Nursing data sharing method and system based on Internet of Things

Abstract

The present invention provides a nursing data sharing method and system based on the Internet of Things. The method comprises: S1: real-time collection of multimodal nursing data through a distributed IoT terminal cluster to form a characteristic signal set; S2: construction of a data coordination engine and execution of federated learning preprocessing operations on the characteristic signal set to form a coordination signal stream; S3: dynamic calculation of data contribution weights and generation of a verification signal chain; S4: real-time monitoring of abnormal vital sign indicators in the coordination signal stream to generate a trigger signal vector; S5: updating incentive mechanism parameters based on the contribution credentials in the verification signal chain, optimizing the service quality strategy based on the network resource allocation scheme in the trigger signal vector, and forming a closed-loop control signal. This invention can solve the data island effect problem that exists in the process of medical nursing data sharing.

Description

Technical Field

[0001] The present invention relates to the intersection of Internet of Things technology and medical care data processing, and in particular to a nursing data sharing method and system based on the Internet of Things. Background Art

[0002] The healthcare sector is currently facing an explosion of data driven by the proliferation of IoT devices, but traditional data sharing models have significant flaws. Inconsistent data standards between medical institutions create information silos, preventing timely synchronization of critical nursing data when patients seek treatment across institutions, leading to inefficient diagnosis and treatment. Existing encryption technologies often employ static strategies, making them less adaptable to the dynamic sensitivity of nursing data. Traditional centralized storage architectures also present single points of failure. A data breach on a medical cloud platform in 2020 exposed the security risks of centralized management.

[0003] Regarding data transmission, conventional network scheduling algorithms cannot distinguish the urgency of nursing data. Delayed transmission of critical vital signs in emergency scenarios can impact treatment outcomes. Regarding device compatibility, the communication protocols used by IoT terminals from different manufacturers vary significantly. Statistics from one tertiary hospital show that the 43 types of devices it connects to involve seven different protocol standards, resulting in high data integration costs. The lack of a data ownership confirmation mechanism leads to a low willingness to share data. Research shows that as many as 67% of medical institutions refuse to share data due to unclear ownership. Summary of the Invention

[0004] In view of the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide a nursing data sharing method and system based on the Internet of Things, which is used to solve the problems of data island effect, privacy and security risks, and low efficiency of emergency data transmission in the process of medical nursing data sharing. The present invention constructs a collaborative sharing mechanism that integrates federated learning and blockchain technology, implements multimodal data standardization processing at the edge computing layer, and uses a dynamic contribution evaluation model to stimulate data sharing. At the same time, it establishes an emergency channel priority transmission system based on smart contracts, combines zero-knowledge proof and dynamic spectrum allocation technology, and realizes safe and efficient data flow. In the specific implementation process, the data collected by the distributed terminal cluster is encapsulated by the characteristic signal, and the edge node completes the data cleaning and alignment under privacy protection. The blockchain network generates a verifiable contribution certificate through quantitative evaluation. When a critical indicator is detected, the cross-domain priority transmission protocol is triggered. Finally, the overall system performance is optimized through closed-loop feedback, and the sharing efficiency is improved while ensuring data security.

[0005] The present invention provides a nursing data sharing method based on the Internet of Things, comprising:

[0006] S1: Collect multimodal nursing data in real time through distributed IoT terminal clusters and form a set of characteristic signals;

[0007] S2: Build a data coordination engine and perform federated learning preprocessing operations on the feature signal set to form a coordination signal stream. The coordination signal stream contains encrypted standardized data blocks and associated quality assessment matrices. The preprocessing operations include abnormal data cleaning, spatiotemporal dimension alignment, and feature vector normalization.

[0008] S3: Deploy a shared incentive mechanism based on the blockchain network, analyze and coordinate the quality assessment matrix in the signal flow through smart contracts, dynamically calculate the data contribution weight and generate a verification signal chain;

[0009] S4: Real-time monitoring of abnormal vital signs in the coordination signal stream. When a preset critical threshold is detected, a trigger signal vector is generated. The trigger signal vector includes an encrypted emergency digital pass and an optimized network resource allocation plan.

[0010] S5: Update the incentive mechanism parameters based on the contribution certificate in the verification signal chain, optimize the service quality strategy based on the network resource allocation plan in the trigger signal vector, and form a closed-loop control signal. The closed-loop control signal is synchronized to the data coordination engine to achieve collaborative optimization.

[0011] In one embodiment of the present invention, in step S1, the process of generating a characteristic signal set includes: deploying a dynamic perception adapter in a distributed IoT terminal cluster, the adapter automatically matches the data encapsulation protocol according to the terminal device type, and extracting features of multimodal nursing data through parallel processing threads. The extraction process integrates the physical fingerprint of the device and the biometric characteristics of the patient to generate a composite identification code, and at the same time establishes a binding relationship between the timestamp sequence and the spatial coordinates, wherein the data sensitivity classification tag is dynamically generated by real-time analysis of the correlation between the data content and the clinical context, and the encryption level is automatically increased when the core physiological parameters involving the patient's privacy are monitored. The composite identification code includes a unique device identifier generated by an irreversible hash operation and a derivative code after the patient's identity characteristics are blurred.

[0012] In one embodiment of the present invention, in step S2, the specific implementation method of abnormal data cleaning is: constructing a dynamic rule knowledge graph in the data coordination engine, which integrates clinical diagnosis and treatment standards, equipment technical parameters and historical data distribution characteristics, and performs outlier detection on the feature signal set through a sliding time window. The detection process adopts an adaptive threshold algorithm based on density clustering. When an abnormal value beyond the physiological reasonable range is identified, the data repair mechanism is triggered. The repair mechanism includes missing value interpolation, mutation value smoothing and context correlation correction. The repaired data block generates a quality assessment matrix with a version identification, which records the confidence of the original data and the correction operation trajectory.

