A laundry machine and a home health monitoring method based on laundry machine effluent

Abstract

The embodiment of the application discloses a kind of based on laundry machine waste liquid's family health monitoring method and laundry machine, the method comprises: in the process of the drainage stage after the main washing stage of laundry machine is executed, by drainage structure, the waste liquid of discharge is introduced into the storage bin;The waste liquid is detected by waste liquid detection module, and original multi-modal data is obtained;The operating parameter of the main washing stage of laundry machine is acquired, and based on operating parameter and original multi-modal data, the actual value of metabolite output is obtained;A family health baseline model is acquired in advance, and based on family health baseline model, the predicted value of metabolite output at the current time is output;According to the difference between the predicted value of metabolite output and the actual value of metabolite output, generate health analysis report.Washing waste liquid is converted into quantifiable metabolite output, so that laundry machine becomes the entrance of family health data collection, realizes the deep integration of health monitoring and daily washing.

Description

Technical Field

[0001] This invention relates to the field of health monitoring technology, and in particular to a home health monitoring method based on washing machine wastewater and a washing machine. Background Technology

[0002] Current home health monitoring mainly relies on specialized devices such as smart bracelets and smart toilets. These devices monitor health by collecting physiological parameters such as heart rate, blood pressure, weight, and sleep quality. These devices generally use fixed thresholds for health assessment and cannot adapt to fluctuations in individual physiological indicators caused by seasonal changes and changes in daily routines.

[0003] In existing technologies, the functions of ordinary smart washing machines mainly focus on automating the washing program, optimizing energy consumption, and basic clothing material recognition. They only focus on physical parameters during the washing process (such as water level, temperature, and spin speed) to achieve closed-loop control of the washing process. Their drainage system discharges the washing wastewater, which usually contains a large amount of biological information reflecting the physiological state of family members. Existing washing machines directly discard the washing wastewater and do not involve health monitoring based on the washing wastewater. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide a home health monitoring method and a washing machine based on washing machine waste liquid that overcomes or at least partially solves the above problems.

[0005] To address the aforementioned problems, a first aspect of the present invention provides a method for home health monitoring based on washing machine wastewater, wherein the washing machine includes a storage compartment disposed in a drainage structure and a wastewater detection module, and the method includes: During the drainage phase following the main wash cycle of the washing machine, the discharged waste liquid is introduced into the storage compartment through the drainage structure. The waste liquid is detected by the waste liquid detection module to obtain raw multimodal data; The operating parameters of the washing machine during the main wash phase are obtained, and the actual value of metabolite production is obtained based on the operating parameters and the original multimodal data. A pre-constructed family health baseline model is obtained, and based on the family health baseline model, a predicted value of metabolite production at the current time is output; the family health baseline model is trained using historical data of metabolite production. A health analysis report is generated based on the difference between the predicted and actual values ​​of the metabolite production.

[0006] According to a second aspect of the present invention, a washing machine is provided, the washing machine including a storage compartment disposed in a drainage structure and a waste liquid detection module, the washing machine comprising: The controller is used to guide the discharged waste liquid into the storage chamber through the drainage structure during the drainage phase after the main wash stage of the washing machine; detect the waste liquid through the waste liquid detection module to obtain raw multimodal data; acquire the operating parameters of the washing machine during the main wash stage, and obtain the actual value of metabolite production based on the operating parameters and the raw multimodal data; acquire a pre-constructed family health baseline model, and output the predicted value of metabolite production at the current moment based on the family health baseline model; the family health baseline model is trained using historical data of metabolite production; and generate a health analysis report based on the difference between the predicted value and the actual value of metabolite production.

[0007] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method for home health monitoring based on washing machine wastewater and a washing machine. The washing machine includes a storage compartment and a wastewater detection module disposed in the drainage structure. The method includes: during the drainage phase after the main wash phase of the washing machine, introducing the discharged wastewater into the storage compartment through the drainage structure; detecting the wastewater through the wastewater detection module to obtain raw multimodal data; acquiring the operating parameters of the washing machine during the main wash phase, and obtaining the actual value of metabolite production based on the operating parameters and the raw multimodal data; acquiring a pre-constructed home health baseline model, and outputting the predicted value of metabolite production at the current moment based on the home health baseline model; the home health baseline model is trained using historical data of metabolite production; and generating a health analysis report based on the difference between the predicted value and the actual value of metabolite production. By incorporating a storage compartment and waste liquid detection module into the washing machine's drainage structure, waste liquid is automatically collected and subjected to multimodal detection during the main wash cycle. This transforms the washing machine into a health monitoring portal, extracting various biological information from the waste liquid. This fills the technological gap in existing washing machines' inability to perform home health monitoring, realizing the functional transformation of waste liquid from a useless discharge to a carrier of health information, and achieving deep integration of health monitoring with daily washing behavior. Users do not need to wear any devices or perform any additional operations; they can complete the collection and analysis of health data during normal washing machine use. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the steps of a home health monitoring method based on washing machine waste liquid provided in an embodiment of the present invention; Figure 2This is a flowchart of another home health monitoring method based on washing machine waste liquid provided by an embodiment of the present invention; Figure 3 This is a schematic flowchart of a home health monitoring method based on washing machine waste liquid provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the storage compartment of a washing machine provided in an embodiment of the present invention.

[0009] Explanation of reference numerals in the attached drawings: storage compartment 30, floor drain compartment 31, vent valve 32, micro-spectral analysis unit 301, microfluidic biodetection unit 302, basic physical property detection unit 303, filter screen 304, ultrasonic oscillator 305. Detailed Implementation

[0010] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0011] In existing technologies, the functions of ordinary smart washing machines mainly focus on automating the washing program, optimizing energy consumption, and basic clothing material recognition. They only focus on physical parameters during the washing process (such as water level, temperature, and spin speed) to achieve closed-loop control of the washing process. Their drainage system discharges the washing wastewater, which usually contains a large amount of biological information reflecting the physiological state of family members. Existing washing machines directly discard the washing wastewater and do not involve health monitoring based on the washing wastewater.

[0012] One of the core concepts of this invention lies in the automatic collection and multimodal detection of waste liquid during the main wash cycle by incorporating a storage compartment and waste liquid detection module into the washing machine's drainage structure. This transforms the washing machine into a health monitoring entry point, extracting various biological information from the waste liquid. This fills the technological gap in existing washing machines' inability to perform home health monitoring, realizing the functional transformation of waste liquid from a useless discharge to a carrier of health information, and achieving a deep integration of health monitoring and daily washing behavior. Users can complete the collection and analysis of health data during normal use of the washing machine without wearing any devices or performing any additional operations.

[0013] Reference Figure 1 This diagram illustrates a flowchart of a home health monitoring method based on washing machine wastewater provided by an embodiment of the present invention. The washing machine includes a storage compartment and a wastewater detection module disposed in the drainage structure. The method specifically includes the following steps: Step 101: During the drainage stage after the main wash stage of the washing machine, the discharged waste liquid is introduced into the storage compartment through the drainage structure. In this embodiment of the invention, a miniature three-way valve is installed after the washing machine's drain pump and before the drain pipe enters the floor drain, leading out a bypass pipe connected to a specially designed storage compartment. This storage compartment adopts a fully enclosed design and is built into an unused space on the side of the washing machine base to avoid any risk of leakage.

[0014] During the draining phase following the main wash cycle, when the main control board issues a drain command, the three-way valve briefly opens the bypass according to a preset timing sequence (usually for the first 10-30 seconds after draining begins). Using the pressure of the drain pump itself, approximately 50-100 ml of the initial high-concentration waste liquid is introduced into the buffer tank of the storage compartment. The three-way valve is connected to the washing machine's main control board via a simple digital signal line. When the main control board issues a drain command, it simultaneously sends a precisely timed control signal. Upon receiving the signal, the three-way valve opens the bypass according to preset logic. Shielded cables are used between the three-way valve and the main control board to avoid electromagnetic interference.

[0015] A normal washing machine cycle includes a main wash, a rinse, and a spin-dry. The wastewater from the main wash contains a large amount of human metabolites (sebum, protein, cortisol metabolites, etc.), which are highly valuable for health monitoring. The wastewater from the rinse mainly contains residual detergent, with low levels of metabolites, and therefore has low value for health monitoring. The washing and rinsing stages should be distinguished, and sampling should only be conducted during the main wash drainage stage. The wastewater from the main wash contains richer health-related biological information, while sampling during the rinse stage is not very helpful for health monitoring. The washing machine's main control board determines whether to trigger sampling based on the washing program status (main wash / rinse). During the main wash drainage process, the main control board sends a sampling signal; during the rinse drainage process, the main control board does not send a sampling signal.

[0016] In this embodiment of the invention, during the drainage process after the main wash stage, the washing machine uses a reversing structure such as a three-way valve in the drain pipe to introduce the high-concentration waste liquid flowing out in the initial stage of drainage (e.g., the first 10-30 seconds) into a storage chamber, instead of directly discharging it into the sewer. The waste liquid from the main wash stage contains sebum, shed epidermal cells, sweat components, cortisol and other metabolites washed off clothing during the washing process, as well as environmental microorganisms, making it the stage with the richest biological information in the waste liquid. In contrast, the waste liquid from the subsequent rinsing stage has been significantly diluted, with extremely low concentrations of metabolites, and has no health monitoring value. Therefore, sampling is only performed during the main wash drainage stage, ensuring sufficient waste liquid samples with analytical value while avoiding interference with the normal drainage process of the washing machine. The storage chamber is a closed container located in the empty space on the base or side of the washing machine, and integrates a waste liquid detection module. After the waste liquid is introduced into the storage chamber, it enters a state awaiting detection, and the waste liquid detection module subsequently performs multimodal analysis on the waste liquid.

