A distributed household kitchen garbage intelligent data processing system

Abstract

The present application relates to the technical field of intelligent environmental protection and Internet of Things, in particular to a distributed household kitchen garbage intelligent data processing system, comprising: a data acquisition module, which is deployed in a residential unit, and acquires original discarding records containing spectral characteristics, time stamps and weight curves in real time; a component analysis module, which aggregates and analyzes multi-terminal spectral data to generate a cross-household kitchen component distribution heat map; a behavior modeling module, which extracts discarding periodic patterns using time stamps; a load prediction module, which combines weight curves and component heat maps to predict regional garbage degradation oxygen demand and odor generation coefficients; and a scheduling optimization module, which matches discarding patterns with degradation oxygen demand to generate a cleaning and transportation path containing dynamic priority and service time window. The present application realizes source identification and regional monitoring of garbage components, and through the fusion of behavior and biochemical data, changes the cleaning and transportation scheduling from a fixed mode to a dynamic and accurate response.

Description

Technical Field

[0001] This invention relates to the fields of intelligent environmental protection and Internet of Things technology, and in particular to a distributed intelligent data processing system for household kitchen waste. Background Technology

[0002] Currently, the management of urban kitchen waste generally relies on traditional centralized collection and post-processing models. The front end typically uses smart trash cans with simple weighing or overflow sensors, while the back end relies on fixed collection routes or the experience of sanitation workers for collection. At the data analysis level, management often only obtains macro-level statistics such as the total amount of waste at each collection point and the frequency of collection, lacking real-time understanding of the waste's composition, and even less ability to correlate residents' disposal behavior with the biochemical characteristics of the waste itself.

[0003] Existing technological solutions have shortcomings. Relying solely on weight and volume data cannot accurately identify the chemical composition of kitchen waste, making it impossible to assess its specific degradation difficulty and environmental pollution potential. Static collection and scheduling methods based on fixed schedules cannot respond to dynamic changes in waste generation rates and composition, easily leading to waste accumulation and fermentation at some collection points, producing odors, while collection resources are wasted at others. This extensive management model is inefficient and cannot support refined assessments of waste reduction and sorting policies.

[0004] There is a need for a technological solution that can perceive the composition of kitchen waste in real time at the source and dynamically optimize the collection strategy accordingly. The key lies in how to obtain high-dimensional waste composition data and how to deeply integrate this composition data with residents' waste disposal behavior patterns, so as to achieve accurate prediction and proactive intervention of waste generation, evolution, and collection needs. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a distributed intelligent data processing system for household kitchen waste.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a distributed intelligent data processing system for household kitchen waste, comprising: The data acquisition module is a distributed kitchen waste data acquisition terminal deployed in each residential unit. It collects raw kitchen waste disposal records in real time. The raw kitchen waste disposal records include spectral characteristic data of waste components, input timestamps, and weight change curves. The component analysis module, based on the spectral feature data collected by multiple distributed kitchen waste data collection terminals, performs aggregate analysis through a cloud-based component recognition engine to generate a heat map of kitchen waste component distribution across households. The behavior modeling module uses the input timestamp data to construct a time series analysis model and extract the food waste disposal cycle pattern of each residential unit. The food waste disposal cycle pattern includes high-frequency disposal periods, low-frequency disposal periods, and time windows of abnormal disposal events. The load prediction module calculates the accumulation rate of kitchen waste per unit time based on the weight change curve, and predicts the oxygen demand for waste degradation and the potential odor generation coefficient in each area by combining the food categories in the kitchen waste component distribution heat map. The scheduling optimization module performs spatiotemporal matching between the kitchen waste disposal cycle pattern and the oxygen demand for waste degradation to generate a dynamic scheduling path for waste collection vehicles. The dynamic scheduling path includes collection priority, expected loading capacity, and service time window.

[0007] As a further aspect of the present invention, the step of aggregating and analyzing the spectral feature data collected by multiple distributed kitchen waste data collection terminals through a cloud-based component recognition engine to generate a heat map of kitchen waste component distribution across households includes: The heat map of food waste composition distribution marks the areas of high discard intensity for different food categories; The spectral feature data uploaded by each distributed kitchen waste data acquisition terminal is subjected to noise filtering to extract the effective spectral bands; The effective spectral bands are matched and identified with a preset food spectral database to determine the specific food categories and proportions in each discard event. Based on the geographic information system, the spatial location coordinates of each residential unit are converted, and a mapping relationship between spatial location and food waste categories is established. The frequency and weight of discarded food items of different categories at each spatial location within a preset time period are statistically analyzed to form a spatial distribution matrix of food categories. The spatial distribution matrix is ​​visualized using a thermal rendering algorithm to generate a heat map of the food waste component distribution, where color intensity represents the intensity of waste disposal.