[0013] In one embodiment of the present invention, in step S2, a method for implementing spatiotemporal dimension alignment includes: establishing a unified spatiotemporal coordinate system at the edge computing node, correcting the time deviation and spatial offset of the data collected by the device through the indoor positioning system and geo-fence technology, using a sliding window correlation analysis method to match the sampling timing of different devices in the alignment process, implementing time domain compensation for data streams with communication delays, and establishing a spatial mapping model based on the device deployment topology structure, converting the discrete spatial data collected by heterogeneous sensors into a continuous distributed data set under a standard coordinate system, and synchronously injecting the spatiotemporal correction parameters generated by the alignment operation into the metadata area of the coordinated signal stream.

[0014] In one embodiment of the present invention, in step S3, the dynamic calculation process of the data contribution weight includes: constructing a multidimensional evaluation function in the blockchain network, which integrates the integrity index in the quality assessment matrix, the timeliness coefficient of the timestamp sequence and the clinical value evaluation parameters, wherein the integrity index is jointly calculated by the data field missing rate and the check code matching degree, the timeliness coefficient dynamically decays according to the interval between the data generation time and the sharing request time, and the clinical value evaluation parameters are matched with the demand priority of the current medical scenario through knowledge graph retrieval. The output value of the evaluation function is verified by the node consensus to generate an unalterable contribution certificate, which forms a non-linear mapping relationship with the data access permission token.

[0015] In one embodiment of the present invention, in step S4, the generation mechanism of the emergency digital pass includes: when it is detected that the abnormal vital sign index exceeds the preset critical threshold, a multi-factor identity verification process is initiated, the verification process integrates the medical institution digital certificate, the device biometric signature and the patient's temporary authorization code, and verifies the legitimacy of the identity through a zero-knowledge proof protocol without leaking sensitive information. After the verification is passed, an encrypted pass with a time validity period is generated. The pass contains the emergency care data decryption key fragment and the cross-institutional access permission statement. The key fragment must be combined with the key fragment held by the recipient to decrypt the core medical data.

[0016] In one embodiment of the present invention, in step S4, the operating logic of the network resource allocation scheme includes: when the priority transmission channel is activated, real-time monitoring of the wireless channel load status and network topology structure, calculating the optimal spectrum resource allocation scheme through a game theory model, the allocation process takes into account both transmission delay requirements and channel interference suppression requirements, and implements dual guarantees of bandwidth reservation and frequency hopping communication for emergency data transmission. At the same time, a temporary relay node routing table is established, and the data transmission path is dynamically adjusted according to the degree of network congestion. The allocation scheme includes channel switching timing diagram and modulation and coding strategy combination parameters.

[0017] In one embodiment of the present invention, in step S5, the service quality policy optimization process includes: constructing a policy decision tree at the data sharing application layer, which dynamically adjusts the data transmission priority queue according to the network resource allocation plan in the closed-loop control signal. The optimization process uses a reinforcement learning algorithm to simulate system performance indicators under different strategies, and implements end-to-end transmission guarantee for emergency care data streams, including establishing a dedicated cache area, increasing retransmission priority and dynamically adjusting error correction coding strength, while balancing the service quality requirements of conventional data streams to prevent system shock caused by excessive preemption of network resources.

[0018] In one embodiment of the present invention, in step S5, the implementation method of collaborative optimization includes: a closed-loop control signal establishes a two-way feedback channel between the data coordination engine and the blockchain network, the feedback channel transmits the federated learning model gradient update amount and network resource configuration strategy parameters, the coordination engine adjusts the feature extraction rules and data cleaning thresholds according to the feedback signal, and the blockchain network synchronously updates the contribution calculation function and consensus verification mechanism in the smart contract. The two-way feedback channel uses differential privacy technology to process transmission parameters to ensure that the original data distribution characteristics and individual contribution details are not leaked during the optimization process.

[0019] The present invention also provides a nursing data sharing system based on the Internet of Things, comprising:

[0020] The acquisition module collects multimodal nursing data in real time through a distributed IoT terminal cluster and forms a characteristic signal set;

[0021] Coordination module, which builds a data coordination engine and performs federated learning preprocessing operations on the feature signal set and forms a coordination signal flow;

[0022] Verification module, which deploys a shared incentive mechanism based on the blockchain network, uses smart contracts to parse the quality assessment matrix in the coordination signal flow and generate a verification signal chain;

[0023] The early warning module monitors abnormal vital signs in the coordinated signal flow in real time. When a preset critical threshold is detected, a trigger signal vector is generated. The trigger signal vector contains an encrypted emergency digital pass and an optimized network resource allocation plan;

[0024] Optimization module, the optimization module updates the incentive mechanism parameters according to the contribution certificate in the verification signal chain, and optimizes the service quality strategy based on the network resource allocation plan in the trigger signal vector, and forms a closed-loop control signal. The closed-loop control signal is synchronized to the data coordination engine to achieve collaborative optimization.

