A 5G and blockchain-based big health AI prevention management system and method

Abstract

This invention discloses a big health AI prevention management system and method based on 5G and blockchain, belonging to the field of health prevention technology. The invention includes the following steps: real-time collection of physiological and environmental data from individuals via wearable devices; encrypted transmission of raw data to edge nodes using a military-grade 5G slicing network, and execution of motion artifact filtering and local feature calculation; hashing and storing individual identities, managing individual adaptation profiles and model parameters; fusing real-time features with on-chain profiles to dynamically calculate and classify the acute mountain sickness risk index; triggering physical execution of oxygen therapy equipment control, drone-delivered medication, or follow-up evacuation plans based on the risk level; aggregating encrypted gradients to update the global prediction model, which is then synchronized to edge nodes after blockchain verification. This invention, through real-time collection of individual physiological data via a 5G network and the construction of a risk prediction model, enables early and accurate warning of the physiological decompensation process of acute mountain sickness.

Description

Technical Field

[0001] This invention belongs to the field of health prevention technology, and in particular relates to a big health AI prevention management system and method based on 5G and blockchain. Background Technology

[0002] Acute mountain sickness early warning is a crucial link in ensuring the health and safety of personnel during work, tourism, and scientific research activities in high-altitude areas. Its core objective is to promptly identify early signs of the disease, gaining a window of opportunity for intervention and reducing the risk of the condition progressing to severe illness. In the high-altitude environment, low pressure and low oxygen conditions pose a direct challenge to human physiological functions. Acute mountain sickness is characterized by both sudden onset and gradual progression. Missing the early intervention window can not only harm the health of personnel but also disrupt the normal conduct of related activities.

[0003] The occurrence and development of acute mountain sickness is essentially a gradual process of physiological function decompensation over time after the body is exposed to a low-pressure and low-oxygen environment. Its early signs are not reflected in the absolute values ​​of static indicators, but are hidden in the abnormal response of physiological parameters to stressors such as activity load, such as the slow recovery rate of blood oxygen saturation after activity and changes in the regulation pattern of heart rate variability.

[0004] In reality, existing early warning technologies struggle to continuously and in real-time capture and analyze the temporal dynamics exhibited during these resting-active state transitions; these subtle characteristics, however, are crucial for early warning. This directly leads to delayed warning signals, failing to provide effective alerts within the early window of reversible physiological changes, thus delaying intervention and allowing the condition to progress to a more severe stage, posing a potential threat to human health and safety. Therefore, the following solutions are proposed to address these issues. Summary of the Invention

[0005] The purpose of this invention is to provide a big health AI prevention and management system and method based on 5G and blockchain. It collects multimodal physiological data of individuals in high-altitude environments in real time through 5G networks and constructs a risk prediction model based on time-series dynamic feature fusion. This enables early and accurate warning of the physiological decompensation process of acute altitude sickness, solving the problem that existing technologies rely on static indicators and subjective scales, which cannot effectively capture the dynamic changes in physiological state, resulting in serious lag in early warning.

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] This invention relates to a big health AI-based prevention and management method based on 5G and blockchain, the management method specifically including the following steps:

[0008] Step S1, Multimodal sensing and acquisition module: Real-time acquisition of personnel physiological data and environmental data through wearable devices;

[0009] Step S2, 5G Edge Transmission Module: Utilizes a military-grade 5G slicing network to encrypt and transmit raw data to edge nodes, and performs motion artifact filtering and local feature calculation;

[0010] Step S3, Blockchain Evidence Storage Module: Stores personnel identities in hash form, manages individual adaptation files and model parameters, and achieves tamper-proof evidence storage and multi-signature consensus for early warning events;

[0011] Step S4, Dynamic Risk Assessment Module: Integrates real-time features and on-chain archives to dynamically calculate and classify the Acute Mountain Sickness Risk Index (AMS-RI);

[0012] Step S5, Early Warning Intervention Execution Module: Based on the risk level, trigger the physical execution of oxygen therapy equipment control, drone drug delivery, or follow-up delivery plans;

[0013] Step S6, Federated Learning Update Module: Aggregate encrypted gradients to update the global prediction model, and synchronize it to the edge nodes after blockchain verification.

