System and Method for Detecting Infected Individuals Using Pathogen Sensing Data and Occupant Information

Abstract

The present invention provides a system and method for determining whether an infected person has entered by combining pathogen sensor data and occupant information, and provides a multi-variable-based probabilistic infected person identification algorithm rather than a simple threshold comparison. Additionally, the invention provides a system and method for determining an infected person that improves identification accuracy by learning data patterns using a machine learning or deep learning-based model.

Description

Technology Field

[0001] The present invention relates to a technology for determining the presence of an infected person entering an indoor or enclosed environment, and more specifically, to a system and method for determining the possibility of an infected person entering in real time by fusing airborne pathogen sensing data with occupant-related information. Furthermore, the present invention relates to an infected person identification technology that calculates the probability of the presence of an infected person using a machine learning or deep learning-based prediction model. Background Technology

[0003] As it has recently become known that airborne pathogens (viruses, bacteria, fungi, etc.) are the primary cause of indoor infection spread, the importance of technology for early detection of danger signals at the initial stage of infected entry is increasing.

[0004] While some environmental monitoring systems determine air quality levels by measuring fine dust, CO2 concentration, temperature, and humidity, they cannot directly determine the influx of infected individuals based solely on this data. Furthermore, although pathogen detection sensor technology has recently advanced, the following limitations exist.

[0005] i) Difficulty in determining infection status based solely on sensor signals: An increase in pathogen sensor signals can occur not only due to the influx of infected individuals but also due to external variables such as a simple increase in population, increased activity levels, and changes in airflow.

[0006] ii) Lack of correlation analysis between data: There is a lack of technology to identify infected individuals by combining data on changes in pathogen concentration with occupant information (number of people, density, duration of stay, etc.).

[0007] iii) Absence of AI-based prediction models: Existing technologies are simple rule-based and use only fixed thresholds, and AI-based infection identification technology that performs complex pattern analysis has not yet been established.

[0008] Accordingly, there is a need for technology capable of identifying infected individuals with high accuracy by fused analyzing pathogen sensing data and occupant information. Prior art literature

[0010] Patent Registration No. 10-2216258 (Norovirus detection sensor, and electrochemical sensing method using the same) Patent Registration No. 10-2376338 (Coronavirus detection kit equipped with a dual interdigital capacitor sensor chip) Patent Registration No. 10-2796888 (Virus detection device for real-time detection of viruses in the air) Patent Registration No. 10-2663756 (Virus detection device for real-time detection of viruses in the air) The problem to be solved

[0011] Accordingly, the present invention has been devised to resolve the aforementioned problems, and the first objective of the present invention is to provide a system and method for determining whether an infected person has entered by combining pathogen sensor data and occupant information, wherein the invention provides a multi-variable-based probabilistic infected person identification algorithm rather than a simple threshold comparison.

[0012] In addition, the second objective of the present invention is to provide a system and method for identifying infected persons with improved identification accuracy through data pattern learning using a machine learning or deep learning-based model.

[0013] According to an additional objective of the present invention, an infected person identification system and method are provided that include an analysis of occupant behavioral characteristics such as density, dwell time, activity level, and cough / vocalization events in addition to the number of people, and enable the fusion of environmental sensor data (CO2, temperature and humidity, VOC, etc.) to enhance the adaptability of the AI ​​model, and provide a platform capable of outputting infected person identification results in real time or linking with air conditioning control and alarm systems. means of solving the problem

[0015] A system for determining an infected person using pathogen sensing data and occupant information according to the first aspect of the present invention for achieving the above-described purpose is a system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, and is characterized by comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for determining an infected person from the fusion data of the data fusion unit (230); and an infected person determination module (250) for calculating the probability of the presence of an infected person by inputting the features for determining an infected person.

[0016] Preferably, the invention further includes a judgment unit (260) that determines whether an infected person has entered by comparing the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), with a predefined standard (Po).

[0017] More preferably, the system further includes an output unit (270) that issues an alarm or outputs a judgment result signal to an external system when the judgment unit (260) determines that an infected person has entered.

[0018] Additionally, preferably, the pathogen sensing data input unit (210) receives pathogen sensing data input from a surface-enhanced Raman scattering (SERS) sensor of a real-time pathogen measurement system in air, and the pathogen sensing data is characterized by being composed of one or more of a SERS spectrum, peak intensity, peak ratio, baseline drift, and signal change rate.

[0019] More preferably, the pathogen sensing data is characterized by being preprocessed by one or more preprocessing processes among noise filtering, background signal removal, peak detection and quantification, and time-based spectral change rate calculation.

[0020] Additionally, preferably, the data fusion unit (230) combines the pathogen sensing data from the pathogen sensing data input unit (210) and the occupant information from the occupant information input unit (220) by time synchronization or normalization, and the data fusion unit (230) is characterized by generating a fusion index after matching the pathogen sensing data and the occupant information to the same time axis.

[0021] Additionally, preferably, the feature for identifying an infected person extracted by the feature extraction unit (240) from the fusion data is characterized as being a core feature related to an infected person that includes at least one of a pathogen feature, an infected person behavior feature, and an environment feature.

[0022] Additionally, preferably, the infected person identification module (250) calculates the probability of the existence of an infected person using a machine learning or deep learning-based model, and is characterized by calculating the probability of the existence of an infected person (P) through a machine learning / deep learning-based probability prediction model.

[0023] More preferably, the infected person identification module (250) is characterized by including an Autoencoder-based anomaly detection model (251) or a Density-based anomaly detection model (252) to detect anomalies corresponding to the influx of infected persons compared with a normal occupancy pattern.

[0024] Furthermore, the above-mentioned judgment unit (260) is characterized by determining a stage-level infection risk grade when the probability of an infected person (P) exceeds at least one of two multi-stage threshold values.

[0025] Additionally, preferably, the occupant information input to the occupant information input unit (220) is characterized as being non-video information implemented solely based on Wi-Fi or sensors without a camera.

[0026] Additionally, preferably, the infected person identification system operates in conjunction with multiple spaces and is characterized by being able to track the movement path of infected persons by integrating data from multiple indoor spaces to continuously identify infected persons.

[0028] A method for determining an infected person using pathogen sensing data and occupant information according to a second aspect of the present invention for achieving the above-described purpose comprises: (a) receiving the pathogen sensing data at the pathogen sensing data input unit (210) (S10); (b) receiving the occupant information at the occupant information input unit (220) (S20); (d) after steps (a) and (b), the data fusion unit (230) combines the pathogen sensing data and the occupant information to generate fusion data (S40); (e) after step (d), the feature extraction unit (240) extracts features for determining an infected person from the fusion data (S50); and (f) after step (e), the infected person determination module (250) calculates the probability of the presence of an infected person (S60).

