Search

CN-122025138-A - Biological feature analysis-based health intention judging system, certificate and interaction method

CN122025138ACN 122025138 ACN122025138 ACN 122025138ACN-122025138-A

Abstract

The invention provides a health intention judging system, a certificate and an interaction method based on biological feature analysis, wherein the judging system comprises a multi-mode biological feature acquisition module, an environment parameter acquisition module, a personalized baseline modeling module, a dynamic trend judging module, an abnormal marking and confidence degree module, an XR black box certificate storage module, an intention token generation module and an intelligent agent crossing interaction interface module. The invention realizes risk early warning before a static threshold value by analyzing trend changes of biological characteristics in real time, particularly monitoring instantaneous slope of physical signs, triggers image evidence storage of an original physical layer signal at a critical moment when judging to occur to form an untampereable evidence chain, and upgrades health perception from a reading tool to a standard entrance capable of auditing, overtaking responsibility and large-scale performance by taking an intention token as a unified trust interface crossing intelligent body performance.

Inventors

  • CHEN HUICHONG

Assignees

  • 武汉华创全息影像技术有限公司

Dates

Publication Date
20260512
Application Date
20260203

Claims (10)

  1. 1. A health intent determination system based on biometric analysis, characterized in that the determination system comprises the following modules: The multi-mode biological feature acquisition module is used for acquiring at least one biological feature original signal, wherein the original signal comprises an infrared array thermal signal/thermal pixel matrix, spectrum sampling and PPG original waveform original temperature sampling value, and can further extract or infer derived features of heart rate, respiratory rhythm, face redness index and skin temperature; the environment parameter acquisition module is used for acquiring environment variables including temperature and humidity, wind speed, indoor and outdoor states, air conditioner air outlet influence and motion state inference; The individuation baseline modeling module is used for constructing a user history health image or a short-term baseline so as to provide the individual difference correction capability of the user; The dynamic trend judging module is used for carrying out time sequence analysis on the biological characteristics, calculating the change slope and the deviation degree, and outputting a health evaluation score S by combining with environment compensation to further generate an abnormal intention mark; The abnormal marking and confidence coefficient module is used for outputting an abnormal level, confidence coefficient and suggested action type and is used as one of the evidence storage triggering conditions; the XR black box evidence storage module is used for triggering an original signal mirror image, a logic signature and a hash chain package at the moment related to abnormality judgment to form a tamper-proof package; The intention Token generation module is used for packaging the abnormal judgment result into an Intent Token containing evidence hash references and providing an A2A standard output interface; And the cross-agent interaction interface module is used for interfacing with a third party agent and realizing automatic performance or wind control auditing based on an Intent Token.
  2. 2. The biometric analysis-based wellness intent determination system of claim 1, wherein the wellness intent determination system is further configured to perform a wellness intent determination method comprising: s1, acquiring original signals of biological characteristics of a user through at least one biological characteristic sensing unit, and forming a time sequence ; S2, calculating a time sequence generalized change rate index of the biological characteristics The generalized change rate index comprises a first-order difference/slope, a second-order difference, a window regression slope, a window variance change rate, a prediction residual change rate and a cumulative offset statistic, and is used as a core control point for anomaly judgment; S3, based on environment variables Generating an environmental compensation factor Wherein The dynamic weight can be changed along with the environment, and a constant 1 can be taken under the ideal environment; S4, individuation baseline based on users Generating individual baseline mapping factors Wherein The dynamic mapping function can be adopted, and a constant 1 can be adopted under the default condition; s5, based on the generalized change rate index And/or At least one of which generates a health assessment score and outputs an abnormal intent identification.
  3. 3. The biometric analysis-based health intention determination system according to claim 2, wherein the biometric sensing unit in S1 includes: an infrared array or a contact temperature acquisition unit integrated in the handheld device; an optical/infrared sensing unit or a facial thermal feature acquisition unit integrated with the headset; the far-field heat sensing array, the mattress array or the room environment sensing unit is integrated in the intelligent household equipment; the PPG/heart rate/body temperature comprehensive sensing unit is integrated in the wearable equipment; and the heat radiation sensing unit is integrated in a vehicle or a public space.
  4. 4. The system for determining the health intention based on the biological feature analysis according to claim 2, wherein the system for determining the health intention performs time stamp synchronization and normalization resampling on multi-modal asynchronous signals from different sensors through the sensing unit before performing the evaluation, and the weight coefficient is provided with a dynamic adjustment mechanism.
  5. 5. The system for determining the healthy intention based on the biological feature analysis according to claim 4, wherein the system for determining the healthy intention evaluates the confidence of each perception dimension in real time and automatically increases the weight of a high confidence mode; The health intent determination system evaluates the signal-to-noise ratio SNR by monitoring the variance or fundamental frequency of the sensor signal.
  6. 6. The biometric analysis-based health intention determination system according to claim 4, wherein the health intention determination system enforces XR black box certification at an associated time T at which an abnormality determination is generated; The health intention judging system also establishes a pre-storage mechanism, continuously circularly caches an original physical signal of N frames or N seconds in front of the sensor, and when judging that the time T triggering the evidence storage instruction occurs, the system derives an original physical signal mirror image containing a [ T-N, T+M ] time interval, wherein N is a backtracking window before triggering, and M is an extended window after triggering; the system adopts a circular buffer to manage pre-stored signals, and if abnormality judgment is not triggered, old data is automatically covered by new samples, so that the resource occupation on a low-power-consumption terminal is ensured to be minimized.
  7. 7. A certification method executed by the health intention judgment system according to claim 1, wherein when the associated time T of generating abnormal intention identification, health early warning or medical action suggestion by the health intention judgment system arrives, the following flow is triggered to be executed: Image capturing, namely extracting and storing original signals of a physical layer of a sensor in windows [ T-N, T+M ] before and after a moment T, wherein the original signals of the physical layer comprise at least one original voltage/current/thermal pixel matrix/spectrum sampling/PPG waveform; a logic signature, namely packaging the original signal abstract, a judging result, an algorithm version number, environment parameters and a key reasoning abstract; Hash chain encapsulation, namely generating a unique hash chain code for encapsulated content, and encrypting and storing the unique hash chain code to realize non-tamperability; and outputting a verified evidence index or evidence hash reference for accident restoration and law tracing.
  8. 8. The method of claim 7, wherein the original physical layer signal is a sensor ADC sampling sequence or an unfiltered pixel matrix/waveform, and the generation time is earlier than the feature extraction, filtering and calibration, the image capture is a retrospective capture based on a determination time T, and at least comprises an anomaly induction interval of [ T-N, T ], wherein N is not less than an algorithmic reasoning delay time.
  9. 9. An interaction method executed in a health intention judgment system according to claim 1, characterized in that the interaction method comprises the following procedures: the method comprises the steps of packaging an abnormal intention mark into an intention token, wherein the token at least comprises an abnormal grade, a confidence level and an evidence hash reference bound with XR black box evidence Bao Jiang, taking the intention token as a unique legal trigger source for waking up a third-party agent to execute a performing action, and triggering at least one action by the third-party agent based on the intention token to record and verify the consumption behavior of the token so as to ensure that a performing link can be audited and can be overtaken.
  10. 10. The interaction method according to claim 9, wherein the action triggered by the intention token comprises buying medicine distribution, medical registration, consultation of family doctor, family notification, insurance claim or environmental control adjustment, and wherein a causal logic mapping based on evidence hash or unique identifier is established between the abnormal intention judgment result and the subsequent performing action, and the external presentation form of the mapping comprises an explicit intention token data packet, a private protocol, memory sharing and inter-process communication; The intention token comprises a time stamp, a disposable random number Nonce and a validity period TTL, and the third party agent verifies that the Nonce is not consumed before performing the performance and is in the validity period of the TTL and records a token consumption receipt after the performance is completed.