[0014] Further, step S1, the multimodal sensing acquisition module specifically includes the following steps:

[0015] Step S11: The person wears an integrated smart garment and helmet, whose built-in sensors collect blood oxygen saturation (SpO2) once per second, collect electrocardiogram signals at a sampling rate of 250Hz (for calculating heart rate variability HRV), and simultaneously calculate gait stability indicators in real time using a triaxial accelerometer. Among them, S g This represents the gait stability quantization value, where n is the total number of sampling points. These represent the acceleration values ​​of the i-th sampling point on the X, Y, and Z axes, respectively; for environmental data, air pressure P (in hPa) and GPS-assisted altitude H are collected once per minute;

[0016] Step S12: Transmit the above multimodal data to the edge computing node in real time through military 5G private network slicing (ensuring transmission latency ≤20ms and bandwidth ≥100Mbps), and simultaneously transmit high-definition motion video streams to meet the needs of subsequent motion analysis, ensuring the real-time performance and integrity of data transmission in plateau patrol scenarios.

[0017] Furthermore, step S2, the 5G edge transmission module specifically includes the following steps:

[0018] Step S21: The edge computing node first performs data cleaning. Based on the motion intensity judgment threshold synthesized by the triaxial accelerometer (|a|>4g), physiological data during periods of intense motion (such as crawling or running) are removed to eliminate motion artifacts.

[0019] Step S22: Extract time-series features from the effective data segments and dynamically calculate three core indicators:

[0020] The rate of decrease in blood oxygen saturation (ΔSpO2 / Δt, where ΔSpO2 is the change in blood oxygen saturation per unit time and Δt is the time interval) quantifies the magnitude of the decrease in blood oxygen saturation per unit time.

[0021] The nonlinear complexity of heart rate variability is analyzed by sample entropy SampEn(m,r,N) (where m is the embedding dimension, r is the similarity tolerance, and N is the length of the HRV sequence).

[0022] Gait Disorder Index (GDI) In the formula, σ(S) g ) is S g The standard deviation of the sequence, μ(S) g ) is S g The mean of the sequence), using the gait stability sequence S g The ratio of the standard deviation σ to the mean μ represents the degree of abnormality in motor coordination.

[0023] Furthermore, step S3, the blockchain evidence storage module specifically includes the following steps:

[0024] Step S31: Generate an irreversible anonymous hash value for each person's unique identifier (ID) using the national cryptographic SM3 hash algorithm and write it to the private blockchain node; simultaneously, associate and store the encrypted individual high-altitude acclimatization file, including historical records of acute mountain sickness (AMS) attacks, high-altitude medication usage data (such as acetazolamide dosage and duration), and recent physical fitness test results; in addition, dynamically calculate the daily ambient air pressure threshold P. crit =580±Δh (where Δh is a correction based on real-time altitude changes), and this threshold is used as a key model parameter to be synchronously stored on the blockchain;

[0025] Step S32: Aggregate the raw data of multimodal sensors (including blood oxygen SpO2, ECG HRV, gait acceleration and ambient air pressure) every 10 minutes at the edge node, construct a Merkle tree and generate a unique root hash value; after signing the root hash, data acquisition timestamp and device physical ID with a digital certificate, submit them to the blockchain consensus network, complete the batch notarization through the PBFT consensus mechanism of the private chain, and ensure data integrity and operation traceability.

[0026] Furthermore, step S4, the dynamic risk assessment module specifically includes the following steps:

[0027] Step S41: The central AI server receives preprocessed multimodal time-series feature data from each edge computing node, and initiates a query request to the blockchain network based on the unique encrypted identifier of the currently monitored person, calling and verifying the individual high-altitude adaptation file stored on the chain, which contains the person's historical health records and personalized model parameters.

[0028] Step S42: The server then performs dynamic data fusion calculations and conducts a comprehensive assessment using the Acute Mountain Sickness Risk Index (AMS-RI) model formula:

[0029]

[0030] In the formula, AMS-RI(t) is the acute mountain sickness risk index at time t, and α, β, γ, and δ are dynamic weighting coefficients. SampEn is the rate of decrease in blood oxygen. HRV Heart rate variability sample entropy, G d The gait disorder index is P, where P is the current real-time air pressure. crit This is the critical air pressure threshold. This refers to the air pressure deviation.

[0031] In this process, the weighting coefficients α (blood oxygen decline rate weight), β (heart rate variability nonlinear characteristic weight), γ (gait disorder index weight), and δ (environmental air pressure deviation weight) in the formula are not fixed values, but are dynamically adjusted and assigned by the individual adaptation profile obtained in step S41, thereby achieving personalized and accurate prediction.

[0032] Furthermore, step S5, the early warning intervention execution module specifically includes the following steps:

[0033] Step S51: The system automatically classifies the risk based on the real-time calculated Acute Mountain Sickness Risk Index (AMS-RI(t)): When RI < 0.3, it is determined to be low risk, and the system only stores the result of the current normal health status as a record in the blockchain for auditing; when 0.3 ≤ RI < 0.6, it is determined to be medium risk, and the sound and light alarm of the medical station terminal in the outpost is immediately triggered, and an AI-generated individualized intervention suggestion is pushed to the person's smart terminal, such as suggesting that they immediately supplement oxygen therapy for 15 minutes; when RI ≥ 0.6, it is determined to be high risk.

