Intelligent energy cup food intake monitoring system and method

Abstract

The present application relates to the technical field of food intake monitoring system, in particular to an intelligent energy cup food intake monitoring system and method thereof, a multi-sensor data acquisition module; a data preprocessing module electrically connected with the multi-sensor data acquisition module; an AI analysis module data connected with the data preprocessing module; according to the liquid type and the eating behavior, combined with the preset health standard, the real-time food intake of the user is calculated; based on the real-time food intake and the user historical data, the blood glucose curve of the user is predicted; an intelligent intervention module is signal connected with the AI analysis module, whether the preset health threshold is exceeded is judged; when the judgment result exceeds the health threshold, the corresponding intervention instruction is generated; a feedback execution module is control connected with the intelligent intervention module, a data storage and communication module is data connected with the multi-sensor data acquisition module, the AI analysis module and the intelligent intervention module respectively; the data is uploaded to the remote server through the wireless network, the cloud backup and analysis of the data are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of food intake monitoring systems, in particular to a smart energy cup food intake monitoring system and method thereof. Background Art

[0002] As people increasingly prioritize healthy lifestyles, smart drinking cups are gaining popularity as a new health management tool. Traditional smart drinking cups primarily serve as reminders to drink regularly or simply record water intake. However, these products lack functionality and intelligence, failing to meet users' demands for precise health management.

[0003] Currently, the more advanced smart drinking cups on the market typically use a single liquid level sensor or weight sensor to monitor water intake. While this method can roughly record a user's water intake, it has several limitations. First, the measurement accuracy of a single sensor is limited and is easily affected by external factors such as cup tilt and temperature changes, resulting in inaccurate data. Second, these products cannot distinguish between different types of liquids and therefore cannot provide users with personalized recommendations for different beverages. Furthermore, existing products generally lack the ability to deeply analyze user drinking behavior, making it difficult to identify subtle behavioral patterns such as rapid drinking and intermittent drinking, thereby failing to provide truly personalized health guidance.

[0004] When it comes to data processing and analysis, existing technologies often use simple statistical methods or preset fixed thresholds to determine a user's hydration status. This approach ignores individual differences and fails to adapt to the physical characteristics and lifestyles of different users. Furthermore, due to a lack of advanced artificial intelligence algorithms, existing products struggle to learn and predict long-term hydration patterns, and are even unable to correlate hydration behavior with other health indicators (such as blood sugar levels) for comprehensive analysis.

[0005] Furthermore, in terms of user interaction and intervention, most smart drinking cups offer only simple drinking reminders, lacking specificity and flexibility. This single interaction method can easily lead to user neglect or resistance, making it difficult to effectively and sustainably improve drinking habits. Furthermore, existing products generally lack effective protection for user privacy, posing security risks during data transmission and storage.

[0006] In view of the above problems, there is an urgent need for a smart energy cup system that can accurately monitor drinking behavior, intelligently analyze health status, and provide personalized intervention. Summary of the Invention

[0007] This invention addresses this need by proposing an innovative smart energy cup food intake monitoring system and method. By leveraging innovative approaches such as multi-sensor fusion technology, deep learning algorithms, and personalized intervention strategies, this system effectively addresses numerous challenges in existing technologies, providing users with a comprehensive, accurate, and intelligent health management solution.

[0008] The present invention proposes a smart energy cup food intake monitoring system, comprising:

[0009] Multi-sensor data acquisition module for:

[0010] Collect the liquid level and weight information of the liquid in the cup;

[0011] Obtain RFID identification information of the cup body;

[0012] Collect user's geographic location information and environmental data;

[0013] The data preprocessing module is electrically connected to the multi-sensor data acquisition module and is used to:

[0014] Receiving various sensor data sent by the multi-sensor data acquisition module;

[0015] performing noise reduction, standardization and feature extraction processing on the sensor data;

[0016] The AI ​​analysis module is connected to the data preprocessing module and is used to:

[0017] Based on the preprocessed sensor data, using a deep learning algorithm to identify liquid type and eating behavior;

[0018] Calculate the user's real-time food intake based on the liquid type and eating behavior, combined with preset health standards;

[0019] Predicting the user's blood sugar curve based on the real-time food intake and the user's historical data;

[0020] The intelligent intervention module is connected to the AI ​​analysis module signal and is used to:

[0021] Receive food intake data and blood sugar prediction data sent by the AI ​​analysis module;

[0022] Determining whether the food intake data and blood sugar prediction data exceed a preset health threshold;

[0023] When the judgment result exceeds the health threshold, generating corresponding intervention instructions;

[0024] The feedback execution module is connected to the intelligent intervention module and is used to:

[0025] receiving an intervention instruction sent by the intelligent intervention module;

[0026] selectively executing a sound reminder, a vibration warning, or a heating intervention operation according to the intervention instruction;

[0027] The data storage and communication module is respectively connected to the multi-sensor data acquisition module, the AI ​​analysis module and the intelligent intervention module for:

[0028] Store users' historical eating data, health indicators and intervention records;

[0029] The data is uploaded to a remote server via a wireless network to achieve cloud-based backup and analysis of the data.

