Pulse diagnosis training device based on Internet of Things and pulse condition simulation method

Abstract

The invention discloses a pulse diagnosis training device based on the Internet of Things and a pulse condition simulation method, and relates to the crossing field of traditional Chinese medicine education and the Internet of Things technology.The pulse diagnosis training device is composed of pulse sensor hardware and a pulse sensing APP, wireless connection and instruction interaction are achieved through an ESP8266 chip, a pulse sensor adopts three sets of independent amplitude-adjustable motors to simulate pulse of the cun, guan and chi, and the pulse sensor hardware is connected with the pulse sensing APP; in combination with PID closed-loop control and a waveform interpolation algorithm, the bit, number, shape and potential characteristics of 28 kinds of standard pulse conditions are precisely restored; the pulse awareness APP provides staged learning, random testing and learning data management functions to form a complete learning closed loop of theoretical learning-simulation practice-effect evaluation; the pulse diagnosis training device has the advantages of being light in weight, low in cost and convenient to operate, pulse diagnosis learning and practice requirements of traditional Chinese medicine students, enthusiasts and researchers can be met, the market blank of personal pulse diagnosis training equipment is filled, and the pulse diagnosis learning efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] This invention relates to the intersection of traditional Chinese medicine education and Internet of Things (IoT) technology, specifically to an IoT-based pulse diagnosis training device and pulse simulation method. Background Technology

[0002] Pulse diagnosis, one of the four diagnostic methods in Traditional Chinese Medicine (TCM) – observation, auscultation, inquiry, and palpation – is a core diagnostic technique that uses pressure on the cun, guan, and chi arteries to perceive changes in the pulse and assess health status. It boasts a long history and a profound theoretical foundation. However, learning pulse diagnosis presents a significant challenge: "easy to understand in the mind, difficult to apply to the fingers." Learners need extensive practice to develop a sensory understanding of the pulse's "location, rate, shape, and force," while traditional teaching methods suffer from limited clinical practice opportunities, difficulty in quantifying pulse characteristics, and a lack of objective evaluation of learning outcomes.

[0003] Existing pulse diagnosis training equipment has the following main shortcomings:

[0004] 1. The equipment is bulky and expensive: For example, the XM-MXII pulse training instrument from Shanghai Xinman Educational Equipment Co., Ltd. is divided into teacher and student models. It is large and heavy, and the price of a single unit is tens of thousands of yuan. It is only suitable for institutional purchases and cannot meet the needs of individual learning.

[0005] 2. Limited functionality and lack of interactivity: Most devices can only simulate a few pulse patterns, lack theoretical learning resources and self-testing functions, and cannot form a closed loop of "learning-practice-evaluation";

[0006] 3. Low simulation accuracy and lack of feedback mechanism: The simulated pulse is made using a single motor or mechanical structure, which cannot distinguish the differences between the three pulse positions of Cun, Guan, and Chi, and lacks pressure feedback adjustment, resulting in a large deviation between the simulated pulse and the real clinical pulse.

[0007] 4. Lack of wireless connectivity: It relies on wired connections or independent LCD screens, making operation complex and unable to enable learning anytime, anywhere, thus falling behind the current development trend of IoT technology.

[0008] The public's demand for TCM health management has also increased significantly. Against this backdrop, there is an urgent need for a lightweight, low-cost, high-precision pulse diagnosis training device with interactive learning functions to address the pain points of traditional learning and fill the market gap for personal pulse diagnosis training equipment. Summary of the Invention

[0009] The purpose of this invention is to provide an Internet of Things-based pulse diagnosis training device and pulse simulation method to solve the problems of existing equipment being bulky, costly, having limited functions, and insufficient accuracy, and to meet the personalized pulse diagnosis learning needs of TCM students, enthusiasts, and researchers.

[0010] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0011] A pulse simulation method based on the Internet of Things (IoT) specifically includes the following steps:

[0012] Step 1, Pulse Data Standardization: Establish a multi-source synchronous acquisition model to obtain the pressure signal P(t) and the electrocardiogram signal E(t), and the combined signal is... Where T is the collection duration, and the timestamp alignment error is <1ms; the location, number, shape, and momentum characteristics of 28 pulse patterns were labeled by senior TCM doctors to construct a database;

[0013] Step 2, Waveform Analysis and Feature Extraction: Perform Discrete Fourier Transform on the time-domain signal x(t):

[0014]

[0015] Where X(k) is the frequency domain signal of the time domain signal x(n) after the Discrete Fourier Transform (DFT); k is the frequency domain sampling point index; n is the time domain sampling point index; x(n) is the discretized time domain signal; and N is the total number of signal sampling points. The complex exponential basis function is the core operator of the discrete Fourier transform, used to decompose a time-domain signal into sine and cosine components of different frequencies, where j is the imaginary unit. 2 =-1;

[0016] Extracting the main frequency f dom Harmonic component Hk=|Xk)| 2 With energy distribution;

[0017] Among them, the main frequency f dom X(k) represents the main frequency component calculated by X(k), corresponding to the pulse frequency.