[0017] Step 102: The waste liquid is detected by the waste liquid detection module to obtain the original multimodal data; The waste liquid detection module is an integrated detection unit assembly within the storage compartment, used for multi-dimensional analysis of the introduced waste liquid. Its core function is to convert the biochemical information in the waste liquid, which cannot be directly observed, into raw multimodal data reflecting the composition of the waste liquid.

[0018] In this embodiment of the invention, the waste liquid detection module comprises three components: a micro-spectral analysis unit, a microfluidic biodetection unit, and a basic physical property detection unit. The micro-spectral analysis unit measures the absorption of light at a specific wavelength by the waste liquid, outputting absorbance spectral data to reflect the content of substances such as proteins, lipids, and heme in the waste liquid. The microfluidic biodetection unit reacts the waste liquid with a pre-placed dry biochemical reagent to produce a specific colorimetric reaction, measures the absorbance of the reaction products, and outputs the concentration values ​​of specific biomarkers such as cortisol and fungal metabolites. The basic physical property detection unit uses electrodes and sensors directly immersed in the waste liquid to output basic physicochemical parameters such as pH, conductivity, and temperature. These three detection units operate independently and in parallel, simultaneously analyzing the same waste liquid sample from three dimensions: spectral characteristics, biomarker concentration, and basic physicochemical properties, achieving a comprehensive understanding of the waste liquid's composition.

[0019] Step 103: Obtain the operating parameters of the washing machine during the main wash stage, and based on the operating parameters and the original multimodal data, obtain the actual value of metabolite production. In this embodiment of the invention, the operating parameters of the washing machine during the main wash phase are obtained, and the raw multimodal data obtained from the detection are processed based on these parameters to finally obtain the actual value of metabolite production. The operating parameters include the weight of the clothes and the amount of water used in this wash, which determine the degree of dilution of metabolites in the waste liquid. The more clothes and the larger the amount of water, the more thoroughly the same amount of metabolites are diluted in the waste liquid, and the lower the detected concentration. Therefore, the concentration values ​​directly measured in the raw multimodal data (such as protein concentration measured by spectral analysis and cortisol concentration measured by microfluidic detection) cannot be directly compared between different washing batches.

[0020] Based on the weight of the clothes and the amount of water used in this wash, a dynamic dilution factor is calculated. This factor reflects the difference in dilution between this wash and standard washing conditions (e.g., 50L of standard water and 3kg of standard clothes). The measured concentration is then corrected using the dilution factor to obtain the corrected concentration value. Finally, the corrected concentration is multiplied by the amount of water used in this wash and divided by the weight of the clothes to obtain the metabolite production per unit of clothing. The metabolite production per unit of clothing is measured in units of "total metabolites carried per kilogram of clothing" (e.g., μg / kg or mg / kg), and its physical meaning is the total amount of metabolites carried on the clothing.

[0021] Step 104: Obtain a pre-constructed family health baseline model, and based on the family health baseline model, output the predicted value of metabolite production at the current time; the family health baseline model is trained using historical data of metabolite production. The family health baseline model is a personalized time series analysis model trained using historical test data. Its core function is to learn the family's unique metabolic level, trends, and cyclical fluctuations based on the family's long-term historical data on metabolite production, thereby outputting a predicted value at the current moment as a personalized reference benchmark to determine whether the actual test value is abnormal.

[0022] The construction of the family health baseline model is based on time series decomposition theory. It assumes that the observed value of each metabolic indicator is composed of the superposition of three non-observable components: a long-term trend term, a seasonality term, and a random noise term. The long-term trend term is a slowly changing value over time, reflecting the overall direction of change of the metabolic indicator over a longer period (such as a continuous increase or decrease due to aging or changes in lifestyle). The seasonality term is a numerical sequence that repeats at fixed time periods (such as a week or a year), reflecting regular fluctuations caused by factors such as differences in weekday and weekend schedules and seasonal changes; the sum of the seasonality term is zero within a complete cycle. The random noise term is a random variable with a mean of zero, representing accidental fluctuations and measurement errors that cannot be explained by the trend and seasonality terms.

[0023] In the initial stage, at least 20 valid washing data sets are accumulated during the first use of the washing machine. Offline training is performed using a time series decomposition algorithm to obtain initial estimates of the long-term trend and seasonality parameters. After initial training, the model enters an online update mode. Whenever new detection data arrives, the system updates the model parameters in real time through a recursive mechanism of "prediction-comparison-correction": a predicted value is calculated based on the current long-term trend and the corresponding seasonality at the current moment; then, the actual detection value is compared with the predicted value to calculate the deviation; finally, based on the magnitude of the deviation, the long-term trend and seasonality are moderately adjusted at a preset learning rate, enabling the model to slowly track natural physiological changes in individuals (such as the slowing of metabolism due to aging) while maintaining its resistance to random fluctuations.

[0024] In this embodiment of the invention, a pre-trained family health baseline model is acquired, and the predicted value of metabolite production at the current moment is output using the family health baseline model. The family health baseline model is trained using historical data on metabolite production accumulated over a long period of time by the family. Before each new detection data arrives, the family health baseline model calculates an expected value based on the family's historical variation patterns; that is, under normal circumstances, considering long-term trends and periodic fluctuations, the value that the metabolite production at the current moment should be close to. The family health baseline model trains the historical detection data into two parameter components—a long-term trend term and a seasonality term—using a time series decomposition algorithm, and obtains the predicted value of metabolite production at the current moment based on these two parameter components. The predicted value is an "expected estimate" calculated based on historical patterns. The actual detection data is used to subsequently compare with the predicted value to measure the degree to which the actual metabolic level deviates from the expected value, and to update the model parameters to make it more accurately reflect the family's current true metabolic state.

[0025] Step 105: Generate a health analysis report based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output.

[0026] In this embodiment of the invention, the actual value of metabolite output (i.e., the measured metabolite output per unit of clothing) is compared with the predicted value (i.e., the expected value at the current moment calculated by the model based on historical patterns), the difference between the two is calculated, and a health analysis report is generated based on this difference. Fuzzy inference is triggered based on the difference between the predicted and actual values ​​of metabolite output. After triggering inference, the deviation is fuzzified and mapped to fuzzy concepts such as "low," "normal," "high," and "persistently high." These fuzzy concepts are then input into a preset fuzzy rule base for inference. The fuzzy rule base contains multiple rules built based on medical common sense (e.g., "If cortisol is persistently high and protein levels are simultaneously high, it indicates excessive stress load"). Based on the triggered rules and their confidence levels, a natural language template is invoked to generate the final health analysis report.

[0027] The final health analysis report is presented to the user for reference only and does not generate any device control commands. Users can view the report content through mobile apps or other devices.

[0028] Reference Figure 2 This diagram illustrates a flowchart of another home health monitoring method based on washing machine wastewater provided by an embodiment of the present invention. The washing machine includes a storage compartment disposed in the drainage structure and a wastewater detection module. The storage compartment includes a filter and an ultrasonic oscillating plate. The method specifically includes the following steps: Step 201: When identifying the drainage stage after the washing machine performs the main wash stage, control the drainage structure to introduce the waste liquid into the storage compartment via a bypass. In this embodiment of the invention, upon detecting the drainage stage following the main wash phase of the washing machine, the drainage structure is controlled to redirect the wastewater into the storage compartment via a bypass. A complete washing process typically includes a main wash phase and multiple rinse phases. The main wash phase is the core stage where detergent is added to clean the clothes. The wastewater discharged during this phase contains human sebum, shed epidermal cells, sweat components, cortisol and other metabolites washed off the clothes, as well as environmental microorganisms, making it the richest in biological information. In contrast, the wastewater discharged during subsequent rinse phases is significantly diluted, with extremely low concentrations of metabolites, offering no health monitoring value. Therefore, this application only performs sampling during the drainage process of the main wash phase.

[0029] A miniature three-way valve is installed after the washing machine's drain pump and before the drain pipe enters the floor drain. One inlet of the three-way valve connects to the drain pump outlet, one outlet connects to the main drain pipe leading to the floor drain, and the other outlet connects to the storage compartment via a bypass pipe. The three-way valve is physically connected to the washing machine's main control board via a digital signal line. When the main control board executes the main wash cycle to its end and begins draining, it sends a drain command and a control signal to the three-way valve. Based on this signal, the three-way valve switches to the bypass conduction state, using the drain pump's own pressure to introduce waste liquid into the storage compartment. After the preset sampling time has elapsed, the control signal terminates, and the three-way valve resets to the main drain pipe conduction state. Subsequent waste liquid is then discharged normally into the sewer. The entire sampling process has no impact on the washing machine's normal drainage process.