[0030] Preferably, the multi-sensor data acquisition module includes:

[0031] Liquid level sensor, used to detect the height change of liquid in the cup in real time;

[0032] Weight sensor, used to accurately measure the weight of the liquid in the cup;

[0033] An RFID reader is used to identify the unique identifier of the cup body and obtain the cup body capacity information associated with the identifier;

[0034] GPS module, used to obtain the user's real-time geographic location information;

[0035] Environmental sensors are used to collect ambient parameters such as temperature, humidity, and light intensity;

[0036] The data from the liquid level sensor and the weight sensor are used for cross-validation to improve the accuracy of liquid quantity measurement.

[0037] Preferably, the data preprocessing module includes:

[0038] Data cleaning unit, used to remove outliers and noise from sensor data;

[0039] Data normalization unit, used to convert different types of sensor data into a unified numerical range;

[0040] A feature extraction unit is used to extract time series features, statistical features, and frequency domain features from raw sensor data;

[0041] The data fusion unit is used to time align and comprehensively analyze the data from multiple sensors to generate a multi-dimensional feature vector.

[0042] Preferably, the AI ​​analysis module adopts a multi-layer convolutional neural network structure, including:

[0043] An input layer, configured to receive the preprocessed multidimensional feature vector;

[0044] Multiple convolutional and pooling layers to automatically extract high-level features of liquid type and eating behavior;

[0045] The fully connected layer is used to comprehensively analyze the extracted features and output the liquid type recognition results and eating behavior classification;

[0046] The regression output layer is used to predict the real-time food intake and blood glucose curve based on the identified liquid type and eating behavior combined with the user's personal information.

[0047] Preferably, the intelligent intervention module includes:

[0048] A threshold judgment unit, configured to compare the food intake data and blood sugar prediction data with a preset health threshold;

[0049] a personalized parameter adjustment unit, configured to dynamically adjust the health threshold according to the user's age, gender, weight, and health condition;

[0050] An intervention strategy generation unit, used to formulate graded intervention strategies based on the degree and duration of exceeding the threshold;

[0051] The learning optimization unit is used to record the user's response to different intervention strategies and continuously optimize the intervention effect through reinforcement learning algorithms.

[0052] Preferably, the feedback execution module includes:

[0053] An audio reminder unit, used to issue a gentle audio reminder in case of a slight overdose;

[0054] Vibration warning unit, used to generate obvious vibration warning in case of moderate overdose;

[0055] A heating intervention unit, which is used to prevent further drinking by heating the liquid when there is a serious overdose and the user does not respond to the first two levels of reminders;

[0056] LED display unit, used to intuitively display the current food intake status and health advice;

[0057] The heating intervention unit also has a heat preservation function, which can automatically adjust the optimal heat preservation temperature according to the type of liquid.

[0058] Preferably, the data storage and communication module includes:

[0059] A local storage unit, used for temporarily storing collected data in an offline state;

[0060] A data encryption unit, used to encrypt the user's sensitive health data;

[0061] Wireless communication unit, supporting Wi-Fi, Bluetooth and 4G / 5G communication methods, used to achieve real-time data upload;

[0062] The data synchronization unit is used to ensure the consistency of local data and cloud data, and automatically synchronize offline data after the network is restored.

[0063] As an advantage, it also includes:

[0064] Power management module for:

[0065] Monitor the energy consumption of each module in the system;

[0066] Automatically adjust power consumption mode according to usage scenarios;

[0067] Support wireless charging and fast charging technology;

[0068] The power management module also includes an energy recovery unit, which can convert the kinetic energy generated by the user's drinking behavior into electrical energy, thereby extending the use time of the system.

[0069] As an advantage, it also includes:

[0070] Self-cleaning module for:

[0071] Detect the degree of stains on the inner wall of the cup;

[0072] When the soiling exceeds the preset threshold, the automatic cleaning program is started;

[0073] Use ultrasonic technology and special coating materials to achieve deep cleaning of the cup body;

[0074] Among them, the self-cleaning module is data-connected with the AI ​​analysis module, and can automatically adjust the cleaning intensity and method according to the identified liquid type.

[0075] The method for monitoring food intake of a smart energy cup based on the system includes the following steps:

[0076] S1. Through the multi-sensor data acquisition module, the liquid level information and weight information of the liquid in the cup body, as well as the RFID identification information of the cup body, the user's geographic location information and environmental data are collected;

[0077] S2. Using a data preprocessing module, the collected sensor data is subjected to noise reduction, standardization, and feature extraction processing;

[0078] S3. In the AI ​​analysis module, based on the preprocessed sensor data, a deep learning algorithm is used to identify the liquid type and eating behavior, and calculate the user's real-time food intake;

[0079] S4. Based on the real-time food intake and user historical data, the user's blood sugar curve is predicted in the AI ​​analysis module;

[0080] S5. In the intelligent intervention module, the food intake data and blood glucose prediction data sent by the AI ​​analysis module are received and compared with the preset health threshold;

[0081] S6. When the comparison result shows that the health threshold is exceeded, the intelligent intervention module generates a corresponding intervention instruction;

[0082] S7. The feedback execution module selectively executes a sound reminder, a vibration warning, or a heating intervention operation according to the intervention instruction;

[0083] S8. The user's eating data, health indicators, and intervention records are stored locally and uploaded to a remote server through the data storage and communication module;

[0084] S9. Regularly analyze users’ long-term eating data to generate personalized health reports and improvement suggestions;

[0085] S10. Based on user feedback and long-term data analysis results, continuously optimize the system's recognition algorithms, prediction models, and intervention strategies.

