A toilet cleaning precision calling system and method based on multi-sensor fusion

Abstract

The present application relates to the field of calling technology, and discloses a toilet cleaning precise calling system and method based on multi-sensor fusion, which comprises an analysis unit, an attenuation unit, a diffusion unit, a global analysis unit, a pollution unit and a cleaning calling unit. The technical solution breaks through the limitation of traditional odor sensor relying on odor judgment, calculates the global attenuation coefficient by combining the behavior mode of the toilet event and the pollution load value, dynamically analyzes the diffusion law of pollutants, can avoid the interference of disinfectant, dynamically adjusts the stable confirmation time length and the state recovery time length, prevents the pollution from triggering the calling system temporarily, avoids the calling system not being withdrawn in time after the pollution is relieved, solves the calling failure problem caused by disinfectant covering up the odor in the hospital toilet, and improves the timeliness and accuracy of the cleaning calling response.

Description

Technical Field

[0001] This invention relates to the field of call technology, specifically to a precise call system and method for toilet cleaning based on multi-sensor fusion. Background Technology

[0002] In existing toilets, the call for cleaning is usually made through sensors. For example, sensors can detect the odor of feces or urine, triggering a call to have a cleaner come and clean.

[0003] However, the above-mentioned calling method still has the following drawbacks in hospital restrooms: Hospital restrooms often need to be disinfected frequently, requiring the frequent use of strong oxidizing disinfectants such as chlorine and peroxides. After these disinfectants evaporate, they themselves are strong interfering gases, which can mask the odor of feces or urine, making it impossible for the sensor to distinguish between the odor of the disinfectant in the environment and the actual odor. As a result, the calling function cannot be triggered in time when the restroom is dirty, resulting in poor calling effect. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a toilet cleaning precision call system and method based on multi-sensor fusion, which solves the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] A toilet cleaning precision call system based on multi-sensor fusion includes:

[0007] The analysis unit is used to acquire real-time time-series data of the status changes of multiple sensors in each toilet stall within the target call location. Based on the time-series data of status changes, it identifies the complete toilet use event sequence, analyzes the content of the toilet use event, and obtains the pollution load value of the toilet use event. The target call location is the toilet in a hospital.

[0008] The attenuation unit is used to discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data, and obtain the global attenuation coefficient.

[0009] The diffusion unit is used to inject the corresponding pollution load value into the grid at the location of the new toilet event when a new toilet event occurs. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field.

[0010] The global analysis unit is used to identify pollution focal points in a dynamic pollution situation field and to spatially integrate the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity.

[0011] The pollution unit is used to analyze the number, total energy intensity, and spatial concentration of all current pollution focal points to generate a cleaning urgency level.

[0012] The cleaning call unit is used to integrate the overall situation intensity with the urgency of cleaning to generate cleaning call instructions.

[0013] Furthermore, based on the time-series data of state changes, complete toilet-use event sequences are identified, and the content of each toilet-use event is analyzed to obtain the pollution load value of that toilet-use event, including:

[0014] Analyze the time sequence of state changes of the door magnetic sensor and the flushing sensor, and map them into a multi-dimensional behavioral mode vector;

[0015] Based on environmental data, dynamic environmental stress factors are calculated;

[0016] By analyzing the patterns, intensity, and behavioral mode vectors of flushing events, the airborne migration potential is obtained.

[0017] The pollution load value is generated by fusing behavioral modal vectors, environmental stress factors, and airborne migration potential.

[0018] Furthermore, by analyzing ventilation and environmental data, the global attenuation coefficient is obtained, including:

[0019] The three-dimensional spatial layout of the target call area is obtained, and the ventilation data is combined with the three-dimensional spatial layout to calculate the proportion of low-speed airflow area and the wind speed gradient of mainstream area, thereby generating the effectiveness of airflow organization.

[0020] Based on environmental data, we analyze the spatial distribution differences and temporal fluctuation characteristics to generate a microclimate eddy coefficient that represents the intensity of natural convection and turbulence caused by uneven temperature and humidity.

[0021] By combining the effectiveness of airflow organization with the microclimate eddy coefficient in a counter-cyclical manner, pollutant retention potential is generated.

[0022] By obtaining the frequency and intensity of toilet use events and combining them with pollutant retention potential, the environment's ability to resist pollution and restore its cleanliness can be dynamically assessed, thus obtaining the environment's self-purification potential.

[0023] Furthermore, analysis of ventilation and environmental data yields the global attenuation coefficient, which also includes:

[0024] By integrating the airflow organization effectiveness with the microclimate eddy coefficient, a benchmark attenuation constant is generated. The environmental self-purification potential is then combined with the benchmark attenuation constant to calculate the instantaneous eddy attenuation rate.

[0025] The instantaneous eddy current attenuation rate is spatiotemporally smoothed and calibrated according to different functional areas within the target call site to obtain the global attenuation coefficient.

[0026] Furthermore, when a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it undergoes exponential decay in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field, including:

[0027] When a new toilet incident occurs, an initial pollution plume with a dominant direction is generated at the location of the incident, based on its pollution load value and the effectiveness of the current airflow organization.

[0028] Based on the dominant direction of the initial pollution plume and the global attenuation coefficient, the attenuation of pollutants and the migration driving force within each grid cell are calculated to generate a convection-diffusion intensity field.

[0029] Furthermore, when a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it undergoes exponential decay in each calculation cycle based on the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field. This also includes:

[0030] Based on the convection-diffusion intensity field, the flux of pollutants propagating from each grid to its neighboring grids is calculated, generating neighborhood coupling flux;

[0031] The initial pollution plumes of all toilet events are propagated to the pollution load values ​​of each grid through neighborhood coupling flux and then superimposed to generate the instantaneous pollution field strength.

[0032] By integrating the instantaneous pollution field strength across the entire field, a dynamic pollution situation field is obtained.

[0033] Furthermore, spatial integration is performed on the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the overall situation intensity, including:

[0034] Identify the spatial distribution characteristics of pollution load in a dynamic pollution situation field, calculate its gradient field intensity and high load area aggregation degree, and generate a pollution distribution heterogeneity index;

[0035] Based on the global attenuation coefficient, the instantaneous load of each grid point in the dynamic pollution situation field is corrected in a timely manner to generate a spatiotemporal attenuation correction factor.

[0036] Based on the different functional areas within the target call center, calculate the functional area sensitivity weight field for each grid.

[0037] The dynamic pollution situation field, the spatiotemporal attenuation correction factor and the functional area sensitivity weight field are fused to generate the perception weighted pollution field. The perception weighted pollution field is then fused with the pollution distribution heterogeneity index to generate the instantaneous global load integral.

[0038] Based on the ratio of the instantaneous global load integral to the pollution distribution heterogeneity index, the dynamic intensity calibration coefficient is calculated, and the instantaneous global load integral is processed according to the dynamic intensity calibration coefficient to obtain the global situation intensity.

[0039] Furthermore, based on the analysis of the number, total energy intensity, and spatial concentration of all current pollution hotspots, a cleaning urgency level is generated, including:

[0040] Analyze the spatial relationships of all pollution focal points to generate pollution focal point clusters;

[0041] The energy intensity and spatial distribution of each focus are calculated to generate the focus cluster potential field;

[0042] Based on the potential field of the focal cluster, the interaction force generated between any two pollution focal points due to the superposition of potential fields is calculated, and the focal interaction strength is generated.

[0043] Analyze the overall distribution of pollution focus clusters within the target call area, calculate the degree of dispersion or concentration of their distribution, and generate spatial distribution density.

[0044] The cleaning urgency is obtained by combining the intensity of focal interaction, the density of spatial distribution, the number of focal points, and the total energy intensity.

