Medicine box with prompting function and supporting Internet of Things

Abstract

The invention discloses a medicine box with a prompting function and supporting the Internet of Things, and relates to the field of medicine boxes. According to the scheme, the medicine box comprises a medicine box body and a prompting mechanism, the medicine box body comprises a box body used for storing tablets, and one face of the medicine box body is open and blocked through a box cover; the box cover is fixedly connected with the box body through the connecting piece; the accelerometer is used for collecting n acceleration values and acceleration change frequency of the medicine box within the time length i after the prompting mechanism generates the medicine taking prompting information; the judgment module is used for judging the normal medicine taking probability of the user based on the n acceleration values and the acceleration change frequency; the reminding module is used for comparing the medicine taking probability with a preset medicine taking probability threshold value and judging whether a reminding instruction is generated or not. According to the method, the medicine box reminding information is monitored, the state of the medicine box is detected after the medicine box reminding information is monitored, the probability of normal medicine feeding is judged, and therefore whether the reminding information is sent to terminal personnel or not is judged.

Description

Technical Field

[0001] The present invention relates to the field of medicine boxes, and in particular to an Internet of Things-enabled medicine box with a reminder function. Background Art

[0002] Currently in hospitals, medications for patients are basically distributed to the patients, and doctor's orders are marked on the medicine packaging boxes or sealed bags. For tablet medications, the daily usage frequency and the number of tablets are indicated, as well as whether to take them before or after meals. For medications taken by dripping or applying, the interval time and the daily frequency are also indicated;

[0003] For patients with memory loss or cognitive impairment, it is easy to forget to take medications, especially those with a high frequency of administration; for medicine bottles that look similar, misuse may occur, which may delay treatment or exacerbate the condition. It is difficult to control the frequency and interval. Some medications require strict intervals, and it is difficult for patients to adhere to or execute due to the pace of life. Taking medications at night is easily overlooked; when long-acting and short-acting medications are combined, it is easy to confuse the time nodes. For elderly or child patients who need assistance from others for reminders, it is difficult to set the time by themselves;

[0004] Moreover, current medicine boxes only have the function of storing medications, and even if there is a timing reminder function, it cannot ensure whether patients take the medications properly. Summary of the Invention

[0005] (I) Object of the Invention

[0006] In view of this, the object of the present invention is to provide an Internet of Things-enabled medicine box with a reminder function to achieve self-filling.

[0007] (II) Technical Solution

[0008] To achieve the above technical object, the present invention provides an Internet of Things-enabled medicine box with a reminder function:

[0009] It includes a medicine box and a reminder mechanism. The reminder mechanism is used to generate medicine-taking reminder information. It includes a medicine box and a reminder mechanism. The medicine box includes: a box body for storing tablets, one side of the medicine box is open and sealed by a box cover; a connecting member, and the box cover is fixedly connected to the box body through the connecting member; an accelerometer for collecting n acceleration values and the acceleration change frequency of the medicine box within a duration i after the reminder mechanism generates the medicine-taking reminder information; a discrimination module for determining the normal medicine-taking probability of the user based on the n acceleration values and the acceleration change frequency, where both n and i are integers greater than 1; a reminder module for comparing the medicine-taking probability with a preset medicine-taking probability threshold to determine whether to generate a reminder instruction.

[0010] Preferably, the acceleration change frequency is the change frequency of the n acceleration values of the medicine box within a duration i after the reminder mechanism generates the medicine-taking reminder information.

[0011] Preferably, it is characterized in that the method for the discrimination module to determine the probability of the user taking medicine normally includes: calculating the standard deviation of n acceleration values of the medicine box within the duration i, comparing the standard deviation with a preset standard deviation threshold to determine whether it is a medicine-taking action; comparing the acceleration change frequency with a preset acceleration change frequency threshold to determine whether it is a motion state;

[0012] Assign preset weights to the medicine-taking action and the motion state to calculate the medicine-taking probability.