[0086] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0087] First, this invention utilizes multi-sensor fusion technology to significantly improve the accuracy and comprehensiveness of data collection. By integrating liquid level sensors, weight sensors, RFID readers, GPS modules, and environmental sensors, the system not only accurately measures liquid volume but also identifies liquid type and monitors user location and environmental factors. This multi-dimensional data collection lays a solid foundation for subsequent intelligent analysis, enabling the system to fully understand the user's drinking behavior and environmental context.

[0088] Secondly, this invention leverages advanced AI analysis technology to achieve a deep understanding and prediction of user drinking behavior. By employing a multi-layer convolutional neural network architecture, the system automatically extracts high-level features of liquid type and eating behavior, accurately identifying different drinking patterns. This intelligent analysis not only improves monitoring accuracy but also enables the prediction of a user's blood sugar profile, providing a valuable health management tool for patients with chronic diseases such as diabetes.

[0089] Furthermore, the intelligent intervention module of the present invention dynamically adjusts health thresholds based on personalized parameters and employs a graded intervention strategy. This approach accounts for individual user differences and can provide more accurate and effective health recommendations. By continuously adjusting the intervention strategy through the learning optimization unit, the system can gradually adapt to user preferences and habits, improving the acceptance and effectiveness of interventions.

[0090] Furthermore, the feedback execution module of the present invention uses multimodal interaction, including sound reminders, vibration warnings, heating intervention, and LED displays. This diverse feedback method can adapt to different usage scenarios, enhance user experience, and improve the continuity and effectiveness of health management.

[0091] In terms of data security, the invention effectively protects user privacy information through advanced encryption algorithms and data synchronization mechanisms. At the same time, innovative power management technology and self-cleaning functions greatly improve the practicality and user-friendliness of the product.

[0092] In summary, the intelligent energy cup food intake monitoring system and method of the present invention have achieved significant breakthroughs in multiple aspects, including data acquisition accuracy, intelligent analysis capabilities, personalized intervention, user interaction, and data security. These innovations not only solve many problems in the prior art, but also provide users with a comprehensive, accurate, and intelligent health management tool. Through the present invention, users can better understand their eating habits and obtain personalized health advice, thereby effectively preventing overeating, controlling calorie intake, and managing chronic diseases, ultimately achieving the goal of improving overall health. The implementation of this invention will bring revolutionary progress to the field of intelligent health management, with broad application prospects and significant social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 It is a logic block diagram of the entire system of the present invention.

[0094] Figure 2 This is a logic block diagram of the multi-sensor data acquisition module of the present invention.

[0095] Figure 3 This is a logic block diagram of the data preprocessing module of the present invention.

[0096] Figure 4 This is a logical block diagram of the AI ​​analysis module of the present invention.

[0097] Figure 5 This is a logic block diagram of the intelligent intervention module of the present invention.

[0098] Figure 6 This is a logic block diagram of the feedback execution module of the present invention.

[0099] Figure 7 This is a logic block diagram of the data storage and communication module of the present invention.

[0100] Figure 8 This is a logic block diagram of the power management module of the present invention.

[0101] Figure 9This is a logic block diagram of the self-cleaning module of the present invention. DETAILED DESCRIPTION

[0102] See Figure 1-9 The present invention provides a smart energy cup food intake monitoring system and method. This system utilizes multi-sensor fusion technology, artificial intelligence algorithms, and an intelligent feedback mechanism to accurately monitor a user's dietary behavior and provide personalized health management. The technical solutions of the present invention are described in detail below.

[0103] The smart energy cup food intake monitoring system includes a multi-sensor data acquisition module 1, a data preprocessing module 2, an AI analysis module 3, an intelligent intervention module 4, a feedback execution module 5, and a data storage and communication module 6. These modules work together to form a complete intelligent monitoring and intervention system.

[0104] The multi-sensor data acquisition module 1 is used to collect the liquid level and weight information of the cup, obtain the cup's RFID identification information, and collect the user's geographic location and environmental data. This module uses multiple sensors to provide the system with comprehensive and accurate raw data. Preferably, the liquid level information collection frequency can be set to 10 times per second to capture rapid drinking behavior; the weight information collection frequency can be set to 5 times per second to balance accuracy and power consumption.

[0105] The data preprocessing module 2 is electrically connected to the multi-sensor data acquisition module 1 and is used to receive and preprocess the sensor data. This preprocessing process includes noise reduction, standardization, and feature extraction. For example, for liquid level data, a moving average filter algorithm can be used for noise reduction. The filter window size can be set to 5 sampling points. This effectively removes short-term fluctuations while retaining the true consumption trend.