[0045] Furthermore, by integrating the overall situational awareness and the urgency of cleaning, cleaning call instructions are generated, including:

[0046] The dynamic decision-making benchmark value is generated by combining the overall situation intensity, the urgency of cleaning, and the environmental self-purification potential.

[0047] By conducting a pollution synergistic analysis of the overall situation intensity and the urgency of cleaning, a global-local situation coupling factor is generated.

[0048] Based on the rate of change of environmental self-purification potential and global situation intensity, the stable confirmation time required for call triggering is calculated.

[0049] Based on the decay trend of environmental self-purification potential and cleaning urgency, the state recovery time required for call cancellation is calculated.

[0050] The global-local situation coupling factor is compared with the dynamic decision benchmark value to generate a cleaning call instruction.

[0051] Furthermore, a method for precise toilet cleaning call based on multi-sensor fusion, applied to the aforementioned precise toilet cleaning call system based on multi-sensor fusion, includes:

[0052] Step S1: Real-time acquisition of state change time-series data of multiple sensors in each toilet stall within the target call location; based on the state change time-series data, identification of complete toilet use event sequences; analysis of the content of toilet use events; and obtaining the pollution load value of the toilet use event. The target call location is a hospital toilet.

[0053] Step S2: Discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data, and obtain the global attenuation coefficient.

[0054] Step S3: When a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the location of the event. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field.

[0055] Step S4: In the dynamic pollution situation field, identify the pollution focus, and perform spatial integration on the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity.

[0056] Step S5: Analyze the number, total energy intensity, and spatial concentration of all current pollution focal points to generate a cleaning urgency level.

[0057] Step S6: Integrate the overall situation intensity with the urgency of cleaning to obtain a cleaning call instruction.

[0058] In summary, the present invention has the following main beneficial effects:

[0059] The system analyzes the state change time-series data of door magnetic sensors and flushing sensors to identify complete toilet use event sequences. It then analyzes the content of these events to obtain pollution load values, avoiding interference from disinfectant gases on sensor signals. Next, the attenuation unit comprehensively analyzes ventilation and environmental data to generate a global attenuation coefficient that accurately reflects the natural attenuation of pollutants. Simultaneously, the diffusion unit injects the pollution load value into the grid and analyzes its exponential decay and spatial diffusion effects over different calculation cycles, generating a dynamic pollution situation field that dynamically reflects the spatiotemporal distribution characteristics of pollutants. Based on this, the global analysis unit calculates the global situation intensity through spatial integration to understand the overall dirtiness of the toilet. The pollution unit analyzes the number of pollution focal points, total energy intensity, and spatial concentration to generate cleaning urgency, achieving accurate identification of local high-risk areas. Finally, the cleaning call unit integrates the global situation intensity and cleaning urgency to generate the final cleaning call instruction. This solution maintains accurate perception and timely response to toilet hygiene conditions even in complex environments with strong interfering gases such as disinfectant volatilization, improving the accuracy and reliability of cleaning calls and effectively optimizing the allocation efficiency of cleaning resources in hospital toilets. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of a toilet cleaning precision call system based on multi-sensor fusion according to the present invention;

[0061] Figure 2 This is a flowchart of a toilet cleaning precision call method based on multi-sensor fusion according to the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] refer to Figure 1 and Figure 2 A toilet cleaning precision call system based on multi-sensor fusion includes:

[0064] The analysis unit is used to acquire real-time time-series data of the status changes of multiple sensors in each toilet stall within the target call location. Based on the time-series data of status changes, it identifies the complete toilet use event sequence, analyzes the content of the toilet use event, and obtains the pollution load value of the toilet use event. The multiple sensors are: door magnetic sensor and flush sensor. The target call location is a hospital toilet.

[0065] The attenuation unit is used to discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data to obtain the global attenuation coefficient. The ventilation data includes: fan start / stop, speed, wind speed, etc., and the environmental data includes: temperature, humidity, etc.

[0066] The diffusion unit is used to inject the corresponding pollution load value into the grid at the location of the new toilet event when a new toilet event occurs. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field.

[0067] The global analysis unit is used to identify pollution focal points in a dynamic pollution situation field, and to spatially integrate the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity. The global situation intensity is mainly used to quantify the overall dirtiness of the toilet.

[0068] The pollution unit is used to analyze the number, total energy intensity, and spatial concentration of all current pollution focal points to generate a cleaning urgency level.

[0069] The cleaning call unit is used to integrate the overall situation intensity with the urgency of cleaning to generate cleaning call instructions.

[0070] By acquiring time-series data on state changes through door magnetic sensors and flushing sensors, the analysis unit identifies toilet event sequences and calculates pollution load values, avoiding reliance on odor monitoring and mitigating interference from disinfectant gases at the source. The attenuation unit combines ventilation and environmental data to generate a global attenuation coefficient, which the diffusion unit uses to construct a dynamic pollution situation field, thereby understanding the pollution diffusion pattern. Furthermore, the global analysis unit and the pollution unit obtain the global situation intensity that quantifies the overall level of dirtiness and the cleaning urgency that reflects the local urgency through spatial integration and pollution focus analysis, respectively. Then, the cleaning call unit merges these data to generate a cleaning call command. This solution does not rely on odor perception and can accurately identify the true dirtiness of toilets in hospital environments where disinfectants are frequently used, triggering cleaning in a timely manner and reducing call delays caused by disinfectant gases masking odors.

[0071] In one embodiment, based on state change time-series data, a complete toilet-use event sequence is identified, and the content of the toilet-use events is analyzed to obtain the pollution load value of the toilet-use event, including:

[0072] The state change time series of the door magnetic sensor and the flush sensor are analyzed and mapped to a multi-dimensional behavioral mode vector. The state change time series includes: the time stamp sequence of door state (open or closed) switching recorded by the door magnetic sensor, and the time stamp and duration sequence of flush action state (flushing or stopped) switching recorded by the flush sensor. Specifically, the door magnetic sensor is marked as closed (1) and open (0), and the flush sensor is marked as triggered (1) and not triggered (0). The time difference between the door magnetic sensor changing from 1 to 0 is taken as the total door closure time for a single toilet use. The time difference between the door magnetic sensor changing to 1 and the first flush sensor changing to 1 is taken as the delay time from door closure to the first flush. The number of times the flush sensor changes to 1 during the period when the door magnetic sensor is 1 is counted to obtain the total flush volume for a single toilet use. If there are multiple flushes, the time difference between two adjacent flush sensor changes to 1 is calculated, and the average of the time differences is taken as the single flush interval.

[0073] The total time for a single toilet door closure, the delay time, the total flush volume for a single toilet visit, and the flush interval are normalized to the 0-1 range and arranged in the current order to form a four-dimensional behavioral modal vector.

[0074] Based on environmental data, dynamic environmental stress factors are calculated, specifically including: setting baseline values ​​for toilets in different hospital departments, where the baseline temperature for the infectious disease department is 21℃ and the baseline humidity is 45%; and the baseline temperature for the general department is 23℃ and the baseline humidity is 55%; calculating the absolute values ​​of the temperature and humidity differences between the current target call location and the corresponding department's baseline values; calculating the sum of squares of the absolute values ​​of the temperature and humidity differences, setting the weight for the infectious disease department to 1.2 and the weight for the general department to 0.8; multiplying the sum of squares by the corresponding department's weight and dividing by the corresponding department's weight, and then normalizing the result to the 0-1 range, which is the dynamic environmental stress factor;

[0075] Among them, the excrement and secretions of patients in the infectious disease ward toilets have high residual levels in toilet bowls, floors, and air. The survival and spread of pathogens are more sensitive to changes in temperature and humidity. Therefore, the weight is higher at 1.2 to amplify the stress effect of environmental fluctuations in the infectious disease ward toilets on the spread of pollution, so as to facilitate timely cleaning to block the transmission chain. On the other hand, the pathogen load in the pollutants of users in the general ward toilets is low, and the driving effect of temperature and humidity changes on the spread of pollution in the toilet is weak. Therefore, the weight is lower at 0.8 to avoid overreaction, reduce unnecessary toilet cleaning frequency, and balance resource input.