[0030] Step 202: The waste liquid is detected by the waste liquid detection module to obtain the original multimodal data; In this embodiment of the invention, the waste liquid detection module comprises three components: a micro-spectral analysis unit, a microfluidic biodetection unit, and a basic physical property detection unit. The micro-spectral analysis unit measures the absorption of light at a specific wavelength by the waste liquid, outputting absorbance spectral data to reflect the content of substances such as proteins, lipids, and heme in the waste liquid. The microfluidic biodetection unit reacts the waste liquid with pre-placed dry biochemical reagents to produce a specific colorimetric reaction, measures the absorbance of the reaction products, and outputs the concentration values ​​of specific biomarkers such as cortisol and fungal metabolites. The basic physical property detection unit uses electrodes and sensors directly immersed in the waste liquid to output basic physicochemical parameters such as pH, conductivity, and temperature.

[0031] In some embodiments, the waste liquid detection module includes a miniature spectral analysis unit; the storage chamber has opposing transparent windows on both sides; the miniature spectral analysis unit includes a light source and an optical fiber probe located at the transparent windows on both sides of the storage chamber; step 203 may include the following sub-steps: Sub-step S11: The absorbance spectral data of the waste liquid in a preset wavelength range is obtained through the micro-spectral analysis unit, and the absorbance of metabolites is extracted from the absorbance spectral data; Sub-step S12: Determine the concentration value of the first metabolite based on the preset spectral standard curve and the absorbance of the metabolite; the spectral standard curve is a curve describing the quantitative relationship between the concentration value and absorbance of each metabolite, obtained by fitting the concentration value of each metabolite on the x-axis and the absorbance value of each metabolite on the y-axis.

[0032] The hardware components of the miniature spectral analysis unit include a broadband pulsed xenon lamp light source (covering 200nm-1100nm) and a high-sensitivity miniature fiber optic spectrometer (spectral resolution <1.5nm). The fiber optic probes of the light source and spectrometer are located at transparent windows on both sides of the storage compartment, forming a transmission optical path. The transparent windows, being in long-term contact with washing waste liquid, do pose a risk of scaling and biofilm adhesion. Preventive measures include using a special anti-fouling coating (such as a hydrophobic coating) on ​​the transparent windows and designing the window surface to be slightly tilted to reduce contaminant adhesion. After each sampling, an automatic window cleaning operation is performed, spraying a small amount of clean water through a miniature nozzle to remove surface contaminants.

[0033] The detection principle involves light emitted from a light source passing through a waste liquid sample in a buffer pool. Organic and inorganic substances in the sample selectively absorb light of specific wavelengths. The spectrometer receives the transmitted light and generates a complete absorption spectrum curve. Different substances have characteristic absorption peaks. For example, proteins have characteristic absorption around 280 nm; DNA / RNA absorbs at 260 nm; lipids in sebum have characteristic absorption in the near-infrared band of 900-950 nm; fluorescent whitening agents in detergents have strong absorption at 350-400 nm; and heme has extremely strong absorption at 410-420 nm (Soray peak). The data output includes absorbance spectral data across the entire wavelength range (200 nm-950 nm).

[0034] The miniature spectral analysis unit outputs absorbance spectral data across the entire wavelength range (200nm-950nm), meaning a complete spectral curve with one absorbance value for each wavelength. The algorithm extracts absorbance values ​​at specified characteristic wavelengths from the spectrum: proteins (280nm), nucleic acids (260nm), heme (415nm), and lipids (900-950nm integral value). These extracted absorbance values ​​are then substituted into a standard curve to calculate the concentration of the corresponding substance. For example, with proteins: light is emitted from a light source and passes through a waste liquid sample in a buffer pool. Proteins in the waste liquid absorb light at a wavelength of 280nm (this is the characteristic absorption peak of proteins). The spectrometer receives the light after it passes through the waste liquid, generating a complete spectrum. The algorithm finds the absorbance value corresponding to 280nm (e.g., absorbance = 0.5). The algorithm consults the standard curve, knowing that "absorbance 0.5 ≈ protein concentration 10mg / L". The output protein concentration is 10mg / L.

[0035] In this embodiment of the invention, the waste liquid detection module includes a miniature spectral analysis unit. Opposite transparent windows are provided on both sides of the storage compartment, allowing light beams in the ultraviolet-visible-near-infrared band to pass through. The miniature spectral analysis unit includes a light source and an optical fiber probe located at the transparent windows on both sides of the storage compartment. The light source is a broadband pulsed xenon lamp, covering a broad spectrum from 200 nm to 1100 nm; the optical fiber probe is connected to a miniature fiber optic spectrometer with a spectral resolution better than 1.5 nm. The light emitted by the light source passes through the waste liquid sample between the transparent windows in the storage compartment, is received by the optical fiber probe on the other side, and transmitted to the spectrometer for analysis.

[0036] A miniature spectroscopic analysis unit acquires absorbance spectral data of the waste liquid within a preset wavelength range, and extracts the absorbance of metabolites from this data. Specifically, broadband light emitted from a light source passes through the waste liquid sample in the storage chamber. Organic and inorganic substances in the waste liquid selectively absorb light at specific wavelengths. The transmitted light is received by a fiber optic probe and transmitted to a spectrometer, which generates absorbance spectral data across the entire wavelength range (e.g., 200nm-950nm). The absorbance values ​​at characteristic wavelengths reflecting the content of specific metabolites are extracted from this full-band spectral data. For example, proteins have a characteristic absorption peak near 280nm; extracting the absorbance value at 280nm from the spectrum reflects the protein content. DNA / RNA has an absorption peak at 260nm; heme has a very strong Solvay peak absorption at 410-420nm; and lipids have characteristic absorption in the near-infrared band at 900-950nm.

[0037] The concentration of the metabolite is determined based on a pre-set spectral standard curve and the absorbance of the metabolite. The spectral standard curve is pre-established and stored in the system as follows: a series of standard solutions of known concentrations are prepared using pure or high-purity standard substances of the target metabolite and pure water. The absorbance of each standard solution at a characteristic wavelength is measured using a micro-spectral analysis unit. A quantitative relationship curve is then obtained by fitting algorithms such as the least squares method with the concentration of the standard solution on the x-axis and the measured absorbance on the y-axis. During actual detection, the extracted absorbance value is substituted into the standard curve equation to calculate the concentration of the corresponding metabolite in the waste liquid. By setting transparent windows on both sides of the storage chamber and configuring a light source and fiber optic probe, in-situ transmission measurement of the waste liquid in the storage chamber can be performed without additional sampling, simplifying the detection process. The pre-established spectral standard curve quantitatively converts the absorbance value into a concentration value, thereby achieving rapid, accurate, and quantitative detection of metabolites in the waste liquid.

[0038] In some embodiments, the waste liquid detection module further includes a microfluidic biodetection unit; the microfluidic biodetection unit is located at the bottom of the storage chamber, and the microfluidic biodetection unit includes a microchannel network and a light-emitting device and a light-receiving device respectively disposed at both ends of each channel of the microchannel network. Each channel of the microchannel network is pre-filled with dry biochemical reagents for different metabolites, and different metabolites in the waste liquid undergo different colorimetric reactions with the dry biochemical reagents; step 203 may further include the following sub-steps: In sub-step S21, a light beam is emitted through the light-emitting device and passes through the waste liquid that has undergone a colorimetric reaction. The intensity of the transmitted light of the metabolites is detected by the light-receiving device, and the detected intensity of the transmitted light is converted into an absorbance value. Sub-step S22: Determine the concentration value of the second metabolite based on the pre-calibrated standard curve and the absorbance value; the standard curve is a pre-established curve that describes the quantitative relationship between the absorbance value and the concentration value of each metabolite by fitting the concentration value of each metabolite to the x-axis and the measured absorbance value to the y-axis.

[0039] The microfluidic biodetection unit integrates a microchannel network, dry reagents, and a high-precision photoelectric detection array. The dry biochemical reagents have a limited lifespan and are replaceable. Waste liquid from the storage chamber is diverted into multiple parallel microchannels within the microfluidic biodetection unit. Each channel is pre-filled with a different dry biochemical reagent; after the waste liquid dissolves the reagent, a specific colorimetric reaction occurs, with the reaction depth proportional to the target analyte concentration. A photodiode array below reads the absorbance changes of each channel at a specific wavelength in real time, outputting the concentration value of the specific biomarker.

[0040] Waste liquid in the storage chamber is diverted within a microchannel network, simultaneously entering multiple parallel microchannels. The waste liquid flows through the microchannels and comes into contact with dry reagents. If the waste liquid contains the target biomarker, it will react specifically with the reagent, producing a color change. Taking channel 1 for cortisol detection as an example: as the waste liquid flows through, the cortisol in the waste liquid reacts with the dry reagent; the more vigorous the reaction, the deeper the solution color (from colorless to blue), and the color depth is directly proportional to the cortisol concentration. Each channel has a light-emitting diode (LED) (emitting light of a specific wavelength) and a photodetector (receiving transmitted light). The LED emits light of a fixed wavelength, which passes through the colorimetric solution in the reaction chamber. The photodetector measures the intensity of the transmitted light; the more colorimetric product in the solution (i.e., the higher the concentration of the target analyte), the higher the concentration. The intensity of the transmitted light measured by the photodetector is converted into an absorbance value: absorbance = log(incident light intensity / transmitted light intensity). The absorbance is then converted into a concentration value using a pre-calibrated standard curve.