[0106] The AI ​​analysis module 3 is connected to the data preprocessing module 2 and is the core part of the system. This module uses deep learning algorithms to identify liquid types and eating behaviors, calculate real-time food intake, and predict the user's blood sugar curve. The present invention uses an improved convolutional neural network model, whose structure can be expressed as:

[0107] f(x)=softmax(W2·ReLU(W1·Conv(x)+b1)+b2),

[0108] Where x represents the input multidimensional sensor data, Conv represents the convolution operation, W1 and W2 represent weight matrices, b1 and b2 represent bias vectors, ReLU is the activation function, and softmax is used to output the classification probability. The intelligent intervention module 4 is signal-connected to the AI ​​analysis module 3 and is responsible for determining whether intervention is necessary based on the analysis results. This invention utilizes an adaptive threshold algorithm that dynamically adjusts the intervention threshold based on the user's individual characteristics. The threshold adjustment formula is as follows:

[0109] T=T0+k1·BMI+k2·Age+k3·ActivityLevel,

[0110] Where T is the adjusted threshold, T0 is the baseline threshold (for example, for water intake, this can be set to 500 ml per hour), BMI is the body mass index, Age is age, Activity Level is activity level, and k1, k2, and k3 are adjustment coefficients. Typical values ​​are: k1 = 10, k2 = -0.5, and k3 = 50.

[0111] The feedback execution module 5 is controlled and connected to the intelligent intervention module 4 to perform specific intervention operations. The present invention has designed a set of progressive intervention strategies, including sound reminders, vibration warnings and heating interventions. For example, when a mild overdose (exceeding the threshold by less than 10%) is detected, the system will emit a gentle reminder sound; when a moderate overdose (exceeding 10% to 30%) is detected, a vibration warning will be triggered; when a severe overdose (exceeding 30% or more) occurs and the user does not respond to the first two levels of reminders, the system will activate the heating function and heat the liquid to about 40°C to prevent the user from continuing to drink.

[0112] The Data Storage and Communication Module 6 is responsible for local data storage and cloud synchronization. This module uses the AES-256 encryption algorithm to protect user privacy and supports an incremental synchronization strategy, uploading only changed data, thereby improving synchronization efficiency and saving data traffic.

[0113] The intelligent energy cup food intake monitoring system of this invention achieves comprehensive monitoring and intelligent intervention of users' eating behaviors through the collaborative operation of the above modules. The system not only accurately records the user's food intake but also provides personalized health advice based on individual characteristics. This is of great significance for preventing overeating, controlling calorie intake, and managing chronic diseases.

[0114] Furthermore, the multi-sensor data acquisition module 1 of the present invention includes a liquid level sensor 11, a weight sensor 12, an RFID reader 13, a GPS module 14, and an environmental sensor 15. The coordinated operation of these sensors ensures the comprehensiveness and accuracy of data acquisition.

[0115] The liquid level sensor 11 uses ultrasonic ranging technology, achieving a measurement accuracy of ±1mm. The weight sensor 12 uses a high-precision strain gauge sensor with an accuracy of ±0.1g. The RFID reader 13 supports the ISO15693 protocol and has a read range of up to 5cm. The GPS module 14 adopts a low-power design, achieving a positioning accuracy of ±3m in open areas. The environmental sensor 15 integrates temperature, humidity, and light sensors, with measurement accuracies of ±0.5°C, ±3%RH, and ±5%, respectively.

[0116] The cross-validation mechanism of these sensor data significantly improves system reliability. For example, by comparing changes in liquid level and weight, the system can accurately determine whether drinking has actually occurred, thus avoiding misjudgments caused by factors such as cup tilt.

[0117] The data preprocessing module 2 includes a data cleaning unit 21, a data normalization unit 22, a feature extraction unit 23, and a data fusion unit 24. These units are designed to improve the efficiency and accuracy of subsequent AI analysis.

[0118] The data cleaning unit 21 uses a median filter algorithm to remove outliers, and the filter window size can be set to 7 sampling points. The data normalization unit 22 uses the Z-score normalization method to convert different types of data to a distribution with a mean of 0 and a standard deviation of 1. The feature extraction unit 23 not only extracts time domain features (such as mean, variance, and kurtosis), but also extracts frequency domain features through fast Fourier transform (FFT) to capture periodic drinking patterns. The data fusion unit 24 uses the Kalman filter algorithm to achieve optimal fusion of multi-sensor data.

[0119] This multi-level data preprocessing solution of the present invention can effectively improve the quality of raw data, laying a solid foundation for subsequent AI analysis. By removing noise, unifying data scales, and extracting key features, the system can more accurately identify users' drinking behaviors and liquid types, thereby providing more precise health recommendations. The AI ​​analysis module 3 of the present invention adopts a multi-layer convolutional neural network structure, including an input layer 31, multiple convolutional layers and pooling layers 32, a fully connected layer 33, and a regression output layer 34. This deep learning architecture can automatically extract high-level features of liquid type and eating behavior, achieving accurate identification and prediction.

[0120] The input layer 31 is responsible for receiving the pre-processed multi-dimensional feature vector. In one embodiment of the present invention, the input vector contains 128 features, covering the time domain, frequency domain, and statistical features. These features are normalized to ensure the training stability of the network.

[0121] Multiple convolutional layers and pooling layers 32 are the core of the network, which are used to automatically extract high-level features of liquid type and eating behavior. The present invention adopts an innovative residual connection structure, which can be expressed as:

[0122] H l =F(H l-1 )+H l-1 ,

[0123] Among them, H lDenotes the output of layer 1, and F represents the convolution operation. This structure effectively alleviates the vanishing gradient problem of deep networks and improves model performance. Preferably, the present invention uses three convolutional blocks, each containing two convolutional layers and one max pooling layer. The convolution kernel size is 3x3, and the initial number of channels is 32, which is doubled after each block.