[0018] Harmonic component H(k)=|Xk)| 2 The square of the magnitude of the frequency domain signal Xk reflects the energy intensity of different frequency components;

[0019] Wavelet transform noise reduction technology is used to remove high-frequency noise, and then cubic spline interpolation algorithm is used to compensate for waveform details in the sampling interval, ensuring that the root mean square error of the waveform RMSE < 0.05V and the main peak time error < 5ms.

[0020] Step 3, PID closed-loop control:

[0021] Real-time pressure P acquired by a pressure sensor real (k and preset pressure P) set The error e(k) is used to calculate the motor control quantity Δu(k) using an incremental PID algorithm, and its update formula is as follows:

[0022] Δu(k=K p(e(ke(k-1))+K i e(k+K d (e(k-2e(k-1)+e(k-2))

[0023] Where e(k=P) set -P real (k is the pressure error, K) p K i K d These are the proportional, integral, and derivative coefficients, respectively; the motor drive voltage is dynamically adjusted so that P... real (k tracking P) set The steady-state error is <5%; simultaneously, a nonlinear mapping model between pressure P and the duty cycle D of the motor drive signal is established:

[0024] P = k·D 2 +b

[0025] Where k and b are calibration coefficients fitted using pressure sensor data to ensure that the linearity of the drive signal is >95%;

[0026] Step 4, Multi-motor Phase Coordination: By adjusting the amplitudes A1, A2, A3 and phase differences φ1, φ2, φ3 of the three sets of motors, the transmission differences of the three parts (inch, gate, and ruler) are simulated. The phase differences φ1, φ2, φ3 are adjustable from 0-180°. The motor vibration signal is:

[0027]

[0028] Cubic spline interpolation is used to compensate for waveform details, with key feature point time errors <5ms; where f is the pulse frequency; v cun (t) is a function of the vibration displacement of the cun-bu motor as a function of time, reflecting the pulsating state of the cun-bu pulse; v guan (t) is the vibration displacement function of the joint motor, corresponding to the characteristics of the joint pulse; v chi (t) is the vibration displacement function of the ulnar motor, which characterizes the ulnar pulse.

[0029] Step 5, APP-based evaluation: The APP sends commands to drive the motor, and after the test, a report containing accuracy and error analysis is generated.

[0030] As a further preferred embodiment of the pulse simulation method based on the Internet of Things in this invention, the pressure and motor drive signal satisfy the nonlinear model P = k·D 2 +b; where D is the duty cycle, k and b are calibration coefficients, and linearity is >95%.

[0031] As a further preferred embodiment of the pulse simulation method based on the Internet of Things of the present invention, in step 4, the interpolation function S(x) of the waveform interpolation algorithm is in the interval [x...]. i,x i+1 The condition S(x) = a is satisfied within the range. i +b i (xx i )+c i (xx i ) 2 +d i (xx i ) 3 The continuity of the second derivative is ensured by solving a system of linear equations, and the time error of the key feature points, including the main peak and diabetic wave, is <5ms.

[0032] Where S(x) is a cubic spline interpolation function used to compensate for waveform details between discrete sampling points, making the simulated pulse waveform more continuous and smooth, closely resembling the continuous changes of real arterial pulsation; x is the interpolation variable; [x i ,x i+1 [x] represents the i-th interpolation interval. i and x i+1 For two adjacent discrete sampling points; a i b is the constant term for the i-th interpolation interval; i c is the coefficient of the linear term in the i-th interpolation interval; i d represents the coefficient of the quadratic term in the i-th interpolation interval; i Let be the coefficient of the cubic term in the i-th interpolation interval.

[0033] A pulse diagnosis training device based on the Internet of Things pulse simulation method includes a pulse sensor hardware module and a pulse diagnosis APP software module, wherein the pulse sensor hardware module and the pulse diagnosis APP software module are linked through WiFi wireless communication.

[0034] The pulse sensor hardware module includes a signal receiving and processing unit, a drive control unit, a pulse simulation unit, and a pressure feedback unit; the pulse sensor APP software module includes a learning center module, a practical center module, and a personal center module.

[0035] Among them, the signal receiving and processing unit is used to receive pulse control commands sent by the Zhimai APP software module;

[0036] The drive control unit is used to provide motor drive;

[0037] The pulse simulation unit includes three sets of miniature hollow cup vibration motors, corresponding to the cun, guan, and chi positions respectively, and the vibration intensity is controlled by PWM pulse width modulation technology.