[0041] In this embodiment of the invention, the waste liquid detection module further includes a microfluidic biodetection unit located at the bottom of the storage chamber. This unit includes a microchannel network and light-emitting devices and light-receiving devices respectively disposed at both ends of each channel. Each channel is pre-filled with dry biochemical reagents (such as cortisol, fungal metabolites, etc.) for different metabolites. When the waste liquid enters the channel, different metabolites undergo specific colorimetric reactions with their corresponding dry biochemical reagents, and the reaction depth is proportional to the concentration of the target substance. The light-emitting device emits a light beam that passes through the waste liquid undergoing the colorimetric reaction, and the light-receiving device detects the intensity of the transmitted light and converts it into an absorbance value. The absorbance value is then converted into a metabolite concentration value according to a pre-calibrated standard curve, which is a quantitative relationship curve pre-fitted with metabolite concentration values ​​on the x-axis and corresponding absorbance values ​​on the y-axis.

[0042] By pre-loading dry biochemical reagents for different metabolites into a microchannel network, parallel detection of multiple specific biomarkers such as cortisol and fungal metabolites can be achieved with only a trace amount of waste liquid, greatly improving detection throughput and efficiency. The combined use of a luminescent device and a light-receiving device accurately quantifies the intensity of the colorimetric reaction into absorbance values, which are then converted into concentration values ​​using a standard curve, enabling precise quantification of specific biomarkers.

[0043] In some embodiments, the waste liquid detection module further includes a basic physical property detection unit; the basic physical property detection unit is disposed inside the storage chamber; the basic physical property detection unit includes a pH electrode, a conductivity electrode, and a temperature sensor; step 203 may further include the following sub-steps: Sub-step S31 involves detecting the pH value, conductivity, and temperature of the waste liquid using the pH electrode, the conductivity electrode, and the temperature sensor.

[0044] The basic physical property detection unit is integrated within the storage compartment, including a miniature pH electrode, conductivity electrode, and temperature sensor. The detection principle directly measures the basic physicochemical properties of the waste liquid, outputting pH value, conductivity (μS / cm), and temperature (°C). These data are primarily used to assist in calibrating readings from other sensors and as supplementary information in the analysis report.

[0045] In this embodiment of the invention, the waste liquid detection module further includes a basic physical property detection unit, which is located inside the storage chamber and includes a pH electrode, a conductivity electrode, and a temperature sensor. After the waste liquid is introduced into the storage chamber, the pH value, conductivity, and temperature of the waste liquid are detected in real time by directly immersing the pH electrode, conductivity electrode, and temperature sensor in the waste liquid, and the basic physicochemical parameters are output. These detections are performed in parallel with spectral analysis and microfluidic detection, without the need for additional sampling or pretreatment, and can be completed within seconds. By integrating the basic physical property detection unit, these parameters can be quickly obtained. On the one hand, the validity of the waste liquid sample can be preliminarily judged (e.g., a pH value that is too low or too high may indicate an abnormality in the sample). On the other hand, this data can be directly incorporated into the health analysis report as auxiliary information to help users gain a more comprehensive understanding of the overall condition of the washing waste liquid.

[0046] Step 203: Obtain the operating parameters of the washing machine during the main wash stage, and based on the operating parameters and the original multimodal data, obtain the actual value of metabolite production. In this embodiment of the invention, the operating parameters of the washing machine during the main wash stage are acquired, and the raw multimodal data obtained are processed based on these parameters to finally obtain the actual value of metabolite production. A dynamic dilution factor is calculated based on the weight of the clothes and the amount of water used in this wash. The dilution factor reflects the difference in dilution between this wash and standard washing conditions (e.g., 50L of standard water and 3kg of standard clothes). The measured concentration is then corrected using the dilution factor to obtain the corrected concentration value. Finally, the corrected concentration is multiplied by the amount of water used in this wash and divided by the weight of the clothes to obtain the metabolite production per unit of clothing. The metabolite production per unit of clothing is measured in units of "total metabolites carried per kilogram of clothing" (e.g., μg / kg or mg / kg), and its physical meaning is the total amount of metabolites carried on the clothing.

[0047] In some embodiments, the operating parameters include the weight of the clothes and the amount of washing water, and step 203 may include the following sub-steps: Sub-step S41: Determine the dynamic dilution factor based on the standard water volume, the washing water volume, the standard weight of the clothes, and the weight of the clothes; Sub-step S42: Correct the original multimodal data according to the dynamic dilution factor to obtain corrected multimodal data; Sub-step S43: Based on the corrected multimodal data, the washing water volume, and the weight of the clothing, obtain the actual value of the metabolite production.

[0048] Based on the weight of the laundry, water level, and washing mode of this wash, a dynamic dilution factor is calculated. This factor is designed based on the principle of mass conservation: for the same amount of dirt, more water results in a lower concentration, thus requiring an upward correction of the measured concentration; conversely, less water results in a higher concentration, requiring a downward correction. The correction coefficient is proportional to the amount of water per unit of laundry. To further eliminate the influence of varying laundry loads in each wash, the algorithm multiplies the corrected concentration by the water level of the wash and then divides it by the weight of the laundry to obtain the final comparable index. The unit of this index is "milligrams of metabolites per kilogram of laundry," which physically represents the total amount of metabolites carried per kilogram of laundry. Theoretically, this index is independent of the specific parameters of the wash and only reflects the total amount of metabolites on the clothing, ensuring long-term comparability of test data from different times and batches.

[0049] Dilution factor calculation principle: According to the principle of mass conservation, for the same amount of metabolites, the more water used, the lower the concentration, requiring upward correction of the measured concentration; conversely, the less water used, the higher the concentration, requiring downward correction. The dynamic dilution factor is used to standardize metabolite concentrations under different washing conditions to a comparable level. Calculation formula: Dynamic dilution factor = (Standard water volume / Current washing water volume) × (Standard garment weight / Current garment weight). Correction process: Corrected concentration = Measured concentration × Dynamic dilution factor.

[0050] For example: standard water volume is 50L (typical water volume of washing machine), standard laundry weight is 3kg (typical laundry weight of washing machine). Assuming the washing parameters for this wash are: laundry weight = 2kg, water volume = 40L, dynamic dilution factor = (50 / 40) × (3 / 2) = 1.25 × 1.5 = 1.875, and the measured cortisol concentration is 10μg / L, then the corrected concentration = 10 × 1.875 = 18.75μg / L.

[0051] To eliminate the impact of varying laundry loads in each wash, the corrected concentration is converted into "metabolic output per unit of laundry," which is the total amount of metabolites carried per kilogram of laundry. The calculation formula is: Final comparable index (i.e., the actual value of metabolite output) = Corrected concentration × Wash water volume / Laundry weight, where mg / kg is the unit of metabolites per kilogram of laundry.

[0052] Example: Assuming the corrected cortisol concentration = 18.75 μg / L, the amount of water used in this wash = 40L, and the weight of the clothes in this wash = 2kg, the final comparable index = 18.75 × 40 / 2 = 375 μg / kg. This index represents the total amount of metabolites carried per kilogram of clothing. Theoretically, it is unrelated to the specific parameters of this wash and only reflects the total amount of metabolites loaded on the clothing, making the test data from different times and different batches comparable over a long period of time.

[0053] In this embodiment of the invention, a dilution factor is calculated based on the standard water volume, the current wash water volume, the standard garment weight, and the current garment weight, using the formula "Dynamic dilution factor = (standard water volume / current wash water volume) × (standard garment weight / current garment weight)". The dilution factor reflects the degree of dilution difference between the current wash and standard conditions. This dynamic dilution factor is multiplied by the measured concentration of each metabolite in the original multimodal data to obtain a corrected concentration value, making the concentration data under different washing conditions comparable under standard conditions. The corrected concentration is multiplied by the current wash water volume and divided by the current garment weight to obtain the metabolite output per unit of garment (i.e., the actual value of metabolite output, in mg / kg of garment), which physically represents the total amount of metabolites carried per kilogram of garment. By introducing a dynamic dilution factor to correct the measured concentration and further calculating the metabolite output per unit of garment, the incomparability of data caused by differences in garment weight and water volume in each wash is effectively eliminated, avoiding misjudgments due to differences in washing conditions.

[0054] Step 204: Obtain a pre-constructed family health baseline model, and based on the family health baseline model, output the predicted value of metabolite production at the current time; the family health baseline model is trained using historical data of metabolite production. In this embodiment of the invention, a pre-trained family health baseline model is obtained, and the predicted value of metabolite production at the current moment is output using the family health baseline model. The family health baseline model is trained using historical data on metabolite production accumulated over a long period of time by the family. Before each new detection data arrives, the family health baseline model calculates an expected value based on the historical change patterns of the family; that is, under normal circumstances, after considering long-term trends and periodic fluctuations, the value that the metabolite production at the current moment should be close to. The family health baseline model trains the historical detection data into two parameter components, a long-term trend term and a seasonality term, using a time series decomposition algorithm, and obtains the predicted value of metabolite production at the current moment based on these two parameter components.

[0055] In some embodiments, the family health baseline model includes a long-term trend term and a seasonality term; the long-term trend term characterizes the trend of metabolite changes over a longer time scale; the seasonality term characterizes the regular fluctuations of metabolites that recur at fixed periods; step 204 may include the following sub-steps: Sub-step S51: Based on the long-term trend term of the family health baseline model, determine the first predicted value of metabolite output at the current moment, and based on the seasonality term of the family health baseline model, determine the second predicted value of metabolite output at the current moment. Sub-step S52: Based on the first predicted value and the second predicted value, determine the target predicted value of metabolite production at the current time.

