Automatic tea making machine control method and system based on multi-source data

Abstract

This application relates to the field of automatic tea brewing machine control, and particularly to an automatic tea brewing machine control method and system based on multi-source data. The method includes: fitting multi-source data of historical tea bag brewing throughout the entire cycle to obtain the concentration decay curve corresponding to each historical tea bag to characterize the change trend of tea bag concentration with cumulative water output; extracting the inflection point distribution characteristics in the concentration decay curve; clustering based on the characteristics of the concentration decay curve and the inflection point distribution characteristics to generate multiple clusters and a benchmark concentration decay curve corresponding to each cluster; calculating the fit degree and individual matching degree to determine the comprehensive score; and controlling the tea brewing machine according to the control parameter corresponding to the highest comprehensive score. This application can achieve intelligent brewing with high precision and high stability throughout the entire cycle.

Description

Technical Field

[0001] This application relates to the field of automatic tea brewing machine control, and in particular to an automatic tea brewing machine control method and system based on multi-source data. Background Technology

[0002] The core control objective of an automatic tea maker is to achieve precise control of tea concentration and consistent flavor throughout the brewing process by adjusting key parameters such as brewing water temperature, brewing time, and single-batch water volume. For built-in tea bags that support multiple infusions, the concentration decays due to complex non-linear characteristics during repeated brewing. Since a single data dimension cannot fully characterize this decay pattern, it is necessary to integrate multi-source data, including cumulative water volume, tea concentration, brewing parameters, and tea bag type, to support dynamic and precise parameter control.

[0003] However, existing automatic tea brewing machines typically use fixed brewing parameters and rely on a single, static concentration decay curve as the control basis, failing to personalize the process by considering the characteristics of the tea bag type and the user's brewing preferences. This results in an inability to accurately track the actual concentration decay dynamics of the tea bag during multiple consecutive brewing processes, leading to large fluctuations in brewed concentration, poor flavor consistency, and difficulty in maintaining brewing quality over the long term. Therefore, there is an urgent need for an automatic tea brewing machine control method based on multi-source data fusion, with dynamic matching and adaptive update capabilities, to achieve intelligent brewing with high precision and high stability throughout the entire process. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides an automatic tea brewing machine control method and system based on multi-source data.

[0005] Firstly, this application provides an automatic tea brewing machine control method based on multi-source data, employing the following technical solution: An automatic tea brewing machine control method based on multi-source data includes: fitting multi-source data of historical tea bag brewing throughout the entire cycle to obtain the concentration decay curve corresponding to each historical tea bag to characterize the change trend of tea bag concentration with cumulative water output; extracting inflection point distribution features; performing clustering based on the features of the concentration decay curve and the inflection point distribution features to generate multiple clusters and a baseline concentration decay curve corresponding to each cluster; during real-time brewing, retrieving the corresponding cluster according to the current tea bag category; if it is the first brewing, filtering historical first brewing data based on the user-set single water output and adjusting the control parameters according to the filtering results. Set the control parameters for the current brewing process; if it is not the first brewing, fit a real-time concentration decay curve segment based on previous brewing data, calculate the fit between the curve segment and the baseline concentration decay curve of each cluster, select several suitable clusters, and calculate the individual matching degree between the tea bag and the historical tea bag within the suitable cluster. Combine the fit degree and the individual matching degree to determine the comprehensive score, and control the tea machine according to the control parameter corresponding to the highest comprehensive score; after each brewing is completed, if the actual tea concentration deviates from the baseline concentration beyond the preset allowable range, re-determine the control parameters for the next brewing.

[0006] Optional multi-source data includes: cumulative water output, single water output, brewing control parameters, tea concentration, tea bag type information, and inflection points indicating concentration mutations; brewing control parameters include brewing water temperature and brewing time; tea bag type information includes tea variety and weight.

