Lower limb exercise load wearable monitoring equipment based on flexible pressure sensor

Abstract

A lower limb exercise load wearable monitoring device based on a flexible pressure sensor comprises a flexible sensing leg sleeve, a flexible sensing insole, a data processing module and a tibia load estimation module. According to the lower limb exercise load wearable monitoring device based on the flexible pressure sensor, a signal detected by the flexible sensing insole is used for predicting the net torque of an ankle joint, and the flexible sensing leg sleeve is used for predicting muscle force; therefore, lower limb tibia load estimation is realized. The equipment is small and portable, is suitable for various environments, and does not need professionals. Intelligent sensing insoles and intelligent sensing leg sleeves developed on the basis of flexible pressure sensors are provided with corresponding hardware circuits and algorithm models, and lower limb exercise load estimation is achieved. According to data monitored by the sensor, the tibia load change curve of the wearer is obtained, the long-term motion state of the wearer is known in an auxiliary mode, and rehabilitation nursing of a patient is facilitated.

Description

Technical Field

[0001] This invention relates to the field of smart wearable technology, and more specifically to a wearable monitoring device for lower limb movement load based on a flexible pressure sensor. Background Technology

[0002] As people's living standards improve, more and more people are paying attention to their physical health, especially physical activity. Improving physical fitness requires scientific training methods and technical support. In this process, exercise load monitoring has become an indispensable part, because only a scientifically reasonable exercise load can achieve optimal results while avoiding sports injuries.

[0003] Monitoring muscle activity load in the laboratory is challenging and remains a significant unresolved challenge in applications outside the laboratory. In the laboratory, musculoskeletal loads (e.g., forces, torques, stresses, strains) are typically estimated using one of two methods: (1) direct monitoring of tissue stress or strain using implanted or percutaneous devices (e.g., Komi, 1990; Lanyon et al., 1975) or (2) indirect estimation of tissue loads using non-invasive instruments (motion capture systems and pressure plates) through musculoskeletal modeling. While both approaches have driven decades of musculoskeletal biomechanical research, the former is invasive, and the latter requires large instruments and extensive expertise. Wearable devices typically monitor human movement by tracking spatiotemporal metrics (e.g., frequency, velocity, steps, time, etc.), body part motion (acceleration or direction), or the interaction forces between the person and the environment (e.g., ground reaction forces).

[0004] Existing exercise load estimation equipment is mostly large laboratory equipment, requiring the wearing of professional clothing or the application of marking points, making it difficult to operate outside of an indoor environment and with the assistance of professional personnel. Summary of the Invention

[0005] The purpose of this invention is to provide a wearable monitoring device for lower limb exercise load based on a flexible pressure sensor, comprising: a flexible sensing leg sleeve, a flexible sensing insole, a data processing module, and a tibial load prediction module.

[0006] Flexible tactile sensors are arranged on both the flexible sensing leg sleeves and the flexible sensing insoles.

[0007] The flexible sensing leg sleeve uses a flexible tactile sensor to collect muscle movement signals of the leg muscles of the monitored person.

[0008] The flexible sensing insole uses a flexible tactile sensor to collect force signals on the monitored person's foot.

[0009] The data processing module analyzes the collected muscle movement signals of the leg muscles to obtain the predicted muscle force.

[0010] The data processing module analyzes the collected force signals on the foot to obtain the predicted net force of the ankle joint.

[0011] The tibial load prediction module estimates the tibial load of the monitored person's lower limb based on the predicted net ankle force and muscle force, and obtains a real-time bone load curve.

[0012] Furthermore, the step of analyzing the collected muscle movement signals of the leg muscles to obtain the predicted muscle force is as follows:

[0013] A1 constructs a three-dimensional coordinate system and calculates the surface stress of the muscle in the cross-section, as shown below:

[0014]

[0015] In the formula, t represents time. F1(t) represents the surface stress of the muscle in the cross-section. σ xx (t) represents the longitudinal stress on the cross-section. A represents the cross-sectional area of ​​the muscle.

[0016] Substituting a2, representing the surface stress of the muscle in the cross-section, into the viscoelastic constitutive equation, we get the following:

[0017]

[0018] In the formula, τ σ τ ε These are the stress relaxation and delay time constants, respectively. F is the muscle force. υ ⊥ These represent the equilibrium elastic modulus along the fiber direction, the transient transverse elastic modulus, and Poisson's ratio, respectively. The x-axis is defined along the muscle's long axis, and the horizontal plane is defined as the yz plane. ε xx (t) represents the internal strain along the x-axis. For internal strain ε xx The differential of (t). σ yy Let σ be the stress along the y-axis. zz The stress is along the z-axis.

[0019] a3 sets the internal strain ε xx (t) and surface stress σ surf (t) satisfies the spatial weighting relation as shown below:

[0020]

[0021] In the formula, τ represents time. ε xx (x,y,t) represents the internal strain along the x-direction at position (x,y) at time t. σ surf(x,y,τ) represents the surface stress of the muscle at position (x,y) at time τ. G(x,y,t-τ) is the spacetime Green's function.

[0022] The predicted muscle force is calculated using a4, as shown below:

[0023]

[0024] In the formula, F(t) represents the predicted muscle force. K1(x,y) and K2(x,y) are both spatial weighting coefficients. τ1 and τ2 are both characteristic time constants.

[0025] Furthermore, the steps for analyzing the collected foot force signals to obtain the predicted ankle joint net force are as follows:

[0026] b1. Establish a foot coordinate system and determine the point of application r of the ground force using a weighted average method. a,grf As shown below:

[0027]

[0028] In the formula, i is the sensor unit index. N is the total number of sensors. i Let p be the position coordinates of the i-th sensor. i Let A be the pressure value at the i-th sensor. i Let be the sensing area at the i-th sensor. Let T be the center of pressure on the sole of the foot. T is the homogeneous coordinate transformation matrix corresponding to the transformation from the foot coordinate system to the global coordinate system.