[0124] The fully connected layer 33 comprehensively analyzes the extracted features and outputs liquid type recognition and eating behavior classification. The present invention incorporates dropout technology into the fully connected layer, with a dropout rate set to 0.5, to effectively prevent overfitting. Liquid type recognition uses a softmax classifier, capable of distinguishing 10 common beverages, including water, tea, coffee, and carbonated beverages. Eating behavior classification utilizes a multi-label classification approach, capable of simultaneously identifying various behavioral patterns, such as rapid drinking, slow sipping, and intermittent drinking.

[0125] The regression output layer 34 predicts real-time food intake and blood glucose profiles based on the identified liquid type and eating behavior, combined with the user's personal information. This invention employs an innovative multi-task learning framework to simultaneously optimize both food intake prediction and blood glucose profile prediction. Its loss function can be expressed as:

[0126] L=αL intake +βL glucose +λΩ(W),

[0127] Among them, L intake and L glucose They represent the losses of food intake prediction and blood glucose curve prediction, respectively. Ω(W) is the regularization term, and α, β, and λ are weight coefficients. The empirical values ​​can be set as: α = 0.6, β = 0.3, λ = 0.1.

[0128] The intelligent intervention module 4 of the present invention includes a threshold determination unit 41, a personalized parameter adjustment unit 42, an intervention strategy generation unit 43, and a learning optimization unit 44. These units work together to implement an intelligent and personalized intervention mechanism.

[0129] Threshold determination unit 41 is responsible for comparing food intake data and predicted blood glucose data with preset health thresholds. The present invention utilizes a fuzzy logic system to categorize food intake and blood glucose levels into three levels: low, normal, and high. For example, for water intake, hourly intake of less than 200ml can be defined as low, 200-800ml as normal, and over 800ml as high. This fuzzy classification method better reflects the continuous changes in human physiological characteristics than simple fixed thresholds.

[0130] The personalized parameter adjustment unit 42 dynamically adjusts the health threshold according to the user's age, gender, weight and health status. The present invention proposes a threshold adjustment algorithm based on user profile:

[0131] T adjusted =T base ·(1+k1·Age+k2·BMI+k3·HealthScore),

[0132] Among them, T adjusted is the adjusted threshold, T b ase is the baseline threshold, Age is the standardized age value, BMI is the standardized body mass index value, and HealthScore is the health score (a floating point number between 0 and 1). The coefficients k1, k2, and k3 can be dynamically optimized using machine learning methods.

[0133] The intervention strategy generation unit 43 formulates a graded intervention strategy based on the degree and duration of the threshold being exceeded. The present invention designs a three-level intervention mechanism: mild intervention (such as a gentle reminder), moderate intervention (such as a strong warning), and strong intervention (such as physical prevention). The intervention level is selected based on the following formula:

[0134]

[0135] Where I is the intervention level (1-3), E is the amount exceeding the threshold, T is the threshold, S is the sensitivity parameter (adjustable according to user preference), and the clip function limits the results to between 1 and 3.

[0136] The learning optimization unit 44 records the user's response to different intervention strategies and continuously optimizes the intervention effect through the reinforcement learning algorithm. The present invention adopts the Q-learning algorithm, and its update formula is:

[0137]

[0138] Among them, Q(s t ,a t ) means in state s t Next take action a t The value of r t +1 is the immediate reward, α is the learning rate, and γ is the discount factor. This adaptive learning mechanism enables the system to gradually adjust the intervention strategy to achieve the best user response effect.

[0139] The feedback execution module 5 of the present invention includes a sound reminder unit 51, a vibration warning unit 52, a heating intervention unit 53 and an LED display unit 54. These units provide multimodal feedback methods to ensure that the user can receive the system's intervention information in a timely manner.

[0140] The audio reminder unit 51 emits a gentle sound reminder when a slight overdose occurs. The present invention has designed a progressive sound reminder mechanism. The initial volume is set to 1.2 times the ambient noise level. If the user does not respond within 30 seconds, the volume increases by 5 decibels every 10 seconds until it reaches the preset maximum volume (usually 2 times the ambient noise level). This design can attract the user's attention without causing excessive disturbance.

[0141] The vibration warning unit 52 generates a noticeable vibration warning when a moderate overdose occurs. The vibration pattern adopts an innovative rhythm design, which can be expressed as follows:

[0142] V(t)=A·sin(2πft)·e -λt ,

[0143] Where A is the amplitude, f is the frequency, and λ is the attenuation coefficient. This invention sets f to 200 Hz, the frequency range in which the human body is most sensitive to touch. The vibration duration is 0.5 seconds, with an interval of 1.5 seconds, creating a rhythm that is easily recognizable but not annoying.

[0144] The heating intervention unit 53 is activated when the amount is seriously exceeded and the user does not respond to the first two levels of warning. This unit uses a PID control algorithm to accurately control the heating process:

[0145]

[0146] Among them, u(t) is the control output, e(t) is the error between the target temperature and the current temperature, K p , K i , K d The target temperature of the present invention is set to 40°C, which is high enough to make most drinks less tasty but not high enough to cause burns.

[0147] The LED display unit 54 is used to visually display the current food intake status and health recommendations. The present invention uses a color coding system: green indicates normal, yellow indicates approaching the threshold, and red indicates exceeding the threshold. The brightness of the color is proportional to the degree of exceeding the threshold, which can be expressed by the following formula:

[0148]

[0149] Among them, B is brightness, B0 is the basic brightness, B max is the maximum brightness, E is the amount exceeding the threshold, and T is the threshold.