[0007] Optionally, the multi-source data of historical tea bag brewing throughout the entire cycle is fitted to obtain the concentration decay curve corresponding to each historical tea bag. This includes the following steps: performing normalization preprocessing on the collected multi-source data of historical tea bag brewing throughout the entire cycle, and sorting it according to the brewing cycle time sequence; based on the time sequence data of cumulative water output and tea concentration of each historical tea bag category, the Gauss-Newton method is used to fit and generate a nonlinear concentration decay curve that fully represents the change of concentration with cumulative water output.

[0008] Optionally, extract the inflection point distribution features, and perform clustering based on the features of the concentration decay curve and the inflection point distribution features to generate multiple clusters and the corresponding baseline concentration decay curve for each cluster. This includes the following steps: using the second-order difference method to obtain the detected inflection points in the concentration decay curve, recording the inflection point coordinates, and equally dividing the cumulative water output range of this type of tea bag into a preset number of intervals, counting the number of inflection points in each interval, and constructing an N-dimensional inflection point distribution vector according to the interval order, where the... Dimension element represents the first dimension. The number of inflection points within each interval; constructing a decay curve feature vector based on the concentration decay curve; concatenating the decay curve feature vector with the inflection point distribution vector to form a feature vector for clustering; according to the clustering algorithm, using the Euclidean distance between the feature vectors corresponding to any two historical tea bags as a similarity measure, and determining the number of clusters by the silhouette coefficient method, dividing multiple historical tea bags under the same tea bag category into several clusters.

[0009] Optionally, the baseline concentration decay curve can be set as follows: for each cluster, based on the complete concentration decay curves of all historical tea bags within the cluster, a polynomial fitting algorithm is used to fit the curves and generate the baseline concentration decay curve.

[0010] Optionally, if it is the first brewing, the control parameters for the current brewing are set based on the user-set single water output volume, filtering historical first brewing data, and the control parameters for the current brewing are set according to the control parameters corresponding to the filtering results. This includes: normalizing the user-set single water output volume, filtering brewing data in historical first brewing where the absolute value of the difference in single water output volume does not exceed a preset threshold, and taking the average value of the brewing water temperature and brewing time as the control parameters for the current brewing.

[0011] Optionally, the method for calculating the fit between the curve segment and the baseline concentration decay curve of each cluster is as follows: the real-time concentration decay curve segment obtained by fitting the current tea bag based on the previous brewing data is intercepted with the baseline concentration decay curve of any cluster within the same cumulative water output range, and this range is taken as the common range representing the intersection range of the cumulative water output range; equidistant sampling is performed within the common range to obtain several discrete points and construct a concentration sequence; the Euclidean distance is calculated and the Euclidean distance is mapped to the fit through a monotonically decreasing function.

[0012] Optionally, the method for calculating the individual matching degree includes the following steps: based on the cumulative water output of the current tea bag before this brewing, select the brewing round with the closest cumulative water output among the historical tea bags in the matching cluster as the comparison state; construct individual feature vectors for the current tea bag and the historical tea bag in their respective comparison states, wherein the individual feature vectors include at least the normalized cumulative water output, the previous tea concentration, the brewing water temperature, and the brewing time; calculate the Euclidean distance between the two individual feature vectors, and convert the Euclidean distance into an individual matching degree through a monotonically decreasing function.

[0013] Optionally, the comprehensive score can be calculated by weighting the fit of the fit clusters and the individual matching degree to obtain the comprehensive score.

[0014] Secondly, this application provides an automatic tea brewing machine control system based on multi-source data, which adopts the following technical solution: the automatic tea brewing machine control system based on multi-source data includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the automatic tea brewing machine control method based on multi-source data described above is implemented.

[0015] The beneficial effect is that the above-mentioned automatic tea maker control method based on multi-source data is used to generate a computer program and store it in the memory so that it can be loaded and executed by the processor. Thus, the system can be made according to the memory and the processor, which is convenient to use.

[0016] This application has the following technical effects: 1. By integrating the characteristics of tea bag decay curves and inflection point distributions to perform two-dimensional clustering, multiple refined clusters and corresponding baseline concentration decay curves are constructed. This breaks through the limitations of traditional single decay models, enabling control parameters to more accurately match the nonlinear concentration decay law of different tea bags throughout the entire brewing cycle. It accurately tracks the real concentration decay dynamics of tea bags, resulting in smaller fluctuations in brewing concentration, consistent flavor, and long-term maintenance of good brewing quality.