[0034] Step S52: For high-risk situations, the system automatically generates an encrypted high-priority early warning event. This event must be confirmed by multi-signatures from consensus nodes deployed on the blockchain (including the front-line command center and the rear medical center) before it can be finally executed. Once approved, the system immediately activates the emergency plan, including the emergency delivery of special medicines by drones and the initiation of the transfer process. At the same time, all key decision data of this early warning event (including time, geographical coordinates, accurate AMS-RI(t) value and the core feature hash value on which the calculation is based) are completely written into the blockchain to form an immutable and fully traceable decision evidence.

[0035] Furthermore, step S6, the federated learning update module specifically includes the following steps:

[0036] Step S61: Each outpost edge node encrypts the locally trained model gradients (rather than the original sensitive physiological data) and transmits them to the command center node via the 5G private network every day.

[0037] Step S62: The command center node aggregates the gradient update amounts of all edge nodes and automatically verifies the legality and consistency of the data source through a smart contract deployed on the blockchain. After ensuring that there is no malicious node interference, the weighted average algorithm is used to update the parameters of the global acute mountain sickness risk prediction model.

[0038] Step S63: The updated model parameters are confirmed by the blockchain consensus mechanism and automatically synchronized to each edge node to achieve collaborative evolution of full-area early warning capabilities and privacy protection.

[0039] A big health AI prevention and management system based on 5G and blockchain, the management system includes a multimodal sensing and acquisition module, a 5G edge transmission module, a blockchain evidence storage module, a dynamic risk assessment module, an early warning and intervention execution module, a federated learning and update module, a medical center node, a command center node, a sentry post oxygen therapy equipment, and a drone delivery system;

[0040] The multimodal sensing acquisition module, 5G edge transmission module, blockchain evidence storage module, dynamic risk assessment module, and early warning intervention execution module are connected in sequence. The output of the 5G edge transmission module is unidirectionally connected to the dynamic risk assessment module. The dynamic risk assessment module is bidirectionally connected to the federated learning and update module. The output of the federated learning and update module is unidirectionally connected to the blockchain evidence storage module. The blockchain evidence storage module is bidirectionally connected to the medical center node and the command center node, respectively. The output of the early warning intervention execution module is unidirectionally connected to the blockchain evidence storage module, the sentry post oxygen therapy equipment, and the drone delivery system, respectively.

[0041] Furthermore, the multimodal sensing acquisition module is used to collect personnel physiological data and environmental data in real time through wearable devices for high-altitude environments; the 5G edge transmission module is used to encrypt and transmit raw data to edge nodes using a military 5G slicing network, and to perform motion artifact filtering and local feature calculation; the blockchain evidence storage module is used to store personnel identities in hash form, manage individual adaptation files and model parameters, and achieve tamper-proof evidence storage and multi-signature consensus for early warning events; the dynamic risk assessment module is used to integrate real-time features and on-chain files to dynamically calculate the acute mountain sickness risk index; and the early warning intervention execution module is used to trigger oxygen therapy equipment control based on the risk level. The system includes: physical execution and tiered management of drone-based medicine delivery or evacuation plans; a federated learning update module for aggregating encrypted gradients to update the global prediction model, which is then synchronized to edge nodes after blockchain verification; a medical center node for providing professional medical judgment on acute mountain sickness intervention plans and authorizing the execution of high-risk warnings; a command center node for coordinating and scheduling medical and logistical resources at all high-altitude outposts; outpost oxygen therapy equipment for providing quantitative and timed supplemental oxygen therapy to designated personnel to alleviate symptoms; and a drone delivery system for autonomously delivering emergency medicines or first-aid supplies to target outposts after receiving a high-risk warning verified by blockchain.

[0042] The present invention has the following beneficial effects:

[0043] 1. This invention leverages the low latency and high concurrency characteristics of 5G networks to continuously and in real-time collect and transmit multimodal physiological data of users in specific high-risk environments. The system utilizes a dynamic fusion algorithm to deeply extract and analyze the temporal physiological characteristics of users during the transition from resting to active states. This ability to capture the trajectory of dynamic physiological changes enables the system to identify abnormal decline trends in physiological states before obvious clinical symptoms appear, thereby achieving early warning of acute disease risks.

[0044] 2. In this invention, all collected raw physiological data and their derived features are generated with cryptographic hash values ​​and anchored to the blockchain immediately after generation, ensuring data integrity and immutability. Based on a smart contract-based access control mechanism, different roles are given access and usage permissions for the data. All data flow activities are recorded on a distributed ledger, forming an irrefutable operation log. This design, while protecting user privacy, constructs a trusted data collaboration framework in a multi-party collaborative and mutually distrustful environment, providing a solid technical foundation for compliant sharing of health data, clear definition of medical responsibility, and transparent auditing of model decisions.