[0076] The airborne migration potential is obtained by analyzing the pattern, intensity, and behavioral mode vector of flushing events. Specifically, this includes: recording flushing when the door is closed as closed flushing, calculating its proportion of the total number of flushing events, and using this proportion as the closed coordination coefficient; taking the reciprocal of the single flushing interval as the pulse intensity; and subtracting the delay time in the behavioral mode vector from 1 to obtain the immediacy index.

[0077] The weights of the closed synergy coefficient, pulse intensity, and immediacy index are set to 0.4, 0.3, and 0.3 respectively. The closed synergy coefficient, pulse intensity, and immediacy index are multiplied by their respective weights and summed. The sum is then multiplied by the environmental stress factor and then by the corresponding department weight. The calculation result is normalized to the 0-1 range, which is the airborne migration potential.

[0078] The closed-loop synergy coefficient is weighted at 0.4 because when the toilet door is closed, the space is sealed, and the odor produced by flushing is difficult to diffuse, which will accumulate locally and directly increase the concentration of airborne pollution and the risk of cross-infection. This is the core premise for determining whether airborne pollution is easy to accumulate. The pulse intensity and immediacy index are each weighted at 0.3 because both are parameters that amplify the risk. The pulse intensity is used to reflect the pulse-like pollution superposition effect of multiple flushes in a short period of time, and the immediacy index is used to reflect the rapid activation effect of immediate flushing on pollutants. However, the risk contribution of both depends on the closed-loop synergy coefficient, so they are given the same weight.

[0079] The pollution load value is generated by fusing behavioral modal vectors, environmental stress factors, and airborne migration potential. Specifically, this involves: assigning weights to the four-dimensional features of the behavioral modal vector: total flush volume per toilet visit (0.3), total door closure time per toilet visit (0.2), delay time (0.3), and flush interval (0.2); multiplying each of the four feature parameters in the behavioral modal vector by its corresponding weight and then summing the results to obtain the baseline behavioral value; calculating the mean of the environmental stress factors and airborne migration potential to obtain the synergistic amplification coefficient; multiplying the baseline behavioral value by the synergistic amplification coefficient and then by the corresponding department weight, and normalizing the result to the 0-1 interval, which is the pollution load value.

[0080] The weighting of total flush volume per toilet visit is 0.3 because flush volume directly relates to the flushing effect of contaminants. Insufficient flush volume can easily lead to the residue of feces and other contaminants, while excessive flush volume may indicate a large amount of contaminants; both significantly increase the pollution load. The weighting of delay time is 0.3 because the longer the first flush is delayed after the door is closed, the longer the contaminants are exposed in the toilet bowl, making them more likely to form stubborn residues during this period, which is a key factor in the accumulation of pollution load. The weighting of total time the toilet door is closed per visit is slightly lower at 0.2 because the toilet visit duration reflected needs to be combined with the delay time to be more meaningful. If the time is long but flushing is timely, the risk of contaminant residue will be reduced. The weighting of flush interval is 0.2 because the length of the interval mainly affects the frequency of aerosol generation, which is an indirect risk, while the core of the pollution load is the amount of contaminant residue, which is a direct risk.

[0081] The generated pollution load value effectively avoids interference from disinfectant volatilization in hospital toilets on odor monitoring. By analyzing behavioral characteristics such as the total duration of a single toilet door closure and delay time, combined with temperature and humidity benchmarks and departmental weights for infectious disease and general departments, the pollution status can be understood. Among these, the behavioral modal vector is used to capture details of toilet behavior, the environmental stress factor can reflect departmental differences and the impact of temperature and humidity on pollution, and the airborne migration potential can reflect the risk of pollution spread in a timely manner. The pollution load value, which integrates the three factors, can accurately represent the residual and spread trends of pollutants. Even in scenarios with frequent disinfectant use, the degree of toilet dirtiness can still be accurately identified, improving the accuracy of hospital toilet cleaning calls.

[0082] In one embodiment, ventilation data and environmental data are analyzed to obtain a global attenuation coefficient, including:

[0083] The three-dimensional spatial layout of the target call area is obtained, including the location of each cubicle, toilet, and vent. The ventilation data is combined with the three-dimensional spatial layout to calculate the proportion of low-speed airflow area and the wind speed gradient of mainstream area, and to generate the effectiveness of airflow organization. Specifically, when discretizing the space of the target call area into a two-dimensional grid, the origin is set at the intersection of the central axis of the toilet entrance and the ground, the length direction of the toilet is set as the X-axis, and the width direction is set as the Y-axis. A planar coordinate system is established, and the grid is divided into fixed sizes of 0.5 meters × 0.5 meters according to the actual length and width of the toilet.

[0084] The locations of each cubicle, toilet, and vent are mapped to a grid. Based on ventilation data, the wind speed of each grid is determined. Grids with wind speeds < 0.3 m / s are designated as low-speed airflow grids. All grids with low-speed airflow are combined into a low-speed airflow zone. The proportion of low-speed airflow zone grids to the total number of grids is calculated to obtain the low-speed airflow zone percentage. All grids with wind speeds ≥ 0.3 m / s and connected to vents are designated as mainstream zones. The ratio of the wind speed difference between adjacent grids in the mainstream zone to the grid spacing (0.5 meters) is calculated. The mean of all ratios is calculated to obtain the wind speed gradient of the mainstream zone.

[0085] The calculation results are normalized to the 0-1 range by using (1 - proportion of low-speed airflow area) × 0.6 + wind speed gradient of mainstream area × 0.4 to obtain the effectiveness of airflow organization.

[0086] The weight of (1 - proportion of low-speed airflow zone) is 0.6 because hospital toilets have many cubicles and corners, and low-speed airflow zones are prone to becoming pollution dead zones. Odors are difficult to diffuse in these areas and will continue to accumulate, increasing the risk of cross-infection through contact or air transmission. The weight of the wind speed gradient in the mainstream area is 0.4 because it reflects the active diffusion capacity of airflow. The larger the gradient, the faster the airflow can push the pollution from the core area of ​​the toilet to the ventilation opening for discharge. However, if the proportion of low-speed airflow zone is high, even if the gradient in the mainstream area is large, a large amount of pollution will still remain.

[0087] Based on environmental data, the spatial distribution differences and temporal fluctuation characteristics are analyzed to generate a microclimate eddy coefficient representing the intensity of natural convection and turbulence caused by uneven temperature and humidity. Specifically, the real-time temperature and humidity data of each grid are recorded every 20 seconds. The grid directly above and within 1 meter around the toilet stall is the event-related grid, and the rest is the background grid. The average temperature difference and average humidity difference between the event-related grid and the background grid are calculated. The absolute values ​​of the average temperature difference and average humidity difference are taken respectively, and the absolute values ​​of temperature and humidity are added together to obtain the event-background spatial difference.

[0088] For each event-related grid, the temperature and humidity changes within 3 minutes after the toilet use event are statistically analyzed. The sum of the absolute values ​​of temperature changes every 20 seconds is calculated. Similarly, the sum of the absolute values ​​of humidity changes is obtained using the same method as for temperature. The sum of the absolute values ​​of temperature and humidity is added together and divided by 3 minutes to obtain the instantaneous fluctuation value after the event. The event-background spatial difference is added together with the instantaneous fluctuation value after the event, and the calculation result is normalized to the 0-1 range. This result represents the microclimate eddy coefficient, which indicates the intensity of natural convection and turbulence caused by uneven temperature and humidity.