[0029] b2 Calculate the ground reaction force F grf As shown below:

[0030]

[0031] b3 Calculate the ground reaction moment τ grf As shown below:

[0032]

[0033] In the formula, r ref Used as a reference point.

[0034] b4 calculates the net force and net torque at the ankle joint, as shown below:

[0035] F a =m f (a f -g)-F grf (8)

[0036]

[0037] In the formula, F a Net force at the ankle joint. τ a This represents the net torque at the ankle joint. f For foot quality. f F is the acceleration of the foot's center of mass. grf The ground reaction force is denoted by g. g is the acceleration due to gravity. t is time. f Let ω be the moment of inertia of the foot's center of mass. f r represents the angular velocity of the foot. a,grf This is the point of application of the ground force. a τ represents the position of the center of gravity. grf This is the ground reaction torque.

[0038] Furthermore, the wearable monitoring device for lower limb exercise load also includes a data transmission module and a user interface.

[0039] The data transmission module is used for signal transmission between the flexible tactile sensor, the data processing module, and the tibial load prediction module arranged on the flexible sensing leg sleeve and the flexible sensing insole.

[0040] The signal transmission methods include Bluetooth transmission, ZigBee transmission, Wi-Fi transmission, and wired transmission.

[0041] The user interface is integrated into the smart terminal.

[0042] The user interface is used to display the lower limb movement status and real-time bone load curve of the monitored person.

[0043] The lower limb movement states include standing, running, and walking.

[0044] When the load on the tibia of the monitored person's lower limb exceeds the preset fatigue threshold, a lower limb injury warning is displayed on the user interface.

[0045] Furthermore, both the flexible sensing leg sleeve and the flexible sensing insole integrate one or more flexible tactile sensors.

[0046] When multiple flexible tactile sensors are integrated, these flexible tactile sensors are arranged in a distributed manner.

[0047] The distributed layout includes, but is not limited to, array-based distribution and random distribution.

[0048] The integration method includes, but is not limited to, physical bonding, knitting processes, and integrated fabrication of sensors and fabric substrates.

[0049] Furthermore, the flexible tactile sensor includes resistive sensors, piezoelectric sensors, capacitive sensors, triboelectric sensors, and ionoelectric sensors.

[0050] Furthermore, the data processing module stores an ankle joint net force algorithm model.

[0051] The steps for constructing the ankle joint net force algorithm model are as follows:

[0052] c1 simultaneously acquires foot force signals detected by flexible sensor insoles and data detected by laboratory equipment.

[0053] The laboratory equipment includes a motion capture system and a force measuring platform.

[0054] c2 inputs the data from laboratory equipment testing into Opensim to calculate the net ankle joint torque.

[0055] c3 uses the foot force signal detected by the flexible sensor insole as input and the ankle joint net torque as output to establish an ankle joint net force algorithm model.

[0056] Furthermore, the data processing module stores a muscle strength algorithm model.

[0057] The steps for constructing the muscle strength algorithm model are as follows:

[0058] d1 synchronously acquires the muscle movement signals of the leg muscles detected by the flexible sensor leg sleeve and the muscle torque detected by the isokinetic muscle strength machine.

[0059] d2 uses the muscle movement signal of the leg muscles as input and the muscle torque as output to establish a muscle force algorithm model.

[0060] Furthermore, the steps for estimating the tibial load of the monitored person's lower limb are as follows:

[0061] s1 constructs a prediction model that combines time series data prediction methods with an attention mechanism.

[0062] S2 uses a prediction model to extract features and assign weights to the predicted ankle joint net force and muscle force, thereby estimating the tibial load of the monitored person's lower limb.

[0063] Furthermore, the time series data prediction method includes, but is not limited to, convolutional neural networks, long short-term memory networks, and bidirectional long short-term memory networks.

[0064] The technical effectiveness of this invention is undeniable. This invention provides a wearable lower limb exercise load monitoring device based on a flexible pressure sensor. The device consists of a flexible sensing insole, a flexible sensing leg sleeve, hardware circuitry (which may include IMU sensors), and related software algorithms and a user interface. This invention utilizes signals detected by the flexible sensing insole to predict the net ankle joint torque and the flexible sensing leg sleeve to predict muscle force, thereby achieving lower limb tibial load estimation. This device is compact, portable, suitable for various environments, and requires no professional personnel. This invention utilizes intelligent sensing insoles and intelligent sensing leg sleeves developed based on flexible pressure sensors, equipped with corresponding hardware circuitry and algorithm models, to achieve lower limb exercise load estimation.

[0065] The device provided by this invention is small, portable, easy to wear, suitable for various environments, and requires no professional personnel.

[0066] This invention obtains the tibial load change curve of the wearer through data monitored by sensors, which helps to understand the wearer's long-term exercise status and is beneficial to the patient's rehabilitation and care. Attached Figure Description

[0067] Figure 1 This is an overall framework diagram of the present invention;

[0068] Figure 2 This is a schematic diagram of the flexible sensor leg sleeve layout; Figure 2 (a) is a schematic diagram of the sensor layout for the tibialis anterior, extensor digitorum longus, peroneus longus and gastrocnemius muscles on the inner and outer sides of the lower leg. Figure 2 (b) is a schematic diagram of the sensor layout for the inner and outer sides of the gastrocnemius muscle and the soleus muscle of the calf. Figure 2 (c) is a schematic diagram of the sensor layout for the tibialis anterior, the medial and lateral sides of the gastrocnemius muscle, and the soleus muscle.

