Operation state monitoring method and monitoring platform for intelligent coffee machine

Abstract

The invention provides a running state monitoring method and monitoring platform for an intelligent coffee machine, and relates to the technical field of running state monitoring, and the method comprises the steps: connecting a built-in sensor of the coffee machine, and obtaining a self-sensing state parameter; acquiring external monitoring data through external monitoring equipment of the coffee machine; performing internal data grid mapping according to the self-sensing state parameters, and constructing an internal sensing data network; performing alignment segmentation on external monitoring data, and establishing internal and external response data pairs; operating state recognition is conducted, an operating state recognition result is obtained, and coffee machine regulation and control feedback is conducted. The technical problems that in the prior art, due to the fact that operation state monitoring mostly adopts single-point data collection or passive alarm, the timeliness of state monitoring is weak, and the user experience and the service life of the coffee machine are affected are solved, intelligent recognition and regulation and control feedback of the operation state of the coffee machine are achieved by combining external operation monitoring with self-inspection parameters, and the user experience is improved. And the operation efficiency of the coffee machine is improved.

Description

Technical Field

[0001] The present application relates to the technical field of operation status monitoring, and in particular to an operation status monitoring method and monitoring platform for a smart coffee machine. Background Art

[0002] Smart coffee machines use built-in sensors to automatically adjust various parameters in the coffee-making process based on user needs. However, existing smart coffee machines primarily rely on single-point data collection and passive alarms for operational status monitoring. These typically use independent sensors to monitor specific parameters, such as temperature, pressure, or liquid level. This makes it difficult to comprehensively analyze multi-dimensional data, and relying solely on a single data source prevents a comprehensive understanding of the coffee machine's overall operational status. Due to the lack of integrated and collaborative analysis of data from multiple sensors, it is impossible to identify the relationships between different sensors, making it difficult to accurately determine the operating status of the equipment, significantly reducing the timeliness and accuracy of monitoring. Existing methods for monitoring the operational status of smart coffee machines are mostly passive and lack timeliness, meaning alarms are only issued after a fault occurs. This prevents many potential faults or performance degradation issues from being identified and predicted in a timely manner, impacting the user experience and the lifespan of the smart coffee machine.

[0003] In summary, the existing technology has a technical problem that the operating status monitoring is mostly single-point data collection or passive alarm, resulting in poor timeliness of status monitoring, affecting user experience and the life of the coffee machine. Summary of the Invention

[0004] The purpose of this application is to provide an operation status monitoring method and monitoring platform for a smart coffee machine, so as to solve the technical problem in the prior art that the operation status monitoring is mostly single-point data collection or passive alarm, resulting in poor timeliness of status monitoring, affecting user experience and the life of the coffee machine.

[0005] In view of the above problems, the present application provides an operating status monitoring method and monitoring platform for a smart coffee machine.

[0006] In a first aspect, the present application provides an operating status monitoring method for a smart coffee machine, which is implemented by an operating status monitoring platform for the smart coffee machine, wherein the operating status monitoring method for the smart coffee machine includes: connecting to a built-in sensor of the coffee machine to obtain self-sensing state parameters; obtaining external monitoring data through an external monitoring device of the coffee machine; performing internal data grid mapping according to the self-sensing state parameters to construct an internal perception data network; using the internal perception data network to align and segment the external monitoring data to establish internal and external response data pairs to reflect the corresponding relationship between internal parameters and external states at the same time or in the same brewing stage; performing operating status identification according to the internal and external response data pairs to obtain an operating status identification result, and performing coffee machine control feedback according to the operating status identification result.

[0007] Optionally, the self-sensing state parameters include sensor signals from a heater, an electric pump, a motor, a flow detector, a temperature sensor, a liquid level sensor, and a vibration sensor.

[0008] Optionally, the external monitoring data includes: user operation behavior, liquid discharge image sequence, sound feedback signal, and environmental state parameters.

[0009] Optionally, based on the time series, the self-sensing state data of each dimension within every N time points are grouped into data segments; according to the time sliding window, a feature time axis is constructed based on the data segments to capture short-term trends; tensor construction processing is performed on each data segment to obtain a third-order tensor; the third-order tensor is tensor expanded to construct an internal feature map to reflect the temporal and spatial synergistic relationship between multiple parameters, and the internal perception data network is obtained.

[0010] Optionally, the structure of the third-order tensor is: sensor dimension × time dimension × sample number dimension. The third-order tensor is expanded along the sensor dimension to generate a two-dimensional feature matrix, in which the rows represent each sensor channel and the columns represent the corresponding time series data; the two-dimensional feature matrix is input into the feature modeling module to construct the internal feature map, and the feature modeling module is used to extract spatial collaborative features and time evolution trends from the time series data of multiple sensor channels.

[0011] Optionally, the self-sensing state parameters corresponding to each sensor channel in the third-order tensor have a one-to-one mapping relationship with the internal components of the coffee machine, wherein the arrangement order of the sensor channels is set according to the physical connection order of the internal components in heating, pumping, brewing, and water circuits, and is used to generate structural dependencies in tensor expansion and graph construction.

[0012] Optionally, based on the changing trends of each state parameter in the internal perception data network, the key time nodes of the internal state change are identified, and the continuous time series is divided into multiple internal state segments; according to the start and end time of the internal state segments, the external monitoring data with timestamps are aligned and segmented, and the external monitoring data segments corresponding to each internal state segment are extracted; each group of internal state segments is bound to the corresponding external monitoring data segments to construct the internal and external response data pairs.

[0013] Optionally, internal and external state matching is performed based on the internal and external response data pair; when the matching degree does not reach a preset threshold, an abnormal operation state is generated.