[0025] The present invention provides a nursing data sharing method and system based on the Internet of Things. By constructing a collaborative sharing mechanism that integrates federated learning and blockchain technology, it implements multimodal data standardization processing at the edge computing layer, uses a dynamic contribution evaluation model to incentivize data sharing, and establishes an emergency channel priority transmission system based on smart contracts. It combines zero-knowledge proof and dynamic spectrum allocation technology to achieve safe and efficient data flow. In the specific implementation process, the data collected by the distributed terminal cluster is encapsulated with characteristic signals, and the edge nodes complete the data cleaning and alignment under privacy protection. The blockchain network generates verifiable contribution certificates through quantitative evaluation. When critical indicators are detected, the cross-domain priority transmission protocol is triggered. Finally, the overall system performance is optimized through closed-loop feedback, and the sharing efficiency is improved while ensuring data security. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0027] Figure 1 A flowchart of the method for sharing nursing data based on the Internet of Things;

[0028] Figure 2 This is the system architecture diagram of the nursing data sharing system based on the Internet of Things. DETAILED DESCRIPTION

[0029] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0030] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0031] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0032] See Figure 1-2 , showing the present invention's IoT-based nursing data sharing method and system. The IoT-based nursing data sharing method includes: S1: real-time collection of multimodal nursing data through a distributed IoT terminal cluster to form a feature signal set; S2: building a data coordination engine and performing federated learning preprocessing on the feature signal set to form a coordination signal stream. The coordination signal stream contains encrypted standardized data blocks and associated quality assessment matrices. The preprocessing includes abnormal data cleaning, spatiotemporal dimension alignment, and feature vector normalization; S3: deploying a shared incentive mechanism based on a blockchain network, parsing the quality assessment matrix in the coordination signal stream through smart contracts, dynamically calculating data contribution weights, and generating a verification signal chain; S4: real-time monitoring of abnormal vital sign indicators in the coordination signal stream. When a preset critical threshold is detected, a trigger signal vector is generated. The trigger signal vector contains an encrypted emergency digital pass and an optimized network resource allocation plan; S5: updating incentive mechanism parameters based on the contribution credentials in the verification signal chain, optimizing the service quality policy based on the network resource allocation plan in the trigger signal vector, and forming a closed-loop control signal. The closed-loop control signal is synchronized to the data coordination engine to achieve collaborative optimization.

[0033] like Figure 1As shown, the core process consists of five interrelated technical stages, enabling secure data sharing and emergency care response through a step-by-step signal flow transformation and feedback mechanism. In the initial stage, a distributed IoT terminal cluster, serving as the physical foundation for data collection, consists of a variety of heterogeneous medical devices, including but not limited to wearable vital sign monitors, ward environmental sensors, mobile nursing terminals, and medical imaging devices. These terminal devices generate raw data streams using built-in communication modules according to their respective vendor protocols. For example, an ECG monitor uses a medical-specific wireless protocol to transmit waveform data, while a temperature and humidity sensor may upload environmental parameters using a low-power wide area network protocol. During data collection, a lightweight preprocessing module deployed within the terminal cluster performs preliminary encapsulation of the raw data, appending unique device identifiers and timestamps to form basic signal units containing multi-dimensional features. Before transmission to the edge node, these signal units are initially labeled according to a pre-set data sensitivity classification strategy. For example, blood oxygen saturation data may be labeled as high-sensitivity and ward temperature data as low-sensitivity. This classification provides a basis for subsequent differentiated processing. The generation of the feature signal set not only includes the original nursing data itself, but also integrates auxiliary information such as equipment operating status parameters and network transmission quality indicators to form a multi-dimensional data matrix, providing complete contextual information for subsequent processing links.

[0034] After the feature signal set is transmitted to the edge computing node, it enters the processing phase of the data coordination engine. This engine, as the core component of federated learning preprocessing, is deployed on a cluster of edge servers close to the data source and utilizes a containerized architecture to enable flexible scheduling of computing resources. Preprocessing begins by cleaning the feature signal set for abnormal data. A multidimensional detection model is established by loading a dynamically updated knowledge base of clinical rules, such as normal physiological parameter ranges and device measurement error tolerances. The cleaning process utilizes sliding time window analysis techniques to perform pattern recognition on continuous time series data. When outliers outside a reasonable range are detected, a data repair pipeline is triggered. The repair mechanism incorporates multiple processing logics: a context-similarity-based interpolation algorithm is used to address temporary signal loss; sudden abnormal spikes are smoothed using a wavelet transform; and systematic deviations are compensated for by incorporating device calibration parameters. The cleaned and repaired data stream enters the spatiotemporal alignment module, which fuses the spatial coordinate data of the indoor positioning system with the time base of the network time protocol to eliminate data deviations caused by device clock asynchrony or differences in deployment locations. The aligned data undergoes feature vector normalization, converting data of varying dimensions and sampling rates into a unified mathematical representation. For example, converting blood pressure values from millimeters of mercury to Z-scores under a standard normal distribution. A quality assessment matrix generated throughout the preprocessing process details the processing trajectory of each data block, including the confidence score of the original data, the type of repair operation, and the normalization transformation parameters. This matrix is protected by a lightweight encryption algorithm and then injected into the coordination signal stream. Once the coordination signal stream is formed, the blockchain-based shared incentive mechanism implementation phase begins. Blockchain nodes are deployed on a consortium of medical institutions, using a modified Byzantine fault-tolerant consensus mechanism to balance efficiency and security requirements. Smart contracts play a central role in this phase, first parsing the quality assessment matrix in the coordination signal stream and inputting the data integrity indicator, timeliness coefficient, and clinical value parameter contained therein into a multidimensional contribution evaluation model. The integrity indicator is derived by calculating the integrity rate of the data field and the degree of checksum matching. The timeliness coefficient uses an exponential decay function to process the time difference between the data generation time and the sharing request time. The clinical value parameter is determined by querying the medical knowledge graph to match the priority weight of the current application scenario. After the evaluation model outputs the contribution weight of each data provider, the smart contract invokes a verifiable computation framework to generate encrypted contribution certificates. These certificates are organized into a traceable verification signal chain using a Merkle tree structure. Simultaneously, the blockchain network dynamically generates data access tokens based on the contribution weights, employing attribute-based encryption to achieve fine-grained access control. The key innovation of this phase lies in the deep integration of federated learning data processing with blockchain incentive mechanisms. This quantified contribution model transforms the raw data value into tradable digital rights, while ensuring the transparency and auditability of the computational process.