[0069] Figure 3 A schematic diagram of the layout of a flexible sensing insole; Figure 3 (a) is a schematic diagram of the layout of an 8-channel sensor insole; Figure 3 (b) is a schematic diagram of the layout of a 64-channel sensor insole. Detailed Implementation

[0070] The present invention will be further described below with reference to embodiments, but it should not be construed that the scope of the present invention is limited to the following embodiments. Various substitutions and modifications made based on ordinary technical knowledge and common practices in the art without departing from the above-described technical concept of the present invention should be included within the scope of protection of the present invention.

[0071] Example 1:

[0072] See Figures 1 to 3The wearable monitoring device for lower limb exercise load based on flexible pressure sensors includes: flexible sensing leg sleeves, flexible sensing insoles, a data processing module, and a tibial load prediction module.

[0073] Flexible tactile sensors are arranged on both the flexible sensing leg sleeves and the flexible sensing insoles.

[0074] The flexible sensing leg sleeve uses a flexible tactile sensor to collect muscle movement signals of the leg muscles of the monitored person.

[0075] The flexible sensing insole uses a flexible tactile sensor to collect force signals on the monitored person's foot.

[0076] The data processing module analyzes the collected muscle movement signals of the leg muscles to obtain the predicted muscle force.

[0077] The data processing module analyzes the collected force signals on the foot to obtain the predicted net force of the ankle joint.

[0078] The tibial load prediction module estimates the tibial load of the monitored person's lower limb based on the predicted net ankle force and muscle force, and obtains a real-time bone load curve.

[0079] Example 2:

[0080] The wearable monitoring device for lower limb movement load based on a flexible pressure sensor, as described in Example 1, further includes the following steps for analyzing the collected muscle movement signals of the leg muscles to obtain the predicted muscle force:

[0081] A1 constructs a three-dimensional coordinate system and calculates the surface stress of the muscle in the cross-section, as shown below:

[0082]

[0083] In the formula, t represents time. F1(t) represents the surface stress of the muscle in the cross-section. σ xx (t) represents the longitudinal stress on the cross-section. A represents the cross-sectional area of ​​the muscle.

[0084] Substituting a2, representing the surface stress of the muscle in the cross-section, into the viscoelastic constitutive equation, we get the following:

[0085]

[0086] In the formula, τ σ τ ε These are the stress relaxation and delay time constants, respectively. F is the muscle force. υ ⊥These represent the equilibrium elastic modulus along the fiber direction, the transient transverse elastic modulus, and Poisson's ratio, respectively. The x-axis is defined along the muscle's long axis, and the horizontal plane is defined as the yz plane. ε xx (t) represents the internal strain along the x-axis. For internal strain ε xx The differential of (t). σ yy Let σ be the stress along the y-axis. zz The stress is along the z-axis.

[0087] a3 sets the internal strain ε xx (t) and surface stress σ surf (t) satisfies the spatial weighting relation as shown below:

[0088]

[0089] In the formula, τ represents time. ε xx (x,y,t) represents the internal strain along the x-direction at position (x,y) at time t. σ surf (x,y,τ) represents the surface stress of the muscle at position (x,y) at time τ. G(x,y,t-τ) is the spacetime Green's function.

[0090] The predicted muscle force is calculated using a4, as shown below:

[0091]

[0092] In the formula, F(t) represents the predicted muscle force. K1(x,y) and K2(x,y) are both spatial weighting coefficients. τ1 and τ2 are both characteristic time constants.

[0093] Example 3:

[0094] The wearable monitoring device for lower limb exercise load based on a flexible pressure sensor, the main technical contents of which are described in any one of Embodiments 1 and 2, further, the step of analyzing the collected foot force signals to obtain the predicted ankle joint net force is as follows:

[0095] b1. Establish a foot coordinate system and determine the point of application r of the ground force using a weighted average method. a,grf As shown below:

[0096]

[0097] In the formula, i is the sensor unit index. N is the total number of sensors. i Let p be the position coordinates of the i-th sensor. i Let A be the pressure value at the i-th sensor. i Let be the sensing area at the i-th sensor. Let T be the center of pressure on the sole of the foot. T is the homogeneous coordinate transformation matrix corresponding to the transformation from the foot coordinate system to the global coordinate system.

[0098] b2 Calculate the ground reaction force F grf As shown below:

[0099]

[0100] b3 Calculate the ground reaction moment τ grf As shown below:

[0101]

[0102] In the formula, r ref Used as a reference point.

[0103] b4 calculates the net force and net torque at the ankle joint, as shown below:

[0104] F a =m f (a f -g)-F grf (8)

[0105]

[0106] In the formula, F a Net force at the ankle joint. τ a This represents the net torque at the ankle joint. f For foot quality. f F is the acceleration of the foot's center of mass. grf The ground reaction force is denoted by g. g is the acceleration due to gravity. t is time. f Let ω be the moment of inertia of the foot's center of mass. f r represents the angular velocity of the foot. a,grf This is the point of application of the ground force. a τ represents the position of the center of gravity. grf This is the ground reaction torque.

[0107] Example 4:

[0108] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Embodiments 1 to 3. Furthermore, the wearable monitoring device for lower limb exercise load also includes a data transmission module and a user interface.

[0109] The data transmission module is used for signal transmission between the flexible tactile sensor, the data processing module, and the tibial load prediction module arranged on the flexible sensing leg sleeve and the flexible sensing insole.

[0110] The signal transmission methods include Bluetooth transmission, ZigBee transmission, Wi-Fi transmission, and wired transmission.