[0150] This multimodal feedback system can adapt to different usage scenarios and user preferences, effectively improving the success rate of intervention. For example, in a quiet office environment, the system will prioritize vibration and LED display; while in a noisy outdoor environment, it will increase the frequency and intensity of audio reminders.

[0151] As detailed above, the smart energy cup food intake monitoring system of the present invention demonstrates its innovative capabilities in multi-sensor fusion, deep learning analysis, personalized intervention strategies, and human-computer interaction. The organic combination of these technologies not only enables accurate monitoring of user dietary behavior but also provides timely and effective health interventions, offering a novel solution for user health management. The data storage and communication module 6 of the present invention includes a local storage unit 61, a data encryption unit 62, a wireless communication unit 63, and a data synchronization unit 64. The collaborative operation of these units ensures the security, integrity, and accessibility of user data.

[0152] The local storage unit 61 adopts an innovative hierarchical storage structure. In this structure, frequently accessed short-term data (such as the day's diet record) is stored in fast-access flash memory, while long-term statistical data is stored in the EEPROM with lower energy consumption. This design optimizes energy consumption while ensuring data read and write speeds. Preferably, the present invention adopts an adaptive data compression algorithm that can dynamically adjust the compression ratio based on the usage of storage space. The compression ratio R can be expressed as:

[0153]

[0154] Among them, R min and R max are the minimum and maximum compression ratios, S used is the used storage space, S total = Total storage space. This method can maximize storage efficiency while ensuring data integrity. The data encryption unit 62 is responsible for encrypting the user's sensitive health data. The present invention adopts a lightweight encryption scheme based on elliptic curve cryptography, and its encryption process can be expressed as:

[0155] C=E k (M)={kG,M+kH(P)},

[0156] Here, C is the ciphertext, M is the plaintext, k is the randomly selected private key, G is the base point on the elliptic curve, P is the user's public key, and H is the hash function. This encryption scheme maintains security while having low computational complexity, making it suitable for use in resource-constrained embedded devices.

[0157] The wireless communication unit 63 supports multiple communication modes including Wi-Fi, Bluetooth, and 4G / 5G, for real-time data upload. The present invention designs an intelligent communication mode switching algorithm that automatically selects the optimal communication mode based on signal strength, data volume, and battery status. The decision function of the algorithm can be expressed as:

[0158]

[0159] Among them, S is the selected communication mode, W, B, G represent Wi-Fi, Bluetooth and 4G / 5G respectively, Q i is the signal quality, E i is the energy consumption per unit data transmission, T i is the estimated transmission time, w1, w2, and w3 are weight coefficients. This adaptive communication strategy can maximize battery life while ensuring data transmission efficiency.

[0160] The data synchronization unit 64 is responsible for ensuring the consistency of local data with cloud data and automatically synchronizing offline data after the network is restored. The present invention uses a data synchronization algorithm based on version vectors to effectively solve the data consistency problem in distributed systems. Each data record is accompanied by a version vector V, and its update rule is:

[0161] V i [i]=V i [i]+1,

[0162]

[0163] Among them, V i Represents the version vector of device i, V i [j] represents the latest version number of device j known to device i. By comparing version vectors, the system can efficiently identify data that needs to be synchronized and reduce unnecessary data transmission.

[0164] The smart energy cup food intake monitoring system of the present invention also includes a power management module 7 for monitoring the energy consumption status of each module of the system, automatically adjusting the power consumption mode according to the usage scenario, and supporting wireless charging and fast charging technologies.

[0165] The power management module 7 uses an innovative dynamic power consumption adjustment algorithm. This algorithm dynamically adjusts the operating frequency and voltage of each functional module based on the user's usage pattern and remaining power. The core concept can be expressed as follows:

[0166] P=α·f 3 +β·V 2 ·f,

[0167] Among them, P is power consumption, f is operating frequency, V is operating voltage, and α and β are constants related to hardware. By optimizing f and V, the system can minimize energy consumption while ensuring performance. For example, when the user has not used the cup for a long time, the system will automatically enter low-power mode, reduce the frequency of the main processor to the minimum value, and turn off unnecessary sensors. The power management module 7 of the present invention also includes an energy recovery unit 71, which can convert the kinetic energy generated by the user's drinking behavior into electrical energy, thereby extending the use time of the system. The energy recovery unit 71 uses piezoelectric materials, which can convert the vibration of the cup body into electrical energy. Its output power P out It can be expressed as 20

[0168]

[0169] Where k is the electromechanical coupling coefficient, Q m is the mechanical quality factor, P in By optimizing the geometric structure and circuit parameters of the piezoelectric material, the present invention achieves an energy conversion efficiency of up to 30%, significantly extending the service life of the system.

[0170] In addition, the present invention also includes a self-cleaning module 8, which is used to detect the degree of stains on the inner wall of the cup body, and start the automatic cleaning program when the stains exceed a preset threshold, using ultrasonic technology and special coating materials to achieve deep cleaning of the cup body.