[0035] like Figure 1As shown, in a distributed IoT terminal cluster, a dynamic sensing adapter serves as the core middleware, employing a modular design to support plug-and-play integration with multiple communication protocols. The adapter's built-in protocol conversion engine automatically identifies the terminal device's communication protocol. For example, it restructures Bluetooth Low Energy protocol packets into MQTT format, or converts Zigbee sensor data into a JSON structure that can be processed by a RESTful interface. During data encapsulation, a parallel processing thread pool performs real-time feature extraction on multimodal data, such as extracting the coefficient of variation of the RR interval from the ECG signal and calculating respiratory entropy from the respiratory waveform. These characteristic parameters, combined with the raw data, form a characteristic signal unit. To enhance data traceability, a composite identification code is generated by integrating the device's physical fingerprint with the patient's biometrics. The device fingerprint generates a unique identifier by extracting wireless signal characteristics, a hash of the hardware serial number, and firmware version information. The patient's biometrics are collected non-invasively, such as extracting an identity feature vector from the ECG waveform, which is then processed through a fuzzy hash to ensure privacy. Timestamp sequences are bound using a high-precision synchronization protocol, recording the moment data is generated with nanosecond resolution. This information is then combined with the three-dimensional spatial coordinates provided by the indoor positioning system to form a four-dimensional spatiotemporal tag. Data sensitivity classification relies on a real-time content analysis engine, which parses nursing records using natural language processing, combines computer vision technology to identify sensitive areas in medical images, and dynamically adjusts data encryption levels. When critical parameters involving patient privacy are detected, such as HIV test results or genetic disease markers, the system automatically switches to a quantum-resistant encryption algorithm and triggers a distributed storage strategy to shard the data across multiple edge nodes.

[0036] Furthermore, a dynamic rule-based knowledge graph serves as the decision-making core for cleaning operations, integrating three major knowledge sources: a clinical diagnosis and treatment standard library containing thousands of physiological parameter threshold rules developed by medical experts; a device technical parameter library containing metadata such as measurement accuracy and sampling rate for mainstream medical devices; and a historical data distribution feature library analyzing statistical patterns in past nursing data using machine learning models. For outlier detection, the system employs an improved density-based clustering algorithm. This algorithm incorporates a dynamic radius parameter adjustment mechanism to address the strong temporal and multi-dimensional correlation characteristics of medical data. For vital sign data streams, the algorithm calculates local anomaly factors within a sliding time window. Data points are flagged as suspected anomalies when the density difference between a data point and its neighbors exceeds an adaptive threshold. The data repair mechanism employs a hierarchical processing strategy: for sporadic missing data, a predictive model based on a long-short-term memory network is used for interpolation; for persistent anomalies, the device diagnostic protocol is activated, triggering a sensor self-test via remote commands. During the repair process, the system generates a detailed operation log, recording the original outlier value, the rationale for the repair method selected, and the corrected value. This information is desensitized and written into the quality assessment matrix. The matrix utilizes a hierarchical storage structure. The base layer includes quantitative metrics such as integrity scores and timeliness indices. The metadata layer stores key parameters of the repair process. The provenance layer uses hash chain technology to ensure the immutability of operation records. Cleaned and repaired data blocks carry version identifiers in the coordination signal stream, supporting subsequent data traceability and quality verification. The entire cleaning process is completed at the edge computing node, ensuring data processing timeliness while avoiding the privacy risks associated with raw data leaving the local domain.

[0037] Specifically, the spatiotemporal alignment method builds a foundational framework for multi-source data fusion. By establishing a unified spatiotemporal coordinate system at edge computing nodes, it effectively addresses the integration challenges caused by inconsistent spatiotemporal references across heterogeneous devices. This coordinate system relies on the centimeter-level spatial reference provided by the indoor positioning system, combined with the time reference source derived from the Beidou satellite timing system and the Network Time Protocol (NTP), to form a standard reference system with four-dimensional spatiotemporal properties. In implementation, geofencing technology is used to calibrate the boundaries of the device deployment area. By matching received signal strength indicator (RSSI) fingerprints with ultra-wideband (UWB) positioning data, signal drift errors caused by environmental reflections are dynamically corrected. To address time synchronization, the system employs an improved clock offset compensation algorithm, establishing a hierarchical timing mechanism between edge nodes and end devices. End devices periodically transmit beacon frames containing local clock information. Edge nodes calculate a combined estimate of round-trip delay and clock offset, and dynamically adjust using a Kalman filter to keep clock errors within milliseconds. The data streams are time-aligned using a sliding window correlation analysis method. For data from devices with different sampling rates, the sampling rate is first normalized using cubic spline interpolation. The cross-correlation function is then calculated within the sliding time window to determine the optimal phase alignment point. For spatial data mapping, a three-dimensional grid model is constructed based on the physical deployment topology of the medical devices. Discrete sensor readings are converted into a continuous spatial distribution map using radial basis function interpolation. A coordinate transformation matrix is then used to transform the spatial data collected by different positioning systems into a unified coordinate system. The spatiotemporal correction parameters generated during the alignment process, including clock offset compensation, coordinate transformation matrix coefficients, and interpolation function parameters, are encrypted and embedded in the metadata area of the coordination signal stream, providing a traceable spatiotemporal reference for subsequent data analysis. The innovation of this technology lies in its deep integration of indoor positioning error compensation with multi-device clock synchronization. Through a dynamic calibration mechanism, the spatiotemporal deviations between heterogeneous devices are reduced to clinically acceptable levels. For example, the time alignment accuracy of data from an ECG monitor and respiratory monitoring device reaches ±10ms, and the spatial positioning error does not exceed 30 cm, significantly improving the accuracy of multimodal data fusion.