[0111] The user interface is integrated into the smart terminal.

[0112] The user interface is used to display the lower limb movement status and real-time bone load curve of the monitored person.

[0113] The lower limb movement states include standing, running, and walking.

[0114] When the load on the tibia of the monitored person's lower limb exceeds the preset fatigue threshold, a lower limb injury warning is displayed on the user interface.

[0115] Example 5:

[0116] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Embodiments 1 to 4. Furthermore, one or more flexible tactile sensors are integrated in both the flexible sensing leg sleeve and the flexible sensing insole.

[0117] When multiple flexible tactile sensors are integrated, these flexible tactile sensors are arranged in a distributed manner.

[0118] The distributed layout includes, but is not limited to, array-based distribution and random distribution.

[0119] The integration method includes, but is not limited to, physical bonding, knitting processes, and integrated fabrication of sensors and fabric substrates.

[0120] Example 6:

[0121] The main technical contents of the wearable monitoring device for lower limb movement load based on flexible pressure sensor are described in any one of Examples 1 to 5. Further, the flexible tactile sensor includes resistive sensor, piezoelectric sensor, capacitive sensor, triboelectric sensor, and ionoelectric sensor.

[0122] Example 7:

[0123] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Embodiments 1 to 6. Furthermore, the data processing module stores an ankle joint net force algorithm model.

[0124] The steps for constructing the ankle joint net force algorithm model are as follows:

[0125] c1 simultaneously acquires foot force signals detected by flexible sensor insoles and data detected by laboratory equipment.

[0126] The laboratory equipment includes a motion capture system and a force measuring platform.

[0127] c2 inputs the data from laboratory equipment testing into Opensim to calculate the net ankle joint torque.

[0128] c3 uses the foot force signal detected by the flexible sensor insole as input and the ankle joint net torque as output to establish an ankle joint net force algorithm model.

[0129] Example 8:

[0130] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Embodiments 1 to 7. Furthermore, the data processing module stores a muscle strength algorithm model.

[0131] The steps for constructing the muscle strength algorithm model are as follows:

[0132] d1 synchronously acquires the muscle movement signals of the leg muscles detected by the flexible sensor leg sleeve and the muscle torque detected by the isokinetic muscle strength machine.

[0133] d2 uses the muscle movement signal of the leg muscles as input and the muscle torque as output to establish a muscle force algorithm model.

[0134] Example 9:

[0135] The wearable monitoring device for lower limb exercise load based on a flexible pressure sensor, the main technical contents of which are described in any one of Examples 1 to 8, further wherein the step of estimating the tibial load of the monitored person's lower limb is as follows:

[0136] s1 constructs a prediction model that combines time series data prediction methods with an attention mechanism.

[0137] S2 uses a prediction model to extract features and assign weights to the predicted ankle joint net force and muscle force, thereby estimating the tibial load of the monitored person's lower limb.

[0138] Example 10:

[0139] The wearable monitoring device for lower limb movement load based on a flexible pressure sensor, the main technical contents of which are described in any one of Examples 1 to 9, and further, the time series data prediction method includes, but is not limited to, convolutional neural networks, long short-term memory networks, and bidirectional long short-term memory networks.

[0140] Example 11:

[0141] See Figures 1 to 3 The wearable monitoring device for lower limb exercise load based on flexible pressure sensors includes: flexible sensing leg sleeves, flexible sensing insoles, a data processing module, and a tibial load prediction module.

[0142] Flexible tactile sensors are arranged on both the flexible sensing leg sleeves and the flexible sensing insoles.

[0143] The flexible sensing leg sleeve uses a flexible tactile sensor to collect muscle movement signals of the leg muscles of the monitored person.

[0144] The flexible sensing leg sleeve fully covers the calf, and the tactile sensors can be distributed in different positions as needed to monitor pressure data of different muscle groups.

[0145] The flexible sensing leg sleeve is an open elastic band, and its width can be adjusted according to design requirements.

[0146] The flexible sensing insole uses a flexible tactile sensor to collect force signals on the monitored person's foot.

[0147] The data processing module analyzes the collected muscle movement signals of the leg muscles to obtain the predicted muscle force.

[0148] The data processing module analyzes the collected foot force signals to obtain the predicted ankle joint net force. Based on the predicted ankle joint net force and muscle force, the tibial load prediction module estimates the lower limb tibial load of the monitored person, obtaining a real-time bone load curve.

[0149] The net force of the ankle joint is predicted by the signal detected by the flexible sensing insole, the muscle force is predicted by the flexible sensing leg sleeve, and then the tibial load of the lower limb is estimated by the algorithm model.

[0150] Example 12:

[0151] The wearable monitoring device for lower limb movement load based on a flexible pressure sensor, as described in Example 11, further includes the following steps for analyzing the collected muscle movement signals of the leg muscles to obtain the predicted muscle force:

[0152] A1 constructs a three-dimensional coordinate system and calculates the surface stress of the muscle in the cross-section, as shown below:

[0153]

[0154] In the formula, t represents time. F1(t) represents the surface stress of the muscle in the cross-section. σ xx (t) represents the longitudinal stress on the cross-section. A represents the cross-sectional area of ​​the muscle.

[0155] Substituting a2, representing the surface stress of the muscle in the cross-section, into the viscoelastic constitutive equation, we get the following:

[0156]

[0157] In the formula, τ σ τ ε These are the stress relaxation and delay time constants, respectively. F is the muscle force. υ ⊥These represent the equilibrium elastic modulus along the fiber direction, the transient transverse elastic modulus, and Poisson's ratio, respectively. The x-axis is defined along the muscle's long axis, and the horizontal plane is defined as the yz plane. ε xx (t) represents the internal strain along the x-axis. For internal strain ε xx The differential of (t). σ yy Let σ be the stress along the y-axis. zz The stress is along the z-axis.