[0089] The pollutant retention potential is generated by antagonistically fusing airflow organization effectiveness and microclimate eddy coefficient. Specifically, this involves: calculating the absolute value of the difference between airflow organization effectiveness and microclimate eddy coefficient, subtracting the absolute value of the difference from 1 to obtain the coordination degree; and subtracting the airflow organization effectiveness from 1 and multiplying it by (1 minus the microclimate eddy coefficient) to obtain the basic retention value of the combined effect of the two.

[0090] Multiply the base retention value by (2 minus the coordination degree), and normalize the result to the 0-1 range to obtain the pollutant retention potential.

[0091] The frequency and intensity of toilet use events are obtained, and combined with the pollutant retention potential, the environment’s ability to resist pollution and restore a clean state is dynamically assessed to obtain the environment’s self-purification potential. Specifically, this includes obtaining the time interval between two consecutive toilet use events and adding the reciprocals of all time intervals within 10 minutes to obtain the toilet use density.

[0092] Obtain the total flush volume and total door closing time of a single event, and calculate the average total flush volume and average total door closing time of all events; divide the total flush volume and total door closing time of a single event by the average total flush volume and average total door closing time, respectively, to obtain two deviation multiples; take the maximum of the two deviation multiples as the abnormal intensity of a single event; multiply the abnormal intensities of all single events within 10 minutes to obtain the cumulative intensity index;

[0093] Multiply the toilet usage density by the cumulative intensity index to obtain the comprehensive pulse value of pollution input per unit time; multiply 1 - pollutant residence potential by the ventilation runtime percentage and then by the microclimate eddy coefficient to obtain the dynamic purification force, where the ventilation runtime percentage is obtained by dividing the fan running time in the last 10 minutes by 10 minutes; subtract the dynamic purification force from the comprehensive pulse value to obtain the net residual amount during this period, where if the dynamic purification force is greater than the comprehensive pulse value, the net residual amount is 0;

[0094] Divide 1 by (1 + net residue) and normalize the result to the 0-1 range to obtain the environmental self-purification potential.

[0095] In one embodiment, analyzing ventilation data and environmental data to obtain a global attenuation coefficient further includes:

[0096] By fusing airflow organization effectiveness with microclimate eddy coefficient to generate a baseline attenuation constant, and combining environmental self-purification potential with the baseline attenuation constant to calculate the instantaneous eddy attenuation rate, the following steps are taken: multiplying airflow organization effectiveness with microclimate eddy coefficient to obtain the baseline value of their synergistic effect; taking the natural logarithm of the baseline value and adding 1 to obtain the synergistic gain factor, which is used to amplify the attenuation effect when the two are adapted; and multiplying the mean of airflow organization effectiveness and microclimate eddy coefficient by the synergistic gain factor and normalizing the calculation result to the 0-1 interval to obtain the baseline attenuation constant.

[0097] If the environmental self-purification potential is >0.5, multiply the baseline attenuation constant by 1.3 and normalize the calculation result to the 0-1 interval to obtain the instantaneous eddy current attenuation rate; if the environmental self-purification potential is ≤0.5, normalize the calculation result to the 0-1 interval and multiply the baseline attenuation constant by 0.8 to obtain the instantaneous eddy current attenuation rate.

[0098] The instantaneous eddy current attenuation rate is spatiotemporally smoothed and calibrated according to different functional areas within the target call area to obtain a global attenuation coefficient. Specifically, different functional areas within the target call area include toilet areas, handwashing areas, and passageways. A calculation cycle of 20 seconds is used. Temporally, the instantaneous eddy current attenuation rates of the current and the two previous calculation cycles are weighted and summed with weights of 5, 3, and 2 to obtain a temporally smoothed value. Spatially, for each grid, the temporal smoothed values ​​of its eight adjacent grids (up, down, left, right, and diagonally) are weighted according to proximity (e.g., 0.15 for four directly adjacent grids and 0.1 for four diagonally adjacent grids). These eight adjacent grids are multiplied by their respective weights and then summed to obtain a spatial smoothed value. After adding the temporal and spatial smoothed values, the calculation result is normalized to the 0-1 range to obtain the spatiotemporally smoothed attenuation rate.

[0099] The locations of different functional areas within the target call area, such as the toilet area, handwashing area, and passageway area, are obtained and marked on a grid. For each grid, if it belongs to the toilet area, the spatiotemporal smoothing attenuation rate is multiplied by the average of the last 10 pollution load values ​​to obtain the calibration value for that area. If it belongs to the handwashing area, the spatiotemporal smoothing attenuation rate is multiplied by (real-time humidity ÷ corresponding department humidity baseline value) to obtain the calibration value for that area. If it belongs to the passageway area, the proportion of effectively ventilated grids (wind speed ≥ 0.3 m / s is considered effective ventilation) to the total number of grids in the passageway is calculated, and this proportion is multiplied by the spatiotemporal smoothing attenuation rate to obtain the calibration value for that area. The regional calibration values ​​of all grids are multiplied by the corresponding department weights, and the average is taken. The average is then normalized to the 0-1 range to obtain the global attenuation coefficient.

[0100] By combining the three-dimensional spatial layout and ventilation data of the target call area, a 0.5m × 0.5m two-dimensional grid is divided to calculate the effectiveness of airflow organization. This allows for the precise location of pollution dead zones and the assessment of airflow diffusion capabilities. Based on the microclimate eddy coefficient, the intensity of natural convection and turbulence caused by uneven temperature and humidity is reflected, and the self-purification potential of the environment is calculated. The environment's anti-pollution and self-purification capabilities are dynamically assessed. After spatiotemporal smoothing of the instantaneous eddy attenuation rate and functional area calibration for toilet stalls, handwashing areas, and passageways, the final global attenuation coefficient can accurately quantify the spatiotemporal attenuation law of pollution, dynamically adapt to the complex environment of the hospital, and ensure that the toilet's dirt status can still be accurately detected even in disinfectant scenarios, triggering timely cleaning to reduce the risk of cross-infection.

[0101] In one embodiment, when a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field, including:

[0102] When a new toilet-use event occurs, based on its pollution load value and the effectiveness of the current airflow organization, an initial pollution plume with a dominant direction is generated at the event location. Specifically, this involves identifying all grid cells belonging to the mainstream region within a search radius (1.5 meters) centered on the event source location. For each mainstream region grid cell, a base direction vector is generated from the event source location to that grid cell. This base direction vector is formed by drawing a directional arrow between the event source location and each mainstream region grid cell; the direction and length of this arrow constitute the base direction vector. The direction of the base direction vector indicates the direction in which pollutants may propagate from the source point to the grid. The difference between the wind speed at the mainstream grid and the wind speed at the event source location is calculated, and the base direction vector is multiplied by (the wind speed difference divided by the modulus of the base direction vector) to obtain the weighted direction vector. The length (modulus) of the vector represents the straight-line distance between the two locations. The larger the wind speed difference, the stronger the thrust from the event source to the grid. All weighted direction vectors are combined to obtain a composite general direction vector, and the direction of the general direction vector is the dominant direction of the initial pollution plume of this event.

[0103] The final intensity of the initial plume is obtained by multiplying the pollution load value by the airflow organization effectiveness. The higher the airflow organization effectiveness, the better the ventilation, and the faster the pollutants can be carried away and diluted. Therefore, at the source, it manifests as a plume that spreads quickly but whose initial concentration is modulated by the environmental capacity.

[0104] An elliptical diffusion core, centered on the location of the event source and elongated along the dominant direction, is formed where the pollutant's influence range is set to the farthest and its weight decays the slowest in the dominant direction. In the direction perpendicular to the dominant direction, the influence range is closer and its weight decays the fastest. In the completely opposite direction, the influence range is set to the smallest. This forms an initial distribution that is a plume-like structure with its head pointing towards the dominant direction and its tail being long and thin, rather than a circle.

