CN-122025117-A - Deep medical first-aid digital twin method and system based on environmental physiological coupling algorithm

CN122025117ACN 122025117 ACN122025117 ACN 122025117ACN-122025117-A

Abstract

The invention discloses a deep medical first-aid digital twin method based on an environment physiological coupling algorithm, which comprises the following steps of data acquisition and digital twin modeling, first-aid knowledge matching, intelligent first-aid scheme generation, scheme execution and real-time feedback, scheme pushing to a deep rescue terminal through ultralow frequency communication and Beidou short message double links, receiving execution state feedback in real time, updating a digital twin model, and dynamically adjusting a scheme according to environment and physiological data changes to form closed-loop management. The environment and physiology coupling algorithm disclosed by the invention has the beneficial effects that the limitation of the traditional single parameter association is broken through the multi-parameter collaborative modeling of temperature, pressure, gas concentration, humidity and the like, the comprehensive influence of the deep complex environment on the core physiological indexes such as heart rate, blood oxygen and body temperature of a human body can be reflected dynamically and accurately, the problem of disjoint of environment scenes and human physiological states in the traditional scheme is effectively solved, and the emergency scheme can be ensured to fit the real physiological needs of trapped people.

Inventors

WANG HUANHUAN
LIU FULONG
Song Daining
YAN XIANLIANG
WU XIANG

Assignees

深地科学与工程云龙湖实验室
徐州医科大学

Dates

Publication Date: 20260512
Application Date: 20251223

Claims (10)

1. A deep medical first-aid digital twin method based on an environmental physiological coupling algorithm is characterized by comprising the following steps: The method comprises the steps of1, data acquisition and digital twin modeling, deep environment parameter and trapped person physiological data acquisition, construction of a deep environment, an emergency scene and a digital human body trinity digital twin model, wherein the digital human body model integrates an environment and a physiological coupling algorithm, the effect of the deep environment on the physiological state of a human body is dynamically simulated, 2, emergency knowledge matching, entity identification and relation extraction are carried out on the acquired data through a natural language processing technology, a deep exclusive emergency knowledge map is called, an emergency knowledge subset matched with the current scene is matched, the knowledge map is constructed based on a BERT model subjected to deep corpus three-order fine adjustment, 3, intelligent emergency scheme generation is carried out, an environment risk-human body tolerance two-dimensional priority algorithm is adopted, an expert rule base is combined to output a basic emergency framework, the effect of the digital twin model is utilized through BERT sequence decision model optimization step based on deep emergency case training, iterative optimization is carried out until the requirement is met, 4, the scheme is executed and fed back to the deep emergency terminal through an ultra-low frequency communication and Beidou short message dual link, the real-time state is updated, and the dynamic management and the digital twin dynamic management model is formed.
2. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm is characterized in that the geological early warning of the trinity digital twin model in the step 1 comprises three-level thresholds, namely, the primary early warning geological displacement is 0.3-0.5mm/h, the secondary early warning is 0.5-1.0mm/h, the three-level early warning is greater than 1.0mm/h, and the three-level early warning corresponds to different risk grades and physiological simulation responses.
3. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm according to claim 1, wherein the environmental and physiological coupling algorithm in step 1 is a piecewise nonlinear fitting model, and the dynamic association of the quantized multi-environmental parameters and the human core physiological index comprises: the heart rate correlation formula is calculated according to the characteristic intervals of temperature, pressure and oxygen concentration in a sectional manner; the blood oxygen saturation reduction rate association formula is calculated in a sectional manner according to the characteristic interval of the concentration of CO and the concentration of oxygen; The temperature rise rate correlation formula is calculated according to the characteristic interval of temperature and humidity in a segmentation way.
4. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm according to claim 1, wherein the deep proprietary emergency knowledge graph in step 2 comprises a three-layer structure: An entity layer, which contains deep unique environment, symptoms, treatments and equipment entities; relationship layer, including deeply exclusive environment-constraint-treatment association rule; Attribute layer-the labeling device adapts the attributes deeply.
5. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm according to claim 1, wherein the BERT model subjected to deep corpus third-order fine tuning in the step 2 adopts the following method: The method comprises the steps of (1) deep corpus layering pretreatment and feature enhancement, namely layering 15 ten thousand deep corpora, double coding 120 exclusive terms, (2) field self-adaptive pre-training and cross-modal constraint, namely unsupervised pre-training, weak supervision contrast learning, text and numerical characteristics fusion, and (3) task oriented fine tuning and dynamic iterative optimization, namely designing a multi-task joint loss function, and performing incremental training after 10 rescue cases.
6. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm according to claim 1, wherein the calculation formula of the environmental risk-human tolerance two-dimensional priority algorithm in the step 3 is as follows: Wherein ER is an environmental risk, PT is human tolerance, alpha is a dynamic adjustment coefficient, when ER is more than 0.8 or geological early warning is more than or equal to the second level, alpha=0.7, otherwise, alpha=0.3.
7. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm according to claim 6, wherein the ER is calculated as: the grading and scoring of each parameter are set according to deep-ground environment risk characteristics, and the PT is calculated as follows: as a score for the history of the underlying disease, Is a real-time physiological index score.
8. The deep medical emergency digital twin method based on the environmental physiological coupling algorithm of claim 1, wherein the scheme optimizing and deducting mechanism in the step 3 comprises a preset emergency effect threshold library, scheme effects are deduced through a digital twin model, and when the scheme effects are not up to standard, the suitability of the knowledge graph judging equipment is called and a treatment step is newly added.
9. A deep medical emergency treatment intelligent system implementing the environmental physiological coupling algorithm-based deep medical emergency digital twin method of any of claims 1-8, comprising: The system comprises a data acquisition and twinning modeling module, a knowledge matching module, an intelligent scheme generating module, an execution and feedback module, a double-link communication unit, a deep rescue terminal and a real-time monitoring and dynamic adjustment unit, wherein the data acquisition and twinning modeling module comprises a distributed environment sensor, a wearable physiological monitoring terminal and a twinning model construction unit, the knowledge matching module comprises an NLP preprocessing unit, a deep first-aid knowledge map and a BERT knowledge matching unit, the intelligent scheme generating module comprises a two-dimensional priority computing unit, a BERT sequence decision unit and a digital twinning deduction and optimization unit, and the execution and feedback module comprises a double-link communication unit, a deep rescue terminal and a real-time monitoring and dynamic adjustment unit.
10. The deep medical emergency treatment intelligent system of the deep medical emergency digital twin method based on the environmental physiological coupling algorithm of claim 9, wherein the double-link communication unit adopts: Ultra-low frequency communication, namely, an annular magnetic dipole antenna penetrates through a 1000-3000 m rock stratum, and the transmission rate is 200 bps+/-20 bps; and the Beidou short message is used for transmitting 140 bytes at a time and preferentially transmitting physiological data and key steps of the scheme.