[0029] Preferably, the method further comprises (c) a step (S30) of performing at least one preprocessing step among i) noise filtering, ii) background signal removal, iii) peak detection and quantification, and iv) time-based spectral change rate calculation on the pathogen sensing data after step (a); and step (d) is characterized by generating fusion data by combining the received pathogen sensing data from step (a) with the preprocessed pathogen sensing data and occupant information.

[0030] More preferably, the method further comprises: (g) a step (S70) of comparing the probability of an infected person (P), which is the output value of the infected person identification module (250), with a predefined reference value (Po) after step (f); and (h) a step (S80) of determining, if the probability of an infected person (P) is less than the reference value (Po) as a result of the judgment in step (g), returning to the beginning and performing the task again, and conversely, if the output value of the probability of an infected person (P) is greater than or equal to the reference value (P ≥ Po), determining that an infected person has entered and sounding an alarm or outputting the result to a server or an external device.

[0031] Also more preferably, (g1) after step (f), a step (S71) of comparing the probability of an infected person (P), which is the output value of the infected person identification module (250), with a predefined first reference value (P1); (g2) if, as a result of the judgment in step (g1), the probability of an infected person (P) is less than the first reference value (P1) (P < P1), the process is performed again by returning to the beginning, and conversely, if the output value of the probability of an infected person (P) is greater than or equal to the first reference value (P ≥ P1), a step (S72) of comparing the probability of an infected person (P), which is the output value of the infected person identification module (250), with a second reference value (P2) which is greater than the first reference value (P1); (h1) A step (S81) in which, as a result of the judgment in step (g2), if the probability of the presence of an infected person (P) is less than the second reference value (P2) (P < P2), the lowest level is determined as the first infection risk level, and a first level alarm is issued; (g3) A step (S73) in which, as a result of the judgment in step (g2), if the output value of the probability of the presence of an infected person (P) is greater than or equal to the second reference value (P ≥ P2), the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), is compared with the nth reference value (Pn), which is greater than the second reference value (P2); (h2) A step (S82) in which, as a result of the judgment in step (g3), if the probability of the presence of an infected person (P) is less than the nth reference value (Pn) (P < Pn), the intermediate level is determined as the n-1st infection risk level, and an n-1st level alarm is issued; and (h3) if, as a result of the judgment in step (g3) above, the probability of the presence of the infected person (P) is greater than or equal to the nth reference value (Pn) (P ≥ Pn), the highest risk level is determined as the nth infection risk level, and the highest level alarm of the nth level is issued (S83); further characterized by including. Effects of the invention

[0033] As described above, according to the system and method for identifying infected persons using pathogen sensing data and occupant information according to the present invention, a technology is provided to determine whether an infected person has entered by combining pathogen sensor data and occupant information, and a multi-variable probabilistic infected person identification algorithm is provided rather than a simple threshold comparison, and furthermore, it is possible to identify infected persons with improved identification accuracy by learning data patterns using a machine learning or deep learning-based model.

[0034] In addition to the above objectives and effects, other objectives and advantages of the present invention will become apparent through the detailed description of embodiments with reference to the attached drawings. Brief explanation of the drawing

[0036] FIG. 1 is a configuration diagram showing an example of a nanostructure-based airborne pathogen real-time sensing system related to the present invention. FIG. 2 is an overall block diagram of a nanostructure-based airborne pathogen real-time sensing system according to a first embodiment of the related invention of the present invention. FIG. 3 is a schematic diagram showing an example of a pathogen capture unit and a SERS-based detection unit according to a first embodiment of the related invention of the present invention. Figure 4 is a diagram explaining the principle of determining whether a pathogen has entered by comparing a sensor signal with a reference value. FIG. 5 is a flowchart of a nanostructure-based real-time airborne pathogen sensing method according to a first embodiment of the related invention of the present invention. FIG. 6 is an overall block diagram of a nanostructure-based airborne pathogen real-time sensing system according to a second embodiment of the related invention of the present invention. FIG. 7 is a diagram illustrating the principle of identifying whether an infected person has entered by combining a pathogen detection signal and information on the number of people in a space according to a second embodiment of the invention related to the present invention. FIG. 8 is a flowchart of a nanostructure-based real-time airborne pathogen sensing method according to a second embodiment of the related invention of the present invention. FIG. 9 is an overall block diagram of an infected person identification system according to an optimal embodiment of the present invention. FIG. 10 is a flowchart of a method for determining whether an infected person has entered according to an optimal embodiment of the present invention. FIG. 11 is a flowchart of a method for determining whether an infected person has entered according to a modified example of the optimal embodiment of the present invention. Specific details for implementing the invention

[0037] Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings FIGS. 9 to 11, and the description will be made with auxiliary reference to the inventor's related prior art of FIGS. 1 to 8.

[0038] However, it will be readily apparent to those skilled in the art that the attached drawings are provided merely to facilitate the disclosure of the content of the present invention, and that the scope of the present invention is not limited to the scope of the attached drawings.

[0039] First, the prior invention of the present inventors filed on the same date with the title of the invention "Nanostructure-based Real-time Sensing System and Method for Airborne Pathogens" (hereinafter referred to as the "related invention") will be described.

[0041] (First embodiment of the related invention)

[0042] First, with reference to FIGS. 1 to 5, a nanostructure-based real-time sensing system and method for airborne pathogens according to a first embodiment of the related invention of the present invention will be described in detail.

[0043] FIG. 1 is a configuration diagram showing an example of a nanostructure-based airborne pathogen real-time sensing system installed in relation to the present invention, FIG. 2 is an overall block diagram of a nanostructure-based airborne pathogen real-time sensing system according to a first embodiment of the invention related to the present invention, FIG. 3 is a schematic diagram showing a pathogen capture unit and a SERS-based detection unit according to a first embodiment of the invention related to the present invention, and FIG. 4 is a diagram explaining the principle of determining whether a pathogen has entered by comparing a sensor signal with a reference value.

[0044] FIG. 5 is a flowchart of a nanostructure-based real-time airborne pathogen sensing method according to a first embodiment of the related invention of the present invention.

[0046] (Real-time sensing system for airborne pathogens of the first embodiment of the related invention)

[0047] First, the nanostructure-based airborne pathogen real-time sensing system (100) according to the first embodiment of the invention is installed in the ventilation port of an air conditioning system (200) in an indoor space (300) where many people enter and exit, such as a public facility, a medical facility, or a school, as shown in FIG. 1.

[0048] Referring to FIG. 2, the overall configuration of the nanostructure-based airborne pathogen real-time sensing system (100) related to the present invention is described as follows: an air inlet (110) having a flow path structure that stably guides external air into a sensing chamber (100a); a pathogen collection unit (120) including a collection plate (121) formed with a nanostructure-based surface-enhancing material that collects pathogens from the air inlet (110); a signal detection unit (130) that irradiates light onto pathogens collected on the collection plate (121) to sense and measure a signal of reflected light; a signal processing unit (140) that determines whether a pathogen exists by quantifying and analyzing the signal result detected from the signal detection unit (130); and an output unit (160) that transmits the result of determining the presence of pathogens to an external system.

