Spine fatigue detection system
By integrating multimodal data acquisition and intelligent detection modules, and combining them with machine learning models, the problems of subjectivity and individual adaptability in spinal health detection have been solved, achieving high-precision, early identification, and non-invasive assessment of spinal fatigue detection.
Patent Information
- Application Number
- CN202511022436.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-24
- Publication Date
- 2025-11-07
AI Technical Summary
Existing spinal health testing methods suffer from high subjectivity, low accuracy, and poor repeatability. Furthermore, existing electronic testing instruments rely on a single data source, which cannot comprehensively reflect spinal health status. They also have poor individual adaptability, affecting the accuracy and reliability of the tests.
The system employs a pulsed pressure application module, a multimodal data acquisition module, a data processing module, and an intelligent detection module. It integrates pressure sensors, accelerometers, and displacement sensors, and combines them with a machine learning model to achieve multimodal data fusion and adaptive pressure application, outputting spinal fatigue classification results.
It improves the objectivity and accuracy of spinal health detection, enables early identification of intervertebral disc degeneration, reduces human intervention, provides comprehensive diagnostic evidence, improves the repeatability and accuracy of test results, and achieves non-invasive, real-time assessment.
Smart Images
Figure CN120913841A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of spine health detection, in particular to a spine fatigue detection system. BACKGROUND
[0002] Spine health detection is an important part of clinical medicine. Traditional spine detection methods mainly rely on manual palpation by doctors. Doctors press the spine to perceive its rebound speed and hardness, and then judge whether the intervertebral disc is degenerated or the muscle is strained. However, this method has many limitations. First, manual palpation is highly subjective, and the detection results are highly dependent on the experience and technical level of doctors, lacking unified quantitative standards, resulting in large differences in diagnostic results between different doctors. Second, the precision of manual palpation is low, and it is difficult to capture the micro-vibration or early degeneration signal of the spine by touch alone, especially for subtle physiological changes such as changes in intervertebral disc water content, which are almost impossible to identify by manual palpation. In addition, the repeatability of manual palpation is poor, as the intensity and frequency of manual pressure are difficult to standardize, and the results of multiple detections on the same patient may not be consistent, affecting the reliability of the diagnosis.
[0003] With the development of technology, some electronic detection instruments have been introduced into the field of spine health detection. However, most existing electronic detection instruments only use a single pressure sensor to collect spine data, which improves the objectivity of detection to some extent, but due to the single source of data, it cannot fully reflect the health status of the spine. For example, the pressure sensor can only measure the pressure distribution of the spine under pressure, but cannot capture key information such as the rebound characteristics and vibration frequency of the spine. This single data collection method limits the comprehensiveness and accuracy of detection, especially in the identification of early spine degeneration, and the existing technology still has obvious shortcomings.
[0004] In addition, the existing electronic detection instruments also lack standardization in pressure operation. Due to the complex physiological structure of the spine, the spine bearing capacity of different individuals varies greatly, and the existing devices often cannot adaptively adjust according to the physiological characteristics such as body weight and height of individuals, resulting in uneven pressure intensity, which further affects the accuracy of the detection results. Therefore, developing a spine health detection system that can realize multi-modal data fusion, standardize pressure operation, and has high-precision detection capability has become a technical problem to be solved in the current spine health detection field. SUMMARY
[0005] The present application relates to the technical field of spine health detection, in particular to a spine fatigue detection system.
[0006] To solve the above problems, the technical scheme adopted by the present application is as follows:
[0007] A spine fatigue detection system, comprising a pulse pressure applying module, a multi-modal data acquisition module, a data processing module, and an intelligent detection module; the pulse pressure applying module is used to apply pulse pressure to the spine based on the weight of the user being detected; the multi-modal data acquisition module integrates a pressure sensor, an accelerometer, and a displacement sensor, and is used to acquire multi-modal data of the spine under pressure and rebound, including pressure data, rebound acceleration data, and rebound amplitude data; the data processing module is used to process the multi-modal data, including denoising processing and calculation of characteristic parameters, including the elastic coefficient, the damping ratio, and the frequency energy ratio; the intelligent detection module is constructed based on a trained machine learning classification model, and the association between the characteristic parameters and the spine state is established, and it is used to output the spine fatigue classification result according to the characteristic parameters.