[0014] Optionally, feature encoding is performed on the external monitoring data segments in the internal and external response data pairs, and image features, sound features or user operation features are extracted to generate a structured external observation feature vector; external feature prediction is performed based on the internal state segments in the internal and external response data pairs, and the external performance features generated in the state are predicted to obtain an external prediction feature vector; the external prediction feature vector and the structured external observation feature vector are used to calculate the matching degree to obtain the internal and external state matching degree, which is used to measure the degree of consistency between the internal and external states.

[0015] In a second aspect, the present application further provides an operation status monitoring platform for an intelligent coffee machine, which is used to execute an operation status monitoring method for an intelligent coffee machine as described in the first aspect, wherein the operation status monitoring platform for an intelligent coffee machine includes: an internal monitoring module, which is used to connect to the built-in sensor of the coffee machine to obtain self-sensing state parameters; an external monitoring module, which is used to obtain external monitoring data through the external monitoring device of the coffee machine; an internal perception module, which is used to perform internal data grid mapping according to the self-sensing state parameters and construct an internal perception data network; an internal and external correspondence module, which is used to use the internal perception data network to align and segment the external monitoring data and establish internal and external response data pairs to reflect the correspondence between internal parameters and external states at the same time or in the same brewing stage; and an operation status identification module, which is used to perform operation status identification according to the internal and external response data pairs, obtain operation status identification results, and perform coffee machine control feedback according to the operation status identification results.

[0016] One or more technical solutions provided in this application have at least the following beneficial effects:

[0017] By connecting to the coffee machine's built-in sensors, self-sensing state parameters are acquired; external monitoring data is acquired through the coffee machine's external monitoring equipment; internal data grid mapping is performed based on the self-sensing state parameters to construct an internal sensing data network; the external monitoring data is aligned and segmented using the internal sensing data network to establish internal and external response data pairs that reflect the correspondence between the internal parameters and the external state at the same moment or in the same brewing phase; operating state identification is performed based on the internal and external response data pairs to obtain operating state identification results, and coffee machine control feedback is provided based on the operating state identification results. In other words, by acquiring self-sensing state parameters through the built-in sensors, constructing an internal sensing data network, aligning and segmenting the data with the external monitoring data, and performing operating state identification based on the internal and external response data pairs to obtain operating state identification results, comprehensive monitoring and optimized control of the coffee machine's operating state are achieved, thereby improving the coffee machine's operating efficiency.

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0020] Figure 1 This is a flow chart of a method for monitoring the operating status of a smart coffee machine.

[0021] Figure 2 This application provides a structural diagram of an operation status monitoring platform for a smart coffee machine.

[0022] Explanation of the accompanying drawings: internal monitoring module 11, external monitoring module 12, internal perception module 13, internal and external correspondence module 14, operation status identification module 15. DETAILED DESCRIPTION

[0023] This application provides an operating status monitoring method and platform for smart coffee machines, resolving the existing technical issues of poor timeliness, which impacts user experience and the lifespan of the coffee machine due to single-point data collection or passive alarms. By acquiring self-sensing state parameters through built-in sensors, building an internal perception data network, and aligning and segmenting the data with external monitoring data, the operating status is identified based on the internal and external response data to obtain an operating status identification result. This enables comprehensive monitoring and optimized regulation of the coffee machine's operating status, improving its operating efficiency.

[0024] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0025] For example 1, please refer to the attached Figure 1 The present application provides a method for monitoring the operating status of a smart coffee machine. The method is executed by an operating status monitoring platform for a smart coffee machine. The method specifically includes the following steps:

[0026] S100: Connect to the built-in sensor of the coffee machine to obtain self-sensing state parameters.

[0027] The self-sensing state parameters include sensor signals from a heater, an electric pump, a motor, a flow detector, a temperature sensor, a liquid level sensor, and a vibration sensor.

[0028] Specifically, built-in sensors refer to various sensors installed within the coffee machine, used to monitor the operating status of its various components in real time. By connecting multiple sensors to the coffee machine and acquiring their signals in real time, the machine's operating status is collected and monitored. Each sensor operates within its specific functional scope, forming a complete operating status sensing module. The coffee machine is equipped with a variety of sensors, such as heaters, electric pumps, motors, flow meters, temperature sensors, liquid level sensors, and vibration sensors. By establishing connections with the coffee machine's built-in sensors and acquiring their sensor signals, the self-sensing status parameters are derived.

[0029] A temperature sensor monitors the heater's operating temperature in real time. For example, when a coffee machine heats water, the sensor continuously monitors the temperature to ensure it reaches the set point. Assuming the target temperature is 92°C, the temperature sensor's measurement should be close to this temperature range, with an allowable error of ±0.1°C. A liquid level sensor monitors the water tank's liquid level. Suppose the sensor detects a drop to 30% and prompts the user to add water. A flow sensor monitors liquid flow. Suppose the sensor detects a flow rate of 8 ml / s, while the target flow rate is 9-11 ml / s. The flow sensor provides real-time feedback on any deviations and adjusts the electric pump's operating status. By connecting to built-in sensors and collecting real-time self-sensing state parameters, the coffee machine's operating status can be comprehensively monitored, allowing for the timely identification of potential equipment failures, such as heater overheating or insufficient water pump pressure.

[0030] S200: Obtaining external monitoring data through an external monitoring device of the coffee machine.

[0031] The external monitoring data includes: user operation behavior, liquid discharge image sequence, sound feedback signal, and environmental state parameters.