[0049] At this time, the pathogen collection unit (120) further includes a pathogen collector (122) that enables pathogens to be collected at a concentration above a certain level on the collection plate (121). For example, a physical collection means such as optical tweezers is used to apply an electric field, electrostatic field, magnetic field, etc., to the collection plate (121) so that pathogens with a slight electromagnetic field are collected on the collection plate (121), or pathogens are collected and fixed on the surface of the metal collection plate (121). More specifically, the electric field collection means may include a pair of electrodes to allow pathogens to move to and attach to the surface of the collection plate.

[0050] For reference, the aforementioned optical tweezers utilize optical force generated by the interaction between a laser and particles. Since optical force is largely composed of a gradient force, which pulls particles toward the direction of greater electric field strength, and a scattering force, which pushes them in the direction of travel through the transfer of photon momentum, stable trapping is difficult in the gaseous phase because particles move very freely. Therefore, a photothermal trapping method is employed, which involves using a focused laser on a flat metal plate, a nanostructured metal plate, or locally heating a metal plate to form a gas density gradient.

[0051] To explain these in more detail with reference to FIG. 3, the pathogen capture unit (120) may include, for example, a capture plate (121) containing a nanostructure-based surface-enhancing material (for example, gold nanoparticles), and in addition to the nanostructure-based capture plate (121), a physical capturer (122) capable of selectively inducing and capturing pathogens.

[0052] On the surface of the above-mentioned collection plate (121), metal nanoparticles (121a) such as gold, silver, copper, and aluminum, nanorods, nanoshells, nanocubes, nanofilms, nanohole arrays, or nanopattern structures are formed, which can induce surface-enhanced Raman scattering (SERS).

[0053] In addition, it is desirable that the above nanostructure be designed to be capable of signal amplification for viruses (tens to hundreds of nm), bacteria (several μm), and fungal spores (several μm).

[0054] The above physical capture device (122) may be one or more of the following: an electric field capture means using electrophoresis, an electrostatic capture means using static electricity, a magnetic nanoparticle or magnetic field-based magnetic capture means, and optical tweezers, in addition to a nanostructure-based capture plate, as a physical capture means capable of selectively inducing and capturing pathogens. This can maximize the attachment efficiency of pathogens.

[0055] The above physical capture means can be applied individually or in combination of two or more means, for example, by simultaneously applying electric field capture to a nanoplasmonic SERS substrate to further improve the nanostructure surface adhesion rate.

[0056] The above nanostructure-based surface enhancement material may comprise one or more of the following: i) metal nanoparticles (nanospheres, nanorods, nanocubes, etc.); ii) nanopatterned structures (nanohole arrays, nanorib structures, etc.); iii) metal / ceramic / polymer composite nanostructures; and iv) multilayer structures that induce plasmonic resonance.

[0057] That is, in the embodiment of FIG. 3 of the present invention, gold nanoparticles (121a) are described as an example, but this is merely an example and various modifications are possible.

[0058] As such, in the real-time pathogen measurement system of the present invention, the pathogen collection unit (120) is the most important core, because compared to the conventional droplet-type sample method, it is relatively difficult to collect and attach pathogens in the air in the gas phase detection method, and also because separation must be achieved from larger airborne particles such as dust or bacteria (the latter separation will be described later).

[0059] Furthermore, as an example of the signal detection unit (130), a SERS-based spectrum signal detection unit using a laser light source is described in detail. The signal detection unit (130) may include a laser light source (131), an optical condenser (132), and a signal measuring device (133).

[0060] First, the laser light source (131) may use one or more of 532 nm, 633 nm, and 785 nm, and the wavelength may be selected according to the characteristics of the pathogen and the type of nanoplasmonic SERS substrate, and the optical condenser (132) may include a lens, a mirror, and an optical filter for irradiating a laser onto a pathogen on a collection plate and collecting scattered light, and the signal measuring device (133) performs sensing and detection of the collected scattered light, and analyzes the Raman spectrum to determine the presence and concentration of the pathogen.

[0061] Additionally, the signal processing unit (140) quantifies and analyzes the spectrum pattern according to the pathogen to enable detection of various types of airborne pathogens with minimal error.

[0062] For example, FIG. 3 is a schematic diagram showing an example of a pathogen collection unit and a SERS-based detection unit. The SERS signal strength is continuously detected over time. When an infected person enters the room, the SERS signal strength of a specific wavelength on the vertical axis spikes above a specific threshold (Vth) according to the wavelength-specific absorption and / or reflection characteristics of the pathogens collected on the collection plate. Through this, the influx of an infected person can be detected.

[0063] To elaborate, regarding the values ​​on the vertical axis in Figure 3 above, while the SERS signal strength can be directly compared, more accurately, the SERS signal strength is standardized or quantified into a voltage value within a certain range (for example, 0 to 10 V). DC It is more desirable to use the converted value.

[0064] Finally, the output unit (160) performs the function of displaying the measurement value or transmitting it to an external system (e.g., an air conditioning unit) or server.

[0065] Additionally, the real-time pathogen measurement system (100) of the present invention may include a regeneration unit (150), which is a device for regenerating a collection plate by heat treatment or other methods. Since contaminants may accumulate on the surface of the collection plate (121) when used for a long period of time, and thus the detection sensitivity may decrease, to prevent this, the pathogen collection plate can be operated stably for a long period of time by removing contamination from the surface of the collection unit through a regeneration method such as heat treatment.

[0066] The above-mentioned regeneration unit (150) may include one or more of the following methods: i) heat treatment method: heating the collection plate above a certain temperature to decompose organic matter and pathogens; ii) UV sterilization method: removing pathogens by ultraviolet irradiation; iii) plasma treatment method: surface cleaning with low-temperature plasma; and iv) disinfectant or cleaning agent vaporization treatment method.

[0067] In addition, the playback mode of the above-mentioned playback unit (150) can be controlled automatically or manually.

[0068] Optionally, the air inlet (110) may include a microchannel structure for selective separation of pathogens by size or control of the inflow volume. In this case, the microchannel structure may include an inertial impactor structure for preferentially removing bacteria and fungal spores and selectively allowing viruses to pass through.

[0069] Additionally, as described above, the gas phase type pathogen measurement system of the present invention is preferably designed to properly separate larger airborne particles, such as dust or bacteria, and as shown in FIG. 3, a particle separator (170) (e.g., based on an aerodynamic design) is added, which is installed immediately after the air inlet (110).