[0008] As a further aspect of the present invention, the step of constructing a time series analysis model using the input timestamp data to extract the food waste disposal cycle pattern of each residential unit includes: The timestamps of the input over multiple consecutive days for a single residential unit are aggregated to form a histogram of the discard time distribution in hours; Periodic analysis is performed on the discard time distribution histogram to identify discard peaks that occur within a fixed time period each day, and the time periods corresponding to the discard peaks are marked as high-frequency discard periods; Low-density analysis was performed on the histogram of the time distribution of waste disposal to identify time periods in which food waste disposal was rare or absent over several consecutive days, and these time segments were marked as low-frequency disposal periods. Monitor the time points in the discard time distribution histogram that deviate from the historical normal pattern. When the discard frequency or single discard weight at the time point exceeds a preset threshold, define the time point and its buffer zones before and after it as the time window of the abnormal discard event. By integrating the time windows of the high-frequency disposal period, low-frequency disposal period, and abnormal disposal events, the food waste disposal cycle pattern of the residential unit is generated.

[0009] As a further aspect of the present invention, the step of calculating the accumulation rate of kitchen waste per unit time based on the weight change curve, and combining this with the food categories in the kitchen waste component distribution heatmap to predict the oxygen demand for waste degradation and the potential odor generation coefficient for each area includes: The weight change curve is differentiated to calculate its instantaneous slope, and the instantaneous slope is used as the real-time accumulation rate. The average cumulative rate per unit time is obtained by averaging the multiple real-time cumulative rates within a preset time period. Based on the dominant food category in a specific area of ​​the heat map of food waste composition, query the standard unit mass degradation oxygen demand parameter of the dominant food category. The average cumulative rate per unit time is multiplied by the standard oxygen demand per unit mass of waste to calculate the oxygen demand for waste degradation in a specific area. Based on the typical fermentation products and volatile organic compound release rates corresponding to the main food categories, and combined with ambient temperature data, the potential odor generation coefficient is calculated.

[0010] As a further aspect of the present invention, the step of spatiotemporally matching the kitchen waste disposal cycle pattern with the waste degradation oxygen demand to generate a dynamic scheduling path for waste collection vehicles includes: Extract the end time of the high-frequency disposal period in the food waste disposal cycle pattern of each residential unit; The end time of the high-frequency discarding period is combined with the real-time calculated oxygen demand for waste degradation in the corresponding area to form a pending collection node with time attribute and collection urgency attribute. Based on the geographical coordinates of all nodes to be cleared, the urgency of clearing, and vehicle capacity constraints, a clearing route optimization model is constructed. The proposed waste disposal route optimization model aims to minimize the waste disposal completion time for all nodes and prioritize high-urgency nodes. Solve the waste disposal route optimization model to output the dynamic scheduling route, which includes the vehicle departure order, the access node sequence, and the service time window of each node.

[0011] As a further aspect of the present invention, it also includes a step of dynamically adjusting the collection priority based on real-time odor concentration monitoring: A wireless odor sensor network is deployed at pre-designated garbage collection points to monitor ammonia concentration, hydrogen sulfide concentration, and total volatile organic compound concentration in real time. A correlation analysis was conducted between the real-time concentration data of each monitoring point and the potential odor generation coefficient calculated for the corresponding area to establish an actual odor diffusion model. When the real-time concentration data at a certain monitoring point continuously exceeds the safety threshold, an emergency upgrade command for the removal priority is triggered. The emergency priority increase command for waste collection is injected into the solution process of the waste collection route optimization model to recalculate and generate an emergency response-oriented dynamic scheduling route. Based on the recalculated emergency response-based dynamic scheduling path, the optimal navigation information of the waste collection vehicles is updated.

[0012] As a further aspect of the present invention, visualizing the spatial distribution matrix using a thermal rendering algorithm includes: The spatial distribution matrix is ​​normalized to convert the discard frequency and discard weight of each spatial location point into a standard value between zero and one. Establish color mapping rules to map low standard values ​​to cool tones and high standard values ​​to warm tones; Calculate the gradient rate of change between each spatial location point and its adjacent points based on the standard values ​​of each point. The thermal diffusion radius is determined based on the gradient change rate, and a color diffusion region centered on the spatial location point is generated at each spatial location point. The transparency of each color diffusion area is overlaid to generate a continuous and smooth color transition effect. The color transition effect is rendered on the electronic map base map to form a visual heatmap in which the intensity of discard is represented by the color depth.

[0013] As a further aspect of the present invention, the step of differentiating the weight change curve, calculating its instantaneous slope, and using the instantaneous slope as the real-time accumulation rate includes: Collect a sequence of weight sample values ​​uploaded by the weight sensor at fixed time intervals; The difference between two adjacent weight samples is calculated to obtain the weight change per unit time. Divide the weight change by the sampling time interval to obtain the discrete differential value of the weight change curve; A sliding window filtering algorithm is used to smooth the discrete derivative value to eliminate abrupt changes caused by sensor noise; The smoothed discrete derivative value is used as the instantaneous slope of the weight change curve at that time point; Multiply the instantaneous slope by the time unit conversion factor to convert it into the real-time cumulative rate in standard time units.