[0032] Specifically, the smart coffee machine's external monitoring devices—various monitoring devices installed outside or around the machine—are used to monitor the machine's external environment and user behavior. These external monitoring devices do not directly participate in the machine's core functions, but they can provide external information related to the machine's operating status. External monitoring data captured by these devices reflects the machine's surroundings or the user's interactions with it, providing a comprehensive understanding of the machine's operating status. This external monitoring data includes user actions, liquid discharge image sequences, audio feedback signals, and environmental state parameters. User actions are actions or inputs that the user interacts with the machine, such as pressing different buttons, selecting different coffee modes, or adjusting water volume, temperature, or concentration. Liquid discharge image sequences are image sequences captured by cameras or visual sensors during the coffee machine's dispensing process, helping to determine the quality and flow rate of the coffee. Audio feedback signals are acoustic signals generated during the machine's operation, such as the sound of the pump or water flowing. These acoustic signals are collected using microphones and other devices and used to analyze the machine's operating status. Environmental state parameters are various data related to the machine's operating environment, such as room temperature, humidity, and air pressure.

[0033] Users interact with the coffee machine through the touchscreen, buttons, or voice assistant. Every time a user performs an action (such as selecting espresso or adjusting the water temperature), the action is recorded by external monitoring equipment. For example, when a user presses the espresso mode button, the time and type of operation are recorded and converted into data for processing. A camera installed at the coffee machine's outlet captures the flow of coffee liquid in real time, monitoring the flow rate, color, and foam layer, reflecting the coffee's concentration, quality, and proper function. A microphone installed near the coffee machine captures the sound generated during operation. Different sound frequencies and amplitudes can reflect the operating status of various components within the coffee machine. For example, when the water pump is operating normally, it produces a specific sound frequency. If the water pump malfunctions or is short of water, the sound may become abnormal (such as a rattling sound or abnormal vibration). The sound signals can be used to determine whether the device is experiencing an abnormality. Environmental monitoring sensors (such as temperature sensors, humidity sensors, and air pressure sensors) are used to monitor environmental changes around the coffee machine in real time. For example, excessive room temperature or humidity can affect the coffee machine's heating effect or coffee extraction quality.

[0034] By combining external monitoring data and built-in sensor data, a comprehensive understanding of the working status of the coffee machine can be achieved, which not only relies on the hardware status of the machine, but also takes into account the influence of user behavior and environmental factors. By obtaining environmental status data in real time, the coffee machine can automatically adjust working parameters (such as temperature control or extraction time) according to changes in temperature, humidity, etc., to ensure that the best coffee quality can be provided in different environments. For example, the external monitoring data collected by the coffee machine through external monitoring equipment includes: the user selects the espresso mode at 10s, the flow rate is 5ml / s, the color is dark brown, the water pump emits a normal sound signal with a frequency of 50Hz, the room temperature is 22℃, and the humidity is 45%.

[0035] S300: Perform internal data grid mapping according to the self-sensing state parameters to construct an internal sensing data network.

[0036] Furthermore, the present application S300 includes:

[0037] Based on the time series, the self-sensing state data of each dimension within every N time points are grouped into data segments; according to the time sliding window, a feature time axis is constructed based on the data segments to capture short-term trends; tensor construction processing is performed on each data segment to obtain a third-order tensor; the third-order tensor is tensor expanded to construct an internal feature map to reflect the temporal and spatial synergistic relationship between multiple parameters, thereby obtaining the internal perception data network.

[0038] Specifically, a time series is a series of data points arranged in chronological order, typically used to illustrate the trend of a variable over time. A time series describes the temporal changes in sensor data (such as temperature, pressure, and current) within a coffee machine. Following the time series, the self-sensing state data for each dimension at every N time points is organized into data segments.

[0039] Smart coffee machines have multiple built-in sensors, such as heater temperature sensors, electric pump pressure sensors, motor current sensors, and flow meters. These sensors collect data at a specific sampling frequency. For example, data is collected every 100ms. The sensor data at every N time points (e.g., 10 time points) is combined into a data segment. If each time point has M sensor data (such as temperature, pressure, current, etc.), then the dimensions of each data segment will be N×M, that is, the number of time points × the number of sensor channels. Each data segment can be understood as a two-dimensional array with an N×M structure. The N time points represent the length of the sliding window, for example, 100ms sampling × 10 points = 1s of data.

[0040] By sliding the window's extreme values along the time axis, data segments are slid along, gradually generating new data segments. For example, assuming a sliding window size of 100ms, a new data segment is generated each time the window slides until the complete time series data is processed. Data segments from multiple sliding windows are continuously generated over time to form a feature timeline, which can capture short-term trends and changes. The feature timeline is composed of data segments generated from multiple time windows. These data segments can be used to capture the behavioral characteristics of each coffee machine sensor over a period of time, helping to understand short-term trends.

[0041] Each data segment is processed by tensor construction to obtain a third-order tensor. A tensor is a representation of a multidimensional array, and the rank of a tensor indicates the number of dimensions. A third-order tensor is a three-dimensional array consisting of three dimensions: sensor dimension × time dimension × sample number dimension. This means that each dimension of the tensor represents different information: the first dimension is the sensor type (such as temperature, pressure, current, etc.); the second dimension is time (such as each sampling time point); and the third dimension is the sample number of the data segment (i.e., the position of the sliding window). The sensor dimension represents different sensor channels (such as temperature, pressure, current, flow, etc.), with each sensor providing a data point; the time dimension represents each time point within the time window, with each time point corresponding to a sample value of the sensor. The sample number dimension represents the position of each data segment in the sliding window (i.e., the data segment obtained after each slide).