[0070] The above particle separator (170) can adopt a method of separating particles in an aerosol using the difference in inertial force according to particle size, which is a technology based on the core principle of how much a particle “cannot keep up with the flow” in response to a sudden change in fluid flow, and is a technology used in atmospheric particle measurement, environmental monitoring, bioaerosol analysis, semiconductor cleanroom management, etc.

[0071] Representative inertia-based particle separation methods include the Inertial Impactor, Cascade Impactor, Cyclone Separator, Inertial Microfluidics, and Virtual Impactor.

[0072] That is, as air in the form of aerosol from the air inlet (110) passes through the particle separator (170), bacteria and microparticles with relatively large particle sizes (e.g., several μm) have a high inertial force and travel in a straight line to the first air outlet (180), while viruses and nanoparticles with relatively small particle sizes (e.g., tens to hundreds of nm) have a low inertial force and travel toward the lateral microparticle outlet (185) to be captured by the pathogen collection plate (121). Unexplained reference numeral (190) is the second air outlet.

[0073] To explain the operation of the real-time pathogen measurement system according to the first embodiment of the invention related to the present invention, air containing pathogens is introduced into the sensing chamber (100a) through the air inlet (110) (and optionally the particle separator (170)), the pathogens are selectively captured by the pathogen capture unit (120), the captured pathogens are sensed and detected as reflected light signals by the signal detection unit (130), the analysis results are quantified and analyzed through the signal processing unit (140), the capture plate is regenerated by the regeneration unit (150) if necessary, and the pathogen detection results are transmitted to an external system through the output unit (160).

[0074] More preferably, the output unit (150) may further include an external data transmission unit (not shown) to include a function for transmitting the analysis results to an air conditioning system or an external server.

[0075] Thus, the nanostructure-based airborne pathogen real-time sensing system related to the present invention has the following effects.

[0076] 1) Direct detection of airborne pathogens without a liquid capture process: It enables real-time and continuous measurement, allowing for immediate response during the initial influx of infected individuals.

[0077] 2) Detection of trace amounts of pathogens possible through SERS-based high-sensitivity measurement: It is possible to detect the influx of pathogens in the very early stages before the spread of infection.

[0078] 3) Widely applicable regardless of pathogen type and size differences: Capable of detecting viruses, bacteria, fungal spores, and other microbial-derived substances.

[0079] 4) Improved capture efficiency through the combination of nanostructures and physical capture means: Improved sensitivity and reduced false voices compared to existing filter-type sensors.

[0080] 5) Ease of maintenance through regeneration function: Reduction in collection plate replacement costs and assurance of long-term stability.

[0082] (Real-time sensing method for airborne pathogens of the first embodiment of the related invention)

[0083] Hereinafter, a nanostructure-based real-time sensing method for airborne pathogens according to a first embodiment of the invention related to the present invention will be described with reference mainly to FIG. 5 and with reference secondarily to FIGS. 1 to 4.

[0084] When control of the nanostructure-based airborne pathogen real-time sensing method related to the present invention is initiated, first, the particle separator (170) separates the particles in the air according to size, such that dust or bacteria and microparticles with relatively large particle sizes (e.g., several μm) are discharged to the first air outlet (180) from the air introduced through the air inlet (110), and viruses and nanoparticles with relatively small particle sizes (e.g., tens to hundreds of nm) are directed to the fine particle outlet (185) (S100).

[0085] Subsequently, the pathogen collection unit (120) collects pathogens in the air flowing into the sensing chamber (100a) through the fine particle outlet (185) onto the pathogen collection plate (121) (S200), and the signal detection unit (130) irradiates a laser light (131) onto the pathogen collection plate (121) and senses and detects the reflected light signal (S300).

[0086] Subsequently, the signal processing unit (140) performs signal processing and analysis, such as quantification of the detection result from the signal detection unit (130), to finally detect the pathogen (S400), and then outputs the pathogen detection result to an external system through the output unit (160) (S500).

[0087] Afterward, the need for regeneration is determined based on the contamination level of the pathogen collection plate (121) (S600). If there is no need for regeneration, the process returns to the beginning. If it is determined that there is a need for regeneration, the regeneration unit (150) removes contaminants from the pathogen collection plate (121) (S900), and then terminates the process. For reference, the need for regeneration of the pathogen collection plate (121) is determined primarily based on the contamination level of the collection plate (121). This can be controlled automatically or manually. In the case of automatic regeneration, it may be performed at regular intervals, or it may be performed when the contamination level is measured each time and it is determined that the contamination level has exceeded a certain range.

[0089] (Second embodiment of the related invention)

[0090] Now, with reference to FIGS. 6 to 8, a nanostructure-based airborne pathogen real-time sensing system and method according to a second embodiment of the related invention of the present invention will be described in detail.

[0091] FIG. 6 is an overall block diagram of a nanostructure-based airborne pathogen real-time sensing system according to a second embodiment of the related invention of the present invention, and FIG. 7 is a diagram showing the principle of identifying whether an infected person has entered by combining a pathogen detection signal and information on the number of people in a space according to a second embodiment of the related invention of the present invention.

[0092] FIG. 8 is a flowchart of a nanostructure-based real-time airborne pathogen sensing method according to a second embodiment of the related invention of the present invention.

[0094] (Real-time sensing system for airborne pathogens of the second embodiment of the related invention)

[0095] First, referring to FIGS. 6 and FIGS. 7, the nanostructure-based airborne pathogen real-time sensing system according to the second embodiment of the invention is similar to the first embodiment, but differs in that, in the signal processing unit (140), machine learning or AI deep learning techniques are used to further reduce errors, increase accuracy, and enable detection in various ways.

[0096] That is, with reference to FIGS. 6 and 7, a nanostructure-based airborne pathogen real-time sensing system according to a second embodiment of the related invention of the present invention will be described, focusing on the differences from the first embodiment of FIG. 2.

[0097] The signal processing unit (140) of the real-time pathogen sensing system of the second embodiment may include a machine learning (ML) or deep learning (DL) based signal analysis module (141) to analyze the peak intensity, peak pattern, or spectrum change rate of the spectrum as shown in FIG. 6, and thus the signal analysis module (141) can detect not only the presence of the pathogen from the spectrum but also the type and / or concentration of the pathogen according to artificial intelligence learning.

[0098] That is, the signal processing unit (140) of the second embodiment further includes a model learning module (143) for training the signal analysis module (141) using learning data labeled for pathogen type and concentration, and the model learning module (143) may also update the signal analysis module (141) using spectrum data collected in real time or periodically.

[0099] The signal analysis module (141) can also use a multi-pathogen classification AI to analyze Raman spectrum patterns in the multi-pathogen simultaneous detection mode and output multi-class classification results corresponding to each of the viruses, bacteria, and fungal spores.

