a clear spatiotemporal label.

[0024] In some embodiments, the step of generating the original material state flow includes the coordinated operation of multiple stages. In the warehousing stage of the supply chain, RFID readers and gravity sensors are deployed on warehouse shelves, pallets, and key aisles. RFID readers capture identification information from labels affixed to goods or pallets, and gravity sensors detect weight changes in pallet or shelf areas. These devices work together to capture inbound and outbound events, inventory locations, and real-time weight changes, generating warehousing awareness information. In the transportation stage of the supply chain, temperature and humidity sensors, vibration sensors, and positioning modules are fixedly deployed inside transport vehicles. Temperature and humidity sensors continuously monitor environmental parameters inside the cargo compartment, vibration sensors record the amplitude and frequency of vibrations during transportation, and positioning modules periodically report geographic coordinates. These devices work together to capture environmental parameters, transportation vibration trajectories, and real-time geographic locations of goods during transit, generating in-transit awareness information. In the production or sorting stage of the supply chain, visual recognition cameras and proximity sensors are installed at key nodes of the production line. Visual recognition cameras identify the identification or appearance of passing materials, and proximity sensors detect whether materials have reached designated workstations. These devices work together to capture the flow status of materials, work progress, and abnormal dwell events, generating operational awareness information. The IoT gateway encapsulates data from warehouse sensing, in-transit sensing, and operational sensing according to a unified IoT data protocol. The encapsulation process adds a data acquisition device identifier, acquisition timestamp, and location code to each data entry, forming an initial sequence of sensing data packets. The data processing server performs integrity verification on this initial sequence of sensing data packets. Integrity verification is achieved by verifying the checksum of the data packets. The formula for calculating the checksum is: in: Represents the final checksum value. Represents the first in the data packet byte data, It is a predefined mask constant. It is the total number of bytes in the data packet. This indicates a bitwise XOR operation. After integrity verification and timestamp sorting are completed, the data processing server outputs a sequentially continuous stream of the original material status.

[0025] Optionally, the system receives digital business information sent by business systems from upstream and downstream nodes in the supply chain via open interfaces. This implementation relies on standardized interface definitions and asynchronous communication mechanisms. Essentially, a set of standardized supply chain business data interfaces is predefined, with interface specifications covering the data fields and formats of core business documents such as purchase orders, sales orders, delivery notes, receiving notes, invoices, and settlement statements. The message middleware asynchronously retrieves business data messages conforming to the interface specifications from the supplier's business resource planning system, the distributor's customer relationship management system, and the carrier's transportation management system via message queues or remote procedure calls. The interface adaptation service parses the format of each received business data message according to the template defined in the interface specification. Data validation is then performed, including value range checks and logical correlation checks. After passing the checks, the business entities, business relationships, business events, and key attribute values ​​in the message are extracted, transforming this unstructured content into an instance of a unified data model within the system. After the transformation process, each data model instance is appended with its source system identifier, receiving timestamp, and associated business document number, forming a standard structured business record. The indexing service sorts all successfully received structured business records according to their business occurrence time or logical sequence, and creates an index for the sorted records, ultimately forming a structured business information flow.

[0026] It is understandable that the standardized fusion of the raw material state flow and the structured business information flow is crucial for generating high-quality panoramic data. In practical implementation, the system maintains a material identification mapping rule library. This library defines the correspondence between RFID tag codes, visual identification codes, and material codes in the business system. Based on these mapping rules, the specific material identifiers collected by IoT devices in the raw material state flow are associated and mapped with the material codes recorded in business documents in the structured business information flow. In practical implementation, the system defines a unified spatiotemporal coordinate system. This system uniformly maps the latitude and longitude of the physical world, warehouse shelf codes, and transportation route node codes to virtual supply chain network coordinates. Through coordinate transformation, the device location codes in the raw material state flow and the logistics node codes in the structured business information flow are mapped to the same set of supply chain network coordinates. Based on the established material identification mapping relationship and unified supply chain network coordinates, the fusion engine performs time window alignment on the perception data packet sequence in the original material state stream and the business records in the structured business information stream. Time window alignment is based on the theoretical occurrence time of the business event, extended forward and backward by a fixed time tolerance window. Simultaneously, spatial location matching is performed, requiring the distance between the material's network coordinates and the business node's coordinates to be less than a set threshold. The fusion engine then concatenates the perception data packet sequence and business records that have successfully matched in both time and space, generating a new fusion record. This fusion record contains both the material's physical state data field and the associated business information data field. Each generated fusion record is labeled with its corresponding material identifier, fusion time point, and specific location in the supply chain network coordinates. The system continuously outputs this spatiotemporally labeled panoramic supply chain data stream.

[0027] In one embodiment of the present invention, receiving digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces, and generating a structured business information flow, includes: (see reference) Figure 2This system defines a standardized set of supply chain business data interfaces, covering core business documents such as purchase orders, sales orders, delivery notes, receiving notes, invoices, and settlement statements. Business data messages conforming to the interface specifications are asynchronously received from the supplier's Business Resource Planning (BRP) system, the distributor's Customer Relationship Management (CRM) system, and the carrier's Transportation Management (TRM) system via message queues or remote procedure calls. The received business data messages are parsed and validated, extracting business entities, business relationships, business events, and key attribute values, transforming the unstructured message content into internally unified data model instances. Each data model instance is appended with its source system identifier, receiving timestamp, and associated business document number, forming a standard structured business record. All received structured business records are sorted and indexed according to their business occurrence time or logical sequence, forming a structured business information flow.