[0038] The pressure feedback unit is used to collect the contact pressure of the fingers and eyes in real time and dynamically adjust the motor output;

[0039] The learning center module is divided into six categories: weak, strong, floating, deep, slow, and rapid, providing characteristic descriptions of standard pulse patterns and introductions to the main diseases.

[0040] The Practice Center module supports comprehensive random testing and section-specific random testing, and can automatically record the user's answer accuracy and error type.

[0041] The Personal Center module is used to manage user information, store learning records, and protect privacy. It also supports login with third-party accounts.

[0042] As a further preferred embodiment of the pulse diagnosis training device based on the Internet of Things of the present invention, the pulse sensor hardware module also includes an electrocardiogram (ECG) acquisition unit, which adopts an AD8232 ECG acquisition module with an operating voltage of 3.3V and a sampling frequency of 1000Hz. It outputs an analog voltage signal of 0-3.3V to the A0 pin of the ESP8266 chip to synchronously acquire human ECG signals and assist in verifying the authenticity of the pulse simulation.

[0043] As a further preferred embodiment of the pulse diagnosis training device based on the Internet of Things of the present invention, the signal receiving and processing unit adopts the ESP8266 chip, supports the IEEE 802.11b / g / n standard, and has a transmission delay of <50ms.

[0044] As a further preferred embodiment of the pulse diagnosis training device based on the Internet of Things of the present invention, the three sets of miniature hollow cup vibration motors have a diameter of 8mm and a frequency response of 20-500Hz.

[0045] As a further preferred embodiment of the pulse diagnosis training device based on the Internet of Things of the present invention, the pressure feedback unit adopts a pressure sensor with a range of 0-10N and an accuracy of 0.01N.

[0046] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0047] (1) Lightweight and low cost: The pulse sensor weighs less than 80g, and the mobile APP replaces the independent display screen, reducing hardware costs by more than 60%, making it suitable for individual purchase and filling a market gap;

[0048] (2) High precision and high realism: Combining PID closed-loop control, multi-motor phase coordination and waveform interpolation algorithm, the pulse simulation RMSE < 0.05V, the key feature point time error < 5ms, and the synchronous acquisition of ECG signals further improves the simulation realism;

[0049] (3) Full-process learning loop: Zhimai APP provides the function of "theoretical learning - simulated practice - effect evaluation", supports phased training and personalized reports, and solves the pain point of "no practice and no feedback" in traditional learning;

[0050] (4) Strong scalability and compatibility: Reserved interfaces for physiological signals such as ECG and PPG to support subsequent function upgrades; APP is compatible with iS and Android systems, and the pulse sensor uses universal USB-C power supply to adapt to different usage scenarios;

[0051] (5) Policy and market fit: It conforms to the national policy of modernization of traditional Chinese medicine and talent training, meets the public's demand for health learning of traditional Chinese medicine in the post-epidemic era, and the application scenarios cover education, personal learning and scientific research, with broad market prospects. Attached Figure Description

[0052] Figure 1 This is an overall structural block diagram of the pulse diagnosis training device of the present invention, including the composition and connection relationship of the pulse sensor hardware module and the pulse diagnosis APP software module;

[0053] Figure 2 This is a schematic diagram of the physical structure of the pulse sensor hardware, showing the positions of the three motors (cun, guan, chi), the pressure sensor interface, and the USB-C power supply interface.

[0054] Figure 3 This is a flowchart of the pulse simulation method, showing the complete process of "data acquisition - feature extraction - motor control - APP linkage";

[0055] Figure 4 This is a schematic diagram of a multi-motor phase coordination model, comparing the amplitude and phase differences of the cun, guan, and chi positions for different pulse types (such as wiry pulse and slippery pulse); wiry pulse (like pressing a string, straight and long), slippery pulse (flowing back and forth, like beads rolling on a plate).

[0056] Figure 5 These are screenshots of the ZhiMai APP's functional interfaces, including the pulse characteristics display in the learning center, the testing interface in the practice center, and the learning report in the personal center.

[0057] Figure 6 The graph shows the performance test results, comparing the differences between this invention and traditional pulse trainers in terms of simulation accuracy (RMSE), learning efficiency, and cost. Detailed Implementation

[0058] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings:

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. The purpose and effects of the present invention will become clearer. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0060] A pulse diagnosis training device based on the Internet of Things includes a pulse sensor hardware module and a pulse diagnosis APP software module, which are linked together via WiFi wireless communication; the pulse sensor hardware module includes a signal receiving and processing unit, a drive control unit, a pulse simulation unit and a pressure feedback unit, and the pulse diagnosis APP software module includes a learning center module, a practical center module and a personal center module.