[0038] like Figure 1As shown in the figure, the data contribution weight calculation mechanism establishes a fair and transparent value assessment system. Its core lies in constructing a multidimensional evaluation function to achieve quantitative modeling of data value. The design of this function integrates evaluation indicators from three dimensions: data quality, timeliness, and clinical value. The integrity index is calculated by calculating the weighted harmonic mean of the data field missing rate and the checksum matching degree. The checksum matching degree uses a dual verification mechanism of cyclic redundancy check (CRC) and hash value. The timeliness coefficient is modeled using an exponential decay function. The decay value is calculated according to the formula α·exp(-βΔt) based on the time difference Δt between the data generation time and the sharing request time. α is the basic weight coefficient and β is the decay rate parameter. These parameters are dynamically adjusted according to the type of medical data. For example, the β value for emergency data is set to a small value to slow the decay rate. The clinical value assessment parameter is determined by querying the medical knowledge graph. The knowledge graph pre-constructs the disease-symptom-examination item association network. When the sharing request involves a specific clinical scenario, the system uses a graph traversal algorithm to retrieve the most matching medical entity and extract its preset priority weight. The output value of the evaluation function must be verified by consensus across blockchain nodes, using an improved Practical Byzantine Fault Tolerance (PBFT) algorithm to minimize consensus latency to seconds while ensuring security. Contribution credentials are generated using a Verifiable Delay Function (VDF) to ensure the verifiability and tamper-resistance of the calculation process. Each credential contains a contribution value, an evaluation timestamp, and the signature of the validating node. Data access tokens are designed using attribute-based encryption (ABE), mapping contribution values to attribute conditions within access policies. For example, high-contribution institutions receive more granular data access rights. This mechanism's innovation lies in establishing a dynamic evaluation model that links data quality and clinical value. The blockchain's consensus mechanism ensures the credibility of evaluation results, while attribute encryption technology enables precise control of access rights. For example, in cancer research scenarios, high-contribution research institutions gain access to the complete dataset containing genetic test results, while less-contributing institutions only have access to de-identified statistical features.

[0039] Furthermore, the emergency digital pass generation mechanism establishes a secure and efficient emergency identity verification system, the core of which is to balance the efficiency of identity verification in emergency situations with the need for privacy protection. When the vital signs monitoring system detects that a preset critical threshold has been triggered, the system initiates a multi-factor identity verification process: first, the legitimacy of the terminal device is verified through the device's biometric signature, and a hardware fingerprint extraction technology based on a physically unclonable function (PUF) is used to generate a unique device identifier; second, the compliance of the requesting institution is verified by calling the medical institution's digital certificate chain, which contains a root certificate issued by the health regulatory authority and an institution-specific sub-certificate; finally, a temporary authorization code is obtained from the patient, which is generated instantly through near-field communication (NFC) or biometric recognition (such as fingerprint) and is valid only once. The verification process uses a zero-knowledge proof protocol to convert the verification factors into an arithmetic circuit representation, confirming the legitimacy of the identity through a non-interactive proof method without leaking sensitive information. The encrypted pass generated after verification adopts a layered encryption structure: the outer layer uses the national secret SM4 algorithm to encrypt the agency's access permission statement, and the inner layer uses a threshold cryptography scheme to split the data decryption key into multiple fragments, one of which is embedded in the pass, and the remaining fragments are pre-stored in the key management system of the target emergency agency. The validity period of the pass follows the golden time principle of clinical emergency treatment and is dynamically adjusted according to different types of critical illnesses. For example, the validity period of the myocardial infarction emergency pass is set to 120 minutes, and that of stroke emergency treatment is set to 180 minutes. The innovation of this mechanism is reflected in the combination of zero-knowledge proof and threshold encryption, which not only meets the needs of rapid verification in emergency situations, but also ensures the security and controllability of core medical data. For example, emergency personnel must hold both a digital pass and a hospital key fragment to decrypt the patient's complete medical record, preventing data from being stolen by unauthorized parties during transmission.

[0040] like Figure 1As shown, the dynamic spectrum allocation algorithm establishes an intelligent emergency communication support system, the core of which is to achieve the optimal allocation of limited wireless resources. When the priority transmission channel is activated, the system collects wireless channel state information (CSI) in real time, including channel noise power, multipath fading characteristics, and adjacent channel interference levels, to construct a dynamic spectrum map. The network topology analysis module uses probe messages to obtain the operating status and load of routing nodes and constructs a weighted directed graph model. The spectrum allocation problem is modeled as a non-cooperative game under incomplete information, in which emergency data transmission requirements are the priority participant and regular data flows are the secondary participants. The game model's objective function balances transmission delay minimization with channel interference mitigation, and a Nash equilibrium solver is used to obtain the optimal spectrum allocation solution. In specific implementation, orthogonal frequency division multiplexing (OFDM) subcarrier resource blocks are reserved for emergency data, and high-order modulation schemes (such as 1024-QAM) are configured to improve transmission efficiency. Frequency hopping spread spectrum technology is also used to enhance interference resistance. The construction of the temporary relay node routing table uses a heuristic search algorithm to dynamically select the optimal relay path based on the level of network congestion. When the packet loss rate of a certain path exceeds a threshold, it automatically switches to a backup routing node. The specific parameters of the spectrum allocation scheme include the channel switching time sequence, the modulation and coding strategy combination, and the power control parameters. These parameters are encapsulated as a machine-readable configuration instruction set and loaded into the communication equipment in real time through software-defined radio (SDR) technology. The innovation of this technology lies in the deep integration of game theory models and software-defined networking (SDN) technology to achieve dynamic reconfiguration of spectrum resources in emergency scenarios. For example, in an emergency cardiopulmonary resuscitation scenario, the system can complete dedicated channel allocation within 200ms, stably controlling the transmission delay of critical vital signs data to less than 50ms. At the same time, through intelligent power control, adjacent channel interference is reduced by more than 15dB, ensuring the reliability and real-time transmission of emergency data.