[0042] For example, assuming that the following sensor data are sampled (taking temperature, pressure, current, flow, and water level as examples), the data at each time point are as follows: the temperature at time point t1 is 90°C, the pressure is 3.0 Bar, the current is 1.2 A, the flow rate is 100 ml / s, and the water level is 50%; the temperature at time point t2 is 91°C, the pressure is 3.1 Bar, the current is 1.3 A, the flow rate is 101 ml / s, and the water level is 51%; the temperature at time point t3 is 92°C, the pressure is 3.2 Bar, the current is 1.4 A, the flow rate is 102 ml / s, and the water level is 52%; the temperature at time point t4 is 93°C, the pressure is 3.1 Bar, the current is 1.5 A, the flow rate is 103 ml / s, and the water level is 53%; the temperature at time point t5 is 94°C, the pressure is 3.0 Bar, the current is 1. The temperature at time point t6 is 95°C, the pressure is 2.9Bar, the current is 1.7A, the flow rate is 105ml / s, and the water level is 55%; the temperature at time point t7 is 96°C, the pressure is 2.8Bar, the current is 1.8A, the flow rate is 106ml / s, and the water level is 56%; the temperature at time point t8 is 97°C, the pressure is 2.7Bar, the current is 1.9A, the flow rate is 107ml / s, and the water level is 57%; the temperature at time point t9 is 98°C, the pressure is 2.6Bar, the current is 2.0A, the flow rate is 108ml / s, and the water level is 58%; the temperature at time point t10 is 99°C, the pressure is 2.5Bar, the current is 2.1A, the flow rate is 109ml / s, and the water level is 59%. For every 100ms sampling, a sliding window size of 10 is selected to construct data segments, ultimately building a 10 × 5 matrix. As the window slides, 100 data segments are generated, forming a 100 × 10 × 5 third-order tensor.

[0043] The third-order tensor is expanded along the sensor dimension to generate a two-dimensional feature matrix, where the rows represent the sensor channels and the columns represent the data at the corresponding time points. The two-dimensional feature matrix is input into the feature modeling module to construct an internal feature map. The feature modeling module can use a convolutional neural network structure, a graph neural network structure, or a principal component feature extraction algorithm to analyze the response coupling pattern between different sensor channels. The internal feature map is used to reflect the temporal and spatial synergistic relationship between multiple parameters and obtain an internal perception data network. The internal feature map is a structured representation of a multidimensional state, which is used to characterize the synergistic relationship, abnormal deviation trend, and system stability index of parameters such as temperature, pressure, current, and flow within a sliding time window.

[0044] By constructing the multiple sensor data of the coffee machine into a third-order tensor according to the time and space dimensions, and constructing the internal feature map through tensor expansion and feature modeling modules, the short-term trends, time series changes and spatial coordination relationships between sensors can be efficiently captured, thereby improving the accuracy of identifying the operating status of the coffee machine.

[0045] Furthermore, the present application further comprises the following steps:

[0046] The structure of the third-order tensor is: sensor dimension × time dimension × sample number dimension. The third-order tensor is expanded along the sensor dimension to generate a two-dimensional feature matrix, in which rows represent each sensor channel and columns represent corresponding time series data; the two-dimensional feature matrix is input into the feature modeling module to construct the internal feature map. The feature modeling module is used to extract spatial collaborative features and time evolution trends from the time series data of multiple sensor channels.

[0047] The self-sensing state parameters corresponding to each sensor channel in the third-order tensor have a one-to-one mapping relationship with the internal components of the coffee machine. The arrangement order of the sensor channels is set according to the physical connection sequence of the internal components in the heating, pumping, brewing, and water circuits, and is used to generate structural dependencies during tensor expansion and graph construction.

[0048] Specifically, a third-order tensor is a data representation with a three-dimensional structure, sensor dimension × time dimension × sample number dimension, where the sensor dimension represents the number of sensor channels, including data from different sensors (such as temperature, pressure, current, etc.), and each sensor is sampled at each time point; the time dimension represents different time points within the time window; and the sample number dimension represents different data segments of the time series (i.e., different positions of the sliding window).

[0049] Tensor expansion converts high-dimensional tensor data into lower-dimensional matrix data. This involves expanding the third-order tensor according to the sensor dimensions (i.e., different sensors) to generate a two-dimensional feature matrix. The rows of the two-dimensional feature matrix represent the sensor channels, and the columns represent the corresponding time series data. In other words, the multiple sensor data points at each time point are arranged into a row, and the data from all time points form the columns of the matrix.

[0050] The two-dimensional feature matrix is input into the feature modeling module, which uses a convolutional neural network (CNN) architecture, a graph neural network (GNN) architecture, or a principal component feature extraction algorithm to analyze the response coupling patterns between different sensor channels. Convolutional neural networks extract local features from data and are suitable for extracting spatial features and temporal evolution patterns from time series data. Graph neural networks analyze the synergistic relationships between different sensors, especially when complex relationships exist between data, helping to capture spatial and temporal coupling patterns. Principal component analysis can be used for dimensionality reduction, extracting key features from data, and reducing data redundancy.

[0051] The feature modeling module extracts representative and discriminative features from raw data, identifying potential patterns, trends, or anomalies within the data and enabling more accurate decision-making. Convolutional neural networks extract both local and global features from the data through convolutional, pooling, and fully connected layers. In time series data, convolution operations can help capture local trends and relationships between adjacent time points. A two-dimensional feature matrix is passed as input to the convolutional neural network. Convolutional layers extract spatial features from the time series data, pooling layers reduce the data dimensionality, and finally, a fully connected layer outputs the classification or regression results.

[0052] When processing data with a graph structure, graph neural networks can propagate information within the graph through graph convolution operations, capturing the interactions between nodes (sensors) and their impact on the system state. By treating each sensor as a node in the graph and the interdependencies between sensors as edges, graph neural networks extract spatial collaborative features between different sensor channels through multiple layers of graph convolution.

[0053] Principal component analysis (PCA) extracts key features by identifying the direction of maximum variance in the data. This helps remove redundant information and noise, thereby reducing model complexity. A two-dimensional feature matrix is input into the PCA algorithm, the covariance matrix is calculated, and the principal components are extracted. The first few principal components are then selected as new feature vectors. In the feature modeling module, whether using a convolutional neural network architecture, a graph neural network architecture, or a PCA feature extraction algorithm, the goal is to analyze the data response coupling patterns between different sensor channels and capture the temporal correlations and spatial synergies between sensor data.