[0028] Standardized fusion processing is performed on the raw material state flow and structured business information flow to generate a unified, spatiotemporally labeled panoramic supply chain data flow. This includes: establishing material identification mapping rules to associate the specific material identifiers collected by IoT devices in the raw material state flow with the material codes recorded in business documents in the structured business information flow; establishing a spatiotemporal coordinate system to unify the device location codes in the raw material state flow and the logistics node codes in the structured business information flow, mapping them to the same set of supply chain network coordinates; based on the material identification mapping relationship and the unified supply chain network coordinates, aligning the sensing data packet sequences in the raw material state flow with the business records in the structured business information flow using time windows and matching their spatial locations; concatenating the successfully matched sensing data packet sequences with the business records to generate a fused record, which simultaneously contains the physical state data of the materials and the associated business information data; and labeling each fused record with its corresponding material identifier, fusion time point, and specific location in the supply chain network coordinates, forming a unified, spatiotemporally labeled panoramic supply chain data flow.

[0029] In practical implementation, the data acquisition module of the supply chain platform data management system based on the Internet of Things (IoT) receives and processes digital information from external business systems. This relies on a standardized interface specification and asynchronous communication mechanism. Defining a set of standardized supply chain business data interfaces is a prerequisite. These interfaces clearly define the data structure, field types, constraints, and transmission protocols for core business documents such as purchase orders, sales orders, delivery notes, receiving notes, invoices, and settlement statements. The message middleware establishes connections with the supplier's business resource planning system, the distributor's customer relationship management system, and the carrier's transportation management system via message queues or remote procedure calls, asynchronously listening to or retrieving business data messages conforming to the interface specifications. The interface adaptation service parses the format of each received business data message. Based on the interface-defined schema file, the parser deconstructs the original message and then performs data validation. Data validation includes checks on the completeness of required fields, data type compliance, and business logic consistency. After passing the checks, the business entities, business relationships, business events, and key attribute values ​​are extracted from the message. In practice, the extracted information is instantiated into an object of a unified data model within the system. Each data model instance is appended with its source system identifier, receiving timestamp, and associated business order number, forming a standard structured business record that can be directly used by other modules within the system. The indexing service sorts all received structured business records according to the business occurrence time recorded on their business documents. If the times are the same, they are processed according to the receiving order, and an inverted index is built for this sequence based on a composite key of time range and business order number, ultimately forming a continuous and queryable stream of structured business information.

[0030] In some embodiments, the standardized fusion processing performed on the raw material state flow and the structured business information flow is a core step in aligning physical information with business information. In a specific implementation, the system maintains a material identification mapping rule base. This rule base stores the mapping relationships between RFID tag codes, barcodes, or QR codes and material master data codes in the form of configuration tables. During data processing, this mapping relationship is used to associate and map the specific material identifiers collected by IoT devices in the raw material state flow with the material codes recorded in business documents in the structured business information flow. In a specific implementation, the system defines a globally unified spatiotemporal coordinate system with geographic latitude and longitude as the origin. All logistics node codes, such as warehouses, ports, and sorting centers, are mapped to coordinate nodes in the network. A coordinate transformation service maps the device location codes in the raw material state flow and the logistics node codes in the structured business information flow to the same set of supply chain network coordinates. Based on the established material identification mapping relationship and unified supply chain network coordinates, the fusion engine performs time window alignment and spatial location matching operations. The time window alignment operation centers on the theoretical time point of the business record in the structured business information flow, expanding forward and backward by a configurable time tolerance window ΔT. Spatial location matching requires that the logical distance between the physical perception coordinates and the business node coordinates is less than a threshold θ. The formula for calculating the logical distance D is: in: Represents logical distance. and This represents the horizontal and vertical coordinates of the perceived data packet in the supply chain network coordinate system. and This represents the horizontal and vertical coordinates of a business node within the supply chain network. When both time window alignment and spatial location matching conditions are met, the fusion engine concatenates the successfully matched perception data packet sequence with the corresponding business record to generate a fused record. This fused record contains both the physical status data field of the material and the associated business information data field. The fusion engine labels each successfully generated fused record with its corresponding material identifier, fusion time point, and specific location within the supply chain network coordinates. The system continuously outputs this spatiotemporally labeled panoramic supply chain data stream in a streaming processing manner.

[0031] Optionally, the asynchronous mechanism for receiving data from the business system can be implemented through different message patterns. It can be understood that the message queue adopts a publish-subscribe pattern, where the upstream business system encapsulates business events into messages and publishes them to specific topics. The message middleware of the data acquisition module subscribes to these topics and consumes the messages. Optionally, the remote procedure call method adopts a request-response pattern, where the interface service of the data acquisition module actively initiates a call request to the application interface of the upstream business system to pull data. In some embodiments, data validation of business data messages includes multiple levels: format parsing validation ensures that the message structure conforms to the interface specification definition; value range validation ensures that numeric fields are within a preset range and enumerated fields are valid values; and logical correlation validation ensures that the constraint relationships between different fields are valid. It can be understood that the design of the spatiotemporal coordinate system allows for dynamic expansion; new logistics nodes or transportation routes can be registered and assigned unique network coordinates, thereby ensuring the continuous consistency of the coordinate system as the supply chain network expands.