[0017] 2. Parameter determination strategies are designed for first-time brewing and non-first-time brewing respectively. In non-first-time brewing, a two-layer matching mechanism of adaptability and individual matching degree is introduced. By combining weighted candidate parameters and comprehensive score evaluation, the brewing parameters are adaptively controlled to achieve real-time, personalized and high-precision generation of control parameters.

[0018] 3. Make full use of the value of multi-source data, integrate multi-dimensional data such as cumulative water output, single water output, brewing water temperature, brewing time, tea concentration and tea bag type, and build a structured feature system to significantly improve the reliability and generalization ability of data-driven decision-making. Attached Figure Description

[0019] Figure 1 This is a flowchart of the automatic tea brewing machine control method based on multi-source data in an embodiment of this application.

[0020] Figure 2 This is a flowchart of the clustering method in the automatic tea brewing machine control method based on multi-source data in the embodiments of this application. Detailed Implementation

[0021] This application discloses an automatic tea brewing machine control method based on multi-source data, referring to... Figure 1 The process includes steps S1-S3, as detailed below: S1: Fit multi-source data of historical tea bag brewing throughout the entire cycle to obtain the concentration decay curve corresponding to each historical tea bag to characterize the change trend of tea bag concentration with cumulative water output, extract the inflection point distribution features, and perform clustering based on the features of the concentration decay curve and the inflection point distribution features to generate multiple clusters and the baseline concentration decay curve corresponding to each cluster.

[0022] In one embodiment, multi-source data stored in the data storage unit of the historical database is acquired. This multi-source data includes: cumulative water output, single water output, brewing control parameters, tea concentration, tea bag category information, and inflection points indicating concentration abrupt changes. Brewing control parameters include brewing water temperature and brewing time. Tea bag category information includes tea variety (e.g., black tea, green tea, oolong tea) and weight. Based on the combination of tea bag variety and weight, a unique category is determined for each tea bag. A unique category identifier is assigned to each category. During subsequent clustering or parameter recommendation, a pre-defined control module converts the category identifier into a one-hot encoded vector, concatenates it with other numerical features, and inputs it into the analysis model for subsequent data association and retrieval. An identification sensing subunit is set up, including a reader (e.g., an RFID reader or QR code scanner), to automatically read the attached category identifier when a tea bag is placed and transmit this identifier to the control module for retrieving historical data and the clustering model for the corresponding category.

[0023] For each unique category, multi-source data on the entire brewing cycle of multiple historical tea bags of the same category are collected. Specifically, this includes the complete brewing process records of a single historical tea bag. Each brewing data point constitutes a time-series node, and all time-series nodes are arranged sequentially according to the brewing rounds. Min-Max normalization preprocessing is performed on the raw data corresponding to each brewing round, and then the preprocessed data is sorted according to the time-series order of the brewing rounds to form a multi-source data archive for each historical tea bag.

[0024] In one example, for the qth brewing of a specific historical tea bag under any category, single brewing data is collected and preprocessed. The specific data is as follows: Cumulative water output Q1 before the qth brewing: Statistically calculated by the automatic tea machine's workbench metering unit, representing the total water output from when the tea bag was placed in the machine until the start of this brewing session. The cumulative water output Q1 before the first brewing is 0 ml. Single water output Q2 for the qth brewing: Set by the user according to their drinking needs, or the machine's default standard output (e.g., 200 ml / breath). Control parameters for the qth brewing: Including brewing water temperature T and brewing time M. Tea concentration g after the qth brewing: Obtained immediately after brewing by a concentration sensor preset at the automatic tea machine's outlet. Min-Max normalization preprocessing is then performed on the collected data to obtain single brewing data.