[0054] Convolutional neural network structures use multiple convolution kernels to slide over data, capturing relationships between local time points. For example, changes in a sensor at a certain point in time may affect changes in other sensors at the same or similar time points. Convolutional neural network structures can identify these local correlations and effectively extract features. When graph neural networks construct graphs, sensor data is transferred through graph convolution, effectively capturing complex dependencies between different sensors. For example, if a sensor's reading is abnormal, it may affect the readings of other sensors. Graph neural networks can capture these dependencies through the graph structure. Principal component analysis identifies the direction of the largest changes in the data and projects the data into a low-dimensional space using principal components. This is used to identify the main patterns of coordinated changes between sensors, thereby removing noise and retaining the most important features.

[0055] After extracting the spatial collaborative features and temporal evolution trends, an internal feature map is generated, showing the coupling mode and temporal evolution of each sensor channel. The internal feature map is a structured representation of multi-dimensional states, which is used to characterize the collaborative relationship, abnormal deviation trend and system stability index of parameters such as temperature, pressure, current, flow rate within the sliding time window. The feature modeling module is used to extract spatial collaborative features and temporal evolution trends from the time series data of multiple sensor channels. Spatial collaborative features refer to the interaction relationship between different sensors, that is, how different sensor data change in coordination. The temporal evolution trend refers to the changing pattern of the system state over time, which helps to understand the dynamic behavior of the coffee machine and predict its future state, and determine whether the coffee machine is currently operating abnormally, has aging components, or requires maintenance.

[0056] The self-sensing state parameters corresponding to each sensor channel in the third-order tensor have a one-to-one mapping relationship with an internal component of the coffee machine, namely, each sensor channel corresponds to an internal component of the coffee machine, such as a heater, pump, motor, flow meter, etc. The order of the sensor channels is determined by the physical connection sequence of the internal components in the heating, pumping, brewing, and water circuits, and is used to generate structural dependencies during tensor expansion and graph construction. In other words, the order of the sensor channels is determined based on the physical connection sequence of the internal components in the heating, pumping, brewing, and water circuits, reflecting the physical connections and interactions between the internal components of the coffee machine in order to generate structural dependencies during tensor expansion and graph construction.

[0057] By expanding the third-order tensor along the sensor dimension and constructing a feature matrix, combined with the spatiotemporal collaborative feature extraction of the feature modeling module, the changing trend of the coffee machine status can be accurately captured to determine whether the coffee machine is currently operating abnormally, has aging components, or requires maintenance.

[0058] S400: aligning and segmenting the external monitoring data using the internal perception data network to establish internal and external response data pairs to reflect the corresponding relationship between internal parameters and external states at the same moment or in the same brewing stage.

[0059] Furthermore, the present application S400 includes:

[0060] Based on the changing trends of various state parameters in the internal perception data network, the key time nodes of the internal state changes are identified, and the continuous time series is divided into multiple internal state segments; according to the start and end times of the internal state segments, the external monitoring data with timestamps are aligned and segmented, and the external monitoring data segments corresponding to each internal state segment are extracted; each group of internal state segments is bound to the corresponding external monitoring data segments to construct the internal and external response data pairs.

[0061] Specifically, the internal perception data network reflects the temporal and spatial coordination relationship between multiple parameters inside the coffee machine. According to the change trend of each state parameter in the internal perception data network, the change of each state parameter of the coffee machine (such as temperature, pressure, current, etc.) over time is analyzed. Identify the points where the state changes significantly in the time series. For example, the temperature suddenly rises from 80°C to 90°C, or the pressure suddenly rises from 1.2 bar to 2 bar. These change points may represent turning points of the equipment. Determine the key nodes based on the change trend of the data. For example, the temperature gradually rises within a certain range and exceeds the preset working range. This may be a state change node that needs attention. Through change trend identification, multiple key time nodes are found from the continuous time series. These nodes usually represent changes in equipment status, conversion of working modes, or potential failure points. Key time nodes are usually related to turning points in the operation of the equipment, such as temperature increase, pressure change, flow fluctuation, etc.

[0062] According to the identified key time nodes, the entire continuous time series is divided into multiple smaller internal state segments. The start and end time of each internal state segment is obtained, the external monitoring data with timestamps are aligned and segmented, and the external monitoring data segments corresponding to each internal state segment are extracted. According to the start and end time, the external monitoring data with timestamps are aligned and segmented, and the external monitoring data segments corresponding to each internal state segment are extracted. That is to say, according to the start and end points, the corresponding time is determined in the timestamp and segmented to obtain multiple external monitoring data segments corresponding to the multiple internal state segments, and the time of the two is consistent. Alignment segmentation is to align the external monitoring data with timestamps with the internal state segments, and extract the corresponding external monitoring data according to the time range of the internal state segments.

[0063] Each set of internal state fragments is bound to the corresponding external monitoring data fragment to construct an internal-external response data pair. Each set of internal-external response data pairs contains information on the synchronous changes in the internal state and external monitoring signals, and can demonstrate the relationship between external factors and the internal operating state of the coffee machine under specific working conditions. The internal-external response data pair is a paired data set consisting of each internal state fragment and the corresponding external monitoring data fragment, reflecting the mutual relationship between the internal state and the external monitoring signal. For example, the internal-external response data pair includes: the internal state fragment is the temperature rising from 80°C to 85°C and the pressure rising from 1.5 bar to 2 bar; the external monitoring data fragment is the user pressing the heating button and the liquid flow rate increases. By constructing the internal-external response data pair, the relationship between the internal state of the coffee machine and the external environment is determined, helping to improve the accuracy of condition monitoring. The construction of the internal-external response data pair can reveal which external factors (such as user operations or environmental changes) may cause abnormal operation or failure of the coffee machine, thereby providing early warning and making necessary adjustments or maintenance.