[0008] As a further description of the above technical solution, the pulse pressure waveform formula applied to the spine is:
[0009] F(t)=F0e -at sin(2πft)
[0010] Wherein, F0 is the initial pulse pressure amplitude, based on the principle of energy conservation, to ensure the energy applied Matches the elastic potential energy of the human spine tissue;
[0011] a is the attenuation coefficient, determined by the damping characteristics of the spine-pressure applying driver system, designed by measuring the rebound curve of the bionic spine model of different materials; f is the pulse frequency, designed to avoid resonance damage according to the resonance frequency of the human spine.
[0012] As a further description of the above technical solution, the output of the pressure applying driver is adjusted in real time based on the PID controller.
[0013] As a further description of the above technical solution, the initial pulse pressure applied to the spine ranges from 0.5 to 5 N / cm 2 , adjusted according to the weight of the user being detected, the pulse frequency is 0.5-2Hz, and the duration of a single pulse press is ≤1 second; when the initial pulse pressure is 5N / cm 2 , the pulse frequency is 0.5Hz.
[0014] As a further description of the above technical solution, the maximum pulse pressure applied to the spine is:
[0015] F max =k·BMI+c
[0016] Wherein, k=0.3, c=2N / cm 2 , and BMI is the weight of the user being detected.
[0017] As a further description of the above technical solution, the IEEE 1588 precision time protocol is used to ensure that the clock synchronization error of the pressure sensor, the accelerometer and the displacement sensor is ≤1 ms; for the lagging acquisition data, the missing points are filled by using the cubic spline interpolation method.
[0018] As a further description of the above technical solution, the relationship between the pressure data and the rebound amplitude data is fitted by the least square method to eliminate the deviation between the pressure sensor and the displacement sensor.
[0019] As a further description of the above technical solution, the denoising processing includes: using wavelet transform to eliminate the noise of the pressure data, and using a band-pass filter to retain the effective frequency band of the rebound acceleration data and the rebound amplitude data.
[0020] As a further description of the above technical solution, the elastic coefficient is calculated according to the maximum pressure and the maximum rebound amplitude; and the damping ratio is calculated by the displacement amplitude of the continuous vibration period.
[0021] Compared with the prior art, the beneficial effects of the present application are:
[0022] 1) By integrating the pressure sensor, the accelerometer and the displacement sensor, the fusion acquisition of multi-modal data is realized, and the key characteristic parameters such as the elastic coefficient and the damping ratio (vibration attenuation) of the spine can be quantified comprehensively. Compared with the traditional manual palpation method, the system significantly improves the objectivity and accuracy of the detection, avoids the errors caused by the subjective judgment of the doctor, and provides a reliable quantitative basis for the spine health evaluation.
[0023] 2) By analyzing the main frequency energy distribution of the spine vibration signal, the system can capture the early degeneration characteristics such as the water content change of the intervertebral disc; combined with the sensitivity training of the machine learning model to the low-frequency energy anomaly, the system can realize the early warning of the spine damage, help the clinical doctors to intervene in the early stage of the disease, and delay the progression of the disease.
[0024] 3) An adaptive pressure adjustment method based on BMI is designed, and the stability of the applied pressure is ensured through a real-time force feedback system. This closed-loop control mechanism makes the pressure application operation more standardized, avoids the inconsistency of manual pressure application, and improves the repeatability and reliability of the detection results.
[0025] 4) Through the data fusion of the multi-modal sensors and the intelligent detection model, the system can comprehensively analyze the multi-dimensional data such as the pressure distribution, the rebound characteristics and the vibration frequency of the spine. Based on these data, the system can accurately identify various spine health problems such as muscle strain, intervertebral disc degeneration and local low-pressure depression, and provide a comprehensive reference basis for clinical diagnosis.
[0026] 5) By applying pulse pressure on the spine and monitoring the rebound characteristics in real time, non-invasive and real-time spine health assessment is achieved. Compared with traditional invasive detection methods, this system not only reduces the pain of patients, but also can quickly output the detection results, significantly improving the efficiency of clinical diagnosis.
[0027] 6) A machine learning classification model is used to establish the correlation between feature parameters and spine status, which can automatically output the spine fatigue classification results. This intelligent detection method reduces manual intervention, reduces operation difficulty, and improves detection accuracy and consistency.
[0028] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the following embodiments of the present application are described in detail below, and the accompanying drawings are described as follows. BRIEF DESCRIPTION OF DRAWINGS
[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.
[0030] Figure 1 is a structural schematic diagram of the spine fatigue detection system described in the embodiments of the present application;
[0031] Figure 2 is a whole flowchart of the spine fatigue detection system described in the embodiments of the present application;
[0032] Figure 3 The pulse pressure and spine rebound pressure-time diagram of the embodiments of the present application. DETAILED DESCRIPTION
[0033] In order to make the purposes, technical solutions and advantages of the embodiments of the present application more clear, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments.