[0014] As a further aspect of the present invention, it also includes the step of generating a guidance report on reducing kitchen waste based on historical data: Regularly compile heat maps of the distribution of the aforementioned food waste components over long periods to analyze the long-term trends in the amount of food discarded by each food category. By comparing kitchen waste disposal data among different residential units with the same population structure, abnormal households with significantly higher disposal volumes than the average level were identified. For the aforementioned abnormal households, a personalized waste reduction report is generated based on their food waste disposal cycle pattern, which includes specific categories of food items that are being disposed of in excess and suggested reminders for high-frequency disposal periods. The personalized reduction report is pushed to the smart terminal of the corresponding resident via encrypted communication.

[0015] As a further aspect of the present invention, for the abnormal household, in conjunction with its food waste disposal cycle pattern, a personalized waste reduction report is generated, including specific categories of excessively discarded food items and suggested reminders for high-frequency disposal periods, including: Detailed kitchen waste composition data of the abnormal households within the statistical period are retrieved, and food items whose total discarded amount exceeds the community average level set ratio are identified and marked as specific categories of excessively discarded food items. Analyze the food waste disposal cycle patterns of the abnormal households to identify the time periods with the most concentrated disposal behavior as target intervention periods; Based on the specific categories of excessively discarded ingredients, corresponding storage and cooking suggestions are matched from a pre-set food preservation knowledge base; Based on the target intervention period, generate reminders and suggestions for checking food inventory or planning meals before the target intervention period; The personalized waste reduction report is generated by integrating the specific categories of food items to be discarded, storage suggestions, cooking suggestions, and reminders.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Spectral sensors directly collect characteristic spectral data of kitchen waste at the terminal, enabling real-time source identification of the waste's chemical composition and elevating the monitoring dimension from physical quantities to material composition. Cloud-based aggregation and analysis of massive amounts of terminal data generates a dynamic heat map reflecting the overall composition of kitchen waste in the region and its spatiotemporal variations. This allows management to continuously obtain accurate component percentage data, achieving both macro-level understanding and micro-level source tracing of waste material flow, providing direct evidence for accurately assessing sorting efficiency and processing load.

[0017] The system deeply integrates and matches a behavior cycle model based on disposal timestamps with a degradation oxygen demand (COD) prediction model based on composition and weight. Through algorithms, it simultaneously analyzes residents' behavioral patterns and the biochemical evolution of waste to predict the future environmental risk level and collection urgency of each collection point. Collection routes are dynamically generated based on this fusion prediction result, with parameters responding in real-time to changes in actual demand. Collection scheduling is transformed into following a dynamic field jointly defined by behavioral and biochemical data, achieving precise alignment of collection resources and waste treatment needs in the spatiotemporal dimensions. This reduces the environmental impact of waste accumulation while improving the overall efficiency of collection operations. Attached Figure Description

[0018] Figure 1 This is a timing diagram of the distributed intelligent data processing system for household kitchen waste described in this invention. Figure 2 A flowchart for generating a heatmap of kitchen waste composition; Figure 3 A flowchart for predicting the oxygen demand for waste degradation and the potential odor generation coefficient. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0021] See Figure 1The system comprises a data acquisition module consisting of distributed food waste data acquisition terminals deployed in each residential unit. This module collects raw food waste disposal records in real time, including spectral characteristic data of waste components, input timestamps, and weight change curves. The component analysis module receives spectral characteristic data from multiple distributed food waste data acquisition terminals and aggregates and analyzes this data using a cloud-deployed component recognition engine, ultimately generating a heatmap of food waste component distribution across households. The behavior modeling module specifically processes input timestamp data, extracting the food waste disposal cycle pattern of each residential unit through a time series analysis model. This pattern includes high-frequency disposal periods, low-frequency disposal periods, and time windows for abnormal disposal events. The load prediction module calculates the accumulation rate of food waste per unit time based on the weight change curve and, combined with the food categories identified in the food waste component distribution heatmap, predicts the oxygen demand for waste degradation and the potential odor generation coefficient for each area. The core function of the scheduling optimization module is to spatiotemporally match the food waste disposal cycle pattern with the oxygen demand for waste degradation, thereby generating a dynamic scheduling path for waste collection vehicles. This path includes collection priority, expected loading capacity, and service time window.

[0022] See Figure 2 In one embodiment of the present invention, the cloud-based component recognition engine performs noise filtering on the spectral feature data uploaded by each distributed food waste data collection terminal to extract effective spectral bands. For example, in a community containing one hundred residential units, the distributed food waste data collection terminals upload thousands of spectral records daily. Noise filtering removes invalid band data caused by ambient light interference or momentary sensor anomalies. The effective spectral bands are matched and identified with a pre-set food spectral database to determine the specific food categories and proportions in each discard event. In the data comparison, the matching and identification process compares the spectral features with the standard spectra of categories such as fruits and vegetables, meat, and grains stored in the database, and outputs specific composition results such as "60% leafy green vegetables, 30% leftover food, and 10% bones". Based on a geographic information system, the spatial location coordinates of each residential unit are converted to establish a mapping relationship between spatial location and food discard categories. For example, the unit numbers of each building in the community are converted into latitude and longitude coordinates. The frequency and weight of discarded food items of different categories at various spatial locations within a preset time period are statistically analyzed to form a spatial distribution matrix of food categories. In the example scenario, the frequency of discarding "leafy green vegetables" at each coordinate point within a week is used as the matrix element. Discard weight as matrix element ,in Represents a spatial location index. This represents an index of food categories.