[0100] The machine learning or deep learning-based signal analysis module (141) and model learning module (143) are preferably CNN (Convolutional Neural Network) learning models, and may include, for example, one or more of Random Forest, Support Vector Machine (SVM), XGBoost, 1D-CNN, Recurrent Neural Network (RNN), LSTM, GRU, Transformer, or a combination thereof.

[0101] Additionally, the signal processing unit (140) of the second embodiment may further include an anomaly detection module (142) that learns a spectrum pattern in a normal operating state using an AI for anomaly detection / drift correction, and the anomaly detection module (142) preferably detects an abnormal spectrum caused by contamination of the collection plate, changes in light source output and / or an optical system abnormality, and generates a correction coefficient for the abnormal spectrum or an operation trigger signal for the regeneration unit (150).

[0102] FIG. 7 is a diagram illustrating the principle of identifying whether an infected person has entered by combining a pathogen detection signal and information on the number of people in a space according to a second embodiment of the invention related to the present invention. The model learning module (143) analyzes the video of an indoor CCTV using a CNN technique to determine the number of people in the indoor space and detects the corresponding SERS signal strength. It learns the SERS signal strength according to the number of people in the indoor space when there are no infected people (Level 0), learns the SERS signal strength when the number of infected people is at the lowest level, 'Level 1', learns the SERS signal strength when the number of infected people is at a level one step higher than 'Level 1', and learns the SERS signal strength when the number of infected people is at a level one step higher than 'Level 2' and when the number of infected people is at a level one step higher than 'Level 2', thereby building a learning model.

[0103] Afterward, after sufficient supervised learning is performed, the signal analysis module (141) can determine the current level of infected person inflow (risk level) according to the learning model.

[0104] Of course, by learning the spectrum pattern in a normal operating state through the above-mentioned anomaly detection module (142), and then learning the abnormal spectrum caused by contamination of the collection plate, changes in light source output, or optical system abnormalities, the error in pathogen identification by the above-mentioned signal analysis module (141) is minimized.

[0105] In particular, the real-time sensing system for airborne pathogens according to the second embodiment can selectively operate between a virus-only detection mode and a multiple pathogen simultaneous detection mode that simultaneously detects viruses, bacteria, and fungal spores.

[0107] (Real-time sensing method for airborne pathogens of the second embodiment of the related invention)

[0108] Finally, a nanostructure-based real-time sensing method for airborne pathogens according to a second embodiment of the related invention of the present invention will be described with reference primarily to FIG. 8 and with reference secondarily to FIG. 1, FIG. 6, and FIG. 7.

[0109] When control of the nanostructure-based airborne pathogen real-time sensing method according to the second embodiment of the present invention is started, first, the signal processing unit (140) builds a learning model by learning the SERS signal strength when there is no infected person (Level 0), when the infected person is at the lowest level 'Level 1', when the infected person is at a level one step higher than 'Level 1', and when the infected person is at a level one step higher than 'Level 3', and initializes the iteration judgment parameter (N) (S100').

[0110] Subsequently, the pathogen collection unit (120) collects pathogens in the air flowing into the sensing chamber (100a) on the pathogen collection plate (121) (S200), and the signal detection unit (130) irradiates a laser light (131) onto the pathogen collection plate (121) and detects and analyzes the reflected signal (S300).

[0111] Subsequently, the signal processing unit (140) performs signal processing, such as quantification, through the analysis results from the signal detection unit (130) to detect pathogens (S400), and then outputs the pathogen detection signal to an external system or server through the output unit (160) (S500).

[0112] Subsequently, it is determined whether the repetition judgment parameter (N) has reached a reference value (No) (S700); if not, the repetition judgment parameter (N) is incremented (S750), and the process is repeated by returning to step S200. This is to reduce the error by detecting the SERS signal a sufficient number of times.

[0113] On the other hand, if the repetition judgment parameter (N) reaches the threshold value (No) as a result of the judgment in step S700, the current level of infected person inflow is inferred by referring to the established learning model (S800), and then the regeneration unit (150) removes contaminants from the pathogen collection plate (121) (S900), and the process is terminated.

[0115] (Examples of the present invention)

[0116] Now, with primary reference to FIGS. 9 to 11 and auxiliary reference to the related invention of FIGS. 1 to 8, a system and method for identifying infected persons using pathogen sensing data and occupant information according to an optimal embodiment of the present invention will be described in detail.

[0117] FIG. 9 is an overall block diagram of an infected person identification system according to an optimal embodiment of the present invention, FIG. 10 is a flowchart of a method for determining whether an infected person has entered according to an optimal embodiment of the present invention, and FIG. 11 is a flowchart of a method for determining whether an infected person has entered according to a modified example of an optimal embodiment of the present invention.

[0119] (Infected person identification system of the optimal embodiment of the present invention)

[0120] First, referring to FIG. 9, an infected person identification system using pathogen sensing data and occupant information according to an optimal embodiment of the present invention comprises: a pathogen sensing data input unit (210) into which pathogen sensing data, such as the presence and / or concentration (i.e., pathogen concentration) of airborne pathogens measured in real time by a sensing device in FIG. 2 and FIG. 3, is input; an occupant information input unit (220) into which occupant information, such as the number of occupants and / or behavioral characteristics identified by various sensors such as a camera, IR, ToF, acoustic, Wi-Fi / MAC, BLE, UWB, etc., is input; a data fusion unit (230) that combines the pathogen sensing data and occupant information by time synchronization and / or normalization; a feature extraction unit (240) that extracts pathogen pattern features, population dynamic features, etc., from the fused data; and an infected person identification module (250) that calculates the probability of the presence of an infected person using a machine learning or deep learning-based model, and calculates the probability of the presence of an infected person including a machine learning / deep learning-based probability prediction model. It includes a judgment unit (260) that determines whether an infected person has entered based on a predefined standard or model output value, and / or performs a threshold judgment or multi-stage classification based on an AI output value (P-value); and an output unit (270) that performs an alarm when an infected person is determined to have entered, outputs an alarm and result signal, and performs an alarm and output in conjunction with an external system (GUI, air conditioning system, server, etc.).

[0121] To explain these in more detail, first, the pathogen sensing data input unit (210) receives pathogen sensing data input from a SERS sensor, a surface enhancement-based sensor, or other pathogen detection sensor as shown in FIG. 3. The pathogen sensing data is input in the form of a SERS spectrum, peak intensity, peak ratio, baseline drift, signal change rate, etc. If necessary, the pathogen sensing data may include preprocessing steps such as noise filtering, background signal removal, peak detection and quantification, and time-based spectrum change rate calculation. Thus, the preprocessed data is signal-processed to more accurately reflect the pathogen concentration change pattern resulting from the influx of infected persons.