[0013] Preferably, it is characterized in that the method for determining whether it is a medicine-taking action is: if the standard deviation is greater than the preset standard deviation threshold, it is determined to be a medicine-taking action and marked as 1; if the standard deviation is less than or equal to the preset standard deviation threshold, it is determined to be a non-medicine-taking action and marked as 0.

[0014] Preferably, it is characterized in that the method for determining whether it is a motion state is: if the acceleration change frequency is less than the preset acceleration change frequency threshold, it is determined to be a motion state and marked as 0; if the acceleration change frequency is greater than or equal to the preset acceleration change frequency threshold, it is determined to be a non-motion state and marked as 1.

[0015] Preferably, it is characterized in that the method for determining whether to generate a reminder instruction is: if the medicine-taking probability is less than or equal to the preset medicine-taking probability threshold, it is determined to generate a reminder instruction and send the instruction to the terminal through wireless communication; if the medicine-taking probability is greater than the preset medicine-taking probability threshold, it is determined not to generate a reminder instruction.

[0016] Preferably, it is characterized in that the terminal is a mobile device used by the user-set person.

[0017] Preferably, it is characterized in that the reminder mechanism includes a box body, and the outer surface of the box body is embedded with a multi-color status indicator light, a frequency setting button, a time increase button, a time decrease button, a selection and confirmation button, an information playing function key, a display screen, and a reset button.

[0018] Preferably, it is characterized in that a two-dimensional code for obtaining the method of binding the terminal is attached to the outer surface of the medicine box.

[0019] From the above technical solutions, it can be seen that the present application has the following beneficial effects:

[0020] 1: By monitoring the reminder information of the medicine box and detecting the status of the medicine box after the reminder information of the medicine box is detected, the probability of normal medicine intake is determined, so as to determine whether to send a reminder message to the terminal personnel.

[0021] 2: Set the medicine-taking reminder through the frequency setting button, the time increase button, the time decrease button and the selection and confirmation button, and send a reminder on time through the multi-color status indicator light. Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0023] Figure 1 It is a schematic diagram of the overall structure of a medicine box supporting the Internet of Things with a prompting function provided by the present invention;

[0024] Figure 2 It is a schematic cross-sectional structure diagram of the medicine box of a medicine box supporting the Internet of Things with a prompting function provided by the present invention;

[0025] Figure 3 It is a schematic diagram of the discriminant module structure of a medicine box supporting the Internet of Things with a prompting function provided by the present invention.

[0026] Description of the drawings: 1. Medicine box; 11. Box body; 12. Box cover; 13. Connector; 14. Accelerometer; 2. Prompting mechanism; 21. Multi-color status indicator; 22. Frequency setting button; 23. Time increase button; 24. Time decrease button; 25. Selection and confirmation button; 26. Information playback function key; 27. Display screen; 28. Reset button; 3. QR code. Detailed implementation manners

[0027] The following description is essentially exemplary only and is not intended to limit the present disclosure, its application, and uses. It should be understood that in all these drawings, the same or similar reference numerals indicate the same or similar parts and features. Each drawing only schematically shows the concept and principle of the embodiments of the present disclosure, and does not necessarily show the specific dimensions and their ratios of the embodiments of the present disclosure. In a specific part of a specific drawing, the relevant details or structures of the embodiments of the present disclosure may be illustrated in an exaggerated manner.

[0028] Embodiment 1

[0029] Refer to Figure 1As shown in the figure, an Internet of Things-enabled medicine box with a reminder function includes a medicine box 1 and a reminder mechanism 2. The reminder mechanism 2 is used to generate medicine-taking reminder information. The reminder mechanism 2 includes a box body, and the outer surface of the box body is embedded with a multi-color status indicator light 21, a frequency setting button 22, a time increase button 23, a time decrease button 24, a selection and confirmation button 25, an information playback function key 26, a display screen 27, and a reset button 28. It is worth mentioning that a PCB is installed inside the box body, and a wireless communication module and a logic encoding chip, such as a single-chip microcomputer, are integrated on the PCB. The pins of the multi-color status indicator light 21, the frequency setting button 22, the time increase button 23, the time decrease button 24, the selection and confirmation button 25, the information playback function key 26, the display screen 27, and the reset button 28 are all fixed on the PCB. The specific circuit connection method is a known and publicly disclosed technology, and will not be elaborated here.