[0061] Pulse sensor hardware module: Overall weight <80g, powered by USB-C interface (5V / 1A), battery life ≥4 hours, including:

[0062] The signal receiving and processing unit uses an ESP8266 NdeMCU chip with a built-in WiFi module, supports 2.4GHz band communication, has a transmission delay of <50ms, integrates a 24-bit ADC converter, has a quantization error of <0.1% for ECG and stress signals, supports IEEE 802.11b / g / n standards, and is used to receive pulse control commands sent by the Zhimai APP with a transmission delay of <50ms.

[0063] The drive control unit uses a TB6612FNG motor driver with an output current of 1.2A, supports overcurrent protection, and has a motor start / stop response time of <20ms.

[0064] The pulse simulation unit includes three sets of miniature hollow cup vibration motors (8mm in diameter, frequency response 20-500Hz), corresponding to the cun, guan, and chi positions respectively. The vibration intensity (0.1-5N) is controlled by 10-bit PWM technology, which can simulate 28 standard pulse types such as floating pulse (amplitude 0.5-1.5N) and deep pulse (amplitude 2.0-3.5N).

[0065] The pressure feedback unit uses a pressure sensor with a range of 0-10N and an accuracy of 0.01N to collect the pressure applied by the learner's fingers and eyes in real time. The motor output is dynamically adjusted through a PID algorithm to ensure the authenticity of the simulated pulse.

[0066] The learning center module is divided into six categories: weak, strong, floating, deep, slow, and rapid, providing characteristic descriptions and disease introductions for 28 standard pulse patterns;

[0067] The practical training center module supports comprehensive random testing and segmented random testing, and can automatically record the user's answer accuracy and error type.

[0068] The personal center module enables user information management, learning record storage, and privacy protection, and supports third-party account login.

[0069] The pulse sensor hardware module also includes an electrocardiogram (ECG) acquisition unit, which uses an AD8232 ECG acquisition module with an operating voltage of 3.3V and a sampling frequency of 1000Hz. It outputs an analog voltage signal (0-3.3V) to the A0 pin of the ESP8266 chip to synchronously acquire human ECG signals and assist in verifying the authenticity of the pulse simulation.

[0070] ZhiMai APP software modules: Supports iOS and Android systems, and includes three main functional modules:

[0071] Learning Center: Divided into six categories: weak, strong, floating, deep, slow, and rapid, providing characteristic descriptions of each pulse type (such as short pulse "short at both ends, not reaching three parts"), introductions to the main diseases (such as qi deficiency, qi stagnation), and slow-motion waveform demonstrations (0.1-second playback).

[0072] Practice Center: Supports "Comprehensive Random Test" (random questions for all 28 pulse types) and "Sectional Random Test" (questions based on major pulse categories), automatically records answer accuracy and reaction time, and marks error types (such as "deviation in the location of the pulse point").

[0073] Personal Center: Manage user information and learning records, support privacy settings and data encryption, and provide learning report generation function (such as "Floating pulse recognition accuracy is 78%, and training in cun pulse perception needs to be strengthened").

[0074] A pulse simulation method for a pulse diagnosis training device based on the Internet of Things (IoT) specifically includes the following steps:

[0075] Step 1, Standardized collection of pulse data:

[0076] A multi-source synchronous acquisition model is adopted to simultaneously acquire pulse pressure signals P(t) (via a flexible pressure sensor) and electrocardiogram signals E(t) (via an AD8232 module) from healthy individuals and clinical patients, establishing a joint signal. Timestamp alignment error < 1ms. The collected data were labeled by three or more associate chief physicians of traditional Chinese medicine to determine the "location, number, shape and momentum" characteristics of each pulse type, and a database containing 28 standard pulse types was constructed, with a pulse frequency range of 30-120 beats / minute and an accuracy of ±1 beat / minute.

[0077] Step 2, Waveform Analysis and Feature Extraction: Perform Discrete Fourier Transform (DFT) on the acquired time-domain signal x(t). The calculation formula is as follows:

[0078]

[0079] Where N is the number of signal sampling points. The main frequency f is extracted using DFT. dom (Corresponding to pulse frequency), harmonic component H(k)=|X(k)| 2 (Corresponding waveform shape) and energy distribution (corresponding intensity); wavelet transform noise reduction technology is used to remove high-frequency noise, and then the waveform details of the sampling interval are compensated by cubic spline interpolation algorithm to ensure that the root mean square error (RMSE) of the waveform is <0.05V and the main peak time error is <5ms.

[0080] X(k): The frequency domain signal of the time domain signal x(n) after the Discrete Fourier Transform (DFT), which transforms the pulse fluctuations in the time domain into frequency components in the frequency domain, making it easier to extract the "number" (frequency) features of the pulse.