Description

Biological feature analysis-based health intention judging system, certificate and interaction method Technical Field The invention relates to the fields of artificial intelligence big health, biological feature perception and time sequence trend analysis, trusted evidence security audit and cross-agent (A2A) interaction protocols, in particular to a health intention judging system, an evidence and an interaction method based on biological feature analysis. Background With the development of health monitoring devices and AI agents, health terminals have evolved from "passive reading displays" to "active advice and automatic performance. For example, AI thermometers may automatically recommend medications and connect to a pharmacy for delivery after detecting abnormal body temperature, AI glasses may detect the risk of heatstroke of the wearer and guide to the nearest clinic, smart home systems may find abnormal fluctuations in night body temperature and automatically adjust the environment or notify family members. However, the prior art still has systematic drawbacks in going from "perception" to "decision" in that it is mainly manifested in the following three aspects: The limitation of static threshold determination is that existing devices commonly employ a fixed threshold comparison strategy, such as setting the body temperature threshold to 37.3 ℃ or 38.0 ℃, and outputting "fever" or "anomaly" when the measured value exceeds the threshold. The method ignores three factors of environmental interference, individual difference and early warning missing. Therefore, the prior art has higher false alarm/false alarm rate, and the requirement of 'AI autonomous decision' on reliability is difficult to meet. Legal and liability risks of AI black box decision-making when the health terminal further integrates AI agents and outputs medical related advice (e.g. "advice to take medicine immediately", "advice to seek doctor" and "risk of possible infection"), the decision-making process is often completed locally or in the cloud by an algorithm model, and the reason why it makes a judgment cannot be restored externally. Once misdiagnosis or improper advice occurs, it will be difficult to determine the source of responsibility. Commercial closed-loop fault that the existing health terminal can only output a numerical value or a prompt even if the abnormality can be detected, and the current health terminal can not directly become a trusted trigger source of third-party commercial services (pharmacies, hospitals, insurance, family doctors and the like). The third party service cannot verify whether the abnormal judgment is credible or not, lacks an evidence certificate which is strongly bound with the judgment, is difficult to meet the requirements of compliance and wind control, and has no standardized interface between a perception layer and a performance layer, so that the butt joint cost is high, the cooperation threshold is high, and the large-scale difficulty is caused. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide a health intention judging system, a certificate and an interaction method based on biological feature analysis so as to solve the problems in the background art, the invention builds a full-link closed loop from physical signals to commercial intention through trend determination (slope control point) -XR black box original signal storage certificate-Intent Token standardized output, and upgrades health perception from a reading tool to a standard entrance which can be audited, overtime and on a large scale. In order to achieve the aim, the invention is realized by the following technical scheme that the health intention judging system based on biological characteristic analysis comprises the following modules: The multi-mode biological feature acquisition module is used for acquiring at least one biological feature original signal, wherein the original signal comprises an infrared array thermal signal/thermal pixel matrix, spectrum sampling and PPG original waveform original temperature sampling values, and derived features such as heart rate, respiratory rhythm, facial redness index, skin temperature and the like can be further extracted or inferred; the environment parameter acquisition module is used for acquiring environment variables including temperature and humidity, wind speed, indoor and outdoor states, air conditioner air outlet influence and motion state inference; The individuation baseline modeling module is used for constructing a user history health image or a short-term baseline so as to provide the individual difference correction capability of the user; The dynamic trend judging module is used for carrying out time sequence analysis on the biological characteristics, calculating the change slope and the deviation degree, and outputting a health evaluation score S by combining with environment compensation to further generate