[0032] In one embodiment of the present invention, the steps of extracting key performance data related to a specific supply chain contract from the supply chain panoramic data stream and automatically verifying the key performance data according to preset smart contract logic include: parsing the digital terms of the supply chain contract, identifying the key performance events, performance conditions, performance standards, and related constraint parameters stipulated in the terms, and converting them into computable state machine logic. From a unified supply chain panoramic data stream with spatiotemporal labels, fused records related to material identification, participants, and logistics paths involved in the supply chain contract are selected. Based on the trigger events defined by the state machine logic, specific performance observation data is further extracted from the selected fused records. The performance observation data includes actual delivery time, actual arrival location, actual environmental indicators, and actual quantity specifications. The extracted performance observation data is automatically compared with the performance standards and constraint parameters stipulated in the state machine logic to determine whether the performance observation data meets the agreed conditions. The detailed process and results of the comparison are recorded as an immutable verification log, which serves as proof of the fulfillment of the key performance data.

[0033] The steps for automatically generating a reliable contract execution event signal when key performance data meets the triggering conditions of the smart contract logic include: First, pre-setting multiple execution stage nodes in the smart contract logic and associating each execution stage node with a set of performance conditions to be met. Second, monitoring the output of the verification log in real time; when the verification log confirms that the key performance data has reached or exceeded the set of performance conditions associated with a certain execution stage node, determining that the triggering condition is met. Third, generating a standard-format contract execution event signal, which at least includes a unique contract identifier, the execution stage node number that triggered the condition, a trigger timestamp, and a credential hash value pointing to the corresponding verification log. Fourth, publishing the generated contract execution event signal to a distributed event bus, and simultaneously storing the contract execution event signal and its associated verification log credential hash value in a tamper-proof distributed storage system.

[0034] In practical implementation, the contract execution module of the supply chain platform data management system based on the Internet of Things extracts and verifies key performance data from the supply chain panoramic data stream. Parsing the digital terms of the supply chain contract is the initial step. The contract parsing engine transforms the contract terms described in natural language or structured text into computable state machine logic. The state machine logic clearly identifies and defines the key performance events, performance conditions, performance standards, and related constraint parameters stipulated in the terms. Key performance events include "goods arrive at the destination warehouse," performance conditions include "temperature between 2-8 degrees Celsius," performance standards include "delivery time no later than the agreed time," and constraint parameters include specific temperature values ​​and time thresholds. In practical implementation, the data filtering unit of the contract execution module filters all fused records related to a specific supply chain contract from the unified, spatiotemporally labeled supply chain panoramic data stream, based on material identification, participant identification, and predefined logistics path node sequences. According to the trigger events defined by the state machine logic, the verification calculation unit further extracts specific performance observation data from the filtered fused records. Performance observation data includes the actual delivery timestamp, the actual latitude and longitude coordinates of arrival, the actual environmental indicator readings, and the actual quantity and specifications.

[0035] In some embodiments, the extracted performance observation data is automatically compared with the performance standards and constraint parameters agreed upon in the state machine logic. This process involves numerical calculation and logical judgment. The verification calculation unit compares the performance observation data with the contract standard values ​​to determine whether the performance observation data meets the agreed condition range, such as whether the actual temperature reading is continuously within the temperature range specified in the contract, or whether the actual delivery time is earlier than the latest time specified in the contract. The detailed process and results of the comparison are recorded as an immutable verification log. The verification log includes the contract identifier, comparison time, performance observation data source used, comparison rules, calculation process, judgment result, and snapshot hash value of the data state at the judgment time. This verification log is stored as proof of the satisfaction of key performance data. In specific implementations, when the key performance data meets the triggering conditions of the smart contract logic, a reliable contract execution event signal is automatically generated. The smart contract logic is pre-set with multiple execution stage nodes during deployment. Each execution stage node is associated with a set of performance conditions to be satisfied. The set of performance conditions may contain multiple performance conditions that need to be satisfied simultaneously. The event generation service monitors the output of the verification log in real time. When the verification log confirms that the key performance data has reached or exceeded the performance condition set associated with a certain execution stage node, the event generation service determines that the trigger condition is met.

[0036] It is understandable that generating a standard-format contract execution event signal contains fixed information elements. The event generation service generates a standard-format contract execution event signal, which includes at least a unique contract identifier, the execution stage node number that triggered the event, a trigger timestamp, and a credential hash value pointing to the corresponding verification log. The formula used to calculate the credential hash value H of the verification log pointed to by the contract execution event signal is: in: This represents the hash value of the final generated voucher. Represents a cryptographic hash function. This represents the binary content of the complete verification log associated with the triggered event. This represents the execution phase node number that triggers the fulfillment of the condition. Represents the trigger timestamp. This represents a string concatenation operation. The event dispatch component publishes the generated contract execution event signal to the distributed event bus, which broadcasts the signal to all listening services subscribed to the relevant topic via topic publication. Optionally, the event persistence component also stores the contract execution event signal and its associated verification log credential hash value in a tamper-proof distributed storage. The tamper-proof feature is implemented through distributed ledger technology or a data storage service with integrity protection, ensuring that the record cannot be unilaterally modified afterward.

[0037] Optionally, the decision logic for the performance condition set can be a Boolean combination of multiple conditions. In some embodiments, the performance condition set of an execution phase node may require both the conditions "actual temperature is within the range" and "actual location has been reached" to be true simultaneously. Optionally, the immutability of the verification log can be achieved by anchoring its hash value to a blockchain transaction or by signing the log using digital signature technology. It can be understood that the distribution of contract execution event signals is a driving event of the collaboration of various modules within the system. The distributed event bus, as a central conduit, ensures that the signals can be reliably and timely consumed by subscribers such as the node control module and the credit assessment module.