[0122] Next, the occupant information input unit (220) receives occupant information, such as the number of occupants and activity data, such as camera, IR, ToF, Wi-Fi, and BLE. For example, the occupant information used in the present invention may be one or more of the following:

[0123] i) Personnel count information: Camera-based object recognition, IR sensor, ToF sensor, Wi-Fi / MAC counting, BLE / UWB-based location tracking, etc.

[0124] ii) Density Information: Number of people per unit area

[0125] iii) Duration of Stay Information: Analysis of Duration of Stay by Number of People

[0126] iv) Activity information: Movement frequency, walking speed, gesture analysis, etc.

[0127] v) Information on suspected infection behavior (optional): Cough / sneeze / speaking (vocalization) events

[0128] Furthermore, while the aforementioned occupant information will primarily utilize image data from cameras, it is not necessarily limited to this; acoustic sensors or image-based behavior recognition AI models may also be applied.

[0129] Furthermore, the data fusion unit (230) combines the pathogen sensing data from the pathogen sensing data input unit (210) and the occupant information from the occupant information input unit (220) by time synchronization and / or normalization, and the data fusion unit (230) can generate fusion indicators such as ① a ratio of the number of people to the pathogen signal, ② an activity level correction indicator to the pathogen signal, ③ a pathogen signal slope correction value over time, and ④ a normalized value based on dynamic population fluctuation modeling.

[0130] For example, the data fusion unit (230) is capable of AI feature engineering and representation learning, and subsequently, the fusion indicator is used to calculate the probability of the existence of an infected person.

[0131] Furthermore, the feature extraction unit (240) extracts key features related to the infected person, such as pathogen features, infected person behavior features, and environmental features.

[0132] Using the above features as input variables, the infected person identification module (250) calculates the probability of the existence of an infected person based on an AI model, and examples of the AI ​​model input variables that are the above features are as shown in [Table 1] below.

[0133] Variable types example Pathogen Sensing Characteristics SERS intensity, peak ratio, baseline drift, time derivative, slope Number of people information total occupants, occupancy change rate Behavioral / Activity Characteristics motion index, coughing event frequency, voice activity Environmental characteristics (optional) CO2, temperature, humidity, ventilation rate, etc.

[0134] Meanwhile, the above-mentioned infected person identification module (250) may be composed of one or more of the following artificial intelligence models and performs pre-learning or online learning based on learning data.

[0135] a) Machine Learning (ML) Models: Random Forest, SVM, XGBoost, etc.

[0136] b) Deep Learning (DL) Models: CNN, RNN, LSTM, GRU, Transformer, GNN, etc.

[0137] c) Hybrid model: Combined statistical, ML, and DL models

[0138] d) Transformer-based time series forecasting

[0139] e) GNN-based spatial-temporal model

[0140] f) Autoencoder-based outlier detection

[0141] g) Privacy protection learning based on Federated Learning

[0142] h) Edge AI-based real-time processing

[0143] That is, the infected person identification module (250) may include an RNN, LSTM, GRU, or Transformer-based time series deep learning model to analyze the time change pattern of pathogen sensing data, and may also include an Autoencoder-based anomaly detection model (251) or a Density-based anomaly detection model (252) to detect anomalies corresponding to the influx of infected persons by comparing with normal occupancy patterns.

[0144] Preferably, the infected person identification system may include a function to retrain the infected person identification module (250) using data collected in real time or periodically, via online learning, continuous learning, or a server-based update method.

[0145] More preferably, the infected person identification module (250) may also perform Federated Learning-based distributed learning or Edge AI-based local model execution without storing the original image data.

[0146] On the other hand, the AI ​​model output value of the infected person identification module (250) is the probability of the existence of an infected person (P), and the judgment unit (260) determines whether there is an influx of an infected person by comparing the output value of the infected person identification module with a preset reference value. For example, if the output value 'P' is greater than or equal to the reference value (e.g., 0.7), it can be determined that there is an influx of an infected person.

[0147] Additionally, the above-mentioned judgment unit (260) may also determine a stage-level infection risk level when the probability of an infected person (P) exceeds any one of the multi-stage threshold values ​​(e.g., a 3-stage or 5-stage risk level).

[0148] Finally, the alarm and output unit (270) takes follow-up measures, such as issuing an alarm, controlling an external air conditioning system, or transmitting data to a mobile device, depending on the result of the judgment unit (260).

[0149] Optionally, the infected person identification system of the present invention may be linked with an external server to perform updates or advancements to the infected person identification model, or may use only anonymized occupant information without storing original video data to minimize personal information, or the system may further include a function to estimate the movement path of an infected person by integrating data acquired from multiple spaces, or the infected person identification result may be transmitted to an external device to be linked with an indoor air conditioning system or ventilation control system, or may provide a user interface in the form of an alarm, notification, screen display, text message, or app notification according to the infected person identification result, and it is also possible to simultaneously add one or more of the optional additional functions.

[0150] In other words, additionally, the system of the present invention is also possible with the following modified embodiments.

[0151] ㉠ Non-visual method of identifying infected individuals: Implemented based solely on Wi-Fi or sensors without a camera

[0152] ② Server-based model update: Automatic model improvement based on accumulated training data

[0153] (c) Real-time infected person risk classification: Determination of 3 or 5 risk levels

[0154] ㉣ Multi-space Interconnection: Integrating data from multiple indoor spaces to track the movement paths of infected individuals

[0155] As described above, the infected person identification system using pathogen sensing data and occupant information according to the present invention has the following unique effects.

[0156] 1) Reduces false positives / negatives that are difficult to determine based solely on pathogen sensor signals → Improved accuracy by analyzing personnel information and behavioral data together

[0157] 2) AI-based complex pattern analysis possible → Improved sensitivity and specificity in identifying infected individuals compared to the simple threshold method

[0158] 3) Can be implemented with various sensor combinations → Applicable to non-camera environments and personal information-sensitive spaces

[0159] 4) Contributes to the rapid suppression of infection spread → Alert possible during the initial influx of infected individuals

[0160] 5) Maximizing synergy when combined with other pathogen sensing devices → Formation of a strong mutual protection structure when combined with the aforementioned related invention

[0162] (Method for identifying infected persons in the optimal embodiment of the present invention)

[0163] Now, with reference primarily to FIGS. 10 and 11 and secondarily to FIGS. 1 through 9, a method for determining whether an infected person has entered using pathogen sensing data and occupant information according to an optimal embodiment of the present invention will be described.

[0164] When control of the method for determining an infected person using pathogen sensing data and occupant information related to the present invention is started, first, pathogen sensing data such as the presence and / or concentration of airborne pathogens (hereinafter referred to as 'pathogen concentration') is received from the pathogen sensing data input unit (210) (S10), and occupant information such as the number of occupants and / or behavioral characteristics is received from the occupant information input unit (220) (S20), and preprocessing processes such as noise filtering, background signal removal, peak detection and quantification, and time-based spectrum change rate calculation are performed on the pathogen sensing data, and preprocessing processes such as quantification of the number of people and activity amount calculation are performed on the occupant information (S30).