[0158] a3 sets the internal strain ε xx (t) and surface stress σ surf (t) satisfies the spatial weighting relation as shown below:

[0159]

[0160] In the formula, τ represents time. ε xx (x,y,t) represents the internal strain along the x-direction at position (x,y) at time t. σ surf (x,y,τ) represents the muscle surface stress at position (x,y) at time τ. G(x,y,t-τ) is the spatiotemporal Green's function, characterizing the delay and attenuation of stress transmission.

[0161] The predicted muscle force is calculated using a4, as shown below:

[0162]

[0163] In the formula, F(t) represents the predicted muscle force. K1(x,y) and K2(x,y) are spatial weighting coefficients, determined by the muscle geometry and parameters. τ1 and τ2 are characteristic time constants.

[0164] Example 13:

[0165] The wearable monitoring device for lower limb exercise load based on a flexible pressure sensor, the main technical contents of which are described in any one of Examples 11 to 12, further includes the following steps for analyzing the collected foot force signals to obtain the predicted ankle joint net force:

[0166] b1. Establish a foot coordinate system and determine the point of application r of the ground force using a weighted average method. a,grf As shown below:

[0167]

[0168] In the formula, i is the sensor unit index. N is the total number of sensors. i Let p be the position coordinates of the i-th sensor. i Let A be the pressure value at the i-th sensor. i Let be the sensing area at the i-th sensor. Let T be the center of pressure on the sole of the foot. T is the homogeneous coordinate transformation matrix corresponding to the transformation from the foot coordinate system to the global coordinate system.

[0169] b2 Calculate the ground reaction force F grf As shown below:

[0170]

[0171] b3 Calculate the ground reaction moment τ grf As shown below:

[0172]

[0173] In the formula, r ref As a reference point, it is generally chosen to be the origin of the ankle or foot coordinate system.

[0174] b4 calculates the net force and net torque at the ankle joint, as shown below:

[0175] F a =m f (a f -g)-F grf (8)

[0176]

[0177] In the formula, F a Net force at the ankle joint. τ a This represents the net torque at the ankle joint. f For foot quality. f F is the acceleration of the foot's center of mass. grf The ground reaction force is denoted by g. g is the acceleration due to gravity. t is time. f Let ω be the moment of inertia of the foot's center of mass. f r represents the angular velocity of the foot. a,grf This is the point of application of the ground force. a τ represents the position of the center of gravity. grf This is the ground reaction torque.

[0178] Example 14:

[0179] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Examples 11 to 13. Furthermore, the wearable monitoring device for lower limb exercise load also includes a data transmission module and a user interface.

[0180] The data transmission module is used for signal transmission between the flexible tactile sensor, the data processing module, and the tibial load prediction module arranged on the flexible sensing leg sleeve and the flexible sensing insole.

[0181] The signal transmission methods include Bluetooth transmission, ZigBee transmission, Wi-Fi transmission, and wired transmission.

[0182] The user interface is integrated into the smart terminal.

[0183] The user interface is used to display the lower limb movement status and real-time bone load curve of the monitored person.

[0184] The lower limb movement states include daily movement scenarios such as standing, running, and walking. This device serves as a long-term wearable device for monitoring daily activities.

[0185] The device collects data and feeds it into various model algorithms for calculations, thereby displaying the current state of the tibia in the lower limb (no injury, different injury levels).

[0186] When the load on the tibia of the monitored person's lower limb exceeds the preset fatigue threshold, a lower limb injury warning is displayed on the user interface.

[0187] The changes in tibial load injury in the lower limb conform to the SN curve. Taking running as an example, the maximum peak load for each gait cycle is 107 MPa, and the minimum peak load is 5 MPa. Based on the maximum and minimum loads, the load levels throughout the running process can be divided. By combining finite element analysis with the SN curve, the load level L for each load level can be determined. i The maximum number of cycles that the lower tibia can withstand, Nmax(L) i This step ensures we understand the tibia's tolerance under different load conditions. The results for each load level and the maximum number of cycles tolerable are as follows.

[0188] (1) High load (greater than 85 MPa): Nmax(L1) = 10^5

[0189] (2) Medium to high load (65MPa-85MPa): Nmax(L2)=10^12

[0190] (3) Medium load (45MPa-65MPa): Nmax(L3)=10^20

[0191] (4) Medium and low load (25MPa-45 MPa)

[0192] (5) Low load (less than 25MPa)

[0193] Based on existing data and actual experiments, risk thresholds are set to convert different levels of fatigue damage (Dtotal) into specific risk levels.

[0194] (1) Low risk: When Dtotal<0.2, it indicates that there is less fatigue accumulation in the tibia and the risk is low.

[0195] (2) Moderate risk: When 0.2≤Dtotal<0.5, it indicates that the tibia has been subjected to moderate fatigue accumulation, but has not reached a dangerous state.

[0196] (3) High risk: When 0.5≤Dtotal<0.8, the fatigue accumulation of the tibia increases significantly, and there is a high risk of injury.

[0197] (4) Extremely high risk: When Dtotal≥0.8, the tibia is in a state of fatigue limit and has a very high risk of injury. It is recommended to take immediate intervention measures.

[0198] Example 15:

[0199] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of Examples 11 to 14. Furthermore, one or more flexible tactile sensors are integrated in both the flexible sensing leg sleeve and the flexible sensing insole.

[0200] When multiple flexible tactile sensors are integrated, these flexible tactile sensors are distributed in a distributed manner. By combining human tibial load analysis and muscle group force characteristics, flexible sensing insoles and flexible sensing leg sleeves are developed.