[0023] In some embodiments, the spatial distribution matrix is ​​normalized to convert the discard frequency and discard weight at each spatial location point into standard values ​​between zero and one. The normalization process uses the following formula:

[0024] in: Indicates position Food categories The standard value, Indicates position Food categories The original statistical values, This represents the minimum value among all original statistical values ​​for all locations and all categories. This represents the maximum value among all original statistical values ​​for all locations and categories. A color mapping rule is established, mapping low standard values ​​to cool colors and high standard values ​​to warm colors. For example, in data comparison, a standard value of 0.2 is mapped to blue, and a standard value of 0.8 is mapped to red. The gradient rate of change between each spatial location point and its neighboring points is calculated based on the standard value of that point. The gradient rate of change is obtained by calculating the average difference between the standard values ​​of the center point and its eight surrounding neighboring points.

[0025] Optionally, the thermal diffusion radius is determined based on the gradient change rate, and a color diffusion region centered on each spatial location is generated. With gradient rate of change The gradient rate of change is inversely proportional to the heat diffusion radius; that is, the larger the gradient rate of change, the smaller the heat diffusion radius. The transparency of each color diffusion region is overlaid to generate a smooth, continuous color transition effect. During the transparency overlay process, the transparency of each color diffusion region's edge is set to increase from the center to the edge. The color transition effect is then rendered on the electronic map base, forming a visual heatmap where color intensity represents discard intensity. In the specific example, the red areas on the final heatmap represent buildings with high "meat" discard intensity, and the blue areas represent buildings with low "grain" discard intensity.

[0026] It is understandable that the color mapping rules can be further refined, assigning different hues and base colors to different food categories for easier differentiation. In some embodiments, the thermal rendering algorithm supports interactive operation, allowing users to click on specific areas of the heatmap to view detailed discarded component data comparisons for that location. For example, clicking on a red highlighted area can pop up a window displaying "This week's cumulative meat discarded is 15 kg, accounting for 40% of the total discarded amount in this area." It is also understandable that normalization processing can be performed separately on the discard frequency matrix and the discard weight matrix, and then the two standard value matrices are weighted and fused to generate a comprehensive standard value matrix for rendering.

[0027] In one embodiment of the invention, the timestamps of food waste disposal over multiple consecutive days for a single residential unit are aggregated to form a histogram of disposal time distribution in hours. For example, for residential unit A101, the system collects all food waste disposal timestamps for seven consecutive days, groups and counts these timestamps by hour, and generates a histogram of disposal time distribution covering 168 time slots. The count value of each time slot in the histogram represents the number of disposal events that occurred within the corresponding hour. Periodic analysis of the disposal time distribution histogram is performed to identify disposal peaks that occur within fixed time periods each day. The time periods corresponding to the disposal peaks are marked as high-frequency disposal periods. In data comparison, analysis of the disposal time distribution histogram of residential unit A101 reveals that the count value from 6 PM to 8 PM each day is consistently higher than other time periods, and this time period is marked as a high-frequency disposal period. Low-density analysis is performed on the disposal time distribution histogram to identify time periods with sparse or absent food waste disposal over several consecutive days. These time segments are marked as low-frequency disposal periods. For example, in the disposal time distribution histogram of residential unit A101, the count value from 2:00 AM to 6:00 AM each day is zero or close to zero for several consecutive days, and this time period is marked as a low-frequency disposal period. In some embodiments, periodic analysis uses a sliding window averaging method to compare the count values ​​of the same hour on different dates. If the count value of the same hour exceeds a set level for several consecutive days, the hour is determined to be a high-frequency disposal period. Low-density analysis calculates the proportion of days with a zero count value for each hourly time slot within the observation period. When the proportion exceeds a set threshold, it is marked as a low-frequency disposal period.

[0028] In practice, the system monitors time points in the discard time distribution histogram that deviate from the historical normal pattern. When the discard frequency or single discard weight at a time point exceeds a preset threshold, the time point and its preceding and following buffers are defined as the time window of the abnormal discard event. The preset threshold is calculated using historical data; for example, an abnormal discard frequency threshold can be defined as follows:

[0029] in: Indicates the discard frequency threshold. This represents the average discard frequency for the same historical period. The standard deviation of the frequency of discards during the same historical period. To adjust the coefficients. In the example scenario, historical data for residential unit A101 shows the average abandonment frequency between 9:00 AM and 10:00 AM on weekdays. 0.5 times / hour, standard deviation The value is 0.2, so the adjustment factor is taken. The threshold was calculated. If, for example, two waste disposals are detected between 9:00 AM and 10:00 AM on a given day, the time point is considered to deviate from the historical normal pattern. Optionally, a similar method can be used to set the single-disposal weight threshold, calculated based on the mean and standard deviation of historical single-disposal weights. When the weight of a particular disposal event exceeds the weight threshold, an anomaly detection is triggered. The system integrates the time windows of high-frequency disposal periods, low-frequency disposal periods, and abnormal disposal events to generate a food waste disposal cycle pattern for the residential unit. For example, a document on the food waste disposal cycle pattern for residential unit A101 is generated, listing the high-frequency disposal period as 6:00 PM to 8:00 PM daily, the low-frequency disposal period as 2:00 AM to 6:00 AM daily, and recording the time windows of abnormal disposal events, such as "October 26, 2023, 9:00 AM to 10:00 AM".