[0081] k: Frequency domain sampling point index, corresponding to different frequency components. The value range of k [0, N-1] covers all frequency domain information of the signal.

[0082] n: Time-domain sampling point index, marking the discrete sampling position of the time-domain signal to ensure that the discretization of the signal in the time domain conforms to the Nyquist sampling theorem (the sampling frequency must be greater than twice the highest frequency of the signal).

[0083] x(n): Discretized time-domain signal, obtained by sampling the continuous time-domain signal x(t), such as converting the continuous pressure signal P(t) into a discrete sequence at a sampling frequency of 1000Hz.

[0084] N: Total number of signal sampling points, and sampling frequency (f s The data collection duration (T) and the data collection time satisfy N = f s ×T, for example, when the sampling frequency is 1000Hz and the acquisition time is 10 seconds, N=10000.

[0085] The complex exponential basis function is the core operator of the discrete Fourier transform, used to decompose a time-domain signal into sine and cosine components of different frequencies, where j is the imaginary unit (ji). 2 =-1).

[0086] f dom(Dominant frequency): The main frequency component calculated by X(k), which corresponds to the pulse rate. For example, the dominant frequency of a normal adult pulse is about 1-1.25Hz (60-75 beats / minute), which is a key indicator to distinguish between "slow pulse" (<60 beats / minute) and "rapid pulse" (>90 beats / minute).

[0087] H(k)=|X(k)| 2 (Harmonic components): The square of the modulus of the frequency domain signal X(k) reflects the energy intensity of different frequency components and can characterize the "shape" (waveform morphology) of the pulse. For example, the harmonic components of a slippery pulse are richer and the waveform is smoother; the harmonic components of a string pulse are concentrated and the waveform is steeper.

[0088] RMSE (Root Mean Square Error) < 0.05V: This is an index measuring the similarity between the simulated waveform and the real pulse waveform. It is calculated as follows: (y i For the true value, (These are simulated values). The smaller the error, the higher the simulation accuracy, ensuring that the pulse characteristics encountered by learners are highly consistent with the actual clinical pulse characteristics.

[0089] Step 3, PID closed-loop control and motor drive:

[0090] Real-time pressure P acquired by a pressure sensor real (k) and preset pressure P set The error e(k) is used to calculate the motor control quantity Δu(k) using an incremental PID algorithm, and its update formula is as follows:

[0091] Δu(k)=K p (e(k)-e(k-1))+K i e(k)+K d (e(k)-2e(k-1)+e(k-2))

[0092] Where e(k) = P set -P real (k) represents the pressure error, K p K i K d These are the proportional, integral, and derivative coefficients, respectively. This algorithm dynamically adjusts the motor drive voltage so that P... real (k) Tracking P set The steady-state error is <5%. Simultaneously, a nonlinear mapping model between pressure P and the motor drive signal duty cycle D is established:

[0093] P = k·D 2 +b

[0094] Where k and b are calibration coefficients fitted using pressure sensor data, ensuring that the linearity of the drive signal is >95%.

[0095] Step 4, Multi-motor phase coordination simulation:

[0096] By adjusting the amplitudes A1, A2, and A3 and the phase differences φ1, φ2, and φ3 (adjustable from 0-180°) of the three sets of motors, the transmission differences between the cun, guan, and chi parts are simulated (e.g., the cun part of the chordal pulse leads the guan part by 30° in phase). The motor vibration signal is generated according to the following formula:

[0097]

[0098] Where f is the pulse rate. The above formula allows for independent adjustment of the three pulse positions. cun (t): The vibration displacement of the cun-position motor changes with time, reflecting the pulsating state of the cun-position pulse, such as the v of the cun-position floating pulse. cun (t) The amplitude is relatively large, and the waveform fluctuates significantly. v guan (t): Vibration displacement function of the Guan pulse, corresponding to the characteristics of the Guan pulse, such as the Guan pulse of a wiry pulse. guan (t) The amplitude is usually greater than that of the inch and ruler portions, and the waveform is steeper. v chi (t): Vibration displacement function of the ulnar pulse, characterizing the ulnar pulse pattern, such as the ulnar pulse in patients with kidney deficiency. ci (t) has a small amplitude and weak pulsation.

[0099] The interpolation function of the waveform interpolation algorithm satisfies the condition within the interval. The continuity of the second derivative is ensured by solving a system of linear equations. The time error of the key feature points, including the main peak and diphtheria wave, is <5ms.

[0100] Step 5: App Integration and Learning Outcome Evaluation

[0101] Users select learning or testing modes through the ZhiMai APP. The APP sends control commands (such as "simulate floating pulse, cun position amplitude 1.2N") to the ESP8266 chip via WiFi. The chip retrieves the target pulse parameters from the database and drives the motor to simulate the pulse. After the practical test is completed, the APP compares the user's answers with the standard features and generates a training report that includes accuracy and error point analysis to help users improve in a targeted manner.