[0064] S500: performing operation status identification according to the internal and external response data to obtain an operation status identification result, and performing coffee machine control feedback according to the operation status identification result.

[0065] Furthermore, the present application S500 includes:

[0066] Based on the internal and external response data pairs, internal and external status matching is performed; when the matching degree does not reach a preset threshold, an abnormal operation state is generated.

[0067] Furthermore, the present application further comprises the following steps:

[0068] Perform feature encoding on the external monitoring data segments in the internal and external response data pairs, extract image features, sound features or user operation features, and generate a structured external observation feature vector; perform external feature prediction based on the internal state segments in the internal and external response data pairs, predict the external performance features generated in this state, and obtain an external prediction feature vector; use the external prediction feature vector and the structured external observation feature vector to calculate the matching degree, and obtain the internal and external state matching degree, which is used to measure the degree of consistency between the internal and external states.

[0069] Specifically, feature encoding is performed on the external monitoring data segments in the internal and external response data pairs, and image features, sound features, or user operation features are extracted, and these raw data are converted into structured feature vectors. For image data, a convolutional neural network is used to extract the spatial features of the image. For example, features such as the shape and color changes of liquid flow are extracted through a convolutional neural network and converted into a vector representation. Input image data (such as an image sequence of coffee liquid flow); use a convolutional neural network to convolve the image and extract the feature map in the convolution layer; flatten the convolved feature map into a vector to obtain the final image feature vector.

[0070] For sound data, Fourier transforms are used to extract audio features, including the frequency content, volume, and pitch of the audio signal. Audio data, typically a time series waveform, is acquired from a microphone. A short-time Fourier transform (STFT) is performed on the audio signal to obtain a time-frequency feature representation. The MFCC algorithm is used to extract the audio's spectral features, which reflect characteristics such as pitch, volume, and frequency. These audio features are then converted into a structured feature vector.

[0071] User operation features can be encoded based on user input behaviors (such as button clicks and setting adjustments). For example, the temporal changes in user operations and click frequency can be recorded and converted into structured feature vectors. Operational behavior data can be collected from the coffee machine's user interface (such as the touch screen and physical buttons), and user behaviors (such as the number of clicks and the time interval between operations) can be converted into numerical features. Based on the characteristics of the time series, statistical methods (such as mean, variance, and maximum value) can be used to extract the behavioral characteristics of user operations.

[0072] Generate a structured external observation feature vector based on image features, sound features, or user operation features. Encode the image, sound, and user operation features. These feature vectors are concatenated to form a comprehensive feature vector, the structured external observation feature vector.

[0073] Collect internal and external response data pairs over multiple time periods, including internal state segments and corresponding external monitoring data segments. Extract key features (such as temperature, pressure, and current) from the internal state segments, and simultaneously encode external performance features (such as images, sounds, and user actions) into structured feature vectors. Train the model using supervised learning methods, using internal state segments as input features and external feature vectors as target outputs. Use the trained model to predict new internal state segments and obtain the external features generated by that state.

[0074] The trained model predicts external image features based on the internal state, which may be image feature vectors of the liquid flow state. Based on the internal state fragment, the sound feature vectors that may be generated when the coffee machine is operating are predicted, reflecting the audio changes during the pumping and heating processes. Based on historical data, the user's possible operational behavior characteristics (such as adjusting the concentration, starting and stopping the button, etc.) under specific internal states are predicted. All predicted external features (image, sound, user operation) are spliced according to certain rules to obtain a comprehensive external prediction feature vector. For example, assuming the internal state fragment is temperature = 85°C, pressure = 12 bar, and flow rate = 1.3 L / min, the model predicts: image feature vector: a vector of length 4096, representing the image features of the liquid flow; sound feature vector: a vector of length 40, representing the spectral characteristics of the pumping sound; user operation feature vector: a vector of length 3, representing the characteristics of user operations (such as concentration adjustment).

[0075] The similarity or matching degree between the external prediction feature vector and the actual structured external observation feature vector collected is calculated and measured using Euclidean distance. If the external prediction features are highly consistent with the actual observation features, the model prediction is accurate; otherwise, adjustments and optimization can be made. The Euclidean distance between the external prediction feature vector and the actual observation feature vector is calculated to obtain the internal and external state matching degree, which is used to measure the degree of consistency between the internal and external states. Before calculating the matching degree, the input feature vector is usually standardized to eliminate the influence of different dimensions so that different types of features have the same weight when calculating the matching degree. By selecting an appropriate similarity or distance measurement method, the external prediction feature vector and the actual observation feature vector are calculated to obtain the matching degree between them.

[0076] Based on business needs and experience, a matching threshold (preset threshold) is set. If the matching degree exceeds the preset threshold, the internal and external states are considered to be highly consistent; otherwise, the consistency is poor. If the matching degree falls below the set threshold, it indicates a device failure, sensor anomaly, or external interference. This indicates an abnormal operating state and generates a corresponding operating state identification result.

[0077] The operating status recognition result is calculated based on the matching degree, indicating the current operating status of the coffee machine. If the matching degree falls below a preset threshold, the machine is identified as abnormal. The operating status recognition result can be binary (normal / abnormal) or can include more specific categories (such as mild abnormality, severe abnormality, etc.).

[0078] Different feedback measures are taken for different abnormal conditions. For mild abnormalities, it may only be necessary to issue a warning to the user, prompting the user to pay attention to the operating status of certain equipment. For moderate or severe abnormalities, certain control parameters of the equipment are automatically adjusted to avoid equipment damage. In the case of severe abnormalities, the user is reminded to perform maintenance or replace parts, such as replacing the pumping system or cleaning the sensor. For normal conditions, no excessive intervention is required, and only the existing operating parameters need to be maintained. When an abnormality occurs, detailed data of the abnormality is recorded, including time, abnormality type, and the degree of match between prediction and actual. By collecting and analyzing continuously occurring abnormalities, the system optimizes its own state recognition and feedback model, improving the accuracy of abnormality recognition and feedback response speed in the future.