[0025] To obtain the concentration decay curves for each historical tea bag by fitting multi-source data on the entire brewing cycle, the following steps are taken: Normalizing the collected multi-source data on the entire brewing cycle of historical tea bags and sorting it chronologically by brewing round; using the Gauss-Newton method to fit the cumulative water output and tea concentration time-series data for each historical tea bag category, generating a nonlinear concentration decay curve that fully represents the change in concentration with the cumulative water output. Specifically, the initial parameter values ​​of the fitting polynomial are set to coefficients obtained through linear regression, and the iteration termination condition is set to the norm of the parameter update being less than... Or the rate of change of the sum of squares of residuals is less than .

[0026] For the same unique category (as defined in step S1), two sets of time-series data—cumulative water output Q1 and tea concentration g—are retrieved and sorted sequentially by brewing round. The Gauss-Newton method is used to fit the sorted data to obtain a complete concentration decay curve (non-linear decay model, polynomial decay model can be selected) for the historical tea bag, representing "cumulative water output Q1 - tea concentration g". The complete concentration decay curve has the cumulative water output Q1 on the x-axis and the tea concentration g on the y-axis, accurately reflecting the dynamic process of concentration decay of a single tea bag as the cumulative water output increases.

[0027] Extract the inflection point distribution characteristics, and perform clustering based on the characteristics of the concentration decay curve and the inflection point distribution characteristics to generate multiple clusters and a corresponding baseline concentration decay curve for each cluster. Figure 2 This includes steps S10-S11: S10: The inflection point in the concentration decay curve is obtained by using the second-order difference method. The coordinates of the inflection point are recorded. The cumulative water output range of this type of tea bag is divided into a preset number of intervals at equal intervals. The number of inflection points in each interval is counted, and an N-dimensional inflection point distribution vector is constructed in the order of the intervals.

[0028] In the N-dimensional inflection point distribution vector, the first... Dimension element represents the first dimension. The number of inflection points within each interval.

[0029] For a single complete concentration decay curve, key parameters of its nonlinear decay model are extracted to construct a feature vector of the decay curve. The inflection point distribution vector of the curve is constructed as follows: the inflection points of the complete concentration decay curve are calculated using the second-order difference method, and the coordinates (Q1, g) of each inflection point are recorded; the range of cumulative water output Q1 for this type of tea bag is divided into equal intervals to obtain a preset number (e.g., 20) of cumulative water output intervals; the number of inflection points within each cumulative water output interval is counted, and the statistical results are arranged in ascending order of the cumulative water output intervals to construct the inflection point distribution vector of the curve.

[0030] S11: Construct a decay curve feature vector based on the concentration decay curve; concatenate the decay curve feature vector with the inflection point distribution vector to form a feature vector for clustering; according to the clustering algorithm, use the Euclidean distance between the feature vectors of any two historical tea bags as a similarity measure, and determine the number of clusters by the silhouette coefficient method to divide multiple historical tea bags under the same tea bag category into several clusters.

[0031] The coefficients of the polynomial in the polynomial decay model are used as key parameters. The extracted key parameters are arranged in a fixed order to form a numerical vector, which is the characteristic vector of the decay curve of the tea bag.

[0032] The extracted decay curve feature vectors are concatenated with the inflection point distribution vectors to generate the feature vectors corresponding to each historical tea bag.

[0033] Set clustering parameters (where the number of clusters K is determined by the contour method), use the Euclidean distance between the high-dimensional feature vectors corresponding to any two historical tea bags as the similarity criterion (the smaller the Euclidean distance, the greater the similarity), perform K-means clustering to obtain multiple clusters.

[0034] The method for setting the baseline concentration decay curve is as follows: For each cluster, based on the complete concentration decay curves of all historical tea bags in the cluster, a polynomial fitting algorithm is used to uniformly fit and generate a universal baseline concentration decay curve within the cluster. The baseline concentration decay curve satisfies the one-to-one mapping relationship between "cumulative water output - tea concentration", and the average fitting error between the complete concentration decay curves of all historical tea bags in the cluster and the baseline concentration decay curve is ≤5%, ensuring that the baseline concentration decay curve can accurately characterize the common concentration decay pattern of all tea bags in the cluster.