[0102] Device binding and networking

[0103] Teacher-side operation: Open the Zhimai APP, enter "Teacher Mode", click "Add Student Device", and complete the binding of 10 student-side pulse sensors by scanning the QR code generated by the student-side APP. After binding, the APP interface will display the device numbers (such as S1-S10) and online status of all student-side devices.

[0104] Student operation: After binding, fix the pulse sensor to the wrist with the elastic strap, open the "Student Mode" of the APP, and confirm that the device is in the "waiting for instructions" state and the pressure sensor is zeroed (the pressure reading is 0N).

[0105] Synchronous pulse simulation

[0106] On the teacher's end, select "Willow Pulse" in the "Pulse Database" of the APP, set the parameters as follows: Cun amplitude 1.0N, Guan amplitude 1.2N, Chi amplitude 0.9N, frequency 75 times / minute, phase difference (Cun leads Guan by 20°), and click "Send Synchronously".

[0107] The system automatically records the instruction transmission time: the time difference between when the teacher sends the instruction and when the pulse sensors on all students' ends start to vibrate. It is recorded once every 30 seconds, for a total of 5 times.

[0108] Students should observe the vibration status of the pulse sensor, confirm that the LED indicators for the cun, guan, and chi positions are all constantly lit in green (normal working condition), and press the corresponding positions with their fingers as instructed by the teacher, maintaining a pressure of 1.0-1.5N.

[0109] Remote data acquisition and guidance

[0110] The teacher's app receives real-time pressure feedback data from 10 student devices (sampling frequency 10Hz) and generates a "pressure-time" curve on the interface, marking the pressure fluctuation range of each device (e.g., S1: 1.2±0.1N, S2: 1.4±0.3N).

[0111] For student devices (such as S2) with pressure fluctuation range >0.3N, the teacher clicks "Remote Guidance" and sends the text instruction "The pressure needs to be stable, and it is recommended to reduce the fluctuation of finger force." At the same time, the teacher sends the "Standard Pulse" pressure reference curve to the device for the student to compare and adjust.

[0112] After receiving the instruction, the student adjusts the pressing technique, while the teacher observes the pressure curve changes in real time until the pressure fluctuation range of S2 is ≤0.3N, and records the adjustment time.

[0113] Data Statistics and Analysis

[0114] After the experiment, the teacher exported the following data: average instruction transmission delay, student stress fluctuation compliance rate (fluctuation ≤ 0.3N is considered compliant), and average adjustment time for remote guidance.

[0115] Students fill out a "Pulse Perception Questionnaire" and provide feedback on the consistency between the simulated pulse and the description in the textbook (1-5 points, 5 points for complete consistency).

[0116] Example 1: Device Hardware Assembly and Debugging

[0117] 1. Hardware selection and assembly:

[0118] Pulse simulation unit: A miniature hollow cup vibration motor (model: 1020) with a diameter of 8mm is selected. The three sets correspond to the cun, guan, and chi positions respectively. They are fixed inside a 3D printed shell (size: 60mm×40mm×15mm) with a spacing of 10mm, matching the cun, guan, and chi positions of the human wrist.

[0119] Signal processing unit: The ESP8266 NdeMCU development board is used, which integrates a WiFi module and a 24-bit ADC, and connects the TB6612FNG driver and AD8232 module through DuPont wires;

[0120] Pressure feedback unit: The FSR402 flexible pressure sensor is selected and attached above the motor, aligned with the finger contact position. The output signal is connected to the A1 pin of the ESP8266.

[0121] Power supply unit: Uses a 5V / 1A USB-C interface to connect to a lithium battery (capacity 1000mAh), with a battery life of ≥4 hours.

[0122] 2. Hardware debugging:

[0123] Motor drive test: PWM control code was written using Arduin IDE, and the duty cycle was adjusted (0-1023). The motor amplitude was tested to be 0.1-5N with an error of ±0.08N, which meets the design requirements.

[0124] WiFi communication test: Within an indoor range of 10m, the transmission latency of the APP sending commands to the ESP8266 is <50ms, and the packet loss rate is <1%;

[0125] Pressure feedback test: Simulate the change in pressure applied by the learner (0.5-3N), adjust the motor output using a PID algorithm, with a steady-state error of <5% and a response time of 32ms.

[0126] Example 2: Validation of Pulse Simulation and Learning Functions

[0127] 1. Construction of a standard pulse database:

[0128] Subjects: Twenty healthy adults (10 men and 10 women, aged 20-30 years) and ten clinical patients (including those with Qi deficiency and Qi stagnation syndromes) were selected. The pulse diagnosis was performed by three associate chief physicians of traditional Chinese medicine from the First Affiliated Hospital of Nanjing University of Traditional Chinese Medicine.