[0030] Specifically, the multi-color status indicator light 21 can clearly indicate whether the medicines in the current day's partition are completed according to the setting. Exemplarily, yellow indicates in progress, red indicates not started, and green indicates completed.

[0031] The display screen 27 is used to prompt medicine-taking information, such as medicine name, set number of times of taking medicine, medicine-taking interval, remaining time, remaining number of times, cumulative treatment course, etc., and to display the setting menu.

[0032] In this embodiment, at least one medicine box 1 is provided. For example, four medicine boxes 1 are provided in this embodiment. The four medicine boxes 1 are respectively set with medicine usage times, and all settings of the default time are made through the frequency setting button 22. For example, if three times are selected, the default times are 8:00, 12:00, and 17:00. If time adjustment is required, the time increase button 23 and the time decrease button 24 are selected for time adjustment, and the adjusted time is entered through the selection and confirmation button 25, and the adjusted time is confirmed through the selection and confirmation button 25.

[0033] Press the reset button 28 once briefly to reset the frequency, press it twice continuously to reset the interval, and press and hold the reset button 28 for 5 seconds to reset all information. The display screen 27 prompts whether to perform the corresponding reset, and the time increase button 23 and the time decrease button 24 are used to select yes or no for reset, and the reset is confirmed through the confirmation button 25.

[0034] When the information playback function key 26 is pressed, the set information and status information of the current function area are played in voice. When the set medicine-taking interval time arrives, the multi-color status indicator light 21 prompts the patient to take medicine. In some embodiments, a voice reminder can also be made through a speaker.

[0035] Further, based on the wireless communication module, a two-dimensional code 3 for obtaining the method of binding the terminal is attached to the outer surface of the medicine box 1. The terminal is a mobile device used by the set user, such as a mobile phone APP. A supporting APP can be developed to provide functions such as setting, management, and medication statistics. The current patient and medication information can be uploaded to the management platform through the Internet of Things. Patients or relatives can also scan the two-dimensional code 3 and view and connect to the medicine box through the APP or mini-program for setting and status viewing;

[0036] Exemplarily, APP functions: remotely set the medication plan. Record medication data and provide medication statistics. Remotely view the status of the medicine box. Send reminder notifications to patients or guardians. Data synchronization: The patient's medication information can be uploaded to the cloud management platform. Medical staff or relatives can remotely view the patient's medication situation.

[0037] In this embodiment, the medicine box 1 is managed through manual setting and remote setting. The medication reminder and status viewing are carried out through the reminder mechanism 2. Based on the wireless communication module, it is convenient for patients, medical staff, and relatives to assist in management and status viewing, making the doctor-patient management more direct and close, and also facilitating intelligent management.

[0038] Further, low-power optimization design, sleep and low-power mode. The medicine box enters the low-power sleep mode when it is in a non-working state. The sensor and wireless communication module are only activated during the set time period. Low-power wireless communication, using BLE (low-power Bluetooth) or LoRa for short-distance communication. Using NB-IoT or 4G module for remote data transmission.

[0039] Application scenarios, daily medication management for the elderly or chronic disease patients, children's medication reminder to prevent missed doses, intelligent medication management in hospital wards, batch management in institutions such as rehabilitation centers and nursing homes.

[0040] Embodiment 2

[0041] Refer to Figure 1 、 Figure 2 and Figure 3 As shown in

[0042] A medicine box with a reminder function and supporting the Internet of Things. On the basis of Embodiment 1, the medicine box 1 includes a box body 11, a connecting piece 13, an accelerometer 14, a discrimination module, and a reminder module. Among them, the accelerometer 14, the discrimination module, the reminder module, and the reminder mechanism 2 are connected through a wired and / or wireless network.