[0171] The self-cleaning module 8 uses an innovative stain detection algorithm. This algorithm determines the degree of staining by analyzing the light reflection characteristics of the inner wall of the cup. Specifically, the system emits light of a specific wavelength and measures the intensity and spectral distribution of the reflected light. The stain degree $D$ can be calculated using the following formula:

[0172]

[0173] Where R(λ) is the current reflectance spectrum of the inner wall of the cup, R0(λ) is the standard reflectance spectrum of a clean cup, and λ1 and λ2 are the selected wavelength ranges. When D exceeds a preset threshold (e.g., 0.2), the system initiates an automatic cleaning process.

[0174] During the self-cleaning process, the present invention uses innovative ultrasonic cleaning technology. This technology uses variable-frequency ultrasonic waves to generate a complex sound field distribution, effectively removing different types of stains. The change in ultrasonic frequency f over time can be expressed as:

[0175] f(t)=f0+A·sin(2πωt),

[0176] Where f0 is the center frequency (usually 40kHz), A is the frequency modulation amplitude, and ω is the modulation frequency. This frequency conversion technology can effectively avoid standing waves in the sound field, ensuring that every part of the inner wall of the cup is evenly cleaned.

[0177] The method for monitoring food intake of a smart energy cup of the present invention comprises the following steps:

[0178] S1. Multi-sensor data acquisition module 1 collects information about the level and weight of the liquid in the cup, as well as the cup's RFID identification information, the user's geographic location, and environmental data. Preferably, the frequency of collecting level and weight information can be set to 10 times per second to capture rapid drinking behavior.

[0179] S2. Using the data preprocessing module 2, the collected sensor data is subjected to noise reduction, standardization and feature extraction. The present invention uses wavelet transform for noise reduction, and its denoising function can be expressed as:

[0180]

[0181] in, is the denoised signal, d j ,k is the wavelet coefficient,λ is the threshold,ψ j ,k(t) is the wavelet basis function.

[0182] S3. In AI analysis module 3, based on the pre-processed sensor data, a deep learning algorithm is used to identify the liquid type and eating behavior, and calculate the user's real-time food intake. The present invention uses an innovative attention mechanism that can automatically focus on the most relevant features. The calculation formula of attention weight α is:

[0183] e i =v T tanh(Wh i +b),

[0184] Among them, h i is the hidden state, W, v and b are learnable parameters.

[0185] S4. Based on the real-time food intake and user historical data, the user's blood sugar curve is predicted in the AI ​​analysis module 3. The present invention adopts a blood sugar prediction model based on a recurrent neural network (RNN), and its core formula is:

[0186] h t =tanh(W xh x t +W hh h t-1 +b h ),

[0187] y t =W hy h t +b y ,

[0188] Among them, h t is the hidden state, x t is the input feature, y t is the predicted blood glucose value, W and b are model parameters.

[0189] S5-S7. Intelligent Intervention Module 4 receives the food intake data and blood sugar prediction data from AI Analysis Module 3 and compares them with pre-set health thresholds. If the comparison results indicate that the health thresholds are exceeded, a corresponding intervention instruction is generated. Feedback Execution Module 5 selectively executes an audio reminder, vibration warning, or heating intervention based on the intervention instruction.

[0190] S8-S10. Through the data storage and communication module 6, the user's eating data, health indicators, and intervention records are stored locally and uploaded to a remote server. The user's long-term eating data is regularly analyzed to generate personalized health reports and improvement suggestions. Based on user feedback and long-term data analysis results, the system's recognition algorithms, prediction models, and intervention strategies are continuously optimized.

[0191] The smart energy cup food intake monitoring system and method proposed in this paper utilize innovative technologies such as multi-sensor fusion, deep learning analysis, and personalized intervention strategies to accurately monitor and intelligently intervene in users' dietary behaviors. This system not only accurately records and analyzes a user's dietary habits but also provides customized health recommendations based on individual characteristics. This system has significant practical applications in preventing overeating, controlling caloric intake, and managing chronic diseases.

[0192] It should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. Intelligent energy cup food intake monitoring system, characterized by: include: The multi-sensor data acquisition module is used to: collect the liquid level and weight information of the liquid in the cup; obtain the RFID identification information of the cup; and collect the user's geographic location information and environmental data; a data preprocessing module electrically connected to the multi-sensor data acquisition module, configured to receive various sensor data sent by the multi-sensor data acquisition module; and perform noise reduction, standardization, and feature extraction on the sensor data; and an AI analysis module data-connected to the data preprocessing module, configured to identify liquid type and eating behavior using a deep learning algorithm based on the preprocessed sensor data; According to the liquid type and eating behavior, combined with the preset health standards, the user's real-time food intake is calculated; based on the real-time food intake and the user's historical data, the user's blood sugar curve is predicted; the intelligent intervention module is signal-connected to the AI ​​analysis module, and is used to: receive the food intake data and blood sugar prediction data sent by the AI ​​analysis module; judge whether it exceeds the preset health threshold based on the food intake data and blood sugar prediction data; when the judgment result exceeds the health threshold, generate a corresponding intervention instruction; the feedback execution module is control-connected to the intelligent intervention module, and is used to: receive the intervention instruction sent by the intelligent intervention module; according to the intervention instruction, selectively execute sound reminders, vibration warnings or heating intervention operations; the data storage and communication module is data-connected to the multi-sensor data acquisition module, the AI ​​analysis module and the intelligent intervention module respectively, and is used to: store the user's historical eating data, health indicators and intervention records; upload the data to a remote server via a wireless network to realize cloud-based backup and analysis of data; The multi-sensor data acquisition module includes: a liquid level sensor for detecting the height change of the liquid in the cup in real time; a weight sensor for accurately measuring the weight of the liquid in the cup; an RFID reader for identifying the unique identifier of the cup and obtaining the cup capacity information associated with the identifier; a GPS module for obtaining the user's real-time geographic location information; and an environmental sensor for collecting environmental parameters such as ambient temperature, humidity, and light intensity. The data from the liquid level sensor and weight sensor are used for cross-validation to improve the accuracy of liquid volume measurement. The feedback execution module includes: a sound reminder unit, which is used to issue a gentle sound reminder in case of mild overdose; a vibration warning unit, which is used to generate a clear vibration warning in case of moderate overdose; a heating intervention unit, which is used to prevent further drinking by heating the liquid when there is a serious overdose and the user does not respond to the first two levels of reminders; an LED display unit, which is used to intuitively display the current food intake status and health advice; wherein, the heating intervention unit also has a heat preservation function, which can automatically adjust the optimal heat preservation temperature according to the type of liquid.