[0105] The directional asymmetric diffusion nucleus and its final intensity are superimposed onto the event source location and the surrounding affected grid in the dynamic pollution situation field, thereby generating an initial pollution plume with a dominant direction at the event location.

[0106] Based on the dominant direction of the initial pollution plume and the global attenuation coefficient, the attenuation of pollutants and the migration driving force within each grid cell are calculated to generate a convection-diffusion intensity field. Specifically, this includes: multiplying the pollution load value of the grid by the global attenuation coefficient to obtain the pollutant attenuation of the grid within the current calculation cycle (20 seconds); calculating the angle between the real-time wind direction of the grid and the dominant direction of the initial pollution plume in this event, and calculating the cosine value of the angle; processing the cosine value through the ReLU function to ensure that it is non-negative; and multiplying the pollution load value by the cosine value to obtain the migration driving force.

[0107] Create a new convection-diffusion intensity field of the same size as the grid map, and directly fill the migration driving force into the corresponding grid position in the field. The pollutant decay amount will be used independently to update the pollution load value of each grid. The update process is: subtract the pollutant decay amount from the pollution load value and then add the net flux to obtain the new pollution load value.

[0108] The net flux is calculated as follows: Using this grid as the central grid, multiply the pollution load value of all adjacent grids by the corresponding migration driving force and then by the time step factor (20 seconds) to obtain the individual inflow. Add the individual inflows of all adjacent grids that meet the conditions to obtain the total inflow from all sides into the central grid. Multiply the pollution load value of this grid by the corresponding migration driving force and then by the time step factor (20 seconds) to obtain the total outflow. Subtract the total outflow from the total inflow to obtain the net flux.

[0109] In one embodiment, when a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field. The method also includes:

[0110] Based on the convection-diffusion intensity field, the flux of pollutants propagating from each grid to its neighboring grids is calculated to generate the neighborhood coupling flux. Specifically, this involves: calculating the difference between the migration driving force of the grid and the migration driving force of each neighboring grid, using the difference as a response coefficient, multiplying the pollution load value of the grid by the response coefficient, and then multiplying by a time step of 20 seconds to obtain the basic propagation flux; if the wind direction of the initial pollution plume is in the dominant direction, the weight is 0.9; if the wind direction is in the vertical direction, the weight is 0.6; if the wind direction is in the completely opposite direction, the weight is 0.3; multiplying the basic propagation flux by the corresponding wind direction weight yields the basic flux.

[0111] Calculate the standard deviation of all neighboring grids of the grid. When the standard deviation is greater than 0.15, multiply the base flux by 0.85 to obtain the neighborhood coupling flux; when the standard deviation is less than or equal to 0.15, multiply the base flux by 1.1 to obtain the neighborhood coupling flux.

[0112] The instantaneous pollution field strength is generated by superimposing the pollution load values ​​of the initial pollution plumes of all toilet events propagated to each grid through neighborhood coupling flux. Specifically, for each recorded toilet event, the length of time since the event occurred is multiplied by the global attenuation coefficient to obtain the current effective weight of the pollution load of the event; for the target grid, the pollution load values ​​of all toilet events propagated to the grid through neighborhood coupling flux are obtained, the pollution load values ​​of each event propagated to this grid are multiplied by the current effective weight and summed, and the calculation results are normalized to the 0-1 interval, which is the instantaneous pollution field strength.

[0113] The instantaneous pollution field strength of the entire field is integrated to obtain a dynamic pollution situation field. Specifically, this includes: calculating the field in 20-second cycles; if the grid belongs to the toilet area, multiplying the average instantaneous pollution field strength of the grid over the past 5 cycles by 1.1 to obtain the calibrated pollution field strength; if the grid belongs to the handwashing area, multiplying the instantaneous pollution field strength by the square root of (real-time humidity divided by the department's humidity baseline) to obtain the calibrated pollution field strength; if the grid belongs to the passageway area, multiplying the instantaneous pollution field strength by the proportion of effectively ventilated grids to obtain the calibrated pollution field strength; and then combining the calibrated pollution field strengths of all grids after normalizing them to the 0-1 range to form the dynamic pollution situation field.

[0114] By calculating the airflow organization effectiveness using ventilation data, an initial pollution plume with a dominant direction is generated at the location of the incident. The influence of pollution source intensity and on-site ventilation on the initial diffusion path of pollutants is comprehensively considered. Subsequently, based on the global attenuation coefficient and the dominant direction of the initial pollution plume, the exponential attenuation and migration driving force of pollutants in each grid are calculated by generating a convection-diffusion intensity field. Furthermore, the propagation process of pollutants between grids is understood by calculating the neighborhood coupling flux. Finally, by superimposing the influence of all toilet-use events and integrating them to generate a dynamic pollution situation field, dynamic analysis and quantification of the entire process of pollutant generation, diffusion and attenuation in the toilet space are achieved. This makes the judgment of the degree of dirtiness no longer dependent on the concentration of a single and easily contaminated odor molecule. Even in complex environments with strong interfering gases such as disinfectant volatilization, it can still maintain a high-precision perception and assessment of the actual pollution load, improving the accuracy and reliability of cleaning calls.

[0115] In one embodiment, the pollution load values ​​of all grid points in the dynamic pollution situation field are spatially integrated to obtain the global situation intensity, including:

[0116] To identify the spatial distribution characteristics of pollution load in a dynamic pollution situation field, calculate its gradient field intensity and high load area aggregation, and generate a pollution distribution heterogeneity index, the following steps are taken: For each grid in the dynamic pollution situation field, calculate the absolute value of the difference between its pollution load value and that of all its directly adjacent grids (8 grids), sum these absolute values ​​and divide by the number of adjacent grids to obtain the local gradient intensity of the grid, and calculate the mean of the local gradient intensities of all grids, which is the gradient field intensity.

[0117] Calculate the mean pollution load value of all grids. Grids with pollution load values ​​exceeding 1.5 times the mean are defined as high-load grids. Count the number of continuous areas (i.e., pollution focal points) formed by these high-load grids. Calculate the mean number of high-load grids contained in each pollution focal point; this mean is the average focal point size. Calculate the average Euclidean distance between the centers of all pollution focal points as the average distance between focal points. Divide the average focal point size by the average distance between focal points to obtain the high-load area concentration.

[0118] The gradient field intensity is multiplied by the aggregation degree of the high-load area, and the calculation result is normalized to the 0-1 interval to generate the pollution distribution heterogeneity index.

[0119] Based on the global attenuation coefficient, the instantaneous load of each grid point in the dynamic pollution situation field is corrected for timeliness, generating a spatiotemporal attenuation correction factor. Specifically, this includes: for each grid in the dynamic pollution situation field, obtaining the time of the toilet event corresponding to its current pollution load value; calculating the duration from that time point to the current time, multiplying the duration by the global attenuation coefficient to obtain the time attenuation factor; for all current pollution focal points, for the target grid, calculating the Euclidean distance between it and the center point of each pollution focal point, using this Euclidean distance as the spatial influence weight, summing all the spatial influence weights to obtain the total spatial diffusion influence value of the grid; multiplying the time attenuation factor by the total spatial diffusion influence value, normalizing the calculation result to the 0-1 interval, and obtaining the spatiotemporal attenuation correction factor for each grid.

[0120] Based on the different functional areas within the target call area, the sensitivity weight field of each grid is calculated, specifically including: the toilet area, because it is a direct source of pollution, has a basic sensitivity coefficient of 0.9; the handwashing area, because of its high-frequency contact characteristics, has a basic sensitivity coefficient of 0.7; and the passage area, because of the characteristics of personnel flow, has a basic sensitivity coefficient of 0.5.