[0043] The accelerometer 14 is used to collect n acceleration values and the acceleration change frequency of the medicine box 1 within the duration i after the reminder mechanism 2 generates the medicine-taking reminder information. The acceleration values are marked as A(t), that is, the three-axis acceleration data of the medicine box 1 at time t, A(t) = (ax, ay, az), where ax, ay, and az are the acceleration values in the x, y, and z directions respectively. The acceleration change frequency is marked as HP, and the acceleration modulus is calculated. A mag (t) is the acceleration modulus of the medicine box at time t, n is the number of collected acceleration samples, usually a smaller window is taken, such as collecting 10 times per second, and i is the duration of collecting data, usually taken as 1 - 3 seconds for analyzing the acceleration signal.

[0044] Furthermore, the acceleration change frequency is marked as HP. The acceleration change frequency is for the periodic analysis of the acceleration change. The acceleration change frequency is subjected to a fast Fourier transform (FFT) to calculate the frequency components. The specific calculation formula is a known public technology and will not be elaborated here too much.

[0045] The discrimination module is used to determine the probability of the user taking medicine normally based on n acceleration values and the acceleration change frequency. Both n and i are integers greater than 1.

[0046] The reminder module is used to compare the medicine-taking probability with a preset medicine-taking probability threshold to determine whether to generate a reminder instruction.

[0047] The acceleration change frequency is the change frequency of n acceleration values of the medicine box 1 within the duration i after the reminder mechanism 2 generates the medicine-taking reminder information.

[0048] The method for the discrimination module to determine the probability of the user taking medicine normally includes:

[0049] Calculate the standard deviation of n acceleration values of the medicine box 1 within the duration i, compare the standard deviation with a preset standard deviation threshold to determine whether it is a medicine-taking action. The standard deviation of the acceleration values is marked as JA, that is, the standard deviation of the acceleration values within the duration i. The calculation formula is:

[0050] JA = σ(A mag (t)), where σ is the standard deviation operator to calculate the standard deviation of n acceleration values;

[0051] Compare the acceleration change frequency with a preset acceleration change frequency threshold to determine whether it is a motion state;

[0052] Assign preset weights to the medicine-taking action and the motion state to calculate the medicine-taking probability.

[0053] The method for determining whether it is a medicine-taking action is:

[0054] If the standard deviation is greater than the preset standard deviation threshold, it is determined as a medicine-taking action and marked as 1;

[0055] If the standard deviation is less than or equal to the preset standard deviation threshold, it is determined as a non-medication-taking action and marked as 0.

[0056] The method for determining whether it is a motion state is as follows:

[0057] If the acceleration change frequency is less than the preset acceleration change frequency threshold, it is determined as a motion state and marked as 0;

[0058] If the acceleration change frequency is greater than or equal to the preset acceleration change frequency threshold, it is determined as a non-motion state and marked as 1.

[0059] The method for determining whether to generate a reminder instruction is as follows:

[0060] If the medication-taking probability is less than or equal to the preset medication-taking probability threshold, it is determined to generate a reminder instruction and send an instruction to the terminal through wireless communication;

[0061] If the medication-taking probability is greater than the preset medication-taking probability threshold, it is determined not to generate a reminder instruction.

[0062] Exemplarily, the medication-taking probability is marked as P take , and the calculation formula is:

[0063] P take =ω JA ×JA + ω HP ×HP, where ω JA and ω HP are the preset weights for the medication-taking action and the motion state respectively.

[0064] Example 3

[0065] Based on the above embodiments, in order to improve the accuracy of medication-taking detection of the intelligent medicine box, this embodiment introduces a machine learning algorithm to improve the recognition ability of medication-taking behavior through sensor data collection, feature extraction, model training, and optimization.