[0165] Subsequently, the data fusion unit (230) combines the preprocessed pathogen sensing data and occupant information by time synchronization and / or normalization (S40), and the combined data (pathogen sensing data and occupant information matched to the same time axis) becomes a fusion indicator such as a ratio of the number of people to the pathogen signal, an activity level correction indicator relative to the pathogen signal, a pathogen signal slope correction value over time, a normalized value based on dynamic population fluctuation modeling, etc., as described above.

[0166] Afterwards, the feature extraction unit (240) extracts key features related to the infected person, such as pathogen features, infected person behavior features, and environmental features (S50), and using the key features related to the infected person as input variables, the infected person identification module (250) calculates the probability of the existence of an infected person (P) based on an AI model (S60).

[0167] The above judgment unit (260) determines whether there is an influx of an infected person by comparing the output value of the infected person identification module with a preset reference value (Po) (S70). For example, if the output value 'P' is greater than or equal to the reference value (e.g., Po = 0.7) (e.g., if P ≥ 0.7), it can be determined that there is an influx of an infected person. However, while it is appropriate to perform such comparison and judgment in a separate judgment unit, it is also possible to perform it by adding only a comparator to the infected person identification module (250), so a separate judgment unit is not essential.

[0168] That is, if the output value of the probability of the presence of an infected person (P) is less than the reference value (P < Po) as a result of the judgment in step S70, the process returns to step S10 and is performed again from the beginning; conversely, if the output value of the probability of the presence of an infected person (P) is greater than or equal to the reference value (P ≥ Po), it is determined that an infected person has entered, an alarm is sounded, and output to a server or other external device, thereby controlling the ventilation device or sterilization device of the air conditioning system (200 in FIG. 2) directly (S80).

[0169] Meanwhile, FIG. 11 discloses a modified example of the optimal embodiment of the present method for identifying infected persons. In this modified example, two or more infection risk levels are set, ranging from the lowest level, the first level, to the second level, which is higher than the first level, and to the highest level, the nth level (where n is a natural number greater than or equal to 2), thereby enabling different alarm issuance and control. For reference, if n is 2, the n-1st level is the first level, and thus consists of two levels: the first level and the second level. If n is 3, the n-1st level is the second level, and thus consists of three levels: the first level, the second level, and the third level.

[0170] To explain this with reference to FIG. 11, for example, the case where 'n = 3', steps S10 through S60 are identical to the optimal embodiment of FIG. 10 described above, but in this modified example, after step S60, the judgment unit (260) compares the output value of the infected person identification module with a preset first reference value (P1) (S71), and if the output value 'P' is greater than or equal to the first reference value (e.g., P1 = 0.6) (i.e., P ≥ P1 = 0.6), it can be determined as an infected person inflow stage.

[0171] That is, if the output value of the probability of the presence of an infected person (P) is less than the first reference value (P < P1 = 0.6) as a result of the judgment in step S71, the process returns to step S10 and is performed again from the beginning. Conversely, if the output value of the probability of the presence of an infected person (P) is greater than or equal to the first reference value (P ≥ P1), the output value of the infected person identification module is compared with a pre-set second reference value (P2 = 0.7) (S72). If the output value 'P' is greater than or equal to the first reference value (P1) but less than the second reference value (P2) (i.e., 0.6 = P1 ≤ P < P2 = 0.7), it is determined to be the lowest level, the first infection risk level, and a first level alarm is issued (S81).

[0172] On the other hand, if the output value of the probability of the presence of an infected person (P) is greater than or equal to the second reference value (P ≥ P2 = 0.7) as a result of the judgment in step S72, the output value of the infected person identification module is compared with the pre-set nth reference value (Pn = 0.8) (S73). If the output value 'P' is greater than or equal to the second reference value (P2) but less than the nth reference value (Pn) (i.e., 0.7 = P2 ≤ P < Pn = 0.8), it is determined to be an intermediate stage, the n-1st stage of infection risk, and an n-1st stage alarm is issued (S82).

[0173] Finally, as a result of the judgment in step S73, if the output value of the probability of the presence of an infected person (P) is greater than or equal to the n-th reference value (P ≥ Pn = 0.8), it is determined to be the n-th infection risk level, which is the highest risk level, and the highest level alarm of the n-th level is issued (S83). In this case, the output is sent to a server or other external device, and the ventilation device or sterilization device of the direct air conditioning system (200 in FIG. 2) is also controlled together (S83).

[0175] The foregoing description is merely an illustrative explanation of the technical concept of the present invention, and those skilled in the art to which the present invention pertains will be able to make various modifications and variations within the scope of the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical concept of the present invention, and the scope of the technical concept of the present invention is not limited by these embodiments. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention. Explanation of the symbols

[0177] (Related invention: FIGS. 1–8) 100: Real-time pathogen measurement system 100a: Chamber 110: Air inlet 120: Pathogen capture unit 121 : Pathogen collection plate 121a: Gold nanoparticles 122 : Pathogen collector 130 : Signal detection unit 131 : Laser light source 132 : Optical condenser 133 : Signal meter 140 : Signal processing unit 141 : Signal Analysis Module 142 : Anomaly Detection Module 143 : Model Training Module 150 : Regeneration section 160 : Output section 170 : Particle separator 180: First air outlet 185 : Fine particle outlet 190: Second air outlet 200 : Air conditioning system 300 : Indoors (The present invention: FIGS. 9–11) 210: Pathogen sensing data input section 220 : Occupant Information Input Section 230 : Data Fusion Department 240 : Feature extraction unit 250 : Infected Person Identification Module 251 : Autoencoder-based anomaly detection module 252: Density-based anomaly detection model 260 : Judgment section 270: Alarm and Output Section