[0121] The frequency of use is obtained by summing the number of sensor triggers in each grid within the last 10 minutes and normalizing the sum to the 0-1 range. For each area of ​​the toilet stall area, handwashing area, and passageway area, the duration of stay for personnel in each area is calculated, and the mean of the sum of the durations is normalized to the 0-1 range, which is the personnel exposure coefficient. The infectious disease department is assigned a hygiene standard level coefficient of 1.2, while the general department is assigned a hygiene standard level coefficient of 0.8.

[0122] Multiply the basic sensitivity by the usage frequency coefficient, then by the square root of the personnel exposure coefficient, and finally by the health standard level coefficient to obtain the initial weight. Normalize the initial weights of all grids to the interval between 0 and 1 to obtain the dynamic functional area sensitivity weight field of each grid.

[0123] Among them, the excrement, vomit, and secretions of patients in the infectious disease department contain high concentrations of active pathogens, so the hygiene standard level coefficient is set at 1.2. This can trigger cleaning instructions earlier and more frequently, thereby cutting off the accumulation and transmission chain of pathogens in the toilet environment at the first time. In contrast, the pathogen load in the pollutants of patients in the general ward is low, and the risk of cross-infection is relatively small, so the hygiene standard level coefficient is set at 0.8.

[0124] The dynamic pollution situation field, the spatiotemporal attenuation correction factor, and the functional area sensitivity weight field are fused to generate a perceived weighted pollution field. The perceived weighted pollution field is then fused with the pollution distribution heterogeneity index to generate the instantaneous global load integral. Specifically, this involves: for each grid in the dynamic pollution situation field, multiplying its pollution load value by the spatiotemporal attenuation correction factor and the initial weight to obtain the perceived weighted pollution value of that grid; the perceived weighted pollution values ​​of all grids together constitute the perceived weighted pollution field; summing the perceived weighted pollution values ​​of all grids in the perceived weighted pollution field to obtain the basic global integral; and multiplying the basic global integral by the pollution distribution heterogeneity index to obtain the instantaneous global load integral.

[0125] Based on the ratio of the instantaneous global load integral to the pollution distribution heterogeneity index, a dynamic intensity calibration coefficient is calculated. The instantaneous global load integral is then processed according to this coefficient to obtain the global situation intensity. Specifically, this involves: dividing the instantaneous global load integral by the pollution distribution heterogeneity index to obtain an initial ratio; when the initial ratio > 0.7, a high load-high heterogeneity state is identified, and the dynamic intensity calibration coefficient is set to 1.3; when the initial ratio < 0.3, a low load-low heterogeneity state is identified, and the dynamic intensity calibration coefficient is set to 0.7; when 0.3 ≤ initial ratio ≤ 0.7, the dynamic intensity calibration coefficient is 1; and finally, the instantaneous global load integral is multiplied by the dynamic intensity calibration coefficient, and the product is normalized to the 0-1 interval to obtain the global situation intensity.

[0126] By calculating the pollution distribution heterogeneity index, the non-uniformity and aggregation characteristics of pollution load in spatial distribution were understood. The pollution distribution heterogeneity index integrates the gradient field intensity, which reflects the intensity of local changes, and the aggregation degree of high-load areas, which reflects the concentration of high-pollution areas. Thus, it can accurately identify potential high-risk areas for cross-infection. Secondly, the pollution load value of each grid in the dynamic pollution situation field is corrected in a timely manner by using a spatiotemporal attenuation correction factor to ensure that the assessment of the pollution situation can reflect its true evolution over time. Furthermore, through the generated global situation intensity, even in complex environments with strong interfering gases such as disinfectant volatilization, it can still maintain a stable and accurate perception of the overall dirtiness of toilets.

[0127] In one embodiment, the urgency of cleaning is generated by analyzing the number of all current pollution focal points, their total energy intensity, and the degree of spatial concentration, including:

[0128] Analyzing the spatial relationships of all pollution foci to generate pollution focus clusters involves: using the centroid of each pollution focus (the centroid being the mean of all grid coordinates) as a node; for each pollution focus, calculating the standard deviation of the Euclidean distances from all grids to the centroid as the feature radius of that focus; multiplying the mean of all focus feature radii by 1.5 to obtain the dynamic distance threshold; starting from any focus, finding all adjacent focuses whose Euclidean distance is less than the dynamic distance threshold, merging these focuses into an initial cluster; then, based on this cluster, continuing to find other focuses whose Euclidean distance to any focus within the cluster is less than the dynamic distance threshold, incorporating them into the cluster, repeating this process until no new focuses can be added, thus completing the division of a cluster; subsequently, repeating the above process from the remaining focuses until all focuses are classified, thereby generating a pollution focus cluster.

[0129] The energy intensity and spatial distribution of each focus are calculated to generate the focus cluster potential field. Specifically, this includes: summing the pollution load values ​​of all grids within the focus to obtain the energy intensity of that focus; calculating the Euclidean distance from the centroid of the focus to the farthest grid point as the basic influence radius; and multiplying the energy intensity by the basic influence radius to obtain the dynamic influence radius of that focus.

[0130] Since the toilet space is divided into a 0.5m x 0.5m grid, for each grid point, the Euclidean distance to the centroid of each pollution focus is calculated. For each focus, when the grid point is within its dynamic influence radius, the Euclidean distance is divided by the dynamic influence radius to obtain the relative distance. The energy intensity is then multiplied by... Multiply by the relative distance to obtain the potential field value of the grid; where e is the natural constant; superimpose the potential field values ​​generated by all pollution focal points at each grid point to obtain the complete focal cluster potential field.

[0131] Based on the potential field of the focal cluster, the interaction force generated between any two pollution focal points due to the superposition of potential fields is calculated to generate the focal interaction strength. Specifically, this includes: dividing the line connecting the two centroids into several sampling points with the centroids of the two pollution focal points as endpoints; at each sampling point, obtaining the potential field value generated by the first focal point and the potential field value generated by the second focal point in the potential field of the focal cluster; taking the smaller of these two potential field values ​​as the interaction strength at that point; summing the interaction strengths of all sampling points and dividing the sum by the total number of sampling points to obtain the average interaction strength; multiplying the average interaction strength by the product of the energy intensities of the two focal points, and normalizing the calculation result to the 0-1 interval to obtain the interaction strength between the two focal points.

[0132] The analysis of the overall distribution of pollution focus clusters within the target call area space is performed to calculate the degree of dispersion or concentration of their distribution and generate spatial distribution density. Specifically, this includes: taking a weighted average of the centroid coordinates of all pollution focus clusters according to their total energy intensity to obtain the energy weighted center point of the entire pollution distribution; calculating the Euclidean distance from each pollution focus cluster to the energy weighted center, and then summing these Euclidean distances according to the total energy intensity of their respective clusters and dividing by the total energy intensity of all clusters to obtain the energy weighted average distribution radius.

[0133] Calculate the standard deviation of the Euclidean distance from each pollution focus cluster to the energy weighting center. The standard deviation reflects the dispersion of each cluster around the center point. Add the average distribution radius of the energy weighting to the standard deviation, divide the sum by 1, and normalize the result to the 0-1 interval to obtain the spatial distribution density. The closer the spatial distribution density is to 1, the more concentrated the distribution is. The closer the spatial distribution density is to 0, the more dispersed the distribution is.

[0134] The cleaning urgency is calculated by integrating the intensity of focal interaction, spatial distribution density, number of focal points, and total energy intensity. Specifically, the following steps are taken: the four parameters of number of focal points, total energy intensity, mean intensity of interaction of all focal points, and spatial distribution density are normalized to the 0-1 interval and then multiplied to obtain the basic urgency value; the basic urgency value is multiplied by (number of focal points plus 1), and the square root of the calculation result is taken and normalized to the 0-1 interval to obtain the cleaning urgency.