[0066] Specifically, the machine learning objective is to identify the medication-taking action through sensor data, reduce false alarms and missed reports. Train a model to distinguish states such as normal medication-taking, accidental touch, and abnormal movement. Combine the Internet of Things (IoT) to achieve remote monitoring and intelligent analysis of the intelligent medicine box.

[0067] Data collection step, sensor arrangement, the following sensors are integrated inside the intelligent medicine box, an acceleration sensor (triaxial acceleration, used to detect the movement amplitude), a Hall sensor (detecting the opening and closing state of the medicine box lid), a pressure sensor (optional, used to detect the removal of drugs), an ambient light sensor (optional, detecting whether the medicine box is exposed to light), a temperature and humidity sensor (optional, monitoring the storage environment).

[0068] Data acquisition process, start data acquisition: Set the sampling rate at 10 - 50 Hz per second. When the set medication time is reached, start the sensor for detection. Record data such as acceleration data A(t), Hall sensor signal, and pressure sensor changes. Data preprocessing, perform noise filtering (such as using Kalman filtering).

[0069] Data annotation, manually annotate data (record when the user actually takes the medicine). Automatic annotation (joint judgment based on Hall sensor and pressure sensor). Annotation categories: taking medicine (correctly taking the medicine), accidental touch (opening the box but not taking the medicine), shaking (misoperation, such as moving the medicine box), normal stillness, other abnormal states.

[0070] Data storage, store it in local storage or cloud database. Data format: timestamp, acceleration data, Hall sensor status, pressure sensor value, etc.

[0071] Machine learning model training, training objective, build a classification model to predict whether the user has really taken the medicine. Input features: acceleration change (standard deviation, frequency analysis), Hall sensor status (whether the box is opened), drug weight change (pressure sensor), interaction mode (such as APP remote confirmation). Output category 1 means successful medicine taking, and 0 means not taking the medicine. Select machine learning algorithms such as SVM (Support Vector Machine), RandomForest, LSTM (Long Short-Term Memory Network) or CNN + LSTM;

[0072] Training steps, dataset division, collect more than 10,000 groups of data and divide them according to the ratio: training set (70%), validation set (20%), test set (10%). Feature extraction, time-domain features: maximum / minimum acceleration, standard deviation, mean; frequency-domain features: Fast Fourier Transform (FFT), frequency component energy; statistical features: Root Mean Square (RMS), autocorrelation analysis; use Python (TensorFlow / PyTorch) for training, model optimization, adjust hyperparameters, such as decision tree depth, learning rate, etc., and adopt cross-validation to improve generalization ability.

[0073] Model deployment, deploy a lightweight model (TensorFlowLite) on the medicine box side, cloud training + device inference: The medicine box uploads data, the cloud updates the model, and the device only performs inference.

[0074] Exemplarily, Case 1: Elderly medication monitoring: Elderly person A at 70 years old often forgets to take medicine. Through the intelligent medicine box detection, the system finds that the medicine has not been taken at 8 am, and the device automatically sends a reminder to the family member and pops up a warning on the APP.

[0075] Case 2: Batch management in hospital wards. Medical staff need to manage the drug distribution of multiple wards every day. Through the intelligent medicine box, the system can monitor whether patients take medicine on time. If a patient fails to pick up the medicine, the nurse station will receive a reminder and intervene in a timely manner.

[0076] Case 3: Detection of children accidentally taking medicine. A 6-year-old child accidentally opened the medicine box (detected by Hall sensor). However, the pressure sensor did not detect a decrease in the medicine. The device triggers an alarm to prevent the risk of accidental ingestion.

[0077] Furthermore, for the existing design of the intelligent medicine box, it can be improved from multiple aspects such as algorithm optimization, data optimization, sensor fusion, intelligent analysis, and power consumption optimization to improve the accuracy, real-time performance, and user experience of medicine-taking detection.