Description

Deep medical first-aid digital twin method and system based on environmental physiological coupling algorithm Technical Field The invention relates to the technical field of deep emergency rescue and artificial intelligence crossing, in particular to a deep medical first-aid digital twin method and system based on an environmental physiological coupling algorithm. Background The deep environment has the characteristics of high temperature, high pressure, low oxygen, high humidity, complex geological structure, serious communication signal attenuation and the like, and the medical emergency faces the technical bottleneck for a long time: The environment-physiological coupling simulation is lacking, namely the existing system only focuses on the simulation of geological parameters, and does not establish a dynamic correlation model of temperature, pressure, gas, humidity and other multi-environment parameters, human heart rate, blood oxygen and body temperature, so that the first-aid scheme is disjointed from the physiological state. The adaptability of deep first-aid knowledge is insufficient, the general medical knowledge graph lacks deep exclusive rules, the hidden experience is unstructured, and the knowledge is called to frequently mismatch scenes. The decision algorithm lacks multidimensional synergy, the existing system is singly dependent on physiological indexes, does not integrate environmental risks such as geological collapse, harmful gas diffusion and the like, and is easy to neglect rescue safety in a composite scene. The communication and hardware extreme resistance is poor, 4G/5G fails at the depth of 1000 meters, the tolerance temperature of the existing sensor is lower than 50 ℃, the data drift exceeds 10%, the terminal has no explosion-proof design, and the deep explosion-proof requirement cannot be met. The prior art can not solve the problems, and an integrated deep first-aid intelligent scheme taking environment-physiological coupling as a core is needed to be constructed, so that the safety and the high efficiency of rescue are ensured. Disclosure of Invention The invention mainly solves the technical problem of providing a deep medical emergency digital twin method and a system based on an environmental physiological coupling algorithm, and solves one or more of the problems in the prior art. In order to solve the technical problems, the invention adopts a technical scheme that the deep medical first-aid digital twin method based on the environmental physiological coupling algorithm is characterized by comprising the following steps: The method comprises the steps of1, data acquisition and digital twin modeling, deep environment parameter and trapped person physiological data acquisition, construction of a deep environment, an emergency scene and a digital human body trinity digital twin model, wherein the digital human body model integrates an environment and a physiological coupling algorithm, the effect of the deep environment on the physiological state of a human body is dynamically simulated, 2, emergency knowledge matching, entity identification and relation extraction are carried out on the acquired data through a natural language processing technology, a deep exclusive emergency knowledge map is called, an emergency knowledge subset matched with the current scene is matched, the knowledge map is constructed based on a BERT model subjected to deep corpus three-order fine adjustment, 3, intelligent emergency scheme generation is carried out, an environment risk-human body tolerance two-dimensional priority algorithm is adopted, an expert rule base is combined to output a basic emergency framework, the effect of the digital twin model is utilized through BERT sequence decision model optimization step based on deep emergency case training, iterative optimization is carried out until the requirement is met, 4, the scheme is executed and fed back to the deep emergency terminal through an ultra-low frequency communication and Beidou short message dual link, the real-time state is updated, and the dynamic management and the digital twin dynamic management model is formed. In some embodiments, the geological early warning of the trinity digital twin model in the step 1 comprises three-level thresholds, namely, a first-level early warning geological displacement is 0.3-0.5mm/h, a second-level early warning is 0.5-1.0mm/h, and the three-level early warning is greater than 1.0mm/h, and corresponds to different risk levels and physiological simulation responses. In some embodiments, the environmental and physiological coupling algorithm in step 1 is a piecewise nonlinear fitting model, and the dynamic association of the environmental parameters and the human core physiological index is quantified, and the dynamic association includes: the heart rate correlation formula is calculated according to the characteristic intervals of temperature, pressure and oxygen concentration in a sectional ma