[0038] See Figure 3In the data management system of the supply chain platform based on the Internet of Things, the synergistic evolution trend of dynamic credit score and on-time performance rate intuitively reflects the strong coupling relationship between contract execution quality and the credibility level of supply chain entities. From a time-series perspective, the two curves show a significant co-directional fluctuation characteristic: the peak of the on-time performance rate (box curve) (3-20, approximately 98.5%) and the peak of the dynamic credit score (dot curve) (3-20, approximately 95 points) are highly synchronized; while the trough of the on-time performance rate (3-10, approximately 92%) also corresponds to the stage low point of the credit score (3-10, approximately 85 points). This verifies the core logic in the credit assessment module that "performance behavior characteristics directly drive the generation of dynamic credibility score," that is, historical performance quality (on-time performance rate) is a key input feature for calculating the operational health and risk status of supply chain links. Further analysis of the causes of the fluctuations reveals that from March 1st to March 10th, the on-time performance rate declined from approximately 98% to 92%, and the credit score simultaneously dropped from approximately 92 to 85. This decline can be attributed to the fact that the performance observation data verified by the contract execution module (such as delivery delays and deviations in environmental indicators) did not meet the triggering conditions of the smart contract logic, resulting in the node control module failing to optimize resource status in a timely manner, which in turn lowered the dynamic credibility score output by the credit assessment module. From March 10th to March 20th, the on-time performance rate continued to rise to its peak, and the credit score also climbed to 95, indicating that the key performance data met the multi-stage execution conditions of the smart contract. The contract execution event signals effectively drove the automated control of nodes and the updating of resource status. The high-quality data feature vector generated after deep feature extraction obtained a higher similarity weighted score when compared with the historical performance behavior feature library. Between March 20th and March 30th, both indicators declined: the on-time performance rate dropped from 98.5% to approximately 97%, and the credit score fell from 95 to approximately 91. This fluctuation reflects the real-time nature and sensitivity of the dynamic credit score, meaning the system continuously updates data feature vectors and credibility scores based on the latest performance verification logs and resource status records, enabling dynamic perception of supply chain risks. Overall, the diagram clearly presents the closed-loop effect of the business process: "contract execution → node control → credit assessment." The on-time performance rate is a direct externalization of contract execution quality, while the dynamic credit score is a quantitative abstraction of performance behavior and resource status. The synergistic trend between the two provides reliable data asset support for collaborative decision-making among supply chain alliance nodes (such as safety stock optimization and cross-enterprise allocation).

[0039] In one embodiment of the present invention, the step of triggering automated control commands for preset logistics nodes or warehousing nodes in the supply chain based on trusted contract execution event signals, and synchronously updating the resource status records of the corresponding nodes, includes: maintaining a predefined control rule base, wherein each rule defines a mapping relationship between a specific contract execution event signal and one or more automated control commands; monitoring a distributed event bus to capture published contract execution event signals; using the captured contract execution event signals as input, performing pattern matching in the control rule base to find and trigger matching control rules; executing the matching control rules to generate specific automated control commands with operational semantics, including "opening a designated warehouse door," "allowing outbound shipment," "sorting to a specific channel," and "triggering a payment request"; and distributing the generated automated control commands to designated logistics equipment or warehousing management subsystems for execution via a control network. After the commands are issued and executed, the resource status records of the corresponding logistics nodes or warehousing nodes stored in the supply chain platform database are automatically modified according to the semantic logic of the commands. The resource status records include inventory quantity, warehouse occupancy status, and equipment operating status.

[0040] The steps for deep feature extraction from updated resource status records to generate data feature vectors for supply chain finance assessment include: extracting multi-dimensional time-series data related to the target supply chain financing entity within a continuous time period from the updated resource status records. This multi-dimensional time-series data includes inventory turnover sequences, order delivery on-time rate sequences, and in-transit goods value change sequences. Feature engineering is performed on the multi-dimensional time-series data to calculate statistical features, trend features, stability features, and volatility features, generating a primary feature set. Historical performance quality features, including the number of defaults, performance delay distribution, and condition fulfillment accuracy, are extracted from the verification logs by combining historical contract execution event signals. The primary feature set and historical performance quality features are then fused and input into a pre-trained deep feature encoding network. The deep feature encoding network performs dimensionality reduction and abstraction on the fused features through multiple nonlinear transformations, ultimately outputting a fixed-dimensional, dense data feature vector. This data feature vector is used to characterize the operational health and risk status of the supply chain links.

[0041] In practical implementation, the node control module of the supply chain platform data management system based on the Internet of Things (IoT) triggers control commands based on contract execution event signals. The node control module maintains a predefined control rule base, which is stored in the database in the form of rule tables. Each rule clarifies the mapping relationship between a contract execution event signal and one or more automated control commands. See Table 1 for an example of the control rule base.

[0042] Table 1: Control Rule Mapping Table In practical implementation, the event listener of the node control module continuously monitors the distributed event bus, capturing various contract execution event signals issued by the contract execution module. The event listener uses these captured contract execution event signals as input and performs pattern matching in the control rule base. Pattern matching is based on precise searches using key fields such as the unique contract identifier and the execution stage node number contained in the contract execution event signal. The rule engine executes the successfully matched control rules, generating specific automated control instructions with operational semantics. These automated control instructions include "open the specified warehouse door," "allow outbound shipment," "sort to a specific channel," and "trigger payment application," each with necessary target parameters. The instruction dispatcher distributes the generated automated control instructions to the designated logistics equipment or warehouse management subsystem for execution via a secure control network. After the instructions are issued and executed, the state synchronizer automatically modifies the resource status records of the corresponding logistics node or warehouse node stored in the supply chain platform database according to the semantic logic of the instructions. The resource status records include fields such as inventory quantity, warehouse occupancy status, and equipment operating status.