[0034] Please refer to Figure 1 and Figure 2 , the embodiments of the present application provide a spine fatigue detection system, which comprises a pulse pressure module, a multi-modal data acquisition module, a data processing module and an intelligent detection module.
[0035] In the present embodiment, the pulse pressure module is used to apply standardized pulse pressure on the spine by an electric device based on the weight of the user to be detected, and the specific description is as follows:
[0036] The formula for the pulse pressure waveform applied to the spine is:
[0037] F(t) = F0e -at sin(2πft)
[0038] Where F0 is the initial pulse pressure amplitude, based on the principle of energy conservation, ensuring the applied energy... The system matches the elastic potential energy of the human spinal tissue; 'a' is the attenuation coefficient, determined by the damping characteristics of the spinal-pressure application actuator system, calibrated experimentally (by measuring the rebound curves of biomimetic spinal models made of different materials); 'f' is the pulse frequency, designed to avoid resonance damage based on the resonant frequency of the human spine (typically 1-3Hz). The output of the pressure application actuator is adjusted in real-time using a PID controller.
[0039] The initial pulse pressure applied to the spine ranges from 0.5 to 5 N / cm. 2 The pulse frequency is adjusted according to the user's weight (BMI), ranging from 0.5 to 2 Hz, simulating a doctor's compression rhythm. Each pulse compression lasts ≤1 second to avoid soft tissue damage. The initial pulse pressure is 5 N / cm². 2 At this time, the pulse frequency is 0.5Hz to avoid damage to deep tissues.
[0040] The formula for adjusting the initial pulse pressure F0 is:
[0041] F0 = 0.5 N / cm 2 BMI < 18.5 (underweight)
[0042] F0 = 2.0 N / cm 2 A BMI of 18.5 or less and < 25 is considered normal.
[0043] F0 = 5.0 N / cm 2 BMI ≥ 25 (overweight)
[0044] The formula for the maximum pulse pressure applied to the spine is:
[0045] F max =k·BMI+c
[0046] Where k = 0.3, c = 2 N / cm 2 The results were determined through clinical trials to cover the tolerance range of 95% of the population, with BMI being the weight of the tested users.
[0047] In this embodiment, the multimodal data acquisition module integrates a pressure sensor, an accelerometer, and a displacement sensor to simultaneously acquire multimodal data on spinal compression and rebound, including pressure data, rebound acceleration data, and rebound amplitude data, as detailed below:
[0048] A flexible pressure sensor array covers the spine region to measure the pressure distribution. An accelerometer detects the rebound acceleration. A displacement sensor directly measures the rebound amplitude with high accuracy. All sensors are synchronously sampled at a rate of ≥ 100 Hz with time-stamped alignment.
[0049] The relationship between the pressure data P(x, y, t) and the rebound amplitude data Δx(t) is fitted by the least square method to eliminate the bias between sensors.
[0050] The second integral formula of the acceleration signal a(t) is where Δx0 is the initial displacement, which provides a reference value by the displacement sensor.
[0051] The IEEE 1588 precision time protocol is used to ensure that the clock synchronization error of the pressure sensor, accelerometer, and displacement sensor is ≤ 1 ms; for the collected data with lag, the missing points need to be filled using the cubic spline interpolation method.
[0052] In this embodiment, the data processing module is used to process the multi-modal data, including denoising processing and calculating the characteristic parameters, including the elastic coefficient, damping ratio, and frequency energy ratio, which are specifically explained as follows:
[0053] The wavelet transform is used to eliminate the noise of the pressure data, and the effective frequency band of the rebound acceleration data and the rebound amplitude data is retained by the band-pass filter (0.1-50 Hz).
[0054] The wavelet denoising threshold processing is:
[0055] Hard threshold function
[0056] where the threshold value σ is the noise standard deviation, and N is the signal length.
[0057] The elastic coefficient is the ratio of the maximum pressure and the maximum rebound amplitude, which can reflect the stiffness of the spine. The damping ratio is calculated by the displacement amplitude of the continuous vibration period, which can reflect the energy dissipation speed.
[0058] The calculation of the frequency energy ratio is optimized:
[0059] (1) Power spectral density (PSD) estimation: the Welch method is used for segment averaging to reduce spectral leakage;
[0060]
[0061] where M is the number of segments, x m (t) is the mth segment signal.