[0030] In some embodiments, the aggregation period of the disposal time distribution histogram can be adjusted to 30 days or longer to obtain a more stable periodic pattern. It is understood that the identification of high-frequency and low-frequency disposal periods can be distinguished by combining weekday and weekend patterns, establishing separate weekday and weekend food waste disposal periodic patterns. Optionally, the time window definition for abnormal disposal events considers not only frequency and weight but also abrupt changes in discarded components as an auxiliary judgment condition, such as a sudden appearance of a large number of uncommon food categories in the discarded food waste within a certain time period. It is understood that the output format of the food waste disposal periodic pattern includes a structured data list or a visual timeline chart for subsequent module calls or user viewing.

[0031] See Figure 3In one embodiment of the present invention, a sequence of weight sampling values ​​uploaded by a weight sensor at fixed time intervals is collected. In the example scenario, the weight sensor built into the distributed food waste data acquisition terminal deployed in residential unit C305 uploads weight sampling values ​​at 5-minute intervals, forming a weight sampling value sequence that changes over time. For example, the sequence segment is [time T0: 1.2 kg, time T1: 1.5 kg, time T2: 2.3 kg]. The difference between two adjacent weight sampling values ​​is calculated to obtain the weight change per unit time. According to the above sequence segment, the weight change from time T0 to time T1 is 0.3 kg, and the weight change from time T1 to time T2 is 0.8 kg. Dividing the weight change by the sampling time interval yields the discrete derivative of the weight change curve. The sampling time interval is 5 minutes, or 300 seconds. The discrete derivative from time T1 to time T2 is calculated to be 0.8 kg / 300 seconds = 0.00267 kg / second. A sliding window filtering algorithm is used to smooth the discrete derivative values, eliminating abrupt changes caused by sensor noise. The algorithm employs a moving average with a window size of 5, taking the arithmetic mean of the discrete derivative values ​​at the current time and the previous four time points as the smoothed output value. This smoothed discrete derivative value is used as the instantaneous slope of the weight change curve at that time point. In a specific data comparison, the unsmoothed discrete derivative value at time T2 is 0.00267 kg / s, which may be corrected to 0.0024 kg / s after sliding window filtering. This corrected value is used as the instantaneous slope of the weight change curve at time T2. The instantaneous slope is multiplied by a time unit conversion factor to convert it to a real-time accumulation rate in standard time units. For example, multiplying the instantaneous slope of 0.0024 kg / s by 3600 seconds / hour yields a real-time accumulation rate of 8.64 kg / hour.

[0032] In some embodiments, the average cumulative rate per unit time is obtained by averaging multiple real-time cumulative rates within a preset time period. The preset time period is set to 6 hours. A total of 72 real-time cumulative rates are collected every 5 minutes within this time period. The arithmetic mean of these values ​​is calculated as the average cumulative rate over 6 hours. The calculation formula is expressed as follows:

[0033] in: Indicates the average cumulative rate. This indicates the total number of real-time cumulative rate data points within a preset time period. Indicates the first The real-time cumulative rate is calculated from each data point. It is a data point index.

[0034] In practice, a potential odor generation coefficient is calculated based on typical fermentation products and volatile organic compound (VOC) release rates corresponding to the main ingredient category, combined with ambient temperature data. Typical fermentation products such as ammonia and hydrogen sulfide corresponding to the "fruit and vegetable mixture," along with their unit mass release rate parameters at different temperatures, are obtained from a pre-set parameter database. Simultaneously, real-time ambient temperature data (e.g., 25 degrees Celsius) uploaded by temperature sensors deployed in the specific area is acquired. Based on the ambient temperature data and release rate parameters, a weighted calculation model is used to derive the potential odor generation coefficient. This model is the sum of the products of the release rate of each odor component and an ambient temperature correction factor. Optionally, the ambient temperature correction factor is obtained from a pre-set temperature-release rate correction table using linear interpolation. It is understood that the window size of the sliding window filtering algorithm can be adjusted based on sensor accuracy and historical noise levels. In some embodiments, the standard unit mass degradation oxygen demand (ODD) parameter and VOC release rate parameter can be further subdivided based on the freshness or rough pretreatment state of the ingredients.

[0035] In one embodiment of the present invention, the end time of the high-frequency disposal period in the food waste disposal cycle pattern of each residential unit is extracted. For example, if the high-frequency disposal period is extracted as 18:00-20:00 daily from the food waste disposal cycle pattern of residential unit E01, then the end time is 20:00; if the high-frequency disposal period is extracted as 17:30-19:30 daily from the food waste disposal cycle pattern of residential unit E02, then the end time is 19:30. The end time of the high-frequency disposal period is combined with the real-time calculated oxygen demand for waste degradation in the corresponding area to form a collection node with time attribute and collection urgency attribute. The collection urgency attribute is quantified and determined by the magnitude of the oxygen demand for waste degradation; the higher the oxygen demand for waste degradation, the higher the collection urgency attribute value. Referring to Table 1, an exemplary data comparison containing five collection nodes is shown.