[0035] S2: During real-time brewing, the corresponding cluster is retrieved based on the current tea bag category. If it is the first brewing, historical first brewing data is filtered based on the user-set single water output, and the control parameters for the current brewing are set according to the control parameters corresponding to the filtering results. If it is not the first brewing, the real-time concentration decay curve segment is fitted based on the previous brewing data, the fit of the curve segment with the baseline concentration decay curve of each cluster is calculated, several suitable clusters are obtained, and the individual matching degree of tea bags in the suitable clusters with historical tea bags is calculated. The comprehensive score is determined by combining the fit degree and the individual matching degree, and the tea brewing machine is controlled according to the control parameter corresponding to the highest comprehensive score.

[0036] In the curve segment of real-time concentration decay fitted based on previous brewing data, the previous brewing data should include at least the previous two brewing records of the current tea bag to ensure that the fitted curve segment has sufficient degrees of freedom to reflect the decay trend.

[0037] The automatic tea maker uses a tea bag identification unit (with a built-in RFID reader to read the information stored on the paper label attached to the surface of the tea bag) to collect the tea variety and weight of the tea bag to be brewed in real time, match it with the unique category of the tea bag, determine the corresponding unique category identifier, and retrieve all clusters corresponding to that category from the reference library based on the identifier.

[0038] If this is the first brewing, the user-set single-batch water volume is normalized, and brewing data from historical first brewings where the absolute value of the difference in single-batch water volume does not exceed a preset threshold are selected. The average of the brewing water temperature and brewing time is taken as the control parameters for the current brewing. The control parameters for the first brewing are thus determined.

[0039] Determine that the current brewing cycle is the first brew (i.e., the cumulative water output Q11 = 0ml after the tea bag is placed in the device). Obtain the user-set target single-batch water output Q21 and perform Min-Max normalization. From all retrieved clusters, extract the first brewing data of all historical tea bags within each cluster, and set a first filtering threshold (the first threshold can be set to 0.1, which is based on the standard deviation of the normalized single-batch water output in the historical data, allowing...). The fluctuation range), the screening condition is "the absolute value of the difference between the single water output of the historical first brew and the current target single water output Q21 is ≤ the first threshold". Control parameter calculation and execution: For all historical first brew data that meet the screening conditions, calculate the arithmetic mean of brewing water temperature Tavg and the arithmetic mean of brewing time Mavg respectively. Use (Tavg, Mavg) as the control parameter for this first brew and send it to the execution module of the automatic tea maker to start the first brew process.

[0040] In other embodiments, when a valid category identifier for the current tea bag cannot be identified (e.g., the tea bag is not in a preset category library): In some embodiments, if the automatic tea maker fails to read the category identifier of the tea bag (e.g., the tea bag does not have an RFID chip, the QR code is damaged, or the identifier is not in the preset category library), the system prompts the user to manually select the tea category (such as black tea, green tea, oolong tea, etc.) through the human-computer interaction interface.

[0041] The system retrieves historical tea bag data from multiple sources based on the user-selected category. It reuses the initial brewing parameter determination logic, filtering historical initial brewing data from the historical tea bag data based on the user-defined single-batch water volume. The system then sets the current brewing control parameters based on the average brewing temperature and brewing time corresponding to the filtered results. The historical tea bag data is compiled from the initial brewing data of historical tea bags within that category.

[0042] If this is not the first brewing, the method for calculating the fit between the curve segment and the baseline concentration decay curve of each cluster is as follows: The real-time concentration decay curve segment obtained by fitting the current tea bag based on previous brewing data is intersected with the baseline concentration decay curve of any cluster within the same cumulative water output range, and this range is taken as the common interval representing the intersection of cumulative water output. Within this common interval, equidistant sampling is performed to obtain several discrete points and construct a concentration sequence. The Euclidean distance is calculated and mapped to the fit using a monotonically decreasing function. Specifically, the Euclidean distance is calculated... The Euclidean distance is then mapped to the fitness level using a monotonically decreasing function. Preferably, the monotonically decreasing function is an exponential function. ,in This is an adjustment coefficient, and its value range can be... In this embodiment, we take .