[0041] like Figure 2As shown, the present invention relates to a nursing data sharing system based on the Internet of Things, including an acquisition module, a coordination module, a verification module, an early warning module, and an optimization module. The acquisition module collects multimodal nursing data in real time through a distributed Internet of Things terminal cluster and forms a characteristic signal set; the coordination module builds a data coordination engine and performs federated learning preprocessing operations on the characteristic signal set to form a coordination signal flow; the verification module deploys a shared incentive mechanism based on a blockchain network, parses the quality assessment matrix in the coordination signal flow through a smart contract, and generates a verification signal chain; the early warning module monitors abnormal vital sign indicators in the coordination signal flow in real time, and when a preset critical threshold is detected, generates a trigger signal vector, which contains an encrypted emergency digital pass and an optimized network resource allocation plan; the optimization module updates the incentive mechanism parameters according to the contribution certificate in the verification signal chain, and optimizes the service quality strategy based on the network resource allocation plan in the trigger signal vector, and forms a closed-loop control signal, which is synchronized to the data coordination engine to achieve collaborative optimization.

[0042] like Figure 2As shown in the figure, the technical implementation of the service quality policy optimization process has built a dynamic and adaptive network resource management system. The policy decision engine deployed at the data sharing application layer constructs a multi-dimensional state vector space in real time by analyzing the network resource allocation plan in the closed-loop control signal. This space contains key parameters such as the current network bandwidth utilization, data transmission delay distribution, channel bit error rate statistics, and emergency care data traffic ratio. The decision tree is constructed using the Monte Carlo tree search algorithm. Each decision node corresponds to a specific service quality adjustment strategy, such as priority queue weight coefficient adjustment, error correction code redundancy configuration, or routing path switching threshold setting. The training process of the reinforcement learning model simulates system performance indicators under different strategy combinations, including end-to-end transmission success rate, resource utilization efficiency, and system stability coefficient, and continuously optimizes strategy selection through the deep Q network (DQN) algorithm. Safeguards for emergency care data flows include a three-tiered protection mechanism: First aid packets are allocated a dedicated buffer at the physical layer to prevent them from competing with regular data for storage resources; Automatic Repeat Request (ARQ) priority is increased at the link layer to ensure that critical packets are retransmitted first; and Forward Error Correction (FEC) redundancy is dynamically adjusted at the application layer, adaptively selecting either Reed-Solomon or Low-Density Parity-Check (LDPC) coding schemes based on real-time channel quality. Quality of service (QoS) balancing for regular data flows is achieved through dual control loops: a primary control loop dynamically adjusts bandwidth allocation weights based on network load, while a secondary control loop monitors system oscillation indicators. When resource allocation fluctuations exceed safety thresholds, a fuzzy PID controller is activated for damping adjustments. Optimized policy parameters are distributed to each network device via a software-defined networking (SDN) controller, enabling synchronized updates of network-wide policies within seconds. The innovation of this mechanism lies in its combination of reinforcement learning and classical control theory, showing good adaptability in burst traffic scenarios. For example, when emergency data transmission requests suddenly increase by 300%, the system can complete policy adjustments within 500ms, controlling the transmission delay variance of emergency data packets within 10ms, while maintaining the throughput of regular data flows at no less than 80% of the baseline level.

[0043] like Figure 2As shown, the collaborative optimization mechanism establishes a cross-level dynamic feedback system, achieving full-link coordination of data processing and incentive mechanisms through a bidirectional communication channel. Closed-loop control signals establish a time-stamped feedback data pipeline between the data coordination engine and the blockchain network. This pipeline utilizes a duplex communication mode, with the uplink transmitting the federated learning model's gradient updates and feature importance parameters, and the downlink transmitting network resource allocation policy parameters and contribution evaluation rule change instructions. Differential privacy technology is incorporated into the federated learning model update process. Gaussian noise is injected before gradient aggregation. The noise intensity is dynamically adjusted based on data sensitivity levels to ensure that model updates cannot be reversed to the original training data. Upon receiving the feedback signal, the data coordination engine initiates an adaptive adjustment module: feature extraction rules are ranked based on feature importance, as reflected by gradient updates, and feature selection weights are dynamically optimized. Data cleaning thresholds are flexibly adjusted based on network resource status, for example, increasing the strictness of abnormal data filtering during network congestion to reduce transmission load. The blockchain network's smart contract synchronously updates the contribution calculation function, incorporating the latest clinical value evaluation parameters through an online learning mechanism. The consensus verification mechanism dynamically adjusts the timeout threshold based on network latency metrics. The privacy protection mechanism of the feedback channel utilizes homomorphic encryption technology to ensure that gradient parameters and policy instructions remain encrypted during transmission. Even if intercepted, they cannot be decrypted to obtain valid information. The innovation of this collaborative mechanism lies in the establishment of a positive feedback loop between data processing and incentive mechanisms. For example, when the system detects a surge in demand for sharing a certain type of physiological parameter, it automatically increases the contribution weight coefficient of that data, incentivizing relevant organizations to prioritize sharing of this data. Simultaneously, the coordination engine optimizes the corresponding feature extraction algorithm, forming a virtuous interaction between data value mining and sharing incentives.