[0036] Table 1: Example Attribute Table of Nodes to be Cleaned

[0037] Based on the geographical coordinates of all nodes to be cleared, their urgency attributes, and vehicle capacity constraints, a waste collection route optimization model is constructed. The model aims to minimize the completion time of all nodes and prioritize services for nodes with high urgency. The objective function of the waste collection route optimization model is as follows: It can be represented as:

[0038] in: This represents the overall optimization objective value. This indicates the total number of nodes to be cleared. Indicates from node Drive to the node The time required For 0-1 decision variables, Represents a node The urgency attribute value of waste collection. For 0-1 decision variables, This is the urgency weighting coefficient.

[0039] In some embodiments, a wireless odor sensor network is deployed at preset waste collection points to monitor ammonia, hydrogen sulfide, and total volatile organic compound (TVOC) concentrations in real time. The real-time concentration data from each monitoring point is correlated with the potential odor generation coefficient calculated for the corresponding area to establish an actual odor diffusion model. The correlation analysis uses linear regression to establish a mathematical relationship between real-time concentration and multiple variables such as the potential odor generation coefficient, ambient temperature, and wind speed. When the real-time concentration data at a monitoring point continuously exceeds a safety threshold, an emergency priority adjustment command is triggered. For example, if monitoring point S2 is associated with node N2, and the ammonia concentration monitored by S2 exceeds the safety threshold of 20 ppm for three consecutive sampling cycles, an emergency priority adjustment command for node N2 is generated. This emergency priority adjustment command is injected into the solution process of the waste transport path optimization model, recalculating and generating an emergency response-oriented dynamic scheduling path. The injection process manifests as a temporary and significant increase in the priority of the corresponding node. Urgency value of waste collection Based on the recalculated emergency response-oriented dynamic scheduling route, the optimal navigation information of the waste collection vehicles is updated, and the updated route node sequence and service time window are sent to the smart terminal of the waste collection vehicles via wireless communication.

[0040] Optionally, the data sampling frequency of the wireless odor sensor network can be dynamically adjusted based on historical odor generation coefficients, with higher sampling frequencies set for areas with high potential odor generation coefficients. It is understood that the constraints of the waste collection route optimization model also include hard or soft limitations on vehicle load capacity and node service time windows. In some embodiments, the actual odor diffusion model can further integrate weather forecast data to predict odor diffusion trends over the next few hours and pre-adjust collection priorities. Optionally, emergency-response dynamic scheduling route calculation can enable backup collection vehicles dedicated to handling emergency tasks. It is understood that the quantification of collection urgency attributes can use the normalized value of waste degradation oxygen demand (ODD) or a comprehensive calculation combining historical collection intervals.

[0041] In one embodiment of the invention, a heat map of food waste composition distribution over a long period is periodically compiled to analyze the long-term trend of waste volume for each food category. The system compiles the heat map of food waste composition distribution on a monthly basis, overlays and analyzes the heat map data for twelve consecutive months, and calculates the slope of the waste volume change for each food category over time. For example, in the annual analysis of "Happy Community," the waste volume of "leafy green vegetables" shows an upward trend in summer months, while the waste volume of "staple grains" shows an upward trend in winter months. By comparing food waste waste data under the same population structure in different residential units, abnormal households with waste volumes significantly higher than the average level are identified. The system first groups residential units according to the population information registered with the property management, for example, all households registered as three people are grouped together. Then, the average daily waste waste weight and standard deviation of all households in the group are calculated. When the average daily waste waste weight of a household continuously exceeds "group average + 2 times the standard deviation," the household is marked as an abnormal household. In a specific example, there are 50 households in the group registered as three people, and the average daily waste waste weight of the group is calculated. kilograms, standard deviation kilograms, set deviation threshold Then determine the threshold. The average daily amount of waste discarded by resident H05 over 30 consecutive days was 3.8 kg, exceeding the threshold for judgment. Therefore, resident H05 was identified as an abnormal resident.

[0042] Detailed food waste composition data for abnormal households within the statistical period was retrieved to identify food items whose total discarded amount exceeded the community average by a set percentage. These were then marked as specific categories of excessively discarded food items. For the identified abnormal household H05, detailed food waste composition data for the most recent 30 days was retrieved, and the percentage of each type of food item discarded was calculated relative to the average discarded weight for the same population group within the community. When the percentage of a certain food item category exceeded a preset threshold, it was marked. Data showed that household H05 discarded 15 kg of "fruit," while the community average for "fruit" discarded was 8 kg, representing a percentage of 187.5%. Therefore, "fruit" was marked as a specific category of excessively discarded food item. The food waste disposal cycle pattern of abnormal households was analyzed to identify the time period with the most concentrated discarding behavior as the target intervention period. The food waste disposal cycle pattern document for household H05 was retrieved, showing that its high-frequency discarding period was from 20:00 to 22:00 daily. This period was determined as the target intervention period for reducing the amount of food discarded by household H05. Based on the specific category of excessively discarded ingredients, the system matches corresponding storage and cooking suggestions from a pre-built food preservation knowledge base. The food preservation knowledge base stores data in a structured format. The system matches multiple suggestions based on the "fruit" category. For example, storage suggestions include "bananas, apples and other fruits should be stored separately to delay ripening" and "berry fruits should be dried with kitchen paper and refrigerated after purchase". Cooking suggestions include "slightly overripe fruits can be used to make fruit smoothies or for baking".