[0043] In one embodiment, a second threshold is set (the second threshold can be 0.6; when there is no cluster with a value greater than the second threshold, it indicates that the current tea bag decay characteristics are significantly different from the historical database. In this case, the cluster with the highest fit is selected to ensure control continuity). Clusters with a fit greater than the second threshold are selected to form a set of matching clusters. If there is no cluster with a fit greater than the second threshold, only the cluster with the highest fit is selected as the matching cluster.

[0044] The method for calculating the individual matching degree includes the following steps: Based on the cumulative water output of the current tea bag before this brewing, select the brewing round with the closest cumulative water output among the historical tea bags in the matching cluster as the comparison state; construct individual feature vectors for the current tea bag and the historical tea bag in their respective comparison states, and the individual feature vectors should include at least the normalized cumulative water output, the previous tea concentration, the brewing water temperature, and the brewing time; calculate the Euclidean distance between the two individual feature vectors, and convert the Euclidean distance into the individual matching degree through a monotonically decreasing function.

[0045] In one embodiment, to further improve matching accuracy, the individual feature vector can be expanded to 6 dimensions, adding features of previous single outflow volume and current single outflow volume.

[0046] Real-time multi-source data acquisition and feature extraction: If the current brewing time is the t-th brewing time (t≥2), collect all data from the previous t-1 brewing times, including cumulative water output, single water output, brewing temperature, brewing time, and tea concentration data, and perform Min-Max normalization processing; fit a real-time curve segment of "cumulative water output - tea concentration"; for each cluster of the category to which the tea bag belongs, calculate the Euclidean distance between the real-time curve segment and its baseline concentration decay curve, and further obtain the fit of each cluster. Taking the cumulative water output Q1=0 as a unified starting point, divide the cluster into baseline concentration decay curve segments with the same cumulative water output, and calculate the Euclidean distance d between the real-time curve segment and the baseline concentration decay curve segment of the cluster.

[0047] In one embodiment, the mathematical expression for fitness can be: ;in, This refers to the fitness of the i-th cluster; It refers to the Euclidean distance between the real-time curve segment and the baseline concentration decay curve segment of the i-th cluster; the smaller the Euclidean distance, the more similar the two curve segments are, and the greater the fit.

[0048] For example, the individual matching degree calculation for a single historical tea bag: For each matching cluster, filter historical tea bags within the cluster—taking a single historical tea bag as an example, based on the real-time cumulative water output Q1. t Get and Q1 t For the same historical tea bag, the m-th brewing record is used to construct an individual feature vector as follows: Obtain the individual feature vectors of the historical tea bag and the current tea bag. Calculate the Euclidean distance as the individual matching degree v of this historical tea bag.

[0049] Determination of candidate control parameters for single-fit clusters: For each matching cluster, based on the individual matching degree of historical tea bags, sort them in descending order, select the top 3 historical tea bags as matching historical tea bags, and calculate the weight of each matching historical tea bag based on the individual matching degree. The candidate control parameters for this cluster are obtained by weighting the control parameters (brewing water temperature T, brewing time M) corresponding to the m-th brewing. The mathematical expression of the candidate control parameters can be: ;in, It refers to the weight of the j-th matching historical tea bag in the k-th adaptation cluster; This refers to the individual matching degree of the j-th matching historical tea bag in the k-th fitting cluster. The mathematical expression for the control parameter can be: ;in, , These refer to the brewing water temperature and brewing time (corresponding to the t-th brewing of the current tea bag) among the candidate control parameters of the k-th adaptive cluster. It refers to the weight of the j-th matching historical tea bag in the k-th adaptation cluster; , These refer to the brewing water temperature and brewing time (corresponding to the m-th brewing of the j-th matching historical tea bag) in the control parameters of the k-th matching cluster.

[0050] The overall score is calculated by weighting the fitness of the fit clusters and the individual matching degree. The overall score is then calculated for each candidate control parameter, and the candidate control parameter with the highest overall score is selected as the control parameter for this brewing process. Brewing is then initiated and real-time data is collected. ;in, This refers to the overall score of the k-th fit cluster; This refers to the fitness of the k-th fit cluster; It refers to the individual matching degree of the j-th matching historical tea bag in the k-th matching cluster.