[0056] For each metabolic indicator's time-series data, the family health baseline model consists of the superposition of three unobservable components: a long-term trend term reflecting slowly changing baseline levels, such as the natural slowing of metabolism with age, or long-term upward or downward trends due to changes in lifestyle; a seasonality term reflecting periodic fluctuations, such as the weekend effect (higher activity levels and more metabolites on weekends) or the seasonal effect (dry winters and more dandruff); and a random noise term representing the accidental error of a single measurement and random fluctuations that cannot be explained by any regularity. The long-term trend term itself consists of the current baseline level and the direction and speed of this level's change over time (i.e., the trend slope). Both of these components can change slowly over time, allowing the model to adapt to slow physiological shifts.

[0057] In this embodiment of the invention, the family health baseline model is trained on historical test data using a time series decomposition method, decomposing the historical observations of each metabolic indicator into two parameter components: a long-term trend term and a seasonality term. The long-term trend term is a slowly changing value over time, representing the overall direction of change in metabolites over a timescale of several months or longer (e.g., a continuous increase or decrease due to aging or changes in lifestyle). The seasonality term is a set of numerical sequences bound to a fixed period (e.g., a week or a year), representing the regular fluctuations in metabolites caused by factors such as differences in weekday and weekend schedules and seasonal changes. The two components are independent, describing the long-term direction and periodic fluctuations in the metabolite variation patterns, respectively.

[0058] Based on the current moment, the long-term trend value corresponding to this moment is obtained from the long-term trend term as the first predicted value, and the seasonal adjustment amount corresponding to the cycle phase (e.g., Wednesday) of the current moment is obtained from the seasonality term as the second predicted value. The first and second predicted values ​​are added together to obtain the target predicted value of metabolite production at the current moment. This predicted value is calculated based solely on the time information of the current moment (e.g., what day of the week it is) and is completely independent of the actual detection data for this period; therefore, the calculation can be completed before waste liquid testing is finished. By decomposing the metabolite variation pattern into two independent components—long-term trend and seasonal fluctuation—the predicted value can simultaneously reflect the long-term direction of change and periodic patterns (e.g., weekend effect, seasonal variation) of the household.

[0059] In some embodiments, the method further includes the following steps: Based on the deviation between the predicted and actual values ​​of metabolite production and the preset learning rate, parameter adjustment information is obtained; the long-term trend term and the seasonality term are corrected based on the parameter adjustment information.

[0060] After the initial training phase, the model enters an online update mode. Whenever new detection data arrives, the system adjusts the model parameters in real time using a recursive update algorithm, allowing the baseline to dynamically adapt to the slow changes in individual data. This update mechanism can be understood through three steps: "prediction-comparison-correction." Prediction: Based on past data and established patterns (trends and periodicity), the model makes a predicted estimate of the current detection value. Comparison: The actual detection value is compared with the predicted value, and the difference between the two is calculated. Correction: Based on the magnitude of the difference between the actual and predicted values, the model decides the extent to which it adjusts the original parameters. If the difference is small, it indicates that the model's prediction is accurate, requiring only fine-tuning; if the difference remains consistently large, it indicates that the individual's physiological patterns may have actually changed, and the model will gradually adjust the trend and seasonality terms to make future predictions more accurate. This online update mechanism allows the model to automatically adapt to slow physiological shifts in individuals (e.g., slower metabolism with age) without requiring retraining offline.

[0061] The prediction uses model parameters (long-term trend term and seasonality term) fitted with historical data to calculate the predicted value for the current moment. Example: A family's predicted cortisol level on Wednesday = 14 μg / kg (based on long-term trend and seasonality patterns). The actual measured value is compared with the predicted value, and the difference between the predicted and actual observed values ​​is calculated using the formula: Difference = Actual Observed Value - Predicted Value. Example: Predicted value 14 μg / kg, actual value 18 μg / kg, then the difference = 4 μg / kg. The correction is implemented by updating the model parameters based on the difference between the actual measured value and the predicted value. The update formula is: New Parameter = Old Parameter + Learning Rate × Difference × Model Uncertainty. Model uncertainty update: New Uncertainty = Old Uncertainty × (1 - Learning Rate) + |Difference| × Learning Rate. Long-term trend term: New long-term trend = Old long-term trend + Learning Rate × Difference; Seasonality term: New seasonality = Old seasonality + Learning Rate × Difference × Period Weight; Model uncertainty: New uncertainty = Old uncertainty × (1 - Learning Rate) + |Difference| × Learning Rate. Greater uncertainty indicates a less confident model in its predictions; lower uncertainty indicates greater confidence. When first used, before historical data has been accumulated, the model's uncertainty is set to a relatively large initial value. After initial training, the model's uncertainty is adjusted in real-time through recursive updates whenever new detection data arrives.

[0062] In this embodiment of the invention, after the model outputs a predicted value of the metabolite production at the current moment and compares this predicted value with the actual value obtained from waste liquid detection, the model itself is further updated using the deviation between the two. First, the deviation between the actual value and the predicted value is calculated. Then, this deviation is multiplied by a preset learning rate to obtain parameter adjustment information. Finally, the long-term trend term and seasonality term are corrected based on this parameter adjustment information. The learning rate is a preset constant between 0.01 and 0.1, which controls the model's response speed to new information. A larger learning rate makes the model more sensitive to recent data changes and adapts faster, while a smaller learning rate makes the model more stable and less sensitive to random fluctuations.

[0063] When the detected value is close to the model's prediction, the deviation is small, and the model makes almost no adjustments, maintaining its original pattern recognition. However, when the detected value consistently deviates from the prediction, the deviation accumulates, and the learning rate gradually corrects the long-term trend and seasonality terms. This allows the model to slowly track real-world physiological changes in individuals (such as metabolic slowdown due to aging and changes in lifestyle habits), while avoiding overreaction to single, accidental fluctuations. This online update mechanism enables the baseline model to continuously optimize itself based on actual household metabolic data without requiring offline retraining, achieving long-term adaptability.

[0064] Step 205: Generate a health analysis report based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output.

[0065] In this embodiment of the invention, the actual value of metabolite output (i.e., the measured metabolite output per unit of clothing) is compared with the predicted value (i.e., the expected value at the current moment calculated by the model based on historical patterns), the difference between the two is calculated, and a health analysis report is generated based on this difference. Fuzzy inference is triggered based on the difference between the predicted and actual values ​​of metabolite output. After triggering inference, the deviation is fuzzified and mapped to fuzzy concepts such as "low," "normal," "high," and "persistently high." These fuzzy concepts are then input into a preset fuzzy rule base for inference. The fuzzy rule base contains multiple rules built based on medical common sense (e.g., "If cortisol is persistently high and protein levels are simultaneously high, it indicates excessive stress load"). Based on the triggered rules and their confidence levels, a natural language template is invoked to generate the final health analysis report.

[0066] Based on the triggered rules and their confidence levels, a structured health report is generated using a natural language template. The report includes a core summary (a one-sentence summary of the main findings), trend charts (displaying historical data and dynamic baseline ranges), detailed interpretations (the specific meaning of each indicator), and improvement suggestions (reference suggestions based on a knowledge base). All suggestions are for reference only and do not generate any device control commands.

[0067] In some embodiments, step 205 may include the following sub-steps: Sub-step S61: Obtain the uncertainty value of the current family health baseline model, and determine the prediction range of metabolite output based on the predicted value of the metabolite at the current moment and the uncertainty value; the uncertainty characterizes the accuracy of the family health baseline model in determining the predicted value of the metabolite output. Sub-step S62: If the actual value of the metabolite output exceeds the predicted range, a health analysis report is generated based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output.

[0068] In this embodiment of the invention, the family health baseline model outputs not only the predicted value but also an estimated uncertainty value characterizing the accuracy of the predicted value, and a dynamic prediction interval determined based on this uncertainty value. The uncertainty value of the current family health baseline model is obtained, and then the interval width (typically ±1.96 times the uncertainty value) is determined using the predicted value as the center, thus defining the prediction interval for metabolite production at the current moment. This uncertainty value is obtained through dynamic estimation of the root mean square of historical prediction errors, reflecting the model's grasp of its own prediction results: when the model has a long history of accurate predictions and small historical prediction errors, the uncertainty is low, the prediction interval is narrow, and the judgment criteria are more refined; when the model's predictions are not stable enough and historical prediction errors are large, the uncertainty is high, the prediction interval is widened, and the judgment criteria are more lenient, to accommodate normal physiological fluctuations.

[0069] The actual value of metabolite production is compared with the predicted range. If the actual value exceeds the predicted range, it indicates that the test result significantly deviates from the model's expected normal fluctuation range, thus triggering the generation of a health analysis report. This comparison is not a simple fixed threshold judgment, but a personalized prediction range dynamically calculated by the model based on the family's historical data. It considers the family's long-term trends, periodic patterns, and the model's own uncertainties, thus more accurately identifying truly noteworthy abnormal deviations. By replacing fixed thresholds with dynamic prediction ranges, the criteria for abnormal judgment vary from person to person and from time to time. Different families have different prediction ranges (personalized), and the same family has different prediction ranges at different times (dynamic). Furthermore, the range width adaptively adjusts with the model's prediction accuracy (the more accurate the prediction, the narrower the range; the less accurate, the wider the range), effectively overcoming the shortcomings of existing fixed threshold methods that cannot adapt to individual differences and time-varying characteristics. The introduction of prediction ranges provides a clear judgment boundary for the generation of health analysis reports. Report generation is only triggered when the test value truly exceeds the expected normal fluctuation range for the family's current state, significantly reducing the false alarm rate and improving the accuracy and reliability of health alerts.