[0078] The current detection methods are mainly based on traditional machine learning methods such as random forest or SVM. The optimization directions include: the combination of LSTM+CNN, LSTM (Long Short-Term Memory network): used to process time series data and learn the medicine-taking behavior patterns of users. CNN (Convolutional Neural Network): used to extract the spatial features of acceleration data and improve the ability to recognize motion states. Advantage: It can capture both short-term changes (acceleration fluctuations) and long-term trends (medicine-taking habits) simultaneously. Transformer-based time series model, using TimeSeriesTransformer to process time series signals and improve the ability to detect anomalies. Suitable for large-scale cloud data analysis. AutoML automatically optimizes the hyperparameters of the model, using AutoML (such as GoogleAutoML, Optuna) to automatically adjust hyperparameters such as learning rate, number of layers, optimization function, etc., to reduce the time of manual hyperparameter tuning;

[0079] Using reinforcement learning to optimize the medicine-taking reminder strategy. Problem: Fixed-time reminders may be ignored by users. How to improve the effectiveness of reminders? Solution: Use reinforcement learning (Reinforcement Learning), and optimize the reminder strategy through user feedback (whether the medicine is taken): State: current time, user's historical medicine-taking habits, ambient light, sound. Action: Select the reminder method (voice, light, vibration, APP notification). Reward: +1 for taking medicine in time, -0.5 for taking medicine late, -1 for not taking medicine. Through training, the system can learn the optimal reminder method and improve the medicine-taking compliance.

[0080] Predict the user's future medication behavior through time series analysis (ARIMA, LSTM): Identify the risk of missed doses (e.g., the user fails to take medicine on time multiple times). Provide personalized reminders (e.g., more suitable time points). Combine health data to provide comprehensive analysis and integrate wearable device data (such as heart rate, blood pressure): If the user forgets to take medicine, analyze the impact in combination with health data and provide risk assessment. Cloud AI doctor's advice: Provide personalized health advice according to the user's medication situation and optimize the medication plan.

[0081] Furthermore, the intelligent home medical center combines with smart bracelets / smart weighing scales to build an all-round health monitoring system. Provide a remote health management solution through the AI doctor. The hospital and nursing home are intelligentized. Combining 5G + Internet of Things (IoT), realize the intelligent distribution of hospital drugs. Automatically remind medical staff to distribute drugs and improve management efficiency. Combine with the AI voice assistant, use the voice assistant for voice interaction to remind of taking medicine. Ask about the user's medication status and adjust the reminder strategy.

[0082] In this embodiment, by combining the standard deviation of acceleration and frequency analysis, a preset threshold is used to determine whether the medicine box takes medicine normally. Calculate the probability of taking medicine using weights and determine whether to generate a reminder instruction according to the probability result. This method can effectively distinguish the medicine-taking action from the carrying movement and reduce misjudgment.

[0083] In the above text, the exemplary implementation manners of the solution proposed by the present disclosure are described in detail with reference to the preferred embodiments. However, those skilled in the art can understand that, without departing from the concept of the present disclosure, various modifications and variations can be made to the above specific embodiments, and various combinations of the technical features and structures proposed by the present disclosure can be made, without exceeding the protection scope of the present disclosure. The protection scope of the present disclosure is determined by the appended claims.

Claims

1. An Internet of Things-enabled medicine box with a reminder function, comprising a medicine box (1) and a reminder mechanism (2), wherein the reminder mechanism (2) is used to generate medicine-taking reminder information, and is characterized in that, The medicine box (1) includes: A box body (11) for storing pills, with one side of the medicine box (1) open and sealed by a box cover (12); A connecting member (13), through which the box cover (12) is fixedly connected to the box body (11); An accelerometer (14) for collecting n acceleration values and acceleration change frequencies of the medicine box (1) within a duration i after the reminder mechanism (2) generates a medicine-taking reminder message, marked as HP; A discrimination module for determining the probability of the user taking medicine normally based on the n acceleration values and acceleration change frequencies, where both n and i are integers greater than 1; A reminder module for comparing the medicine-taking probability with a preset medicine-taking probability threshold to determine whether to generate a reminder instruction.