[0129] Data acquisition: Simultaneously acquire pressure signals (sampling frequency 1000Hz) and electrocardiogram signals, label the characteristic parameters of 28 pulse types such as floating, deep, slow, and rapid, and establish a database.

[0130] 2. Pulse simulation verification:

[0131] Three typical pulse patterns were selected for simulation: floating pulse (dominant frequency 70 beats / minute, amplitude 1.2N at the cun position), deep pulse (dominant frequency 65 beats / minute, amplitude 2.8N at the guan position), and wiry pulse (cun ​​position leading guan position by 30° phase).

[0132] Ten third-year students majoring in Traditional Chinese Medicine were invited to participate in a blind test to compare the simulation effect of the present invention with that of the traditional training device. The results showed that the pulse recognition accuracy of the present invention (89%) was significantly higher than that of the traditional training device (62%), and 90% of the students believed that "the differences in the three pulses are more obvious and closer to the real clinical pulse".

[0133] 3. Learning function verification:

[0134] Fifty first-year students majoring in Traditional Chinese Medicine were selected and divided into an experimental group (using this invention) and a control group (using traditional textbooks) to learn pulse diagnosis for four weeks.

[0135] After the experiment, the pulse recognition accuracy of the experimental group (82%) was higher than that of the control group (45%), and the learning time was reduced by 40%, proving that the present invention can significantly improve learning efficiency.

[0136] Example 3: Performance Comparison Test

[0137] Under the same testing environment (indoor 25℃, no wind), the performance of this invention was compared with that of the Shanghai Xinman XM-MXII pulse trainer and the American SP2 pulse oximeter. The results are shown in Table 1:

[0138] Table 1

[0139]

[0140] Test results show that the present invention has significant advantages in terms of lightweight, low cost, simulation accuracy and learning function, and can meet the needs of individual pulse diagnosis learning.

[0141] This embodiment is merely an exemplary implementation. The hardware selection (such as motor model and sensor type) and software functions (such as adding a pulse community communication module) can be adjusted according to actual needs. All modifications and extensions based on the technical solution of this invention fall within the protection scope of this invention.

[0142] Unless otherwise stated, the terms used herein (including technical terms) have the common understanding meaning to those skilled in the art; in addition, it is understood that terms defined by commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and should not be understood to have an idealized or overly formal meaning.

Claims

1. A pulse simulation method based on the Internet of Things, characterized in that, Specifically, it includes the following steps: Step 1, Pulse Data Standardization: Establish a multi-source synchronous acquisition model to obtain the pressure signal P(t) and the electrocardiogram signal E(t), and the combined signal is... Where T is the collection duration, and the timestamp alignment error is <1ms; the location, number, shape, and momentum characteristics of 28 pulse patterns were labeled by senior TCM doctors to construct a database; Step 2, Waveform Analysis and Feature Extraction: Perform Discrete Fourier Transform on the time-domain signal x(t): Where X(k) is the frequency domain signal of the time domain signal x(n) after the Discrete Fourier Transform (DFT); k is the frequency domain sampling point index; n is the time domain sampling point index; x(n) is the discretized time domain signal; and N is the total number of signal sampling points. The complex exponential basis function is the core operator of the discrete Fourier transform, used to decompose a time-domain signal into sine and cosine components of different frequencies, where j is the imaginary unit. 2 =-1; Extracting the main frequency f dom Harmonic component Hk=|Xk)| 2 With energy distribution; Among them, the main frequency f dom X(k) represents the main frequency component calculated by X(k), corresponding to the pulse frequency. Harmonic component H(k)=|X(k)| 2 It is the square of the magnitude of the frequency domain signal X(k), reflecting the energy intensity of different frequency components; Wavelet transform noise reduction technology is used to remove high-frequency noise, and then cubic spline interpolation algorithm is used to compensate for waveform details in the sampling interval, ensuring that the root mean square error of the waveform RMSE < 0.05V and the main peak time error < 5ms. Step 3, PID closed-loop control: Real-time pressure P acquired by a pressure sensor real (k and preset pressure P) set The error e(k) is used to calculate the motor control quantity Δu(k) using an incremental PID algorithm, and its update formula is as follows: Δu(k=K p (e(ke(k-1))+K i e(k+K d (e(k-2e(k-1)+e(k-2)) Where e(k=P) set -P real (k is the pressure error, K) p K i K d These are the proportional, integral, and derivative coefficients, respectively; the motor drive voltage is dynamically adjusted so that P... real (k tracking P) set The steady-state error is <5%; simultaneously, a nonlinear mapping model between pressure P and the duty cycle D of the motor drive signal is established: P=k·D 2 +b Where k and b are calibration coefficients fitted using pressure sensor data to ensure that the linearity of the drive signal is >95%; Step 4, Multi-motor Phase Coordination: By adjusting the amplitudes A1, A2, A3 and phase differences φ1, φ2, φ3 of the three sets of motors, the transmission differences of the three parts (inch, gate, and ruler) are simulated. The phase differences φ1, φ2, φ3 are adjustable from 0-180°. The motor vibration signal is: Cubic spline interpolation is used to compensate for waveform details, with key feature point time errors <5ms; where f is the pulse frequency; v cun (t) is a function of the vibration displacement of the cun-bu motor as a function of time, reflecting the pulsating state of the cun-bu pulse; v guan (t) is the vibration displacement function of the joint motor, corresponding to the characteristics of the joint pulse; v chi (t) is the vibration displacement function of the ulnar motor, which characterizes the ulnar pulse. Step 5, APP-based evaluation: The APP sends commands to drive the motor, and after the test, a report containing accuracy and error analysis is generated.