[0079] Based on the identified operating status (normal or abnormal), the system adjusts the coffee machine's control parameters (such as heater power, pump pressure, and flow rate) to ensure optimal operation. For abnormal conditions, regulatory feedback may include alarms, maintenance reminders, or direct automatic adjustments. Real-time status recognition and feedback ensure that operating parameters are adjusted promptly when an abnormality occurs, preventing more serious failures and improving overall operational efficiency.

[0080] In summary, the method for monitoring the operating status of a smart coffee machine provided in this application has the following beneficial effects:

[0081] By connecting to the coffee machine's built-in sensors, self-sensing state parameters are acquired; external monitoring data is acquired through the coffee machine's external monitoring equipment; internal data grid mapping is performed based on the self-sensing state parameters to construct an internal sensing data network; the external monitoring data is aligned and segmented using the internal sensing data network to establish internal and external response data pairs that reflect the correspondence between the internal parameters and the external state at the same moment or in the same brewing phase; operating state identification is performed based on the internal and external response data pairs to obtain operating state identification results, and coffee machine control feedback is provided based on the operating state identification results. In other words, by acquiring self-sensing state parameters through the built-in sensors, constructing an internal sensing data network, aligning and segmenting the data with the external monitoring data, and performing operating state identification based on the internal and external response data pairs to obtain operating state identification results, comprehensive monitoring and optimized control of the coffee machine's operating state are achieved, thereby improving the coffee machine's operating efficiency.

[0082] In the second embodiment, based on the same inventive concept as the method for monitoring the operating status of a smart coffee machine in the first embodiment, the present application further provides an operating status monitoring platform for a smart coffee machine, as shown in the attached diagram. Figure 2 , the operation status monitoring platform for the smart coffee machine includes:

[0083] The internal monitoring module 11 is used to connect to the built-in sensors of the coffee machine to obtain self-sensing state parameters; the external monitoring module 12 is used to obtain external monitoring data through the coffee machine's external monitoring equipment; the internal perception module 13 is used to perform internal data grid mapping based on the self-sensing state parameters and construct an internal perception data network; the internal and external correspondence module 14 is used to use the internal perception data network to align and segment the external monitoring data and establish internal and external response data pairs to reflect the correspondence between internal parameters and external states at the same time or in the same brewing stage; the operating state identification module 15 is used to identify the operating state based on the internal and external response data pairs, obtain an operating state identification result, and provide coffee machine control feedback based on the operating state identification result.

[0084] Furthermore, the internal monitoring module 11 in the operation status monitoring platform for the smart coffee machine is further used to:

[0085] The self-sensing state parameters include sensor signals from a heater, an electric pump, a motor, a flow detector, a temperature sensor, a liquid level sensor, and a vibration sensor.

[0086] Furthermore, the external monitoring module 12 in the operation status monitoring platform for the smart coffee machine is further used to:

[0087] The external monitoring data includes: user operation behavior, liquid discharge image sequence, sound feedback signal, and environmental state parameters.

[0088] Furthermore, the internal perception module 13 in the operation status monitoring platform for the smart coffee machine is further used to:

[0089] Based on the time series, the self-sensing state data of each dimension within every N time points are grouped into data segments; according to the time sliding window, a feature time axis is constructed based on the data segments to capture short-term trends; tensor construction processing is performed on each data segment to obtain a third-order tensor; the third-order tensor is tensor expanded to construct an internal feature map to reflect the temporal and spatial synergistic relationship between multiple parameters, thereby obtaining the internal perception data network.

[0090] Furthermore, the internal perception module 13 in the operation status monitoring platform for the smart coffee machine is further used to:

[0091] The structure of the third-order tensor is: sensor dimension × time dimension × sample number dimension. The third-order tensor is expanded along the sensor dimension to generate a two-dimensional feature matrix, in which rows represent each sensor channel and columns represent corresponding time series data; the two-dimensional feature matrix is input into the feature modeling module to construct the internal feature map. The feature modeling module is used to extract spatial collaborative features and time evolution trends from the time series data of multiple sensor channels.

[0092] Furthermore, the internal perception module 13 in the operation status monitoring platform for the smart coffee machine is further used to:

[0093] The self-sensing state parameters corresponding to each sensor channel in the third-order tensor have a one-to-one mapping relationship with the internal components of the coffee machine. The arrangement order of the sensor channels is set according to the physical connection sequence of the internal components in the heating, pumping, brewing, and water circuits, and is used to generate structural dependencies during tensor expansion and graph construction.

[0094] Furthermore, the internal and external correspondence module 14 in the operation status monitoring platform for the smart coffee machine is further used to:

[0095] Based on the changing trends of various state parameters in the internal perception data network, the key time nodes of the internal state changes are identified, and the continuous time series is divided into multiple internal state segments; according to the start and end times of the internal state segments, the external monitoring data with timestamps are aligned and segmented, and the external monitoring data segments corresponding to each internal state segment are extracted; each group of internal state segments is bound to the corresponding external monitoring data segments to construct the internal and external response data pairs.

[0096] Furthermore, the operation status identification module 15 in the operation status monitoring platform for the smart coffee machine is further used to:

[0097] Based on the internal and external response data pairs, internal and external status matching is performed; when the matching degree does not reach a preset threshold, an abnormal operation state is generated.