[0201] The distributed layout includes, but is not limited to, array-based distribution and random distribution.

[0202] Its distribution map will be designed based on the distribution of muscle groups or the pressure points on the sole of the foot, and according to the testing requirements (see attached map). Figure 2 - Appendix Figure 3 ).

[0203] The integration method includes, but is not limited to, physical bonding, knitting processes, and integrated fabrication of sensors and fabric substrates.

[0204] Example 16:

[0205] The main technical contents of the wearable monitoring device for lower limb movement load based on flexible pressure sensor are described in any one of Examples 11 to 15. Furthermore, the flexible tactile sensor includes resistive sensor, piezoelectric sensor, capacitive sensor, triboelectric sensor, and ionoelectric sensor.

[0206] Example 17:

[0207] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of embodiments 11 to 16. Furthermore, the data processing module stores an ankle joint net force algorithm model.

[0208] The steps for constructing the ankle joint net force algorithm model are as follows:

[0209] c1 simultaneously acquires foot force signals detected by flexible sensor insoles and data detected by laboratory equipment.

[0210] The laboratory equipment includes a motion capture system and a force measuring platform.

[0211] c2 inputs the data from laboratory equipment testing into Opensim to calculate the net ankle joint torque.

[0212] c3 uses the foot force signal detected by the flexible sensor insole as input and the ankle joint net torque as output to establish an ankle joint net force algorithm model.

[0213] Example 18:

[0214] The main technical contents of the wearable monitoring device for lower limb exercise load based on flexible pressure sensor are described in any one of embodiments 11 to 17. Furthermore, the data processing module stores a muscle strength algorithm model.

[0215] The steps for constructing the muscle strength algorithm model are as follows:

[0216] d1 synchronously acquires the muscle movement signals of the leg muscles detected by the flexible sensor leg sleeve and the muscle torque detected by the isokinetic muscle strength machine.

[0217] The output of an isokinetic muscle strength tester is torque (Nm), which reflects the magnitude of force during muscle contraction.

[0218] d2 uses the muscle movement signal of the leg muscles as input and the muscle torque as output to establish a muscle force algorithm model.

[0219] Example 19:

[0220] The wearable monitoring device for lower limb exercise load based on a flexible pressure sensor, the main technical contents of which are described in any one of Examples 11 to 18, further wherein the step of estimating the tibial load of the monitored person's lower limb is as follows:

[0221] s1 constructs a prediction model that combines time series data prediction methods with an attention mechanism.

[0222] The overall model structure of the prediction model is a CNN-BiLSTM-Attention network. The convolutional layers include three one-dimensional convolutional layers. For the i-th convolutional layer (i = 1, 2, 3), the convolution operation can be represented as:

[0223] C i =ReLU(W i *X+b i (10)

[0224] C iThe output of the convolutional layer is X; the input feature tensor is W. i and b i These are the weights and biases of the i-th convolutional layer. * denotes the convolution operation, and ReLU is the activation function. The input to the BiLSTM layer comes from the output of the convolutional layers, as shown in the formula:

[0225] H t =BiLSTM(C3),t=1,…,time_steps (11)

[0226] H t This is the output of the BiLSTM layer;

[0227] The LSTM has 64 hidden units, and the bidirectional generated hidden state has a dimension of 2*64. The bidirectional LSTM output includes both forward and backward hidden states.

[0228] The attention module has an input dimension of 2*64, and uses a linear transformation to generate an attention score A:

[0229] A = Softmax(W att H+b att (12)

[0230] Among them, W att It is the weight matrix, b att It is a bias term. Then it combines the original features with the attention score H. att :

[0231] H att =Linear([H,A]) (13)

[0232] Finally, the output of the attention module is flattened into two dimensions and mapped through a fully connected layer to obtain the final output O:

[0233] O = Linear(Flatten(H) att (14)

[0234] The model training uses the MSELoss loss function and updates parameters using gradient descent optimization. Assuming the target value is y and the predicted value is... The loss function L is expressed as:

[0235]

[0236] By combining the stress-life curve (SN curve) of the tibia, the maximum number of cycles that the tibia can withstand under that load can be obtained by simulating the application of stress loads over a certain period of time. When this maximum number of cycles is reached, the tibia is considered to be completely fatigued and has failed. However, we hope that the tibial force estimated by wearable devices can further simulate the tibial fatigue process and reliably obtain the current fatigue stage of the tibia. Therefore, we model the overuse injury of the tibia during exercise as a mechanical fatigue phenomenon.

[0237] Fatigue life of a material is defined as the number of repeated load cycles it can withstand before complete failure. This fatigue behavior describes the damage nucleation, damage accumulation, and failure process of a material as a single empirical relationship. For engineering materials, the peak stress magnitude σ is related to the number of failure cycles N. f The relationship between them can be well described by the inverse power law:

[0238] N f =A·σ -b (16)

[0239] Where A is a proportionality constant and b is the slope of the SN curve. When subjected to in vitro cyclic loading, biomaterials exhibit the same typical behavior, with logarithmic or exponential decay curves predicting fatigue life. Small changes in σ typically lead to changes in N. f Significant changes in load magnitude. For load magnitudes associated with operation, a 10% reduction in stress is typically associated with a corresponding increase of 100% or more in the number of failure cycles. If the injury from excessive motion is caused by mechanical fatigue, Equation (16) indicates that the risk of injury in a given motion will increase with load magnitude more rapidly than with load cycles. In conclusion, when examining the risk of injury, it may be insufficient to independently examine the primary effects of load magnitude and load cycles; instead, it may be necessary to consider measurements of their potential interactions.