[0044] Specifically, the verification signal chain construction process establishes an auditable trust transfer system. Its core technology lies in the deep integration of the verifiability of distributed computing with the immutability of blockchain. The verifiable computing framework is deployed on each consensus node in the blockchain network. An arithmetic circuit compiler is used to parse the quality assessment matrix into a computational flow chart composed of basic logic gates. The contribution weight calculation process is then converted into a circuit structure consisting of adders, multipliers, and comparators. Proof of computational correctness is generated using a non-interactive zero-knowledge proof protocol. By encoding the circuit execution trace as a polynomial commitment, a concise proof is generated using the bilinear pairing property. When verifying the proof, consensus nodes only need to perform sublinear computations to confirm the correctness of the entire computation, significantly reducing verification overhead. The verification signal chain's storage structure utilizes a modified Merkle Patricia tree (MPT). Each credential node contains a triple structure consisting of the predecessor node's hash value, the current contribution value, and a computational status snapshot. This status snapshot is encoded using a lightweight serialization protocol and records key intermediate variables in the computation process, such as the hash digest of the quality assessment matrix, evaluation function parameters, and consensus node voting results. The traceability mechanism of the chain structure is implemented by combining skip tables and hash pointers, which supports the rapid location of historical credential records. For example, when querying the contribution trend of an institution within three months, the retrieval time complexity can be reduced to O(log n). The update process of the credential chain adopts sharding parallel processing technology, and the newly generated contribution credentials are stored in shards according to the region of the medical institution. When querying across shards, the target shard is located through the distributed hash table (DHT). The innovation of this verification signal chain lies in the verification of the entire life cycle of the calculation process. For example, when a regulatory agency needs to audit a data sharing transaction, it can confirm the compliance of the contribution assessment without decrypting the original data by verifying the zero-knowledge proof and computing state snapshot on the chain. At the same time, by utilizing the aggregated verification characteristics of the Merkle tree, the time efficiency of batch verification of tens of thousands of credential records is improved by two orders of magnitude.

[0045] The present invention provides a nursing data sharing method and system based on the Internet of Things. By constructing a collaborative sharing mechanism that integrates federated learning and blockchain technology, it implements multimodal data standardization processing at the edge computing layer, uses a dynamic contribution evaluation model to incentivize data sharing, and establishes an emergency channel priority transmission system based on smart contracts. It combines zero-knowledge proof with dynamic spectrum allocation technology to achieve safe and efficient data flow. In the specific implementation process, the data collected by the distributed terminal cluster is encapsulated with characteristic signals, and the edge nodes complete the data cleaning and alignment under privacy protection. The blockchain network generates verifiable contribution certificates through quantitative evaluation. When critical indicators are detected, the cross-domain priority transmission protocol is triggered. Finally, the overall system performance is optimized through closed-loop feedback, and the sharing efficiency is improved while ensuring data security.

[0046] Therefore, the present invention provides a nursing data sharing method and system based on the Internet of Things, which can solve the problems of data island effect, privacy and security risks, and low efficiency of emergency data transmission in the process of medical nursing data sharing.

[0047] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. A nursing data sharing method based on the Internet of Things, characterized in that: include: S1: Collect multimodal nursing data in real time through distributed IoT terminal clusters and form a set of characteristic signals; S2: Build a data coordination engine and perform federated learning preprocessing operations on the feature signal set to form a coordination signal stream, wherein the coordination signal stream includes encrypted standardized data blocks and associated quality assessment matrices. The preprocessing operations include abnormal data cleaning, spatiotemporal dimension alignment, and feature vector normalization. S3: Deploy a shared incentive mechanism based on the blockchain network, parse the quality assessment matrix in the coordination signal flow through smart contracts, dynamically calculate the data contribution weight and generate a verification signal chain; S4: monitoring abnormal vital sign indicators in the coordination signal flow in real time, and generating a trigger signal vector when a preset critical threshold is detected, wherein the trigger signal vector includes an encrypted emergency digital pass and an optimized network resource allocation plan; S5: updating the incentive mechanism parameters according to the contribution certificate in the verification signal chain, optimizing the service quality strategy based on the network resource allocation scheme in the trigger signal vector, and forming a closed-loop control signal, which is synchronized to the data coordination engine to achieve collaborative optimization; In step S2, the method for implementing the spatiotemporal dimension alignment includes: establishing a unified spatiotemporal coordinate system at the edge computing node, correcting the time deviation and spatial offset of the data collected by the device through the indoor positioning system and geo-fencing technology, using the sliding window correlation analysis method to match the sampling timing of different devices in the alignment process, implementing time domain compensation for data streams with communication delays, and establishing a spatial mapping model based on the device deployment topology structure, converting the discrete spatial data collected by heterogeneous sensors into a continuous distributed data set in a standard coordinate system, and synchronously injecting the spatiotemporal correction parameters generated by the alignment operation into the metadata area of the coordination signal stream; In step S3, the dynamic calculation process of the data contribution weight includes: constructing a multidimensional evaluation function in the blockchain network, which integrates the integrity index in the quality assessment matrix, the timeliness coefficient of the timestamp sequence and the clinical value evaluation parameters, wherein the integrity index is jointly calculated by the data field missing rate and the check code matching degree, the timeliness coefficient dynamically decays according to the interval between the data generation time and the sharing request time, and the clinical value evaluation parameters are retrieved through the knowledge graph to match the demand priority of the current medical scenario. The output value of the evaluation function generates an unalterable contribution certificate after node consensus verification, and the certificate forms a non-linear mapping relationship with the data access permission token.

2. The method for sharing nursing data based on the Internet of Things according to claim 1, characterized in that: In step S1, the generation process of the characteristic signal set includes: deploying a dynamic perception adapter in a distributed Internet of Things terminal cluster, the adapter automatically matches the data encapsulation protocol according to the terminal device type, and extracts features of multimodal nursing data through parallel processing threads. The extraction process integrates the device physical fingerprint and the patient's biometric features to generate a composite identification code, and at the same time establishes a binding relationship between the timestamp sequence and the spatial coordinates, wherein the data sensitivity classification mark is dynamically generated by real-time analysis of the correlation between the data content and the clinical context, and the encryption level is automatically increased when the core physiological parameters involving the patient's privacy are monitored. The composite identification code includes a unique device identifier generated by an irreversible hash operation and a derivative code after the patient's identity features are blurred.