[0043] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A distributed intelligent data processing system for household kitchen waste, characterized in that, The system includes: The data acquisition module is a distributed kitchen waste data acquisition terminal deployed in each residential unit. It collects raw kitchen waste disposal records in real time. The raw kitchen waste disposal records include spectral characteristic data of waste components, input timestamps, and weight change curves. The component analysis module, based on the spectral feature data collected by multiple distributed kitchen waste data collection terminals, performs aggregate analysis through a cloud-based component recognition engine to generate a heat map of kitchen waste component distribution across households. The behavior modeling module uses the input timestamp data to construct a time series analysis model and extract the food waste disposal cycle pattern of each residential unit. The food waste disposal cycle pattern includes high-frequency disposal periods, low-frequency disposal periods, and time windows of abnormal disposal events. The load prediction module calculates the accumulation rate of kitchen waste per unit time based on the weight change curve, and predicts the oxygen demand for waste degradation and the potential odor generation coefficient in each area by combining the food categories in the kitchen waste component distribution heat map. The scheduling optimization module performs spatiotemporal matching between the kitchen waste disposal cycle pattern and the oxygen demand for waste degradation to generate a dynamic scheduling path for waste collection vehicles. The dynamic scheduling path includes collection priority, expected loading capacity, and service time window.

2. The distributed intelligent data processing system for household kitchen waste according to claim 1, characterized in that, The spectral feature data collected by multiple distributed kitchen waste data collection terminals is aggregated and analyzed by a cloud-based component recognition engine to generate a heat map of kitchen waste component distribution across households, including: The heat map of food waste composition distribution marks the areas of high discard intensity for different food categories; The spectral feature data uploaded by each distributed kitchen waste data acquisition terminal is subjected to noise filtering to extract the effective spectral bands; The effective spectral bands are matched and identified with a preset food spectral database to determine the specific food categories and proportions in each discard event. Based on the geographic information system, the spatial location coordinates of each residential unit are converted, and a mapping relationship between spatial location and food waste categories is established. The frequency and weight of discarded food items of different categories at each spatial location within a preset time period are statistically analyzed to form a spatial distribution matrix of food categories. The spatial distribution matrix is ​​visualized using a thermal rendering algorithm to generate a heat map of the food waste component distribution, where color intensity represents the intensity of waste disposal.

3. The distributed intelligent data processing system for household kitchen waste according to claim 1, characterized in that, The step of constructing a time series analysis model using the input timestamp data to extract the food waste disposal cycle pattern of each residential unit includes: The timestamps of the input over multiple consecutive days for a single residential unit are aggregated to form a histogram of the discard time distribution in hours; Periodic analysis is performed on the discard time distribution histogram to identify discard peaks that occur within a fixed time period each day, and the time periods corresponding to the discard peaks are marked as high-frequency discard periods; Low-density analysis was performed on the histogram of the time distribution of waste disposal to identify time periods in which food waste disposal was rare or absent over several consecutive days, and these time segments were marked as low-frequency disposal periods. Monitor the time points in the discard time distribution histogram that deviate from the historical normal pattern. When the discard frequency or single discard weight at the time point exceeds a preset threshold, define the time point and its buffer zones before and after it as the time window of the abnormal discard event. By integrating the time windows of the high-frequency disposal period, low-frequency disposal period, and abnormal disposal events, the kitchen waste disposal cycle pattern of the residential unit is generated.

4. The distributed intelligent data processing system for household kitchen waste according to claim 1, characterized in that, The calculation of the accumulation rate of kitchen waste per unit time based on the weight change curve, combined with the food categories in the kitchen waste component distribution heat map, and the prediction of the oxygen demand for waste degradation and the potential odor generation coefficient in each area, includes: The weight change curve is differentiated to calculate its instantaneous slope, and the instantaneous slope is used as the real-time accumulation rate. The average cumulative rate per unit time is obtained by averaging the multiple real-time cumulative rates within a preset time period. Based on the dominant food category in a specific area of ​​the heat map of food waste composition, query the standard unit mass degradation oxygen demand parameter of the dominant food category. The average cumulative rate per unit time is multiplied by the standard oxygen demand per unit mass of waste to calculate the oxygen demand for waste degradation in a specific area. Based on the typical fermentation products and volatile organic compound release rates corresponding to the main food categories, and combined with ambient temperature data, the potential odor generation coefficient is calculated.