[0043] In some embodiments, the credit assessment module performs deep feature extraction on the updated resource status records. The feature extraction module extracts multi-dimensional time-series data related to the target supply chain financing entity over a continuous period from the updated resource status records. This multi-dimensional time-series data includes inventory turnover sequences, order delivery on-time rate sequences, and in-transit goods value change sequences. The feature engineering processor performs feature engineering on the multi-dimensional time-series data, calculating statistical features, trend features, stability features, and volatility features. Statistical features include mean and variance; trend features are represented by the slope of a linear fit; stability features are obtained by calculating the coefficient of variation; and volatility features are obtained by calculating the standard deviation within a certain time window, generating a primary feature set. The feature engineering processor combines historical records of contract execution event signals to extract associated historical performance quality features from the stored verification logs. These historical performance quality features include the number of defaults, performance delay distribution, and condition fulfillment accuracy. The feature fusion component fuses the primary feature set with the historical performance quality features to form a joint feature vector containing both time-series features and performance quality features, which is then input into a pre-trained deep feature encoding network.

[0044] It is understandable that pre-trained deep feature encoding networks abstract features through multiple layers of nonlinear transformations. A deep feature encoding network is a multi-layer neural network whose input layer receives a joint feature vector, which undergoes nonlinear transformations through multiple fully connected layers and activation functions. The hidden layers progressively reduce the dimensionality and abstract the fused features. The final layer outputs a fixed-dimensional, dense data feature vector, which represents the operational health and risk status of the supply chain. The formula for calculating the output of a fully connected layer in a deep feature encoding network is: in: Representing the The output vector of the layer, Represents the activation function. Representing the The weight matrix of the layer, Representing the The output vector of the layer (i.e., the input of this layer). Representing the The bias vector of the layer. The network learns the weights and bias parameters through training, so that the final output data feature vector can effectively retain the key information in the original multidimensional data.

[0045] See Figure 4The heatmap presents the dynamic changes and numerical distribution characteristics of five core resource status indicators—inventory quantity, warehouse occupancy rate, equipment operating status, order fulfillment rate, and goods integrity rate—from stage 1 to stage 6. From the color scheme and numerical distribution of the heatmap, the gradient trend from light to dark is positively correlated with the indicator values, intuitively reflecting that each indicator shows a continuous upward trend as the supply chain progresses. Specifically, equipment operating status reaches its maximum value of 100 in stage 6, corresponding to the darkest shade in the heatmap, representing the optimal level of resource status at this stage. Specific indicator analysis shows that inventory quantity gradually increased from 85 in stage 1 to 98 in stage 6, with steady improvement in each stage, reflecting the continuous optimization and stability of inventory resource allocation at the supply chain warehousing end; warehouse occupancy rate increased from 75 in stage 1 to 96 in stage 6, and although it was on an overall upward trend, the relatively low base in stage 1 reflects the characteristic that the efficiency of warehouse space resource utilization gradually improves with the progress of the supply chain; equipment operating status started from 92 in stage 1 and continuously jumped to 100 in stage 6, making it the only dimension among all indicators to reach the maximum value, and the increase was stable in each stage, highlighting the high reliability and continuous operation of equipment at the production / execution end of the supply chain; order fulfillment rate gradually improved from 65 in stage 1 to 95 in stage 6, with a significant increase in the early stage and a slower growth rate in the later stage, reflecting the transition of the supply chain's contract fulfillment capability from an initial imperfect state to a mature and stable state; and the goods integrity rate increased from 88 in stage 1 to 99 in stage 6, maintaining a high growth rate throughout, reflecting the continuous strengthening of the supply chain's goods quality control capabilities at the logistics end. Heatmaps, through a dual presentation of numerical gradients and color levels, quantitatively reveal the evolution patterns of the five core resource states—inventory, warehousing, equipment, fulfillment, and logistics—at each stage of the supply chain. They verify the overall optimization trend of resource states throughout the entire supply chain process as stages progress, and also provide visualized data support for real-time monitoring of node states, supply chain risk identification, and resource scheduling optimization.

[0046] In one embodiment of the present invention, the step of comparing and analyzing data feature vectors with a historical performance behavior feature database to calculate the dynamic reliability score of the current supply chain link includes: maintaining a historical performance behavior feature database, which stores data feature vectors generated by different supply chain participants in different business scenarios and their corresponding actual performance outcome labels. Calculating the similarity between the currently generated data feature vector and each historical data feature vector in the historical performance behavior feature database. Selecting several historical data feature vectors with the highest similarity as a nearest neighbor sample set. Analyzing the actual performance outcome label corresponding to each sample in the nearest neighbor sample set, where the actual performance outcome label includes "full performance," "partial default," and "serious default." Based on the similarity weight of the nearest neighbor samples, weighted voting is performed on their performance outcome labels to calculate the predicted probability of various performance outcomes occurring in the current supply chain link within a future period. According to a preset scoring model, the predicted probability is mapped to a quantified score, which is the dynamic reliability score of the current supply chain link.