[0045] 3. This invention effectively resolves the contradiction between personalization and universality in health management models by combining blockchain-based individual historical records with a federated learning paradigm. The system can dynamically adjust the parameters and weights of the risk assessment model based on each user's unique historical health records and real-time environmental data, making the warning threshold more aligned with individual physiological characteristics and improving the accuracy of personalized recommendations. At the same time, federated learning ensures that each edge node can collaboratively train without exchanging original sensitive data, only through encrypted model gradient updates. This not only protects data privacy but also aggregates the value of niche scenario data scattered across different regions, driving the continuous evolution of the global prediction model.

[0046] 4. When the system identifies a high-risk situation, the warning event is recorded on the blockchain as a trusted credential verified by multiple parties, and automatically triggers a preset, tamper-proof smart contract. This contract directly coordinates subsequent intervention resources, such as notifying medical units, preparing emergency supplies, or planning evacuation procedures, ensuring the consistency, transparency, and efficiency of key instructions in the transmission and execution process. This design reduces the delays, misunderstandings, or shirking of responsibility that may occur in traditional manual communication, and ensures the automation and reliability of the process from risk discovery to response.

[0047] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart illustrating a big health AI prevention and management method based on 5G and blockchain according to the present invention.

[0050] Figure 2 This is a schematic diagram of the structure of a big health AI prevention management system based on 5G and blockchain according to the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figure 1 As shown, this invention is a big health AI prevention and management method based on 5G and blockchain. The management method includes the following steps:

[0053] Step S1, Multimodal sensing and acquisition module: Real-time acquisition of personnel physiological data and environmental data through wearable devices;

[0054] Step S1, the multimodal sensing acquisition module specifically includes the following steps:

[0055] Step S11: The person wears an integrated smart garment and helmet, whose built-in sensors collect blood oxygen saturation (SpO2) once per second, collect electrocardiogram signals at a sampling rate of 250Hz (for calculating heart rate variability HRV), and simultaneously calculate gait stability indicators in real time using a triaxial accelerometer. Among them, S g This represents the gait stability quantization value, where n is the total number of sampling points. These represent the acceleration values ​​of the i-th sampling point on the X, Y, and Z axes, respectively; for environmental data, air pressure P (in hPa) and GPS-assisted altitude H are collected once per minute;

[0056] Step S12: Transmit the above multimodal data to the edge computing node in real time through military 5G private network slicing (ensuring transmission latency ≤20ms and bandwidth ≥100Mbps), and simultaneously transmit high-definition motion video streams to meet the needs of subsequent motion analysis, ensuring the real-time performance and integrity of data transmission in plateau patrol scenarios.

[0057] Step S2, 5G Edge Transmission Module: Utilizes a military-grade 5G slicing network to encrypt and transmit raw data to edge nodes, and performs motion artifact filtering and local feature calculation;

[0058] Step S2, the 5G edge transmission module specifically includes the following steps:

[0059] Step S21: The edge computing node first performs data cleaning. Based on the motion intensity judgment threshold synthesized by the triaxial accelerometer (|a|>4g), physiological data during periods of intense motion (such as crawling or running) are removed to eliminate motion artifacts.

[0060] Step S22: Extract time-series features from the effective data segments and dynamically calculate three core indicators:

[0061] The rate of decrease in blood oxygen saturation (ΔSpO2 / Δt, where ΔSpO2 is the change in blood oxygen saturation per unit time and Δt is the time interval) quantifies the magnitude of the decrease in blood oxygen saturation per unit time.

[0062] The nonlinear complexity of heart rate variability is analyzed by sample entropy SampEn(m,r,N) (where m is the embedding dimension, r is the similarity tolerance, and N is the length of the HRV sequence).

[0063] Gait Disorder Index (GDI) In the formula, σ(S) g ) is S g The standard deviation of the sequence, μ(S) g ) is S g The mean of the sequence), using the gait stability sequence S g The ratio of the standard deviation σ to the mean μ represents the degree of abnormality in motor coordination.