5. A distributed intelligent data processing system for household kitchen waste according to claim 4, characterized in that, The step of matching the kitchen waste disposal cycle pattern with the oxygen demand for waste degradation in time and space to generate a dynamic scheduling path for waste collection vehicles includes: Extract the end time of the high-frequency disposal period in the food waste disposal cycle pattern of each residential unit; The end time of the high-frequency discarding period is combined with the real-time calculated oxygen demand for waste degradation in the corresponding area to form a pending collection node with time attribute and collection urgency attribute. Based on the geographical coordinates of all nodes to be cleared, the urgency of clearing, and vehicle capacity constraints, a clearing route optimization model is constructed. The proposed waste disposal route optimization model aims to minimize the waste disposal completion time for all nodes and prioritize high-urgency nodes. Solve the waste disposal route optimization model to output the dynamic scheduling route, which includes the vehicle departure order, the access node sequence, and the service time window of each node.

6. A distributed intelligent data processing system for household kitchen waste according to claim 5, characterized in that, It also includes a dynamic adjustment step for collection priority based on real-time odor concentration monitoring: A wireless odor sensor network is deployed at pre-designated garbage collection points to monitor ammonia concentration, hydrogen sulfide concentration, and total volatile organic compound concentration in real time. A correlation analysis was conducted between the real-time concentration data of each monitoring point and the potential odor generation coefficient calculated for the corresponding area to establish an actual odor diffusion model. When the real-time concentration data at a certain monitoring point continuously exceeds the safety threshold, an emergency upgrade command for the removal priority is triggered. The emergency priority increase command for waste collection is injected into the solution process of the waste collection route optimization model to recalculate and generate an emergency response-oriented dynamic scheduling route. Based on the recalculated emergency response-based dynamic scheduling path, the optimal navigation information of the waste collection vehicles is updated.

7. A distributed intelligent data processing system for household kitchen waste according to claim 2, characterized in that, The visualization of the spatial distribution matrix using a thermal rendering algorithm includes: The spatial distribution matrix is ​​normalized to convert the discard frequency and discard weight of each spatial location point into a standard value between zero and one. Establish color mapping rules to map low standard values ​​to cool tones and high standard values ​​to warm tones; Calculate the gradient rate of change between each spatial location point and its adjacent points based on the standard values ​​of each point. The thermal diffusion radius is determined based on the gradient change rate, and a color diffusion region centered on the spatial location point is generated at each spatial location point. The transparency of each color diffusion area is overlaid to generate a continuous and smooth color transition effect. The color transition effect is rendered on the electronic map base map to form a visual heatmap in which the intensity of discard is represented by the color depth.

8. A distributed intelligent data processing system for household kitchen waste according to claim 4, characterized in that, The step of differentiating the weight change curve, calculating its instantaneous slope, and using the instantaneous slope as the real-time accumulation rate includes: Collect a sequence of weight sample values ​​uploaded by the weight sensor at fixed time intervals; The difference between two adjacent weight samples is calculated to obtain the weight change per unit time. Divide the weight change by the sampling time interval to obtain the discrete differential value of the weight change curve; A sliding window filtering algorithm is used to smooth the discrete derivative value to eliminate abrupt changes caused by sensor noise; The smoothed discrete derivative value is used as the instantaneous slope of the weight change curve at that time point; Multiply the instantaneous slope by the time unit conversion factor to convert it into the real-time cumulative rate in standard time units.

9. A distributed intelligent data processing system for household kitchen waste according to claim 1, characterized in that, It also includes the step of generating a guidance report on reducing food waste based on historical data: Regularly compile heat maps of the distribution of the aforementioned food waste components over long periods to analyze the long-term trends in the amount of food discarded by each food category. By comparing kitchen waste disposal data among different residential units with the same population structure, abnormal households with significantly higher disposal volumes than the average level were identified. For the aforementioned abnormal households, a personalized waste reduction report is generated based on their food waste disposal cycle pattern, which includes specific categories of food items that are being disposed of in excess and suggested reminders for high-frequency disposal periods. The personalized reduction report is pushed to the smart terminal of the corresponding resident via encrypted communication.

10. A distributed intelligent data processing system for household kitchen waste according to claim 9, characterized in that, For the aforementioned households exhibiting abnormal behavior, a personalized waste reduction report is generated based on their food waste disposal cycle pattern. This report includes specific categories of excessively discarded food items and suggested reminders for high-frequency disposal periods. Detailed kitchen waste composition data of the abnormal households within the statistical period are retrieved, and food items whose total discarded amount exceeds the community average level set proportion are identified and marked as specific categories of excessively discarded food items. Analyze the food waste disposal cycle patterns of the abnormal households to identify the time periods with the most concentrated disposal behavior as target intervention periods; Based on the specific categories of excessively discarded ingredients, corresponding storage and cooking suggestions are matched from a pre-set food preservation knowledge base; Based on the target intervention period, generate reminders and suggestions for checking food inventory or planning meals before the target intervention period; The personalized waste reduction report is generated by integrating the specific categories of food items to be discarded, storage suggestions, cooking suggestions, and reminders.

Images