[0070] In some embodiments, step 205 may further include the following sub-steps: Sub-step S71: Based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output and a preset threshold range, determine the fuzzy concept to which the actual value of the metabolite output belongs; the fuzzy concept includes at least one of low, normal, high, and persistently high. In some embodiments, step S71 may include the following sub-steps: Sub-step S711: Determine the threshold range to which the difference belongs; different threshold ranges correspond to different fuzzy concepts; Sub-step S712: Determine the fuzzy concept to which the actual value of metabolite production belongs based on the threshold range to which the difference belongs.

[0071] In this embodiment of the invention, after determining that the actual value of metabolite production deviates from the predicted value, a report is not simply generated directly based on the degree of deviation. Instead, the numerical deviation is transformed into a health alert with semantic expression through fuzzification and fuzzy rule reasoning. Based on the difference between the actual and predicted values, a preset membership function is invoked for fuzzification to determine the fuzzy concept to which the difference belongs (including at least one of "low," "normal," "high," and "persistently high"), and the membership value corresponding to each fuzzy concept is obtained. This fuzzification process allows the same difference to belong to multiple fuzzy concepts at different degrees simultaneously (e.g., a difference of 1.2 may simultaneously belong to "normal" with a membership degree of 0.3 and "high" with a membership degree of 0.7), thereby preserving information about boundary cases and avoiding information loss caused by precise set partitioning.

[0072] The system acquires a pre-defined fuzzy rule base and performs inference based on the fuzzy concepts obtained through fuzzification. The fuzzy rule base contains multiple health assessment rules built upon medical common sense (e.g., "If cortisol is persistently high and protein levels are synchronously high, it indicates excessive stress"). Each rule corresponds to a health analysis conclusion and its confidence level. Using the fuzzy concepts of each metabolite as input, the system calculates the trigger strength of each rule and, combined with the rule's confidence level (trained from historical health data), outputs the health analysis conclusion and its credibility corresponding to the trigger rule. Based on this health analysis conclusion and its confidence level, a structured health analysis report is generated using a natural language template and presented to the user through the user terminal. By introducing fuzzy logic, fuzzy rule inference based on multiple indicators allows health alerts to be made not only based on the numerical value of a single indicator but also by comprehensively considering the correlated changes of multiple indicators (such as the synchronous changes in cortisol and protein levels), thus improving the accuracy and reliability of health assessments.

[0073] Sub-step S72: Obtain a preset fuzzy rule base, and perform reasoning based on the fuzzy rule base and the fuzzy concept to obtain the health analysis conclusion and the confidence level corresponding to the health analysis conclusion; Sub-step S73: Generate the health analysis report based on the health analysis conclusion and the confidence level corresponding to the health analysis conclusion.

[0074] In this embodiment of the invention, the difference between the actual value and the predicted value is input into a preset membership function to determine the distribution of the difference within the threshold range corresponding to each fuzzy concept. It should be noted that the threshold range is not a fixed boundary in precise set partitioning, but rather refers to the numerical intervals corresponding to different fuzzy concepts in the membership function. These intervals overlap, allowing the same difference to fall within the threshold ranges of multiple fuzzy concepts simultaneously. Based on the distribution of the difference within each threshold range, the membership value of the difference to each fuzzy concept is calculated, thus obtaining the fuzzy concept to which the actual metabolite yield belongs and its corresponding membership value.

[0075] To illustrate with a specific example: Suppose the difference is +1.2. The threshold range for the concept of "normal" is [-0.5, +0.5] (complete membership) and [-1.5, -0.5)∪(+0.5, +1.5] (partial membership). The threshold range for the concept of "slightly high" is [+0.5, +1.5] (partial membership) and [+1.5, +3.0] (complete membership). The difference of +1.2 falls into both the partial membership intervals of "normal" and "slightly high." Based on this, the membership degree of "normal" is calculated to be 0.3, and the membership degree of "slightly high" is calculated to be 0.7. The fuzzy result "normal (membership degree 0.3), slightly high (membership degree 0.7)" is output. The specific membership degree value is calculated using a trapezoidal membership function based on the precise position of the difference within the transition interval.

[0076] By transforming precise difference values ​​into fuzzy results that "simultaneously belong to multiple fuzzy concepts to varying degrees," the gradual change information at the numerical boundaries is preserved. This avoids information loss and misjudgment that may occur at the boundaries of precise set division (such as directly judging the difference as "too high" or "normal"), thus obtaining more accurate and realistic health assessment conclusions.

[0077] In some embodiments, the fuzzy rule base includes multiple health assessment rules, each rule consisting of preset conditions and corresponding health analysis conclusions; the preset conditions include at least one of the following: persistent abnormality in the production of a single metabolite, combined abnormality in the production of multiple metabolites, abnormal trend in the production of metabolites, or combined abnormality in the production of metabolites and external environmental parameters.

[0078] Reasoning is performed using a fuzzy rule base built upon medical common sense and experimental data. The rules take the form of "if certain conditions are met, then there is a certain degree of confidence in reaching a certain conclusion." Example rules include: Rule R1 (Stress Assessment): If cortisol levels remain elevated for more than two weeks, and protein levels are also elevated, there is a high probability of "excessive stress load; attention to rest quality is recommended." Rule R2 (Dry Skin): If protein levels (reflecting dandruff) remain elevated, while lipid levels (reflecting sebum) remain low, and the external humidity is below 40%, there is a moderate probability of "increasing skin dryness / dandruff; attention to indoor humidity is recommended." Rule R3 (Fungal Risk): If fungal levels remain elevated, and three consecutive tests exceed twice the standard deviation, there is a high probability of "environmental fungal growth risk; checking the washing machine drum and bathroom ventilation is recommended." Rule R4 (Metabolic Recovery): If previously elevated cortisol levels begin to decrease, and two consecutive tests show a downward trend, then "stress levels are decreasing; condition is good."

[0079] The fuzzy rule base contains multiple health assessment rules. Each rule consists of preset conditions and corresponding health analysis conclusions, which are used to map the fuzzy membership values ​​of each indicator into health prompts with semantic expression. The preset conditions include at least one of the following four types: 1) Persistent abnormality in the production of a single metabolite, i.e., the membership degree of a certain metabolic indicator (such as cortisol) being "persistently high" or "persistently low" exceeds a preset threshold, triggering a health alert related to that indicator (e.g., "persistently high fungal indicators → risk of environmental fungal growth"); 2) Joint abnormality in the production of multiple metabolites, i.e., two or more metabolic indicators simultaneously exhibiting abnormal states, triggering a health alert corresponding to the multi-indicator combination pattern (e.g., "persistently high cortisol and synchronously high protein → excessive stress load"); 3) Abnormal trend in the change of metabolite production, i.e., the direction of change of metabolic indicators over time deviates from the expected trend, triggering a health alert corresponding to the trend change (e.g., "previously high cortisol has decreased repeatedly → stress indicators show a downward trend"); 4) Joint abnormality in the production of metabolites and external environmental parameters, i.e., when the abnormal state of metabolic indicators and external environmental parameters (such as humidity) jointly meet the preset conditions, triggering the corresponding health alert (e.g., "persistently high protein and persistently low lipids and environmental humidity below 40% → trend of dry skin / increased dandruff").

[0080] By constructing a fuzzy rule base covering four types of abnormalities—persistent single-indicator anomalies, combined anomalies of multiple indicators, abnormal trend changes, and combined anomalies of indicators and the environment—health assessment is no longer limited to isolated judgments of single indicators. Instead, it can comprehensively consider the synergistic changes of multiple metabolic indicators, their temporal evolution trends, and external environmental factors for joint reasoning, effectively improving the accuracy of health alerts. Furthermore, the rule base design facilitates the addition, deletion, and adjustment of rules based on updates to medical knowledge, exhibiting good scalability and maintainability.

[0081] Step 206: When it is identified that the washing machine is performing other washing stages, the bypass of the drainage structure is controlled to remain closed.

[0082] In this embodiment of the invention, when the washing machine is detected to be performing other washing stages besides the main wash drainage stage (such as rinsing stage, spin-drying stage, etc.), the bypass of the drainage structure is kept closed, and waste liquid sampling is not performed. During the drainage process of the subsequent rinsing stage, the main control board does not send a conduction control signal to the three-way valve. The three-way valve always keeps the main drain pipe open and the bypass closed. All waste liquid in this stage is discharged directly into the sewer through the main drain pipe and does not enter the storage compartment.

[0083] The wastewater from the rinsing stage mainly contains residual detergent, while human metabolites (such as sebum and cortisol) are largely washed away and discharged with the main wash wastewater during the main wash stage. Therefore, the biological information value of the rinsing wastewater is extremely low. By sampling only during the main wash drainage stage and shutting off the bypass in other stages, the collected wastewater samples are ensured to have the highest concentration of metabolites and health analysis value. This avoids wasting detection resources with invalid samples, reduces interference with the original drainage process of the washing machine, lowers system power consumption and pipeline burden, and improves the targeting of the sampling strategy and the efficiency of system operation.