[0064] Step S3, Blockchain Evidence Storage Module: Stores personnel identities in hash form, manages individual adaptation files and model parameters, and achieves tamper-proof evidence storage and multi-signature consensus for early warning events;

[0065] Step S3, the blockchain evidence storage module specifically includes the following steps:

[0066] Step S31: Generate an irreversible anonymous hash value for each person's unique identifier (ID) using the national cryptographic SM3 hash algorithm and write it to the private blockchain node; simultaneously, associate and store the encrypted individual high-altitude acclimatization file, including historical records of acute mountain sickness (AMS) attacks, high-altitude medication usage data (such as acetazolamide dosage and duration), and recent physical fitness test results; in addition, dynamically calculate the daily ambient air pressure threshold P. crit =580±Δh (where Δh is a correction based on real-time altitude changes), and this threshold is used as a key model parameter to be synchronously stored on the blockchain;

[0067] Step S32: Aggregate the raw data of multimodal sensors (including blood oxygen SpO2, ECG HRV, gait acceleration and ambient air pressure) every 10 minutes at the edge node, construct a Merkle tree and generate a unique root hash value; after signing the root hash, data acquisition timestamp and device physical ID with a digital certificate, submit them to the blockchain consensus network, complete the batch notarization through the PBFT consensus mechanism of the private chain, and ensure data integrity and operation traceability.

[0068] Step S4, Dynamic Risk Assessment Module: Integrates real-time features and on-chain archives to dynamically calculate and classify the Acute Mountain Sickness Risk Index (AMS-RI);

[0069] Step S4, the dynamic risk assessment module specifically includes the following steps:

[0070] Step S41: The central AI server receives preprocessed multimodal time-series feature data from each edge computing node, and initiates a query request to the blockchain network based on the unique encrypted identifier of the currently monitored person, calling and verifying the individual high-altitude adaptation file stored on the chain, which contains the person's historical health records and personalized model parameters.

[0071] Step S42: The server then performs dynamic data fusion calculations and conducts a comprehensive assessment using the Acute Mountain Sickness Risk Index (AMS-RI) model formula:

[0072]

[0073] In the formula, AMS-RI(t) is the acute mountain sickness risk index at time t, and α, β, γ, and δ are dynamic weighting coefficients. SampEn is the rate of decrease in blood oxygen. HRV Heart rate variability sample entropy, G d The gait disorder index is P, where P is the current real-time air pressure. crit This is the critical air pressure threshold. This refers to the air pressure deviation.

[0074] In this process, the weighting coefficients α (blood oxygen decline rate weight), β (heart rate variability nonlinear characteristic weight), γ (gait disorder index weight), and δ (environmental air pressure deviation weight) in the formula are not fixed values, but are dynamically adjusted and assigned by the individual adaptation profile obtained in step S41, thereby achieving personalized and accurate prediction.

[0075] Step S5, Early Warning Intervention Execution Module: Based on the risk level, trigger the physical execution of oxygen therapy equipment control, drone drug delivery, or follow-up delivery plans;

[0076] Step S5, the early warning intervention execution module specifically includes the following steps:

[0077] Step S51: The system automatically classifies the risk based on the real-time calculated Acute Mountain Sickness Risk Index (AMS-RI(t)): When RI < 0.3, it is determined to be low risk, and the system only stores the result of the current normal health status as a record in the blockchain for auditing; when 0.3 ≤ RI < 0.6, it is determined to be medium risk, and the sound and light alarm of the medical station terminal in the outpost is immediately triggered, and an AI-generated individualized intervention suggestion is pushed to the person's smart terminal, such as suggesting that they immediately supplement oxygen therapy for 15 minutes; when RI ≥ 0.6, it is determined to be high risk.

[0078] Step S52: For high-risk situations, the system automatically generates an encrypted high-priority early warning event. This event must be confirmed by multi-signatures from consensus nodes deployed on the blockchain (including the front-line command center and the rear medical center) before it can be finally executed. Once approved, the system immediately activates the emergency plan, including the emergency delivery of special medicines by drones and the initiation of the transfer process. At the same time, all key decision data of this early warning event (including time, geographical coordinates, accurate AMS-RI(t) value and the core feature hash value on which the calculation is based) are completely written into the blockchain to form an immutable and fully traceable decision evidence.

[0079] Step S6, Federated Learning Update Module: Aggregate encrypted gradients to update the global prediction model, and synchronize it to the edge nodes after blockchain verification;

[0080] Step S6, the Federated Learning Update Module specifically includes the following steps:

[0081] Step S61: Each outpost edge node encrypts the locally trained model gradients (rather than the original sensitive physiological data) and transmits them to the command center node via the 5G private network every day.

[0082] Step S62: The command center node aggregates the gradient update amounts of all edge nodes and automatically verifies the legality and consistency of the data source through a smart contract deployed on the blockchain. After ensuring that there is no malicious node interference, the weighted average algorithm is used to update the parameters of the global acute mountain sickness risk prediction model.

[0083] Step S63: The updated model parameters are confirmed by the blockchain consensus mechanism and automatically synchronized to each edge node to achieve collaborative evolution of full-area early warning capabilities and privacy protection.