[0098] Furthermore, the operation status identification module 15 in the operation status monitoring platform for the smart coffee machine is further used to:

[0099] Perform feature encoding on the external monitoring data segments in the internal and external response data pairs, extract image features, sound features or user operation features, and generate a structured external observation feature vector; perform external feature prediction based on the internal state segments in the internal and external response data pairs, predict the external performance features generated in this state, and obtain an external prediction feature vector; use the external prediction feature vector and the structured external observation feature vector to calculate the matching degree, and obtain the internal and external state matching degree, which is used to measure the degree of consistency between the internal and external states.

[0100] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The operation status monitoring method for a smart coffee machine and the specific examples in Example 1 are also applicable to the operation status monitoring platform for a smart coffee machine in this embodiment. Through the above detailed description of the operation status monitoring method for a smart coffee machine, those skilled in the art can clearly understand the operation status monitoring platform for a smart coffee machine in this embodiment. Therefore, for the sake of brevity of the specification, it will not be described in detail here.

[0101] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0102] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A method for monitoring the operating status of an intelligent coffee machine, characterized in that: include: Connect to the built-in sensor of the coffee machine to obtain the self-sensing state parameters; Obtain external monitoring data through the external monitoring device of the coffee machine; Perform internal data grid mapping according to the self-sensing state parameters to construct an internal sensing data network; Utilizing the internal perception data network to align and segment the external monitoring data, and establish internal and external response data pairs to reflect the corresponding relationship between internal parameters and external states at the same moment or in the same brewing stage; An operating state is identified based on the internal and external response data to obtain an operating state identification result, and coffee machine control feedback is performed based on the operating state identification result.

2. The method for monitoring the operating status of an intelligent coffee machine according to claim 1, wherein: The self-sensing state parameters include sensor signals from a heater, an electric pump, a motor, a flow detector, a temperature sensor, a liquid level sensor, and a vibration sensor.

3. The method for monitoring the operating status of an intelligent coffee machine according to claim 2, wherein: The external monitoring data includes: user operation behavior, liquid discharge image sequence, sound feedback signal, and environmental state parameters.

4. The method for monitoring the operating status of an intelligent coffee machine according to claim 1, wherein: Performing internal data grid mapping according to the self-sensing state parameters to construct an internal sensing data network includes: Based on the time series, the self-sensing state data of each dimension within every N time points are combined into data segments; According to the time sliding window, a feature time axis is constructed based on the data segments to capture short-term trends; Perform tensor construction processing on each data segment to obtain a third-order tensor; The third-order tensor is tensor-expanded to construct an internal feature map for reflecting the temporal and spatial synergistic relationship between multiple parameters, thereby obtaining the internal perception data network.

5. The method for monitoring the operating status of an intelligent coffee machine according to claim 4, wherein: The third-order tensor is expanded to construct an internal feature map to reflect the temporal and spatial synergy between multiple parameters, including: The structure of the third-order tensor is: sensor dimension × time dimension × sample number dimension. The third-order tensor is expanded along the sensor dimension to generate a two-dimensional feature matrix, where rows represent each sensor channel and columns represent corresponding time series data. The two-dimensional feature matrix is input into a feature modeling module to construct the internal feature map. The feature modeling module is used to extract spatial collaborative features and time evolution trends from the time series data of multiple sensor channels.

6. The method for monitoring the operating status of an intelligent coffee machine according to claim 5, wherein: The self-sensing state parameters corresponding to each sensor channel in the third-order tensor have a one-to-one mapping relationship with the internal components of the coffee machine. The arrangement order of the sensor channels is set according to the physical connection sequence of the internal components in the heating, pumping, brewing, and water circuits, and is used to generate structural dependencies during tensor expansion and graph construction.

7. The method for monitoring the operating status of an intelligent coffee machine according to claim 4, wherein: Using the internal perception data network to align and segment the external monitoring data to establish internal and external response data pairs, including: Based on the change trend of each state parameter in the internal perception data network, identifying the key time nodes of the internal state change, and dividing the continuous time series into multiple internal state segments; Align and segment the external monitoring data with timestamps according to the start and end times of the internal state segments, and extract the external monitoring data segments corresponding to each internal state segment; Each set of internal state fragments is bound to the corresponding external monitoring data fragments to construct the internal and external response data pairs.

8. The method for monitoring the operating status of an intelligent coffee machine according to claim 7, wherein: Performing operation status identification based on the internal and external response data to obtain an operation status identification result includes: Performing internal and external state matching based on the internal and external response data pairs; When the matching degree does not reach the preset threshold, an abnormal operation state is generated.

9. The method for monitoring the operating status of an intelligent coffee machine according to claim 8, characterized in that: Based on the internal and external response data pairs, internal and external state matching is performed, including: Performing feature coding on the external monitoring data segments in the internal and external response data pairs, extracting image features, sound features, or user operation features, and generating a structured external observation feature vector; Performing external feature prediction based on the internal state segment in the internal and external response data pair, predicting the external performance features generated in the state, and obtaining an external prediction feature vector; The external prediction feature vector and the structured external observation feature vector are used to perform matching calculation to obtain the internal and external state matching degree, which is used to measure the consistency degree of the internal and external states.

10. An operating status monitoring platform for an intelligent coffee machine, characterized in that: The steps for implementing the method for monitoring the operating status of a smart coffee machine according to any one of claims 1 to 9, wherein the operating status monitoring platform for a smart coffee machine comprises: Internal monitoring module, used to connect to the built-in sensor of the coffee machine to obtain self-sensing state parameters; An external monitoring module, used to obtain external monitoring data through an external monitoring device of the coffee machine; An internal perception module, configured to perform internal data grid mapping according to the self-sensing state parameters and construct an internal perception data network; an internal-external correspondence module, configured to align and segment the external monitoring data using the internal perception data network, and establish internal-external response data pairs to reflect the correspondence between internal parameters and external states at the same moment or in the same brewing stage; An operating state identification module is used to identify the operating state according to the internal and external response data, obtain an operating state identification result, and perform coffee machine control feedback according to the operating state identification result.

Images