[0062] (2) The meaning of energy ratio: when the intervertebral disc degenerates, the collagen fiber breaks, leading to a decrease in stiffness, and the main frequency f moves to a low frequency, such as from 12 Hz to 8 Hz.
[0063] In the present embodiment, the intelligent detection module is constructed based on a trained machine learning classification model, and the association between the feature parameters and the spine state is established, which is used to output the spine fatigue classification result, such as the grade (normal, mild, severe), according to the feature parameters, and output a visual report, which is specifically described as follows:
[0064] Figure 3 The pulse pressure and the spine rebound pressure-time diagram are shown in FIG. 6, where the horizontal axis (x-axis) represents time t, the unit is second, and the vertical axis (y-axis) represents pressure F. F0 represents the peak pressure, and t1 represents the end time of the pressure.
[0065] In the present embodiment, the rebound time τ represents the time required for the pressure to decay to the initial value, and the greater the τ, the more serious the spine aging.
[0066] In the present embodiment, the energy dissipation rate η is calculated according to the following formula:
[0067]
[0068] The aged spine has a higher energy dissipation rate due to a decrease in elasticity.
[0069] In the present embodiment, a multiple regression formula is established, spine age = a·τ + b·η + c, and based on the clinical data, the coefficients are calibrated.
[0070] The above only describes the preferred embodiments of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
Claims
1. A spinal fatigue detection system, characterized by, The device comprises a pulse pressure applying module, a multi-modal data acquisition module, a data processing module, and an intelligent detection module. The pulse pressure applying module is used to apply pulse pressure to the spine based on the weight of the user being detected. The multi-modal data acquisition module integrates a pressure sensor, an accelerometer, and a displacement sensor, and is used to acquire multi-modal data of the spine under pressure and rebound, including pressure data, rebound acceleration data, and rebound amplitude data. The data processing module is used to process the multi-modal data, including denoising and calculation of characteristic parameters, including the elastic coefficient, the damping ratio, and the frequency energy ratio. The intelligent detection module is based on a trained machine learning classification model, and establishes the correlation between the characteristic parameters and the state of the spine, and is used to output the classification results of the spine fatigue according to the characteristic parameters.
2. The spinal fatigue detection system of claim 1, wherein, The pulse pressure waveform formula applied to the spine is: F(t) = F0e -at sin(2πft) Wherein, F0 is the initial pulse pressure amplitude, based on the principle of energy conservation, to ensure the energy applied Matches the elastic potential energy of the human spinal tissue; a is the attenuation coefficient, which is determined by the damping characteristics of the spine-pressure applying driver system, and is designed by measuring the rebound curves of bionic spine models of different materials; f is the pulse frequency, which is designed to avoid resonance damage according to the resonance frequency of the human spine.
3. The spinal fatigue detection system of claim 2, wherein, The output of the pressure applying driver is adjusted in real time based on the PID controller.
4. The spinal fatigue detection system of claim 1, wherein, The initial impulse pressure applied to the spine ranges from 0.5 to 5 N / cm 2 , adjusted according to the weight of the user being tested, with an impulse frequency of 0.5 to 2 Hz and a single impulse press duration of ≤ 1 second; when the initial impulse pressure is 5 N / cm 2 , the impulse frequency is 0.5 Hz.
5. The spinal fatigue detection system of claim 1, wherein, The maximum pulse pressure formula applied to the spine is: F max = k - BMI + c where k = 0.3, c = 2 N / cm 2 BMI is the weight of the user being tested.
6. The spinal fatigue detection system of claim 1, wherein, The IEEE 1588 precision time protocol is used to ensure that the clock synchronization error of the pressure sensor, the accelerometer, and the displacement sensor is less than or equal to 1 ms. For the lagging acquisition data, the missing points are filled using the cubic spline interpolation method.
7. The spinal fatigue detection system of claim 1, wherein, The relationship between the pressure data and the rebound amplitude data is fitted by the least squares method to eliminate the deviation between the pressure sensor and the displacement sensor.
8. The spinal fatigue detection system of claim 1, wherein, The denoising process includes: using wavelet transform to eliminate the noise of the pressure data, and using a band-pass filter to retain the effective frequency band of the rebound acceleration data and the rebound amplitude data.
9. The spinal fatigue detection system of claim 1, wherein, The elastic coefficient is calculated based on the maximum pressure and the maximum rebound amplitude; and the damping ratio is calculated by the displacement amplitude of the continuous vibration period.