[0084] Please see Figure 2 As shown, the present invention is a big health AI prevention and management system based on 5G and blockchain. The management system includes a multimodal sensing and acquisition module, a 5G edge transmission module, a blockchain evidence storage module, a dynamic risk assessment module, an early warning and intervention execution module, a federated learning and update module, a medical center node, a command center node, a sentry post oxygen therapy equipment, and a drone delivery system.

[0085] The multimodal sensing acquisition module, 5G edge transmission module, blockchain evidence storage module, dynamic risk assessment module, and early warning intervention execution module are connected in sequence. The output of the 5G edge transmission module is unidirectionally connected to the dynamic risk assessment module. The dynamic risk assessment module is bidirectionally connected to the federated learning and update module. The output of the federated learning and update module is unidirectionally connected to the blockchain evidence storage module. The blockchain evidence storage module is bidirectionally connected to the medical center node and the command center node, respectively. The output of the early warning intervention execution module is unidirectionally connected to the blockchain evidence storage module, the sentry post oxygen therapy equipment, and the drone delivery system, respectively.

[0086] The multimodal sensing and acquisition module is used to collect real-time physiological and environmental data of personnel through wearable devices in high-altitude areas; the 5G edge transmission module is used to encrypt and transmit raw data to edge nodes using a military-grade 5G slicing network, and to perform motion artifact filtering and local feature calculation; the blockchain notarization module is used to store personnel identities in hash form, manage individual adaptation files and model parameters, and achieve tamper-proof notarization and multi-signature consensus for early warning events; the dynamic risk assessment module is used to integrate real-time features and on-chain files to dynamically calculate the acute mountain sickness risk index; and the early warning intervention execution module is used to trigger the control of oxygen therapy equipment and unmanned aerial vehicles based on the risk level. The system includes: physical execution and tiered management of medical delivery or evacuation plans; a federated learning update module for aggregating encrypted gradients to update the global prediction model, which is then synchronized to edge nodes after blockchain verification; medical center nodes for providing professional medical judgment on acute mountain sickness intervention plans and authorizing the execution of high-risk warnings; command center nodes for coordinating and scheduling medical and logistical resources at all high-altitude outposts; outpost oxygen therapy equipment for providing quantitative and timed supplemental oxygen therapy to designated personnel to alleviate symptoms; and a drone delivery system for autonomously delivering emergency medicines or first-aid supplies to target outposts after receiving a high-risk warning verified by blockchain.

[0087] One specific application of this embodiment is:

[0088] Background: At a post at an altitude of 4,500 meters, border guard A (code name Z-007) and his squad were ordered to carry out a 3-hour border foot patrol mission. The patrol route ranged in altitude from 4,450 meters to 4,550 meters. Although A was a healthy male, this was his first time being stationed at this altitude since enlisting. His individual high-altitude acclimatization record showed that he had no history of AMS attacks, but his physical fitness test data showed that his hypoxia tolerance was at the average level of the company.

[0089] Implementation steps:

[0090] Step S1: Task Start and Data Acquisition

[0091] A puts on the training uniform and helmet with integrated sensors; the system initializes, and each sensor starts working.

[0092] Within the outpost (baseline altitude H = 4500m, air pressure P ≈ 580hPa), baseline data of the animal's resting state were collected: SpO2 was 92%, HRV sample entropy SampEn0 = 1.25, and gait stability baseline Sg0 = 1.05m / s. 2 ;

[0093] These baseline data, associated with their identity IDs (0x7a3f...c89b after SM3 hashing), serve as the starting point for this task's data recording. They are then packaged and a Merkle root hash is generated, which is transmitted via the 5G network to edge nodes and the blockchain network for storage.

[0094] Step S2: Continuous monitoring and edge computing during patrol

[0095] A began patrolling with the shift team; the 5G network continuously transmitted its physiological data with low latency.

[0096] When patrolling up a steep slope (altitude rises to 4550m, air pressure drops to 575hPa), A's heart rate increases; the edge node calculates its SpO2 decrease slope in real time:

[0097]

[0098] Meanwhile, the edge nodes calculated a slight increase in their gait disorder index from their accelerometer data:

[0099]

[0100] These time-series features are extracted in real time and sent to the central AI server.

[0101] Step S3: Dynamic Risk Index (AMS-RI) Calculation and Blockchain Verification

[0102] The central AI server receives the feature data packet at T+120 minutes. Before calculation, it first queries the individual profile of A stored on the blockchain through a smart contract to verify A's identity and obtain A's personal parameters (the weighting coefficients are adjusted to α=0.4, β=0.3, γ=0.2, δ=0.1).