3. The method for sharing nursing data based on the Internet of Things according to claim 1, characterized in that: In step S2, the specific implementation method of the abnormal data cleaning is: constructing a dynamic rule knowledge graph in the data coordination engine, which integrates clinical diagnosis and treatment standards, equipment technical parameters and historical data distribution characteristics, and performs outlier detection on the feature signal set through a sliding time window. The detection process adopts an adaptive threshold algorithm based on density clustering. When an abnormal value beyond the physiological reasonable range is identified, the data repair mechanism is triggered. The repair mechanism includes missing value interpolation, mutation value smoothing and context correlation correction. The repaired data block generates a quality assessment matrix with a version identification. The quality assessment matrix records the confidence of the original data and the correction operation trajectory.

4. The method for sharing nursing data based on the Internet of Things according to claim 1, characterized in that: In step S4, the generation mechanism of the emergency digital pass includes: when it is detected that the abnormal vital sign index exceeds the preset critical threshold, the multi-factor identity verification process is initiated. The verification process integrates the medical institution digital certificate, the device biometric signature and the patient's temporary authorization code, and verifies the legitimacy of the identity through the zero-knowledge proof protocol without leaking sensitive information. After the verification is passed, an encrypted pass with a time validity period is generated. The pass contains the emergency care data decryption key fragment and the cross-institutional access permission statement. The key fragment must be combined with the key fragment held by the recipient to decrypt the core medical data.

5. The method for sharing nursing data based on the Internet of Things according to claim 1, characterized in that: In step S4, the operating logic of the network resource allocation scheme includes: when the priority transmission channel is activated, real-time monitoring of the wireless channel load status and network topology structure, calculating the optimal spectrum resource allocation scheme through a game theory model, taking into account the transmission delay requirements and channel interference suppression requirements in the allocation process, implementing dual guarantees of bandwidth reservation and frequency hopping communication for emergency data transmission, and establishing a temporary relay node routing table at the same time, dynamically adjusting the data transmission path according to the degree of network congestion. The allocation scheme includes channel switching timing diagram and modulation and coding strategy combination parameters.

6. The method for sharing nursing data based on the Internet of Things according to claim 5, characterized in that: In step S5, the service quality policy optimization process includes: constructing a policy decision tree at the data sharing application layer, which dynamically adjusts the data transmission priority queue according to the network resource allocation plan in the closed-loop control signal. The optimization process uses a reinforcement learning algorithm to simulate system performance indicators under different strategies, and implements end-to-end transmission guarantee for emergency care data streams, including establishing a dedicated cache area, improving retransmission priority and dynamically adjusting error correction coding strength, while balancing the service quality requirements of conventional data streams to prevent system shock caused by excessive network resource preemption.

7. The method for sharing nursing data based on the Internet of Things according to claim 1, characterized in that: In step S5, the implementation method of the collaborative optimization includes: a closed-loop control signal establishes a two-way feedback channel between the data coordination engine and the blockchain network, the feedback channel transmits the gradient update amount of the federated learning model and the network resource configuration strategy parameters, the coordination engine adjusts the feature extraction rules and data cleaning thresholds according to the feedback signal, and the blockchain network synchronously updates the contribution calculation function and consensus verification mechanism in the smart contract. The two-way feedback channel uses differential privacy technology to process the transmission parameters to ensure that the original data distribution characteristics and individual contribution details are not leaked during the optimization process.

8. A nursing data sharing system based on the Internet of Things, which is used to use the nursing data sharing method based on the Internet of Things according to any one of claims 1 to 7, characterized in that: include: An acquisition module, which collects multimodal nursing data in real time through a distributed IoT terminal cluster and forms a characteristic signal set; A coordination module, wherein the coordination module constructs a data coordination engine and performs a federated learning preprocessing operation on the feature signal set to form a coordination signal stream. The implementation method of the spatiotemporal dimension alignment includes: establishing a unified spatiotemporal coordinate system at the edge computing node, correcting the time deviation and spatial offset of the device-collected data through the indoor positioning system and geo-fencing technology, using a sliding window correlation analysis method to match the sampling timing of different devices in the alignment process, implementing time domain compensation for data streams with communication delays, and establishing a spatial mapping model based on the device deployment topology structure, converting the discrete spatial data collected by heterogeneous sensors into a continuously distributed data set in a standard coordinate system, and synchronously injecting the spatiotemporal correction parameters generated by the alignment operation into the metadata area of the coordination signal stream; A verification module, wherein the verification module deploys a shared incentive mechanism based on a blockchain network, parses the quality assessment matrix in the coordination signal flow through a smart contract, and generates a verification signal chain. The dynamic calculation process of the data contribution weight includes: constructing a multidimensional evaluation function in the blockchain network, which integrates the integrity index in the quality assessment matrix, the timeliness coefficient of the timestamp sequence, and the clinical value evaluation parameter, wherein the integrity index is jointly calculated by the data field missing rate and the check code matching degree, the timeliness coefficient dynamically decays according to the interval between the data generation time and the sharing request time, and the clinical value evaluation parameter matches the demand priority of the current medical scenario through knowledge graph retrieval. The output value of the evaluation function generates an unalterable contribution certificate after node consensus verification, and the certificate forms a non-linear mapping relationship with the data access permission token; an early warning module, which monitors abnormal vital sign indicators in the coordination signal stream in real time and generates a trigger signal vector when a preset critical threshold is detected. The trigger signal vector includes an encrypted emergency digital pass and an optimized network resource allocation plan; An optimization module updates the incentive mechanism parameters according to the contribution credentials in the verification signal chain, optimizes the service quality strategy based on the network resource allocation scheme in the trigger signal vector, and forms a closed-loop control signal, which is synchronized to the data coordination engine to achieve collaborative optimization.

Images