2. The smart energy cup food intake monitoring system according to claim 1, characterized in that: The data preprocessing module includes: a data cleaning unit for removing outliers and noise from sensor data; a data standardization unit for converting different types of sensor data into a unified numerical range; a feature extraction unit for extracting time series features, statistical features, and frequency domain features from the original sensor data; and a data fusion unit for time-aligning and comprehensively analyzing data from multiple sensors to generate a multidimensional feature vector.

3. The smart energy cup food intake monitoring system according to claim 1, characterized in that: The AI ​​analysis module adopts a multi-layer convolutional neural network structure, including: an input layer for receiving preprocessed multidimensional feature vectors; multiple convolutional layers and pooling layers for automatically extracting high-level features of liquid type and eating behavior; a fully connected layer for comprehensively analyzing the extracted features and outputting liquid type identification results and eating behavior classification; a regression output layer for predicting real-time food intake and blood sugar curves based on the identified liquid type and eating behavior, combined with user personal information.

4. The smart energy cup food intake monitoring system according to claim 1, characterized in that: The intelligent intervention module includes: a threshold judgment unit for comparing the food intake data and blood glucose prediction data with a preset health threshold; a personalized parameter adjustment unit for dynamically adjusting the health threshold based on the user's age, gender, weight and health status; an intervention strategy generation unit for formulating graded intervention strategies based on the degree and duration of exceeding the threshold; and a learning optimization unit for recording the user's response to different intervention strategies and continuously optimizing the intervention effect through a reinforcement learning algorithm.

5. The smart energy cup food intake monitoring system according to claim 1, characterized in that: The data storage and communication module includes: a local storage unit for temporarily storing collected data in an offline state; a data encryption unit for encrypting the user's sensitive health data; a wireless communication unit that supports multiple communication methods such as Wi-Fi, Bluetooth and 4G / 5G to achieve real-time data upload; a data synchronization unit for ensuring the consistency of local data and cloud data, and automatically synchronizing offline data after the network is restored.

6. The smart energy cup food intake monitoring system according to claim 1, characterized in that: Also includes: The power management module is used to: monitor the energy consumption status of each module in the system; automatically adjust the power consumption mode according to the usage scenario; and support wireless charging and fast charging technologies. The power management module also includes an energy recovery unit that can convert the kinetic energy generated by the user's drinking behavior into electrical energy, thereby extending the system's usage time.

7. The smart energy cup food intake monitoring system according to claim 1, characterized in that: Also includes: The self-cleaning module is used to: detect the degree of stains on the inner wall of the cup body; start the automatic cleaning program when it is detected that the stains exceed the preset threshold; use ultrasonic technology and special coating materials to achieve deep cleaning of the cup body; wherein, the self-cleaning module is data-connected with the AI ​​analysis module and can automatically adjust the cleaning intensity and method according to the identified liquid type.

8. A method for monitoring food intake of a smart energy cup based on the system according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. The multi-sensor data acquisition module collects information about the liquid level and weight of the cup, as well as the cup's RFID identification information, the user's geographic location, and environmental data. S2. The data preprocessing module performs noise reduction, standardization, and feature extraction on the collected sensor data. S3. In the AI ​​analysis module, based on the preprocessed sensor data, a deep learning algorithm is used to identify liquid type and eating behavior, and calculate the user's real-time food intake. S4. The AI ​​analysis module predicts the user's blood glucose profile based on the real-time food intake and user historical data. S5. The intelligent intervention module receives the food intake data and blood glucose prediction data sent by the AI ​​analysis module and compares them with a preset health threshold. S6. If the comparison result indicates that the health threshold is exceeded, the intelligent intervention module generates a corresponding intervention instruction. S7. The feedback execution module selectively executes a sound reminder, a vibration warning, or a heating intervention operation according to the intervention instruction; S8. The user's eating data, health indicators, and intervention records are stored locally and uploaded to a remote server through the data storage and communication module; S9. Regularly analyze users’ long-term eating data to generate personalized health reports and improvement suggestions; S10. Based on user feedback and long-term data analysis results, continuously optimize the system's recognition algorithms, prediction models, and intervention strategies.

Images