[0084] Step 207: During the detection of the waste liquid by the waste liquid detection module, the ultrasonic oscillating plate is controlled to oscillate. Step 208: After the waste liquid is detected by the waste liquid detection module, the drain valve connected to the storage tank is controlled to discharge the waste liquid, and the storage tank is rinsed with clean water.

[0085] The storage chamber has a two-layer structure: the first layer is a micro-stainless steel filter (100-micron pore size) to filter out fiber debris and large particulate impurities that may clog the precision sensor; the second layer is a micro-ultrasonic oscillator at the bottom, which intermittently oscillates at a low frequency during sample static analysis to prevent biological particle sedimentation and ensure sample homogeneity. After analysis, the micro-drain valve at the bottom of the storage chamber opens to drain the waste liquid back into the drainage pipeline, and the three-way valve resets. The entire sampling process has no impact on the main drainage process.

[0086] There is indeed a risk of waste liquid residue and biofilm growth after the storage tank is emptied. Therefore, in addition to the drain valve, a simple flushing process is designed. After emptying, a small amount of clean water is automatically injected to flush the storage tank to ensure that the residue is removed. The flushing water is then discharged into the drain pipe to complete the cleaning. The inner wall of the storage tank is coated with an antibacterial coating (such as a silver ion coating or a photocatalytic coating), and the storage tank material is selected to be less prone to biofilm growth (such as specially treated stainless steel or food-grade plastic).

[0087] In this embodiment of the invention, during the detection of waste liquid in the storage chamber by the waste liquid detection module, a miniature ultrasonic oscillator located at the bottom of the storage chamber's buffer pool is controlled to generate low-frequency intermittent oscillations. During the sample static analysis period (typically within a few-minute window during spectral scanning, microfluidic reactions, and basic property measurements by the detection module), the oscillator oscillates at a preset frequency and interval, keeping biological particles (such as exfoliated cells, bacteria, sebum particles, etc.) in the waste liquid sample in a suspended state, preventing uneven sample concentration due to gravitational sedimentation.

[0088] After the waste liquid is fully tested by the waste liquid detection module, the miniature drain valve connected to the bottom of the storage tank opens, draining the tested waste liquid back into the drainage pipe. After draining, the automatic rinsing process injects clean water, using the water flow to flush the inner wall of the buffer tank, the filter surface, and the pipe interfaces. The rinsing water is then discharged through the drain valve, completing the entire cleaning process. This rinsing process aims to remove any residual waste liquid, particulate matter, and organic matter in the buffer tank, preventing the formation of biofilms or crystals after drying, thus providing a clean testing environment for the next sampling. The low-frequency oscillation of the ultrasonic vibrator during testing effectively ensures the uniformity of the waste liquid sample, avoiding measurement errors caused by particle sedimentation. The automatic drain valve and clean water rinsing mechanism enable rapid self-cleaning of the storage tank, effectively reducing the risk of sample cross-contamination and pipe blockage.

[0089] Reference Figure 3 The diagram illustrates a flowchart of a home health monitoring method based on washing machine wastewater provided by an embodiment of the present invention. Figure 3 This demonstrates the complete process of an embodiment of the present invention. During the drainage stage following the main wash cycle of the washing machine, the discharged wastewater is introduced into a storage chamber via a drainage structure. The wastewater is then analyzed using a multimodal water quality analysis module to obtain raw multimodal data. This module includes a micro-spectral analysis unit, a microfluidic biological detection unit, and a basic physical property detection unit. The operating parameters of the washing machine during the main wash cycle are acquired, and based on these parameters and the raw multimodal data, the actual value of metabolite production is obtained. A pre-constructed family health baseline model is acquired, and a predicted value of metabolite production at the current moment is output based on this model. The fuzzy concept to which the actual metabolite production belongs is determined based on the difference between the predicted and actual values. Reasoning is performed based on a preset fuzzy rule base and the fuzzy concept. Finally, a health analysis report is generated.

[0090] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0091] Reference Figure 4 The diagram shows a structural schematic of the storage compartment of a washing machine according to an embodiment of the present invention. The washing machine includes a storage compartment 30 disposed in the drainage structure and a waste liquid detection module. Specifically, the washing machine includes: The controller is used to introduce the discharged waste liquid into the storage chamber 30 through the drainage structure during the drainage stage after the main wash stage of the washing machine; detect the waste liquid through the waste liquid detection module to obtain raw multimodal data; acquire the operating parameters of the washing machine during the main wash stage, and obtain the actual value of metabolite production based on the operating parameters and the raw multimodal data; acquire a pre-constructed family health baseline model, and output the predicted value of metabolite production at the current moment based on the family health baseline model; the family health baseline model is trained using historical data of metabolite production; and generate a health analysis report based on the difference between the predicted value and the actual value of metabolite production.

[0092] In some embodiments, the operating parameters include the weight of the clothes and the amount of water used for washing; The controller is configured to determine a dynamic dilution factor based on the standard water volume, the washing water volume, the standard garment weight, and the garment weight; correct the original multimodal data based on the dynamic dilution factor to obtain corrected multimodal data; and obtain the actual value of the metabolite production based on the corrected multimodal data, the washing water volume, and the garment weight.

[0093] In some embodiments, the waste liquid detection module includes a micro-spectral analysis unit 301; the storage compartment 30 has opposing transparent windows on both sides; the micro-spectral analysis unit 301 includes a light source and an optical fiber probe located at the transparent windows on both sides of the storage compartment 30. The controller is used to acquire the absorbance spectral data of the waste liquid in a preset wavelength range through the micro spectral analysis unit 301, and extract the absorbance of metabolites from the absorbance spectral data; determine the concentration value of the metabolites according to the preset spectral standard curve and the absorbance of the metabolites; the spectral standard curve is a curve describing the quantitative relationship between the concentration value and absorbance of each metabolite, which is obtained by fitting the concentration value of each metabolite on the x-axis and the absorbance value of each metabolite on the y-axis.

[0094] In some embodiments, the waste liquid detection module further includes a microfluidic biodetection unit 302; the microfluidic biodetection unit 302 is located at the bottom of the storage chamber 30, the microfluidic biodetection unit 302 includes a microchannel network and a light-emitting device and a light-receiving device respectively disposed at both ends of each channel of the microchannel network, each channel of the microchannel network is pre-filled with dry biochemical reagents for different metabolites, and different metabolites in the waste liquid undergo different colorimetric reactions with the dry biochemical reagents; The controller emits a light beam through the light-emitting device, which passes through the waste liquid undergoing the colorimetric reaction. The light-receiving device detects the transmitted light intensity of the metabolites and converts the detected transmitted light intensity into absorbance values. Based on a pre-calibrated standard curve and the absorbance values, the concentration values ​​of the metabolites are determined. The standard curve is a pre-established curve that fits the absorbance values ​​to describe the quantitative relationship between the absorbance values ​​and the concentration values ​​of each metabolite, with the concentration value of each metabolite on the x-axis and the measured absorbance value on the y-axis.

[0095] In some embodiments, the waste liquid detection module further includes a basic property detection unit 303; the basic property detection unit 303 is disposed inside the storage chamber 30; the basic property detection unit 303 includes a pH electrode, a conductivity electrode and a temperature sensor; The controller is used to detect the pH value, conductivity, and temperature of the waste liquid via the pH electrode, the conductivity electrode, and the temperature sensor.

[0096] In some embodiments, the family health baseline model includes a long-term trend term and a seasonal term; the long-term trend term characterizes the trend of metabolite changes over a longer time scale; the seasonal term characterizes the regular fluctuations of metabolites that recur repeatedly with a fixed period. The controller is configured to determine a first predicted value of metabolite output at the current moment based on the long-term trend term of the family health baseline model, and a second predicted value of metabolite output at the current moment based on the seasonality term of the family health baseline model; and to determine a target predicted value of metabolite output at the current moment based on the first predicted value and the second predicted value.

[0097] In some embodiments, the controller is configured to obtain parameter adjustment information based on the deviation between the predicted value of the metabolite output and the actual value of the metabolite output and a preset learning rate; and to correct the long-term trend term and the seasonality term based on the parameter adjustment information.

[0098] In some embodiments, the controller is configured to acquire the uncertainty value of the current family health baseline model, and determine the prediction range of metabolite output based on the predicted value of the metabolite at the current moment and the uncertainty value; the uncertainty characterizes the accuracy of the prediction value of the metabolite output determined by the family health baseline model; and if the actual value of the metabolite output exceeds the prediction range, generate a health analysis report based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output.

[0099] In some embodiments, the controller is configured to determine the fuzzy concept to which the actual value of metabolite production belongs based on the difference between the predicted value of metabolite production and the actual value of metabolite production and a preset threshold range; the fuzzy concept includes at least one of low, normal, high, and persistently high; acquire a preset fuzzy rule base, and perform reasoning based on the fuzzy rule base and the fuzzy concept to obtain a health analysis conclusion and the confidence level corresponding to the health analysis conclusion; and generate the health analysis report based on the health analysis conclusion and the confidence level corresponding to the health analysis conclusion.

[0100] In some embodiments, the controller is configured to determine the threshold range to which the difference belongs; different threshold ranges correspond to different fuzzy concepts; and determine the fuzzy concept to which the actual value of metabolite production belongs based on the threshold range to which the difference belongs.