[0103] The server obtains the current ambient air pressure P = 575 hPa and the individual's critical threshold P. c rit = 579 hPa;

[0104] By combining all the data, the server performs dynamic fusion computation for AMS-RI:

[0105]

[0106] Step S4: Risk Warning and Intervention Implementation

[0107] The system automatically generates a medium-risk warning event; after the event (including time, coordinates, RI value, key data hash 0xef23...d17a) is confirmed by the blockchain consensus mechanism (the digital signatures of the nodes of the medical officer in the command center and the military doctor at the outpost), it is written into the blockchain to form an immutable record.

[0108] Meanwhile, early warning information and intervention suggestions are pushed to:

[0109] The outpost medic's smart terminal displayed: "Warning: Z-007, AMS-RI = 0.44, Recommendation: Administer portable oxygen therapy for 15 minutes at the next rest stop, and closely monitor the patient."

[0110] A's information exchange terminal: issues a yellow warning signal;

[0111] The team leader received instructions from the doctor to urge A to use the portable oxygen cylinder he was carrying to inhale oxygen at the next planned rest point at T+130 minutes.

[0112] Step S5: Post-intervention feedback and model iteration

[0113] Fifteen minutes after oxygen therapy and during subsequent patrols, A's SpO2 rose to 91%, HRV stabilized, and the AMS-RI value dropped to 0.32 (low risk) in subsequent calculations.

[0114] The complete time-series data, RI calculation process, early warning events, and physiological data feedback after intervention for this mission, along with their hash digests, are all recorded on the blockchain.

[0115] That night, during non-task hours, the outpost edge nodes used local new data (after anonymization) to perform federated learning training and generate model gradient update files;

[0116] The gradient file is encrypted and uploaded to the command center node; the command center node aggregates the gradients from multiple outposts, verifies the data validity and consistency through a smart contract, and then updates the global AMS-RI prediction model.

[0117] The updated model parameters are synchronized to the outpost edge nodes via blockchain to improve the accuracy of predictions for soldiers in the area the following day.

[0118] This embodiment demonstrates the complete closed loop of the system from data acquisition, real-time calculation, intelligent decision-making to intervention feedback; it utilizes 5G to achieve ultra-low latency multimodal dynamic data backhaul, and uses blockchain technology to reliably integrate static archives and dynamic data to calculate the risk index, ultimately successfully triggering a traceable early intervention before symptoms appear, demonstrating the practical value and advanced nature of this method in special application scenarios.

[0119] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0120] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A big health AI-based prevention and management method based on 5G and blockchain, characterized in that, The management method specifically includes the following steps: Step S1, Multimodal sensing and acquisition module: Real-time acquisition of personnel physiological data and environmental data through wearable devices; Step S2, 5G Edge Transmission Module: Utilizes a military-grade 5G slicing network to encrypt and transmit raw data to edge nodes, and performs motion artifact filtering and local feature calculation; Step S3, Blockchain Evidence Storage Module: Stores personnel identities in hash form, manages individual adaptation files and model parameters, and achieves tamper-proof evidence storage and multi-signature consensus for early warning events; Step S4, Dynamic Risk Assessment Module: Integrates real-time features and on-chain records to dynamically calculate and classify the acute mountain sickness risk index; Step S5, Early Warning Intervention Execution Module: Based on the risk level, trigger the physical execution of oxygen therapy equipment control, drone drug delivery, or follow-up delivery plans; Step S6, Federated Learning Update Module: Aggregate encrypted gradients to update the global prediction model, and synchronize it to the edge nodes after blockchain verification.

2. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S1, the multimodal sensing acquisition module specifically includes the following steps: Step S11: Personnel wear integrated smart clothing and helmets, whose built-in sensors collect blood oxygen saturation and electrocardiogram signals, and at the same time calculate gait stability indicators in real time through a three-axis accelerometer; in terms of environmental data, air pressure and GPS-assisted altitude are collected. Step S12: Transmit the above multimodal data to the edge computing node in real time through military 5G private network slicing, and simultaneously transmit high-definition motion video streams to meet the needs of subsequent motion analysis, ensuring the real-time performance and integrity of data transmission in plateau patrol scenarios.

3. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S2, the 5G edge transmission module specifically includes the following steps: Step S21: The edge computing node first performs data cleaning, and based on the motion intensity judgment threshold synthesized by the triaxial accelerometer, removes physiological data during periods of intense exercise to eliminate motion artifacts; Step S22: Extract time-series features from the effective data segment and dynamically calculate the rate of decrease in blood oxygen, the nonlinear complexity of heart rate variability, and the gait disorder index.

4. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S3, the blockchain evidence storage module specifically includes the following steps: Step S31: Generate an irreversible anonymous hash value for each person's unique identifier using the national cryptographic SM3 hash algorithm, and write it into the private chain node; The system stores encrypted individual altitude acclimatization profiles, including records of historical acute mountain sickness attacks, data on altitude medication use, and recent physical fitness test results. Dynamically calculate the daily ambient air pressure threshold and synchronously upload this threshold to the blockchain as a key model parameter for evidence storage. Step S32: Aggregate the raw data from the multimodal sensors at the edge nodes every certain period of time, construct a Merkle tree and generate a unique root hash value; after signing the root hash, data acquisition timestamp, and device physical ID with a digital certificate, submit them to the blockchain consensus network, complete batch notarization through the PBFT consensus mechanism of the private chain, and ensure data integrity and operation traceability.

5. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S4, the dynamic risk assessment module specifically includes the following steps: Step S41: The central AI server receives preprocessed multimodal time-series feature data from each edge computing node, and initiates a query request to the blockchain network based on the unique encrypted identifier of the currently monitored person, calling and verifying the individual high-altitude adaptation file stored on the chain, which contains the person's historical health records and personalized model parameters. Step S42: The server then performs dynamic data fusion calculations and conducts a comprehensive assessment using the acute mountain sickness risk index model formula.

6. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S5, the early warning intervention execution module specifically includes the following steps: Step S51: The system automatically classifies the risk based on the acute mountain sickness risk index value calculated in real time; Step S52: Write the key data of the early warning event into the blockchain to generate an immutable certificate.

7. The big health AI prevention and management method based on 5G and blockchain according to claim 1, characterized in that, Step S6, the federated learning update module specifically includes the following steps: Step S61: Each outpost edge node encrypts the locally trained model gradients daily and transmits them to the command center node via the 5G private network. Step S62: The command center node aggregates the gradient update amounts of all edge nodes and automatically verifies the legality and consistency of the data source through a smart contract deployed on the blockchain. After ensuring that there is no malicious node interference, the weighted average algorithm is used to update the parameters of the global acute mountain sickness risk prediction model. Step S63: The updated model parameters are confirmed by the blockchain consensus mechanism and automatically synchronized to each edge node to achieve collaborative evolution of full-area early warning capabilities and privacy protection.

8. A big health AI prevention management system based on 5G and blockchain, characterized in that, The management system includes a multimodal sensing acquisition module, a 5G edge transmission module, a blockchain evidence storage module, a dynamic risk assessment module, an early warning and intervention execution module, a federated learning and update module, a medical center node, a command center node, outpost oxygen therapy equipment, and a drone delivery system. The multimodal sensing acquisition module, 5G edge transmission module, blockchain evidence storage module, dynamic risk assessment module, and early warning intervention execution module are connected in sequence. The output of the 5G edge transmission module is unidirectionally connected to the dynamic risk assessment module. The dynamic risk assessment module is bidirectionally connected to the federated learning and update module. The output of the federated learning and update module is unidirectionally connected to the blockchain evidence storage module. The blockchain evidence storage module is bidirectionally connected to the medical center node and the command center node, respectively. The output of the early warning intervention execution module is unidirectionally connected to the blockchain evidence storage module, the sentry post oxygen therapy equipment, and the drone delivery system, respectively.

9. A big health AI prevention management system based on 5G and blockchain according to claim 8, characterized in that, The multimodal sensing acquisition module is used to collect personnel physiological data and environmental data in real time through wearable devices for high altitudes; the 5G edge transmission module is used to encrypt and transmit raw data to edge nodes using a military 5G slicing network, and perform motion artifact filtering and local feature calculation; the blockchain evidence storage module is used to store personnel identities in hash form, manage individual adaptation files and model parameters, and realize tamper-proof evidence storage and multi-signature consensus for early warning events; the dynamic risk assessment module is used to integrate real-time features and on-chain files to dynamically calculate the acute mountain sickness risk index. The early warning and intervention execution module is used to trigger the physical execution and classification of oxygen therapy equipment control, drone drug delivery, or evacuation plans based on risk levels; the federated learning and update module is used to aggregate encrypted gradients to update the global prediction model, which is then synchronized to edge nodes after blockchain verification; the medical center node is used to provide medical professional judgment on acute mountain sickness intervention plans and authorize the execution of high-risk early warnings; the command center node is responsible for coordinating and scheduling medical and logistical resources of all plateau outposts; the outpost oxygen therapy equipment is used to provide quantitative and timed supplemental oxygen therapy to designated personnel to alleviate symptoms; and the drone delivery system is used to autonomously execute the task of urgently delivering medicines or emergency supplies to the target outpost after receiving a high-risk early warning verified by the blockchain.

Images