[0047] The process involves packaging dynamic trust scores and associated key performance data to generate standardized supply chain data asset packages. These packages are then synchronized to authorized supply chain alliance nodes via a pre-defined secure channel to drive collaborative decision-making within the alliance. The steps include: creating a standard data asset package structure template, comprising a data header, main data, and a data signature. The data header includes a unique identifier for the asset package, generation time, data validity period, the identifier of the supply chain segment it belongs to, and a list of target receiving alliance nodes. The main data encapsulates the dynamic trust score, the key performance data upon which the dynamic trust score is based, and index information for relevant verification logs supporting the key performance data. The hash values ​​of the data header and main data are signed using the data generator's private key, and the signature result is entered into the data signature section, forming a complete supply chain data asset package. Based on the target receiving alliance node list in the data header, the supply chain data asset package is distributed to each authorized supply chain alliance node via a peer-to-peer encrypted communication link or an alliance-permissioned blockchain network. After receiving the supply chain data asset package, alliance nodes can use the trusted data encapsulated within it to drive collaborative decision-making within the node. Collaborative decision-making includes shared adjustments to demand forecasts, joint optimization of safety stock, and automatic allocation of inventory across enterprises.

[0048] In practical implementation, the credit assessment module of the supply chain platform data management system based on the Internet of Things compares and analyzes data feature vectors with a historical performance behavior feature database. This database, a dedicated database or feature storage system, stores data feature vectors generated by different supply chain participants in various business scenarios and their corresponding actual performance outcome labels. The scoring service calculates the similarity between the currently generated data feature vector and each historical data feature vector in the database. Similarity calculation can use cosine similarity, the reciprocal of Euclidean distance, or other metrics. The scoring service selects several historical data feature vectors from the database that have the highest similarity to the current data feature vector; these selected historical data feature vectors form a nearest neighbor sample set. The scoring service then analyzes the actual performance outcome label corresponding to each sample in the nearest neighbor sample set. These labels include categories such as "full performance," "partial default," and "serious default." Based on the similarity weights of neighboring samples, the scoring service performs weighted voting on the fulfillment outcome labels of all samples in the neighboring sample set to calculate the predicted probability of various fulfillment outcomes occurring in the current supply chain segment within a future period. According to a pre-defined scoring model, the scoring service maps the calculated predicted probabilities of various fulfillment outcomes into a single, quantified score, which is the dynamic reliability score of the current supply chain segment.

[0049] In some embodiments, the specific process of calculating similarity involves a metric in vector space. The scoring service calculates the cosine similarity between the current data feature vector and the historical data feature vector. The cosine similarity reflects the difference in direction between the two vectors. The calculation formula is: in: Represents the current data feature vector With a certain historical data feature vector Cosine similarity between them Dot product operation representing vectors, The magnitude of the current data feature vector. This represents the magnitude of the historical data feature vector. The scoring service sorts all historical samples in descending order based on the calculated similarity scores, selecting the top K historical data feature vectors as the nearest neighbor sample set. This can be understood as a weighted voting process assigning a weight to each nearest neighbor sample; the weight is typically determined by the similarity between the sample and the current sample—higher similarity results in a larger weight. Subsequently, a weighted count of each actual performance outcome label in the nearest neighbor sample set is calculated, and this weighted count is normalized to a predicted probability for each performance outcome. The scoring model can be a linear function or a non-linear mapping function; the input is a predicted probability vector, and the output is a score within a predetermined range.

[0050] In practical implementation, the data asset module packages dynamic trust scores and associated key performance data into standardized supply chain data asset packages. The packaging service of the data asset module creates a standard data asset package structure template, defined in memory as a data structure containing a data packet header, body data, and data signature. The packaging service writes the unique asset package identifier, generation time, data validity period, identification of the supply chain segment to which it belongs, and a list of target receiving alliance nodes into the data packet header. The packaging service encapsulates the dynamic trust score, the key performance data on which the dynamic trust score is based, and index information of relevant verification logs supporting the key performance data into the body data. The signature service uses the private key of the data generator to perform a hash operation on the combination of the data packet header and body data, then digitally signs the hash value, and fills the signature result into the data signature section of the data asset package, thus forming a complete supply chain data asset package with integrity and non-repudiation. The distribution service distributes the complete supply chain data asset package to each authorized supply chain alliance node in the list via a peer-to-peer encrypted communication link or a permissioned blockchain network, based on the target receiving alliance node list in the data packet header.

[0051] Optionally, the historical performance behavior feature database is periodically updated with new performance data and its outcomes to ensure the timeliness and accuracy of the comparative analysis. Optionally, the number K of nearest neighbor samples can be dynamically adjusted according to the size of the total sample database. In some embodiments, the scoring model can be a lookup function based on expert rule definitions, mapping the predicted probabilities of different intervals to different score segments. It is understood that the data signature portion of the data asset package may include a signature algorithm identifier, a signature value, and optional data generator certificate information. In a consortium permissioned blockchain network, the distribution service can leverage smart contracts to implement the distribution of asset packages and access control. After receiving the supply chain data asset package, consortium nodes can utilize the trusted data encapsulated within to drive collaborative decision-making within the node. Collaborative decision-making includes shared adjustments to demand forecasts, joint optimization of safety stock, and automatic allocation of cross-enterprise inventory.

[0052] See Figure 5 In the credit assessment system of the supply chain platform based on the Internet of Things (IoT), the monthly trend of the dynamic trust score intuitively reflects the temporal changes in the operational health and performance risks of the supply chain links. Specifically, from February to June 2025, the trust score remained in the range of 64-73 points, showing slight fluctuations. This period corresponds to a stable period of basic supply chain performance status, with a high degree of matching between resource status records and the historical performance behavior feature database, and no significant risk events occurred. From July 2025, the score entered a rapid upward channel, reaching a peak of 88 points in October 2025. This leap is highly correlated with the high-frequency positive triggering of contract execution event signals and the continuous optimization of resource status records, reflecting the role of smart contract automatic verification and node automated control in improving performance quality. After November 2025, the score slightly declined but remained at a relatively high level of 77-86 points, reflecting that the overall trust level of the supply chain data asset package has entered a new stage of stable improvement driven by collaborative decision-making among alliance nodes. This trend curve can serve as a core quantitative basis for supply chain finance assessment, risk warning, and alliance collaboration strategy formulation. Its fluctuation characteristics are directly related to the fusion quality of IoT sensing data and structured business information flow, the execution efficiency of smart contracts, and the effect of deep feature encoding.