2. The pulse simulation method based on the Internet of Things according to claim 1, characterized in that, The pressure and motor drive signal satisfy the nonlinear model P = k·D 2 +b; where D is the duty cycle, k and b are calibration coefficients, and linearity is >95%.

3. The pulse simulation method based on the Internet of Things according to claim 1, characterized in that, In step 4, The interpolation function S(x) of the waveform interpolation algorithm is in the interval [x... i ,x i+1 The condition S(x) = a is satisfied within the range. i +b i (xx i )+c i (xx i ) 2 +d i (xx i ) 3 The continuity of the second derivative is ensured by solving a system of linear equations, and the time error of the key feature points, including the main peak and diabetic wave, is <5ms. Where S(x) is a cubic spline interpolation function used to compensate for waveform details between discrete sampling points, making the simulated pulse waveform more continuous and smooth, closely resembling the continuous changes of real arterial pulsation; x is the interpolation variable; [x i ,x i+1 [x] represents the i-th interpolation interval. i and x i+1 For two adjacent discrete sampling points; a i b is the constant term for the i-th interpolation interval; i c is the coefficient of the linear term in the i-th interpolation interval; i d represents the coefficient of the quadratic term in the i-th interpolation interval; i Let be the coefficient of the cubic term in the i-th interpolation interval.

4. A pulse diagnosis training device based on the pulse simulation method of the Internet of Things according to any one of claims 1 to 3, characterized in that, It includes a pulse sensor hardware module and a pulse detection APP software module, which are linked together via WiFi wireless communication. The pulse sensor hardware module includes a signal receiving and processing unit, a drive control unit, a pulse simulation unit, and a pressure feedback unit; the pulse sensor APP software module includes a learning center module, a practical center module, and a personal center module. Among them, the signal receiving and processing unit is used to receive pulse control commands sent by the Zhimai APP software module; The drive control unit is used to provide motor drive; The pulse simulation unit includes three sets of miniature hollow cup vibration motors, corresponding to the cun, guan, and chi positions respectively, and the vibration intensity is controlled by PWM pulse width modulation technology. The pressure feedback unit is used to collect the contact pressure of the fingers and eyes in real time and dynamically adjust the motor output; The learning center module is divided into six categories: weak, strong, floating, deep, slow, and rapid, providing characteristic descriptions of standard pulse patterns and introductions to the main diseases. The Practice Center module supports comprehensive random testing and section-specific random testing, and can automatically record the user's answer accuracy and error type. The Personal Center module is used to manage user information, store learning records, and protect privacy. It also supports login with third-party accounts.

5. The pulse diagnosis training device based on the Internet of Things according to claim 4, characterized in that, The pulse sensor hardware module also includes an electrocardiogram (ECG) acquisition unit, which uses an AD8232 ECG acquisition module with an operating voltage of 3.3V and a sampling frequency of 1000Hz. It outputs an analog voltage signal of 0-3.3V to the A0 pin of the ESP8266 chip to synchronously acquire human ECG signals and assist in verifying the authenticity of the pulse simulation.

6. The pulse diagnosis training device based on the Internet of Things according to claim 4, characterized in that, The signal receiving and processing unit uses the ESP8266 chip, supports the IEEE 802.11b / g / n standard, and has a transmission delay of <50ms.

7. The pulse diagnosis training device based on the Internet of Things according to claim 4, characterized in that, The three sets of miniature hollow cup vibration motors have a diameter of 8mm and a frequency response of 20-500Hz.

8. The pulse diagnosis training device based on the Internet of Things according to claim 4, characterized in that, The pressure feedback unit uses a pressure sensor with a range of 0-10N and an accuracy of 0.01N.

Images