[0051] S3: After each brewing is completed, in response to the deviation between the actual tea concentration and the reference concentration exceeding the preset allowable range, the control parameters for the next brewing are redefined.

[0052] The data update unit is set up to send the determined real-time control parameters to the execution module (heating device, water outlet control device) of the automatic tea brewing machine, start the brewing process, and simultaneously collect the data of this brewing (the cumulative water output, single water output, actual brewing water temperature, actual brewing time, and tea concentration after brewing) and update the full-cycle brewing data file of the tea bag.

[0053] Based on the deviation between the actual tea concentration of the current brew and the corresponding concentration of the matching historical tea bags, determine whether the control parameters for the next brewing need to be redefined. The specific steps are as follows: Based on the screening, the target cluster to which the control parameters for this brewing belong and three matching historical tea bags were determined. Using the current brewing round t as an index, the tea concentration collected for the m-th brewing of the matching historical tea bag was retrieved from the historical database of the target cluster, and the average value was taken as the baseline concentration. Extract the actual concentration of the tea being brewed. Calculate concentration deviation ; Set the allowable deviation range of concentration Δg0 (e.g., Δg0 = 0.05, this range is determined based on the measurement accuracy of the concentration sensor, usually 2-3 times the sensor accuracy); if Δg ≥ Δg0, update the actual tea concentration of the current tea bag and re-determine the control parameters for the (t+1)th brewing of the current tea bag; otherwise, determine that the current brewing concentration meets the requirements, and obtain the control parameters for the next brewing (t+1) based on matching historical tea bag data, specifically: obtain the target single-batch output volume set by the user for the next brewing, and calculate the cumulative output volume Q1 for the (t+1)th brewing. t+1 Locate the complete concentration decay curves of 3 matching historical tea bags and Q1 t+1 Corresponding tea concentration (g) t+1 Determine whether the tea concentration is the actual collected value (not a curve-fitted value); statistically analyze the tea concentration in g. t and g t+1 The brewing time T and brewing water temperature M are used as effective control parameters. Combined with the individual matching degree of the corresponding historical tea bags, the brewing time T and brewing water temperature M are weighted and calculated to obtain the control parameters for the (t+1)th brewing of the current tea bag.

[0054] This application also discloses an automatic tea brewing machine control system based on multi-source data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the automatic tea brewing machine control method based on multi-source data according to this application is implemented.

[0055] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0056] In this application, the aforementioned memory can be any tangible medium that contains or stores a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), etc., or any other medium that can be used to store desired information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0057] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for controlling an automatic tea maker based on multi-source data, characterized in that, The method comprises the steps of: Fitting multi-source data of historical tea bags in the whole cycle of brewing to obtain concentration decay curves corresponding to each historical tea bag to represent the change trend of tea bag concentration with cumulative water volume, extracting inflection point distribution characteristics, clustering based on the characteristics of the concentration decay curve and the inflection point distribution characteristics, and generating multiple clustering clusters and reference concentration decay curves corresponding to each clustering cluster; During real-time brewing, the corresponding clustering cluster is retrieved according to the category of the current tea bag; If it is the first brewing, the historical first brewing data is screened based on the single water volume set by the user, and the control parameters of the current brewing are set according to the control parameters corresponding to the screening result; If it is not the first brewing, a curve segment of real-time concentration decay is fitted based on the previous brewing data, the fitting degrees of the curve segment and the reference concentration decay curves of each clustering cluster are calculated, a plurality of fitting clustering clusters are screened, the individual matching degrees of the tea bags in the fitting clustering clusters and the historical tea bags are calculated, the comprehensive scores are determined in combination with the fitting degrees and the individual matching degrees, and the tea maker is controlled according to the control parameters corresponding to the highest comprehensive score. After each brewing is completed, the control parameters of the next brewing are re-determined in response to the deviation of the actual tea concentration from the reference concentration exceeding a preset allowable range.