Claims

Claim 1 An infected person identification system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for identifying an infected person from the fusion data of the data fusion unit (230); and an infected person identification module (250) for calculating the probability of the existence of an infected person using the features for identifying an infected person as input; wherein the features for identifying an infected person extracted by the feature extraction unit (240) from the fusion data are core features related to an infected person that include at least one of pathogen features, behavioral features of an infected person, and environmental features. Claim 2 An infected person identification system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for identifying an infected person from the fusion data of the data fusion unit (230); and an infected person identification module (250) for calculating the probability of the presence of an infected person using the features for identifying an infected person as input; and a judgment unit (260) for determining whether an infected person has entered by comparing the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), with a predefined standard (Po). An infected person identification system further comprising, wherein the judgment unit (260) determines a step-by-step infection risk grade when the probability of the existence of an infected person (P) exceeds at least one of at least two multi-step reference values. Claim 3 In claim 2, the infected person identification system further comprises an output unit (270) that issues an alarm or outputs a judgment result signal to an external system when the judgment unit (260) determines that an infected person has entered. Claim 4 In claim 1, the pathogen sensing data input unit (210) receives pathogen sensing data input from a surface-enhanced Raman scattering (SERS) sensor of a real-time pathogen measurement system in the air, and the pathogen sensing data is characterized by being composed of one or more of a SERS spectrum, peak intensity, peak ratio, baseline drift, and signal change rate. Claim 5 An infected person identification system according to claim 4, characterized in that the pathogen sensing data is preprocessed by one or more preprocessing processes among noise filtering, background signal removal, peak detection and quantification, and time-based spectral change rate calculation. Claim 6 An infected person identification system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for identifying an infected person from the fusion data of the data fusion unit (230); and an infected person identification module (250) for calculating the probability of the presence of an infected person using the features for identifying an infected person as input. An infected person identification system comprising, wherein the data fusion unit (230) combines the pathogen sensing data from the pathogen sensing data input unit (210) and the occupant information from the occupant information input unit (220) by time synchronization or normalization, and wherein the data fusion unit (230) generates a fusion index after matching the pathogen sensing data and the occupant information to the same time axis. Claim 7 delete Claim 8 An infected person identification system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for identifying an infected person from the fusion data of the data fusion unit (230); and an infected person identification module (250) for calculating the probability of the presence of an infected person using the features for identifying an infected person as input. An infected person identification system comprising, wherein the infected person identification module (250) calculates the probability of the presence of an infected person using a machine learning or deep learning-based model, calculates the probability of the presence of an infected person (P) through a machine learning / deep learning-based probability prediction model, and includes an Autoencoder-based anomaly detection model (251) or a Density-based anomaly detection model (252) to detect anomalies corresponding to the influx of infected persons by comparing with a normal occupancy pattern. Claim 9 delete Claim 10 delete Claim 11 In claim 1, the occupant information input to the occupant information input unit (220) is non-video information implemented using only Wi-Fi or sensor-based methods without a camera, characterized by an infected person identification system. Claim 12 An infected person identification system for determining whether an infected person has entered an indoor space using pathogen sensing data and occupant information, comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for identifying an infected person from the fusion data of the data fusion unit (230); and an infected person identification module (250) for calculating the probability of the presence of an infected person using the features for identifying an infected person as input; wherein the infected person identification system operates in conjunction with multiple spaces and is characterized by being capable of tracking the movement path of an infected person by integrating data from multiple indoor spaces to continuously identify infected persons. Claim 13 A method for determining an infected person using an infected person determination system comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for determining an infected person from the fusion data of the data fusion unit (230); and an infected person determination module (250) for calculating the probability of the existence of an infected person using the features for determining an infected person as input; wherein the method comprises: (a) receiving the pathogen sensing data at the pathogen sensing data input unit (210) (S10); and (b) receiving the occupant information at the occupant information input unit (220) (S20). (d) after steps (a) and (b), a step (S40) in which the data fusion unit (230) combines the pathogen sensing data and occupant information to generate fusion data; (e) after step (d), a step (S50) in which the feature extraction unit (240) extracts features for identifying infected persons from the fusion data; and (f) after step (e), a step (S60) in which the infected person identification module (250) calculates the probability of the presence of infected persons; and (c) after step (a), a step (S30) in which at least one preprocessing process is performed on the pathogen sensing data, among i) noise filtering, ii) background signal removal, iii) peak detection and quantification, and iv) time-based spectrum change rate calculation; and further comprising, wherein step (d) is characterized by generating fusion data by combining the received pathogen sensing data from step (a) with the preprocessed pathogen sensing data and occupant information. Claim 14 delete Claim 15 In claim 13, the method for determining an infected person further comprises: (g) a step (S70) of comparing the probability of an infected person (P), which is the output value of the infected person determination module (250), with a predefined reference value (Po) after step (f); and (h) a step (S80) in which, as a result of the determination in step (g), if the probability of an infected person (P) is less than the reference value (Po) (P < Po), return to the beginning and perform the process again, and conversely, if the output value of the probability of an infected person (P) is greater than or equal to the reference value (P ≥ Po), determine that an infected person has entered and sound an alarm or output the result to a server or an external device. Claim 16 A method for determining an infected person using an infected person determination system comprising: a pathogen sensing data input unit (210) into which pathogen sensing data, including at least one airborne pathogen concentration measured in real time by a sensing device, is input; an occupant information input unit (220) into which occupant information, including at least one of the number of occupants and behavioral characteristics identified by a sensor, is input; a data fusion unit (230) for generating fusion data by combining the pathogen sensing data and the occupant information; a feature extraction unit (240) for extracting features for determining an infected person from the fusion data of the data fusion unit (230); and an infected person determination module (250) for calculating the probability of the existence of an infected person using the features for determining an infected person as input; wherein the method comprises: (a) receiving the pathogen sensing data at the pathogen sensing data input unit (210) (S10); and (b) receiving the occupant information at the occupant information input unit (220) (S20). (d) after steps (a) and (b), the data fusion unit (230) combines the pathogen sensing data and occupant information to generate fusion data (S40); (e) after step (d), the feature extraction unit (240) extracts features for identifying infected persons from the fusion data (S50); and (f) after step (e), the infected person identification module (250) calculates the probability of the presence of an infected person (S60); and (g1) after step (f), the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), is compared with a predefined first reference value (P1) (S71);(g2) If, as a result of the judgment in step (g1), the probability of the presence of an infected person (P) is less than the first reference value (P1) (P < P1), return to the beginning and perform again; conversely, if the output value of the probability of the presence of an infected person (P) is greater than or equal to the first reference value (P ≥ P1), compare the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), with a second reference value (P2) which is greater than the first reference value (P1) (S72); (h1) If, as a result of the judgment in step (g2), the probability of the presence of an infected person (P) is less than the second reference value (P2) (P < P2), determine the lowest level, the first infection risk level, and issue a first level alarm (S81); (g3) A step (S73) in which, as a result of the judgment in step (g2), if the output value of the probability of the presence of an infected person (P) is greater than or equal to the second reference value (P ≥ P2), the probability of the presence of an infected person (P), which is the output value of the infected person identification module (250), is compared with the nth reference value (Pn), which is greater than the second reference value (P2); (h2) A step (S82) in which, as a result of the judgment in step (g3), if the probability of the presence of an infected person (P) is less than the nth reference value (Pn) (P < Pn), it is determined to be an intermediate stage, the n-1st infection risk stage, and an n-1st stage alarm is issued; and (h3) A step (S83) in which, as a result of the judgment in step (g3), if the probability of the presence of an infected person (P) is greater than or equal to the nth reference value (Pn) (P ≥ Pn), it is determined to be the highest risk stage, the nth infection risk stage, and an nth stage highest stage alarm is issued; A method for identifying an infected person characterized by further including

Images