[0135] By using centroid calculation and dynamic distance thresholds to delineate pollution focus clusters, the system accurately identifies pollution accumulation areas. Next, it constructs a potential field for these clusters based on total energy intensity and dynamic influence radius. This is then overlaid with and analyzed to reflect the overall pollution distribution. Finally, by comparing the potential fields of sampling points, the system calculates the interaction strength between the focus clusters, assessing the synergistic impact of pollution sources. The resulting cleaning urgency level represents an upgrade from passively sensing odors to actively analyzing the nature of pollution. Even when disinfectant odors mask the true smell, the system accurately captures the number, energy intensity, distribution concentration, and interaction level of pollution focus clusters, significantly improving the timeliness and accuracy of cleaning calls and avoiding cleaning delays or resource waste caused by sensor misjudgments.

[0136] In one embodiment, the overall situational awareness and cleaning urgency are fused to obtain a cleaning call instruction, including:

[0137] The dynamic decision-making benchmark value is generated by combining the overall situation intensity, cleaning urgency, and environmental self-purification potential. Specifically, the three parameters are normalized to the range of 0 to 1. The arithmetic mean of the overall situation intensity and cleaning urgency is calculated to obtain the basic cleaning demand value. The basic cleaning demand value is multiplied by (1 minus the environmental self-purification potential) to obtain the adjusted cleaning demand value. The adjusted cleaning demand value is then normalized to the range of 0 to 1 again, which is the dynamic decision-making benchmark value.

[0138] A pollution synergy analysis is performed on the overall situation intensity and the urgency of cleaning to generate an overall-local situation coupling factor. Specifically, this includes: multiplying the overall situation intensity by the urgency of cleaning to obtain a basic synergy value; calculating the absolute difference between the overall situation intensity and the urgency of cleaning; subtracting the absolute difference from 1 to obtain the synergy degree; and multiplying the basic synergy value by the synergy degree to obtain an initial coupling value.

[0139] If the degree of coordination is ≥0.7, the initial coupling value is multiplied by 1.2 and the result is normalized to the 0-1 interval to obtain the global-local situation coupling factor, thereby amplifying the coordination effect; if the degree of coordination is <0.7, the initial coupling value is multiplied by 0.8 and the result is normalized to the 0-1 interval to obtain the global-local situation coupling factor, thereby suppressing inconsistent signals.

[0140] Based on the rate of change of environmental self-purification potential and global situation intensity, the stable confirmation time required for call triggering is calculated. Specifically, this includes: obtaining the global situation intensity at the current moment and the previous four calculation cycles; subtracting the global situation intensity of the first calculation cycle (the first of the four calculation cycles) from the current global situation intensity to obtain the absolute change; dividing the absolute change by the time span; and normalizing the absolute value of the calculation result to the 0-1 range to obtain the standardized rate of change; subtracting the standardized rate of change from the environmental self-purification potential to obtain the net change advantage value; when the net change advantage value is greater than zero, it indicates that self-purification capacity is dominant; when the net change advantage value is ≤ zero, it indicates that pollution aggravation is dominant.

[0141] When the net change advantage value is greater than 0.3, the benchmark confirmation time is set to 300 seconds; when the net change advantage value is between -0.3 and 0.3, the benchmark confirmation time is set to 180 seconds; when the net change advantage value is less than -0.3, the benchmark confirmation time is set to 90 seconds.

[0142] Calculate the fluctuation range of the three most recent standardized rates of change. If the fluctuation range is <0.1, shorten the baseline confirmation time by 25%. If the fluctuation range is ≥0.1, extend the baseline confirmation time by 40%, thereby obtaining the stable confirmation time required for call triggering.

[0143] Based on the decay trend of environmental self-purification potential and cleaning urgency, the state recovery time required for call cancellation is calculated. Specifically, this includes: calculating the growth rate of environmental self-purification potential over the most recent three periods, subtracting the growth rate of the earliest period from the growth rate of the most recent period to obtain the absolute growth amount; calculating the decline rate of cleaning urgency over the most recent three periods, subtracting the decline rate of the most recent period from the decline rate of the earliest period to obtain the absolute decline amount; and adding the absolute decline amount and the absolute growth amount and normalizing to the 0-1 range to obtain the comprehensive improvement index.

[0144] When the overall improvement index is greater than 0.6, the basic recovery time is set to 120 seconds; when the overall improvement index is between 0.3 and 0.6, the basic recovery time is set to 240 seconds; when the overall improvement index is less than 0.3, the basic recovery time is set to 480 seconds.

[0145] The ratio of environmental self-cleaning potential to cleaning urgency is calculated. When the ratio is ≥1.5, the basic recovery time is shortened by 30%; when the ratio is <0.5, the basic recovery time is extended by 50%. In other cases, the basic recovery time remains unchanged. Thus, the state recovery time required for call cancellation can be obtained.

[0146] The global-local situation coupling factor is compared with the dynamic decision benchmark value to generate a cleaning call instruction. Specifically, if the global-local situation coupling factor is continuously greater than or equal to the dynamic decision benchmark value when not in a call state, and the duration of maintenance reaches the stable confirmation duration, then an immediate cleaning call instruction is generated.

[0147] If, while the call is active, the global-local situational coupling factor remains below the dynamic decision baseline value, and the duration of this maintenance reaches the state recovery duration, then the cleaning call instruction is cancelled.

[0148] By generating dynamic decision-making benchmark values, the system incorporates the environment's self-purification capabilities into the identification system, avoiding misjudgments of cleaning needs when disinfectant volatilization masks the true pollution. On the other hand, it comprehensively assesses the synergy between global trends and local pollution through a global-local situation coupling factor, amplifying effective signals and suppressing contradictory data to improve judgment accuracy. Simultaneously, it dynamically adjusts the stability confirmation time based on the rate of change and fluctuation range to prevent false calls triggered by short-term pollution fluctuations, and calculates the state recovery time to avoid resource waste when pollution has been alleviated but calls are not canceled in time. Even in scenarios with strong disinfectant interference, it can still accurately identify the true pollution needs, improving the timeliness and reliability of cleaning calls and reducing invalid calls and cleaning delays.

[0149] In one embodiment, a method for precise toilet cleaning call based on multi-sensor fusion is applied to the aforementioned system for precise toilet cleaning call based on multi-sensor fusion, comprising:

[0150] Step S1: Real-time acquisition of state change time-series data of multiple sensors in each toilet stall within the target call location; based on the state change time-series data, identification of complete toilet use event sequences; analysis of the content of toilet use events; and obtaining the pollution load value of the toilet use event. The target call location is a hospital toilet.

[0151] Step S2: Discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data, and obtain the global attenuation coefficient.

[0152] Step S3: When a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the location of the event. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field.

[0153] Step S4: In the dynamic pollution situation field, identify the pollution focus, and perform spatial integration on the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity.

[0154] Step S5: Analyze the number, total energy intensity, and spatial concentration of all current pollution focal points to generate a cleaning urgency level.

[0155] Step S6: Integrate the overall situation intensity with the urgency of cleaning to obtain a cleaning call instruction.