2. The method of claim 1, wherein, The multi-source data includes cumulative water volume, single water volume, brewing control parameters, tea concentration, tea bag category information, and inflection point representing concentration mutation information. The brewing control parameters include brewing water temperature and brewing time. The tea bag category information includes tea variety and gram weight.

3. The method of claim 2, wherein, Fitting multi-source data of historical tea bags in the whole cycle of brewing to obtain concentration decay curves corresponding to each historical tea bag, comprising the steps of: Performing normalization preprocessing on the collected multi-source data of historical tea bags in the whole cycle of brewing, and sorting according to the brewing round time sequence; Based on the cumulative water volume and tea concentration time sequence data of each historical tea bag of the same tea bag category, a complete nonlinear concentration decay curve representing the change of concentration with cumulative water volume is generated by using the Gauss-Newton method fitting.

4. The method of claim 3, wherein, Extracting inflection point distribution characteristics, clustering based on the characteristics of the concentration decay curve and the inflection point distribution characteristics, and generating multiple clustering clusters and reference concentration decay curves corresponding to each clustering cluster, comprising the steps of: The second-order difference method is used to obtain the detection inflection point in the concentration attenuation curve, the inflection point coordinates are recorded, and the cumulative water discharge range of the tea bag of this category is equally divided into a preset number of intervals, the number of inflection points in each interval is counted, and an N-dimensional inflection point distribution vector is formed in order of intervals, wherein the first dimensional element represents the number of inflection points in the first interval. Constructing a decay curve feature vector according to the concentration decay curve; concatenating the decay curve feature vector and the inflection point distribution vector to form a feature vector for clustering; According to the clustering algorithm, the Euclidean distance between the feature vectors of any two historical tea bags is taken as the similarity measure, the number of clusters is determined by the silhouette coefficient method, and a plurality of historical tea bags under the same tea bag category are divided into a plurality of clustering clusters.

5. The method of claim 2, wherein, The setting method of the reference concentration decay curve is: for each clustering cluster, based on the complete concentration decay curves of all historical tea bags in the clustering cluster, a polynomial fitting algorithm is used for fitting to generate a reference concentration decay curve.

6. The method of claim 2, wherein, If it is the first brewing, the historical first brewing data is screened based on the single water set by the user, and the control parameters of the current brewing are set according to the control parameter corresponding to the screening result, which comprises: after normalizing the single water volume set by the user, the brewing data with an absolute value of the single water volume difference less than a preset threshold in the historical first brewing is screened, and the average of the brewing water temperature and the brewing time is taken as the control parameter of the current brewing.

7. The method of claim 4, wherein, The method for calculating the fit between the curve segment and the baseline concentration decay curve of each cluster is as follows: The real-time concentration decay curve segment obtained by fitting the current tea bag based on the previous brewing data is intercepted with the baseline concentration decay curve of any cluster within the same cumulative water output range, and this range is taken as the common range representing the intersection range of the cumulative water output range. Within a common interval, equidistant sampling is performed to obtain several discrete points and construct a concentration sequence; the Euclidean distance is calculated and mapped to fitness using a monotonically decreasing function.

8. The method of claim 7, wherein, The method for calculating individual compatibility includes the following steps: Based on the cumulative water output of the current tea bag before this brewing, the brewing cycle with the closest cumulative water output among the historical tea bags is selected as the comparison state within the matching cluster. Construct individual feature vectors for the current tea bag and the historical tea bag under their respective comparison states. Each individual feature vector includes at least the normalized cumulative water output, previous tea concentration, brewing water temperature, and brewing time. Calculate the Euclidean distance between the two individual feature vectors and convert the Euclidean distance into an individual matching degree using a monotonically decreasing function.

9. The multi-source data based automatic tea maker control method of claim 1, wherein, The overall score is calculated by weighting the fit of the fit clusters and the individual matching degree to obtain the overall score.

10. An automatic tea making machine control system based on multi-source data, characterized in that, include: The processor and memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the automatic tea-making machine control method based on multi-source data according to any one of claims 1-9.

Images