[0240] Cumulative damage models for mechanical fatigue can capture the complex interaction between load magnitude and load cycles. The simplest of all cumulative damage models is the Palmante-Miner rule. This rule states that a material will fail when a specific amount of damage D (defined by the load-time history and SN curve) has accumulated. If the material experiences m different stress conditions (i.e., walking, running, jumping, etc.), then the failure is determined by the peak stress amplitude σ. i The partial damage caused by the i-th case is simply σ i The cycle number n divided by σ i Number of failure cycles N f Therefore, according to the Palmante-Miner rule, failure will occur when:

[0241]

[0242] Substitute formula (16) into formula (17):

[0243]

[0244] For the case of a single peak stress, formula (18) can be written as:

[0245] D∝ n σ b (19)

[0246] Equation (19) is similar to the stimulus equation for daily load to predict load-induced tissue adaptation. There is considerable evidence that load-induced tissue damage is a stimulus-induced biological remodeling. It should be clearly stated that the measurement of cumulative damage, and therefore the risk of injury from the perspective of mechanical fatigue, is not linearly proportional to the measurement of cumulative load, which is usually defined as the product of the number of load cycles and the peak load, or some pulse equivalent.

[0247] If a series of samples with the same macroscopic structure are cyclically loaded to failure at stress level σ, their measured fatigue strength is expected to change according to the following formula:

[0248] P f =1-exp[-(σ / σ * ) m (20)

[0249] Where P f σ* is the probability that the fatigue strength of a given specimen is less than or equal to σ, where σ* and m are experimentally derived constants. The reference stress σ* is a measure of the material's fatigue strength, defined as σ when the failure probability is 63% for a specific number of loading cycles. The Weibull modulus m defines the degree of dispersion observed in fatigue life measurements, where higher values ​​indicate a lower range of variation. Here, Weibull analysis will be used to estimate the probability of tibial fatigue injury during exercise, a common overuse injury associated with prolonged repetitive jumping or running.

[0250] Based on kinematic data from a relatively long running period, the data is divided according to gait cycles. The peak tibial force Pmax in each gait cycle represents the overall load level throughout the entire gait cycle. Pmax is then further categorized into multiple load levels L based on its magnitude. i (Such as low load, low-medium load, medium load, medium-high load, and high load), low load indicates that the load level of this step is relatively low, and high load indicates that the tibia needs to bear a higher load in this step. The load level borne by the tibia is closely related to walking speed, running posture, and foot landing position.

[0251] Table 1 Tibial stress-life (SN) curve data

[0252] Cycles to Failure (10^x) Stress (MPa) 0 125 0.5 121.5 1 118 1.5 114.75 2 110 2.5 107.5 3 105 3.5 102.5 4 100 4.5 97.5 5 95 5.5 92.5 6 90 6.5 88.5 7 87 7.5 85 8 83 8.5 81.5 9 80 9.5 79 10 78 11 76.5 12 75 13 73.5 14 72 15 71 16 70 17 69 18 68 19 67.5 20 67

[0253] Throughout the running process, the calculated maximum peak load for each gait cycle was 107 MPa, and the minimum peak load was 5 MPa. Based on these maximum and minimum loads, the load levels for the entire running process can be divided. By combining finite element analysis with the SN curves in Table 1, the maximum number of cycles Nmax (L) that the tibia can withstand under each load level Li can be determined. i This step ensures we understand the tibia's tolerance under different load conditions. The results for each load level and the maximum number of cycles tolerable are as follows.

[0254] (1) High load (greater than 85 MPa): Nmax(L1) = 10^5

[0255] (2) Medium to high load (65MPa-85MPa): Nmax(L2)=10^12

[0256] (3) Medium load (45MPa-65MPa): Nmax(L3)=10^20

[0257] (4) Medium and low load (25MPa-45 MPa)

[0258] (5) Low load (less than 25MPa)

[0259] It can be observed that the maximum number of fatigue cycles at or below medium load is difficult to calculate based on the SN curve. As the load level decreases, the number of cycles required to reach fatigue failure increases exponentially. This is because the stress on bone tissue under low load is relatively small, failing to reach the microstructural damage threshold of bone tissue, thus requiring an extremely high number of cycles to induce fatigue damage. At this low stress level, the self-repairing ability of bone tissue can usually offset the accumulation of micro-damage, further reducing the likelihood of fatigue failure. Furthermore, the number of cycles that bones can undergo under low load exceeds the experimentally measurable range, making it difficult to determine accurate fatigue life in actual testing. Therefore, fatigue damage to the tibia will not be considered for gait cycles classified as medium and below load.

[0260] For a certain load level L i Assume that the corresponding number of gait cycles is N. i It can calculate the damage contribution D for the entire load level. i , as in formula (21).

[0261]

[0262] Finally, the damage contributions of all load levels are summed to obtain the overall fatigue damage level Dtotal, calculated by formula (22).

[0263]

[0264] This method not only grades the degree of bone injury risk but also accurately calculates the percentage of tibial fatigue injury during exercise. Based on existing data and actual experiments, a risk threshold is set to convert different degrees of fatigue injury (Dtotal) into specific risk levels.

[0265] (1) Low risk: When Dtotal<0.2, it indicates that there is less fatigue accumulation in the tibia and the risk is low.

[0266] (2) Moderate risk: When 0.2≤Dtotal<0.5, it indicates that the tibia has been subjected to moderate fatigue accumulation, but has not reached a dangerous state.

[0267] (3) High risk: When 0.5≤Dtotal<0.8, the fatigue accumulation of the tibia increases significantly, and there is a high risk of injury.