2. The medicine box supporting the Internet of Things with a prompting function according to claim 1, characterized in that, The acceleration change frequency is the change frequency of the n acceleration values of the medicine box (1) within a duration i after the reminder mechanism (2) generates a medicine-taking reminder message.

3. The medicine box with a prompting function and supporting the Internet of Things according to claim 2, characterized in that, The method by which the discrimination module determines the probability of the user taking medicine normally includes: Calculating the standard deviation of the n acceleration values of the medicine box (1) within a duration i, marked as JA, comparing the standard deviation with a preset standard deviation threshold to determine whether it is a medicine-taking action; Comparing the acceleration change frequency with a preset acceleration change frequency threshold to determine whether it is a motion state; Assigning preset weights to the medicine-taking action and the motion state to calculate the medicine-taking probability.

4. The medicine box with a prompting function and supporting the Internet of Things according to claim 3, characterized in that, The method for determining whether it is a medicine-taking action is: If the standard deviation is greater than the preset standard deviation threshold, it is determined to be a medicine-taking action, marked as 1; If the standard deviation is less than or equal to the preset standard deviation threshold, it is determined to be a non-medicine-taking action, marked as 0.

5. The drug box supporting the Internet of Things with a prompting function according to claim 4, characterized in that, The method for determining whether it is a motion state is: If the acceleration change frequency is less than the preset acceleration change frequency threshold, it is determined to be a motion state, marked as 0; If the acceleration change frequency is greater than or equal to the preset acceleration change frequency threshold, it is determined to be a non-motion state, marked as 1.

6. The medicine box with a reminder function and supporting the Internet of Things according to claim 5, characterized in that, The method for determining whether to generate a reminder instruction is: If the medicine-taking probability is less than or equal to the preset medicine-taking probability threshold, it is determined to generate a reminder instruction and send an instruction to the terminal via wireless communication; If the medicine-taking probability is greater than the preset medicine-taking probability threshold, it is determined not to generate a reminder instruction.

7. The medicine box with a prompting function and supporting the Internet of Things according to claim 6, characterized in that, The terminal is a mobile device used by the user's designated personnel.

8. A medicine box with a prompting function and supporting the Internet of Things according to claim 7, characterized in that The reminder mechanism (2) includes a box body, on the outer surface of which are embedded a multi-color status indicator light (21), a frequency setting button (22), a time increase button (23), a time decrease button (24), a selection and confirmation button (25), an information playback function key (26), a display screen (27), and a reset button (28). A QR code (3) for obtaining the method of binding the terminal is attached to the outer surface of the medicine box (1).

9. The medicine box with a prompting function and supporting the Internet of Things according to claim 8, characterized in that, The method for obtaining the standard deviation of the acceleration value is: The acceleration value is marked as A(t), and the three-axis acceleration data of the medicine box 1 at time t is collected, A(t) = (ax, ay, az), where ax, ay, and az are the acceleration values in the x, y, and z directions respectively; Calculate the acceleration magnitude A mag (t) is the acceleration magnitude of the medicine box at time t. n is the number of collected acceleration samples. Usually, a relatively small window is taken, such as collecting 10 times per second. i is the duration of the collected data; The standard deviation of the acceleration value is marked as JA, that is, the standard deviation of the acceleration value within a duration i, and the calculation formula is: JA = σ(A mag (t)), where σ is the standard deviation operator that calculates the standard deviation of n acceleration values.

10. The drug box supporting the Internet of Things with a prompting function according to claim 8, characterized in that, The way to calculate the probability of taking medicine is: the probability of taking medicine is marked as P take , and the calculation formula is: P take = ω JA × JA + ω HP × HP, where ω JA and ω HP are the preset weights for the medicine-taking action and the motion state, respectively.

Images