[0101] In some embodiments, the fuzzy rule base includes multiple health assessment rules, each rule consisting of preset conditions and corresponding health analysis conclusions; the preset conditions include at least one of the following: persistent abnormality in the production of a single metabolite, combined abnormality in the production of multiple metabolites, abnormal trend in the production of metabolites, or combined abnormality in the production of metabolites and external environmental parameters.

[0102] In some embodiments, the controller is configured to, when recognizing a drainage phase following the main wash phase of the washing machine, control the drainage structure to introduce the waste liquid into the storage compartment 30 via a bypass; and to control the bypass of the drainage structure to remain closed when recognizing other washing phases of the washing machine.

[0103] In some embodiments, the storage chamber 30 includes a filter 304 and an ultrasonic oscillating plate 305. The controller is configured to control the ultrasonic oscillating plate 305 to oscillate during the detection of the waste liquid by the waste liquid detection module; and after the waste liquid is detected by the waste liquid detection module, to control the drain valve connected to the storage chamber 30 to discharge the waste liquid and to rinse the storage chamber 30 with clean water.

[0104] As the apparatus embodiment is basically similar to the method embodiment, it is described in a relatively simple manner. For relevant details, please refer to the description of the method embodiment.

[0105] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0106] Furthermore, it should be noted that the scope of the methods and apparatus in the embodiments of the present invention is not limited to performing functions in the order shown or discussed. It may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0107] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0108] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A method for home health monitoring based on washing machine wastewater, characterized in that, The washing machine includes a storage compartment and a waste liquid detection module disposed in the drainage structure, and the method includes: During the drainage phase following the main wash cycle of the washing machine, the discharged waste liquid is introduced into the storage compartment through the drainage structure. The absorbance spectral data of the waste liquid within a preset wavelength range is obtained, and the absorbance of the metabolites is extracted from the absorbance spectral data. Based on the preset spectral standard curve and the absorbance of the metabolites, the concentration value of the first metabolite is determined. A light beam is emitted by a light-emitting device and passes through the waste liquid where a colorimetric reaction occurs. The intensity of the transmitted light of the metabolite is detected, and the detected intensity of the transmitted light is converted into an absorbance value. The concentration of the second metabolite is determined according to a pre-calibrated standard curve and the absorbance value. The pH value, conductivity, and temperature of the waste liquid were measured. Obtain the operating parameters of the washing machine during the main wash phase; the operating parameters include the weight of the clothes and the amount of water used for washing; A dynamic dilution factor is determined based on the standard water volume, the washing water volume, the standard garment weight, and the garment weight. The concentration values ​​of the first metabolite, the second metabolite, and the pH, conductivity, and temperature of the waste liquid are then corrected based on the dynamic dilution factor. The dynamic dilution factor is equal to the ratio of the standard water volume to the washing water volume, multiplied by the ratio of the standard garment weight to the garment weight. The actual value of metabolite production is obtained based on the corrected concentration values ​​of the first metabolite, the second metabolite, the pH value, conductivity, and temperature of the waste liquid, the washing water volume, and the weight of the clothing. A pre-constructed family health baseline model is obtained, and based on the family health baseline model, a predicted value of metabolite production at the current time is output; the family health baseline model is trained using historical data of metabolite production. A health analysis report is generated based on the difference between the predicted and actual values ​​of the metabolite production.

2. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The waste liquid detection module includes a micro-spectral analysis unit, through which the absorbance spectral data is acquired; the storage chamber has opposing transparent windows on both sides; the micro-spectral analysis unit includes a light source and an optical fiber probe located at the transparent windows on both sides of the storage chamber; the spectral standard curve is a curve describing the quantitative relationship between the concentration value and absorbance of each metabolite, obtained by fitting the concentration value of each metabolite on the x-axis and the absorbance value of each metabolite on the y-axis.

3. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The waste liquid detection module further includes a microfluidic biodetection unit; the microfluidic biodetection unit is located at the bottom of the storage chamber, and the microfluidic biodetection unit includes a microchannel network and light-emitting devices and light-receiving devices respectively disposed at both ends of each channel of the microchannel network. The transmitted light intensity is detected by the light-receiving device. Each channel of the microchannel network is pre-filled with dry biochemical reagents for different metabolites. Different metabolites in the waste liquid undergo different colorimetric reactions with the dry biochemical reagents. The standard curve is a pre-established curve describing the quantitative relationship between absorbance values ​​and concentration values ​​of each metabolite, with the concentration value of each metabolite on the x-axis and the measured absorbance value on the y-axis.

4. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The waste liquid detection module also includes a basic physical property detection unit; the basic physical property detection unit is located inside the storage chamber; the basic physical property detection unit includes a pH electrode, a conductivity electrode and a temperature sensor, and the pH value, conductivity and temperature of the waste liquid are detected by the pH electrode, the conductivity electrode and the temperature sensor.

5. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The family health baseline model includes a long-term trend term and a seasonal term; the long-term trend term characterizes the changing trend of metabolites over a longer time scale. The seasonality term characterizes the regular fluctuations in the recurrence of metabolites over a fixed period; The method of outputting a predicted value of metabolite production at the current moment based on the family health baseline model includes: Based on the long-term trend term of the family health baseline model, a first predicted value of metabolite output at the current moment is determined, and based on the seasonality term of the family health baseline model, a second predicted value of metabolite output at the current moment is determined. Based on the first predicted value and the second predicted value, determine the target predicted value of metabolite production at the current time.

6. The method for home health monitoring based on washing machine wastewater according to claim 5, characterized in that, The method further includes: Based on the deviation between the predicted value and the actual value of the metabolite output and the preset learning rate, parameter adjustment information is obtained; The long-term trend item and the seasonal item are corrected based on the parameter adjustment information.

7. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The process of generating a health analysis report based on the difference between the predicted and actual values ​​of metabolite production includes: Obtain the uncertainty value of the current family health baseline model, and determine the prediction range of metabolite output based on the predicted value of the metabolite at the current moment and the uncertainty value; the uncertainty characterizes the accuracy of the family health baseline model in determining the predicted value of the metabolite output. If the actual value of the metabolite output exceeds the predicted range, a health analysis report is generated based on the difference between the predicted value and the actual value of the metabolite output.

8. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The process of generating a health analysis report based on the difference between the predicted and actual values ​​of metabolite production includes: Based on the difference between the predicted value of the metabolite output and the actual value of the metabolite output and a preset threshold range, the fuzzy concept to which the actual value of the metabolite output belongs is determined; the fuzzy concept includes at least one of low, normal, high, and persistently high. Obtain a preset fuzzy rule base, and perform reasoning based on the fuzzy rule base and the fuzzy concepts to obtain a health analysis conclusion and the confidence level corresponding to the health analysis conclusion; The health analysis report is generated based on the health analysis conclusions and the corresponding confidence levels.

9. The method for home health monitoring based on washing machine wastewater according to claim 8, characterized in that, The step of determining the fuzzy concept to which the actual value of metabolite production belongs based on the difference between the predicted value and the actual value of metabolite production and a preset threshold range includes: Determine the threshold range to which the difference belongs; different threshold ranges correspond to different fuzzy concepts; The fuzzy concept to which the actual value of metabolite production belongs is determined based on the threshold range to which the difference belongs.

10. The method for home health monitoring based on washing machine wastewater according to claim 8, characterized in that, The fuzzy rule base contains multiple health assessment rules, each consisting of preset conditions and corresponding health analysis conclusions; the preset conditions include at least one of the following: persistent abnormality in the production of a single metabolite, combined abnormality in the production of multiple metabolites, abnormal trend in the production of metabolites, or combined abnormality in the production of metabolites and external environmental parameters.

11. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The method further includes: When the main wash cycle of the washing machine is detected, the drainage structure is controlled to bypass the waste liquid and introduce it into the storage compartment. When the washing machine is identified as performing other washing stages, the bypass of the drainage structure is controlled to remain closed.

12. The method for home health monitoring based on washing machine wastewater according to claim 1, characterized in that, The storage compartment includes a filter and an ultrasonic vibrating plate, and the method further includes: During the detection of the waste liquid by the waste liquid detection module, the ultrasonic oscillating plate is controlled to oscillate. After the waste liquid is detected by the waste liquid detection module, the drain valve connected to the storage tank is controlled to discharge the waste liquid, and the storage tank is rinsed with clean water.

13. A washing machine, characterized in that, The method for home health monitoring based on washing machine wastewater as described in any one of claims 1-12, wherein the washing machine includes a storage compartment and a wastewater detection module disposed in the drainage structure, and the washing machine includes: The controller is used to introduce the discharged waste liquid into the storage chamber through the drainage structure during the drainage phase after the main wash stage of the washing machine; detect the waste liquid through the waste liquid detection module to obtain raw multimodal data; acquire the operating parameters of the washing machine during the main wash stage, including the weight of the clothes and the washing water volume; determine a dynamic dilution factor based on the standard water volume, the washing water volume, the standard weight of the clothes, and the weight of the clothes, and correct the raw multimodal data based on the dynamic dilution factor; obtain the actual value of the metabolite output based on the corrected multimodal data, the washing water volume, and the weight of the clothes; acquire a pre-constructed family health baseline model, and output the predicted value of the metabolite output at the current moment based on the family health baseline model; the family health baseline model is trained using historical data of metabolite output; and generate a health analysis report based on the difference between the predicted value and the actual value of the metabolite output.

Images