[0156] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A toilet cleaning precision call system based on multi-sensor fusion, characterized in that, include: The analysis unit is used to acquire real-time time-series data of the status changes of multiple sensors in each toilet stall within the target call location. Based on the time-series data of status changes, it identifies the complete toilet use event sequence, analyzes the content of the toilet use event, and obtains the pollution load value of the toilet use event. The target call location is the toilet in a hospital. The attenuation unit is used to discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data, and obtain the global attenuation coefficient. The diffusion unit is used to inject the corresponding pollution load value into the grid at the location of the new toilet event when a new toilet event occurs. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field. The global analysis unit is used to identify pollution focal points in a dynamic pollution situation field and to spatially integrate the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity. The contamination unit, used to generate cleaning urgency, includes: analyzing the spatial relationships of all contamination focal points and generating contamination focal point clusters; The energy intensity and spatial distribution of each focus are calculated to generate the focus cluster potential field; Based on the potential field of the focal cluster, the interaction force generated between any two pollution focal points due to the superposition of potential fields is calculated, and the focal interaction strength is generated. Analyze the overall distribution of pollution focus clusters within the target call area, calculate the degree of dispersion or concentration of their distribution, and generate spatial distribution density. The urgency of cleaning is obtained by integrating the intensity of focal interaction, the density of spatial distribution, the number of focal points, and the total energy intensity. The cleaning call unit is used to integrate the overall situation intensity with the urgency of cleaning to generate cleaning call instructions.

2. The toilet cleaning precision call system based on multi-sensor fusion according to claim 1, characterized in that, Based on state change time series data, complete toilet use event sequences are identified, and the content of each toilet use event is analyzed to obtain the pollution load value of that toilet use event, including: Analyze the time sequence of state changes of the door magnetic sensor and the flushing sensor, and map them into a multi-dimensional behavioral mode vector; Based on environmental data, dynamic environmental stress factors are calculated; By analyzing the patterns, intensity, and behavioral mode vectors of flushing events, the airborne migration potential is obtained. The pollution load value is generated by fusing behavioral modal vectors, environmental stress factors, and airborne migration potential.

3. The toilet cleaning precision call system based on multi-sensor fusion according to claim 1, characterized in that, Analyzing ventilation and environmental data yields the global attenuation coefficient, including: The three-dimensional spatial layout of the target call area is obtained, and the ventilation data is combined with the three-dimensional spatial layout to calculate the proportion of low-speed airflow area and the wind speed gradient of mainstream area, thereby generating the effectiveness of airflow organization. Based on environmental data, we analyze the spatial distribution differences and temporal fluctuation characteristics to generate a microclimate eddy coefficient that represents the intensity of natural convection and turbulence caused by uneven temperature and humidity. By combining the effectiveness of airflow organization with the microclimate eddy coefficient in a counter-cyclical manner, pollutant retention potential is generated. By obtaining the frequency and intensity of toilet use events and combining them with pollutant retention potential, the environment's ability to resist pollution and restore its cleanliness can be dynamically assessed, thus obtaining the environment's self-purification potential.

4. The toilet cleaning precision call system based on multi-sensor fusion according to claim 3, characterized in that, The global attenuation coefficient is obtained by analyzing ventilation and environmental data, and also includes: By integrating the airflow organization effectiveness with the microclimate eddy coefficient, a benchmark attenuation constant is generated. The environmental self-purification potential is then combined with the benchmark attenuation constant to calculate the instantaneous eddy attenuation rate. The instantaneous eddy current attenuation rate is spatiotemporally smoothed and calibrated according to different functional areas within the target call site to obtain the global attenuation coefficient.

5. A toilet cleaning precision call system based on multi-sensor fusion according to claim 4, characterized in that, When a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field, including: When a new toilet incident occurs, an initial pollution plume with a dominant direction is generated at the location of the incident, based on its pollution load value and the effectiveness of the current airflow organization. Based on the dominant direction of the initial pollution plume and the global attenuation coefficient, the attenuation of pollutants and the migration driving force within each grid cell are calculated to generate a convection-diffusion intensity field.

6. A toilet cleaning precision call system based on multi-sensor fusion according to claim 5, characterized in that, When a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the event location. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field. This also includes: Based on the convection-diffusion intensity field, the flux of pollutants propagating from each grid to its neighboring grids is calculated, generating neighborhood coupling flux; The initial pollution plumes of all toilet events are propagated to the pollution load values ​​of each grid through neighborhood coupling flux and then superimposed to generate the instantaneous pollution field strength. By integrating the instantaneous pollution field strength across the entire field, a dynamic pollution situation field is obtained.

7. A toilet cleaning precision call system based on multi-sensor fusion according to claim 6, characterized in that, Spatial integration of the pollution load values ​​at all grid points in the dynamic pollution situation field yields the global situation intensity, including: Identify the spatial distribution characteristics of pollution load in a dynamic pollution situation field, calculate its gradient field intensity and high load area aggregation degree, and generate a pollution distribution heterogeneity index; Based on the global attenuation coefficient, the instantaneous load of each grid point in the dynamic pollution situation field is corrected in a timely manner to generate a spatiotemporal attenuation correction factor. Based on the different functional areas within the target call center, calculate the functional area sensitivity weight field for each grid. The dynamic pollution situation field, the spatiotemporal attenuation correction factor and the functional area sensitivity weight field are fused to generate the perception weighted pollution field. The perception weighted pollution field is then fused with the pollution distribution heterogeneity index to generate the instantaneous global load integral. Based on the ratio of the instantaneous global load integral to the pollution distribution heterogeneity index, the dynamic intensity calibration coefficient is calculated, and the instantaneous global load integral is processed according to the dynamic intensity calibration coefficient to obtain the global situation intensity.

8. A toilet cleaning precision call system based on multi-sensor fusion according to claim 7, characterized in that, By integrating the overall situational awareness and the urgency of cleaning, cleaning call instructions are generated, including: The dynamic decision-making benchmark value is generated by combining the overall situation intensity, the urgency of cleaning, and the environmental self-purification potential. By conducting a pollution synergistic analysis of the overall situation intensity and the urgency of cleaning, a global-local situation coupling factor is generated. Based on the rate of change of environmental self-purification potential and global situation intensity, the stable confirmation time required for call triggering is calculated. Based on the decay trend of environmental self-purification potential and cleaning urgency, the state recovery time required for call cancellation is calculated. The global-local situation coupling factor is compared with the dynamic decision benchmark value to generate a cleaning call instruction.

9. A method for precise toilet cleaning call based on multi-sensor fusion, applied to the toilet cleaning precise call system based on multi-sensor fusion as described in any one of claims 1-8, characterized in that, include: Step S1: Real-time acquisition of state change time-series data of multiple sensors in each toilet stall within the target call location; based on the state change time-series data, identification of complete toilet use event sequences; analysis of the content of toilet use events; and obtaining the pollution load value of the toilet use event. The target call location is a hospital toilet. Step S2: Discretize the space of the target call location into a two-dimensional grid, acquire the ventilation data and environmental data of the target call location in real time, analyze the ventilation data and environmental data, and obtain the global attenuation coefficient. Step S3: When a new toilet-use event occurs, the corresponding pollution load value is injected into the grid at the location of the event. For the pollution load value of each grid point in the target call area, it is exponentially decayed in each calculation cycle according to the global decay coefficient, and its spatial diffusion effect is analyzed to generate a dynamic pollution situation field. Step S4: In the dynamic pollution situation field, identify the pollution focus, and perform spatial integration on the pollution load values ​​of all grid points in the dynamic pollution situation field to obtain the global situation intensity. Step S5, generating cleaning urgency, includes: analyzing the spatial relationships of all contamination focal points and generating contamination focal point clusters; The energy intensity and spatial distribution of each focus are calculated to generate the focus cluster potential field; Based on the potential field of the focal cluster, the interaction force generated between any two pollution focal points due to the superposition of potential fields is calculated, and the focal interaction strength is generated. Analyze the overall distribution of pollution focus clusters within the target call area, calculate the degree of dispersion or concentration of their distribution, and generate spatial distribution density. The urgency of cleaning is obtained by integrating the intensity of focal interaction, the density of spatial distribution, the number of focal points, and the total energy intensity. Step S6: Integrate the overall situation intensity with the urgency of cleaning to obtain a cleaning call instruction.

Images