[0053] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A supply chain platform data management system based on the Internet of Things, characterized in that, include: The data acquisition module collects real-time physical status information of physical materials in the supply chain and generates a panoramic data stream of the supply chain. The contract execution module extracts key performance data related to specific supply chain contracts from the supply chain panoramic data stream, and performs automatic verification on the key performance data according to the preset smart contract logic. When the key performance data meets the triggering conditions of the smart contract logic, a reliable contract execution event signal is automatically generated. The node control module, based on the trusted contract execution event signal, triggers automated control commands for preset logistics nodes or warehousing nodes in the supply chain, and synchronously updates the resource status records of the corresponding nodes. The credit assessment module performs deep feature extraction on the updated resource status records to generate data feature vectors for supply chain finance assessment. The data feature vectors are compared and analyzed with the historical performance behavior feature database to calculate the dynamic credibility score of the current supply chain link. The data asset module packages the dynamic trust score and the associated key performance data to generate a standardized supply chain data asset package, and synchronizes the supply chain data asset package to authorized supply chain alliance nodes through a preset secure channel to drive collaborative decision-making within the alliance.

2. The supply chain platform data management system based on the Internet of Things as described in claim 1, characterized in that, The real-time acquisition of physical state information of physical materials in the supply chain to generate a panoramic supply chain data stream includes: The physical status information of physical materials in the supply chain is collected in real time through IoT sensing devices, and the original material status stream is generated. It receives digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces and generates structured business information flow. Standardized fusion processing is performed on the raw material state flow and the structured business information flow to generate a unified supply chain panoramic data flow with spatiotemporal labels; The step of collecting physical state information of physical materials in the supply chain in real time through IoT sensing devices and generating the original material state flow specifically includes: In the warehousing process of the supply chain, RFID readers and gravity sensors are deployed on warehouse shelves, pallets and key aisles to capture inbound and outbound events of goods, inventory location and real-time weight changes, and generate warehousing perception information. In the transportation link of the supply chain, temperature and humidity sensors, vibration sensors and positioning modules are deployed inside the transportation vehicle to capture environmental parameters, transportation vibration trajectory and real-time geographical location of goods during the journey, and generate in-transit perception information. In the production or sorting process of the supply chain, visual recognition cameras and proximity sensors are deployed at key nodes of the production line to capture the flow status of materials, work progress and abnormal stop events, and generate work perception information. The warehouse sensing information, in-transit sensing information, and operation sensing information are encapsulated according to a unified Internet of Things data protocol, and each encapsulated data is added with a data acquisition device identifier, data acquisition timestamp, and location code to form an initial sensing data packet sequence. The initial sensing data packet sequence is subjected to integrity verification and timestamp sorting to form a time-sequential raw material state flow.

3. The supply chain platform data management system based on the Internet of Things as described in claim 2, characterized in that, The step of receiving digital business information sent by business systems from upstream and downstream nodes of the supply chain through open interfaces and generating a structured business information flow includes: Define a set of standardized supply chain business data interfaces, which cover core business documents such as purchase orders, sales orders, delivery notes, warehouse receipt notes, invoices, and settlement notes; Through message queues or remote procedure calls, business data messages conforming to the interface specifications are received asynchronously from the supplier's business resource planning system, the distributor's customer relationship management system, and the carrier's transportation management system. The received business data messages are parsed and validated, and the business entities, business relationships, business events and key attribute values ​​are extracted. The unstructured message content is then transformed into an internally unified data model instance. Each data model instance is appended with its source system identifier, receiving timestamp, and associated business order number to form a standard structured business record; All received structured business records are sorted and indexed according to their business occurrence time or logical sequence to form a structured business information flow.

4. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 3, characterized in that, The steps of performing standardized fusion processing on the raw material state flow and the structured business information flow to generate a unified, spatiotemporally labeled supply chain panorama data flow include: Establish material identification mapping rules to associate and map the specific material identifications collected by IoT devices in the original material status flow with the material codes recorded in business documents in the structured business information flow. Establish a spatiotemporal coordinate system, unify the equipment location codes in the raw material state flow and the logistics node codes in the structured business information flow, and map the equipment location codes in the raw material state flow and the logistics node codes in the structured business information flow to the same set of supply chain network coordinates; Based on the material identification mapping relationship and the unified supply chain network coordinates, the time window alignment and spatial location matching are performed on the perception data packet sequence in the original material state flow and the business records in the structured business information flow. The successfully matched perception data packet sequence is concatenated with the business record to generate a fusion record, which contains both the physical state data of the material and the associated business information data. Each fusion record is labeled with its corresponding material identifier, fusion time point, and specific location in the supply chain network coordinates, forming a unified, spatiotemporally labeled panoramic supply chain data stream.

5. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 4, characterized in that, The steps of extracting key performance data related to specific supply chain contracts from the supply chain panoramic data stream and automatically verifying the key performance data according to preset smart contract logic include: Analyze the digital terms of supply chain contracts, identify key performance events, performance conditions, performance standards and related constraint parameters stipulated in the terms, and convert them into computable state machine logic; From the unified, spatiotemporally labeled supply chain panoramic data stream, filter out the fused records related to the material identification, participants, and logistics routes involved in the supply chain contract; Based on the triggering events defined by the state machine logic, specific performance observation data is further extracted from the selected fusion records. The performance observation data includes actual delivery time, actual arrival location, actual environmental indicators, and actual quantity specifications. The extracted performance observation data is automatically compared with the performance standards and constraint parameters agreed in the state machine logic to determine whether the performance observation data meets the agreed condition range. The detailed process and results of the comparison are recorded as an immutable verification log, which serves as proof of the fulfillment of the key performance data.

6. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 5, characterized in that, The step of automatically generating a reliable contract execution event signal when the key performance data meets the triggering conditions of the smart contract logic includes: In the smart contract logic, multiple execution stage nodes are preset, and a set of performance conditions to be satisfied is associated with each execution stage node; The output of the verification log is monitored in real time. When the verification log confirms that the key performance data has reached or exceeded the performance condition set associated with a certain execution stage node, the triggering condition is determined to be met. Generate a standard format contract execution event signal, which includes at least a unique contract identifier, the execution stage node number that was triggered, the trigger timestamp, and a credential hash value pointing to the corresponding verification log. The generated contract execution event signal is published to the distributed event bus, and the contract execution event signal and its associated verification log credential hash value are stored in a distributed storage with tamper-proof characteristics.

7. The supply chain platform data management system based on the Internet of Things as described in claim 6, characterized in that, The steps of triggering automated control commands for preset logistics or warehousing nodes in the supply chain based on the trusted contract execution event signal, and synchronously updating the resource status records of the corresponding nodes, include: Maintain a predefined control rule base, where each rule defines the mapping relationship between a specific contract execution event signal and one or more automated control instructions; Listen to the distributed event bus and capture published contract execution event signals; Using captured contract execution event signals as input, the system performs pattern matching in the control rule base to find and trigger matching control rules. Execute the matching control rules to generate specific automated control instructions with operational semantics. The automated control instructions include "open the specified warehouse door", "allow outbound", "sort to a specific channel", and "trigger payment application". The generated automated control commands are sent to designated logistics equipment or warehouse management subsystems for execution via the control network. After the instruction is issued and executed, the resource status records of the corresponding logistics node or warehousing node stored in the supply chain platform database are automatically modified according to the semantic logic of the instruction. The resource status records include inventory quantity, warehouse occupancy status, and equipment working status.

8. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 7, characterized in that, The step of performing deep feature extraction on the updated resource status record to generate a data feature vector for supply chain finance assessment includes: From the updated resource status records, extract multi-dimensional time-series data related to the target supply chain financing entity over a continuous period of time, including inventory turnover sequence, order delivery on-time rate sequence, and in-transit goods value change sequence; The multidimensional time-series data is subjected to feature engineering processing to calculate statistical features, trend features, stability features, and volatility features, generating a primary feature set; By combining the historical records of the contract execution event signals, relevant historical performance quality characteristics are extracted from the verification log, including the number of defaults, performance delay distribution, and accuracy of condition fulfillment. The primary feature set is fused with the historical performance quality features and input into a pre-trained deep feature encoding network; The deep feature encoding network performs dimensionality reduction and abstraction on the fused features through multi-layer nonlinear transformation, and finally outputs a fixed-dimensional, dense data feature vector, which is used to characterize the operational health and risk status of the supply chain links.

9. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 8, characterized in that, The steps of comparing and analyzing the data feature vectors with the historical performance behavior feature database to calculate the dynamic reliability score of the current supply chain link include: Maintain a historical performance behavior feature database, which stores data feature vectors generated by different supply chain participants in different business scenarios and their corresponding actual performance outcome labels. Calculate the similarity between the currently generated data feature vector and each historical data feature vector in the historical performance behavior feature database; Select the feature vectors of several historical data with the highest similarity as the nearest neighbor sample set; Analyze the actual performance outcome label corresponding to each sample in the nearest neighbor sample set. The actual performance outcome label includes "full performance", "partial breach", and "serious breach". Based on the similarity weight of neighboring samples, a weighted vote is performed on the performance outcome labels to calculate the predicted probability of various performance outcomes occurring in the current supply chain link in the future. According to the preset scoring model, the predicted probability is mapped to a quantified score, which is the dynamic reliability score of the current supply chain link.

10. The supply chain platform data management system based on the intelligent Internet of Things as described in claim 9, characterized in that, The process of packaging the dynamic trust score and associated key performance data into a standardized supply chain data asset package, and then synchronizing the supply chain data asset package to authorized supply chain alliance nodes through a preset secure channel to drive collaborative decision-making within the alliance, includes: Create a standard data asset package structure template, which includes three parts: data packet header, body data, and data signature; The header of the data packet contains the unique identifier of the asset package, the generation time, the data validity period, the identifier of the supply chain link to which it belongs, and the list of target receiving alliance nodes. The main data encapsulates the dynamic trust score, the key performance data on which the dynamic trust score is calculated, and index information for the relevant verification logs that support the key performance data. The hash values ​​of the data packet header and the main data are signed using the private key of the data generator. The signature result is then filled into the data signature part to form a complete supply chain data asset package. Based on the target receiving alliance node list in the data packet header, the supply chain data asset package is distributed to each authorized supply chain alliance node through a peer-to-peer encrypted communication link or an alliance permissioned blockchain network. After receiving the supply chain data asset package, the alliance node can use the trusted data encapsulated therein to drive collaborative decision-making within the node. The collaborative decision-making includes shared adjustments to demand forecasts, joint optimization of safety stock, and automatic allocation of inventory across enterprises.

Images