[0268] (4) Extremely high risk: When Dtotal≥0.8, the tibia is in a state of fatigue limit and has a very high risk of injury. It is recommended to take immediate intervention measures.

[0269] In risk assessment, relying solely on fatigue damage values ​​may not fully reflect the specific circumstances of each individual. Therefore, the following individual differences should be comprehensively considered to scientifically fine-tune the risk threshold. Appropriate weights can be assigned to each factor based on the actual situation, thereby more accurately adjusting the risk threshold for each person. Below are some key individual differences and their specific adjustment suggestions for risk thresholds.

[0270] (1) Age: Bone density and endurance decline with age, so older runners can use a lower threshold to assess risk. Specific adjustment recommendations are as follows:

[0271] • 30-40 years old: Maintain the standard risk threshold without making any additional adjustments.

[0272] • 40-50 years old: Lower the risk threshold by 5%, which is equivalent to multiplying the threshold by 0.95.

[0273] • 50-60 years old: Lower the risk threshold by 10%, which is equivalent to multiplying the threshold by 0.90.

[0274] • Age 60 and above: Lower the risk threshold by 15%, which is equivalent to multiplying the threshold by 0.85.

[0275] (2) Exercise history and bone adaptability: Individuals who frequently participate in running or other high-intensity exercises generally have higher bone adaptability and enhanced tolerance, and the risk threshold can be appropriately increased. Adjustment suggestions are as follows:

[0276] • Less than 1 year of running experience: lower bone adaptability, no threshold adjustment.

[0277] • 1-5 years of running experience: Increase the risk threshold by 5%, which is equivalent to multiplying the threshold by 1.05.

[0278] • 5-10 years of running experience: Further increase the risk threshold by 10%, that is, multiply the threshold by 1.10.

[0279] • Over 10 years of running experience: Stronger bone adaptability, risk threshold increased by 15%, which is equivalent to multiplying the threshold by 1.15.

[0280] (3) Weight: Runners with higher body weight experience greater stress on their tibia during running, therefore the risk threshold needs to be lowered to reflect their higher skeletal load. Adjustment recommendations are as follows:

[0281] • For those weighing less than 60kg: Increase the risk threshold by 5%, which is equivalent to multiplying the threshold by 1.05.

[0282] • Weight between 60-80kg: Maintain the standard risk threshold and do not make any adjustments.

[0283] • For those weighing 80-100kg: lower the risk threshold by 5%, which is equivalent to multiplying the threshold by 0.95.

[0284] • Weight greater than 100kg: Lower the risk threshold by 10%, which is equivalent to multiplying the threshold by 0.90.

[0285] (4) Gender: When considering gender differences, men generally have higher bone density and muscle torque, enabling them to better adapt to running loads, while women, due to lower bone density and hormonal influences, have a higher risk of bone injury. Adjustment suggestions are as follows:

[0286] • For men: the risk threshold remains unchanged, or is appropriately increased by 3%, which is equivalent to multiplying the threshold by 1.03.

[0287] • Women: Lower the risk threshold by 5%, which is equivalent to multiplying the threshold by 0.95.

[0288] (5) Running posture and gait analysis: Good running posture and gait can reduce the impact on the tibia and decrease fatigue accumulation, thereby reducing the risk of injury. If runners frequently adjust their gait and optimize their posture, or if gait analysis shows that their posture is good, their risk threshold can be appropriately increased. Adjustment suggestions are as follows:

[0289] • Excellent gait analysis results: Increase the risk threshold by 5%, which means multiplying the threshold by 1.05.

[0290] • Gait analysis results are average: maintain the standard risk threshold and do not make any adjustments.

[0291] • Gait analysis results are unsatisfactory: Lower the risk threshold by 5%, which is equivalent to multiplying the threshold by 0.95.

[0292] (6) Bone health status (bone mineral density, bone strength, etc.): Bone mineral density and bone strength directly reflect the health of the bones and are usually assessed through bone mineral density tests (such as DEXA scans). For individuals with good bone health, the threshold can be appropriately increased; while for individuals with low bone mineral density or osteoporosis, the threshold needs to be lowered. Adjustment suggestions are as follows:

[0293] • Good bone health (normal bone density): Increase the risk threshold by 5%, which is equivalent to multiplying the threshold by 1.05.

[0294] Mild bone mineral density loss: Maintain standard risk threshold.

[0295] • Moderate bone mineral density loss: Lower the risk threshold by 10%, which is equivalent to multiplying the threshold by 0.90.

[0296] • Severe bone mineral density loss (osteoporosis): Lower the risk threshold by 20%, which is equivalent to multiplying the threshold by 0.80.

[0297] (7) Recovery Capacity and Fatigue Recovery Cycle: Individuals exhibit differences in muscle and bone recovery capacity, particularly in the speed of fatigue recovery after high-intensity training. Individuals with stronger recovery capabilities accumulate fatigue more slowly; therefore, the risk threshold can be appropriately increased. Adjustment suggestions are as follows:

[0298] • Strong resilience: Increase the risk threshold by 5%, which is equivalent to multiplying the threshold by 1.05.

[0299] • Moderate resilience: Maintain standard risk threshold.

[0300] • Poor recovery ability: Lower the risk threshold by 5%, which is equivalent to multiplying the threshold by 0.95.

[0301] (8) Training Intensity and Frequency: Frequent high-intensity training can exacerbate fatigue accumulation, leading to increased stress on bones. During the assessment, the threshold can be further fine-tuned based on the runner's training intensity and frequency. Adjustment suggestions are as follows:

[0302] • High training frequency (more than 5 times per week) and high intensity: Lower the risk threshold by 10%, which is equivalent to multiplying the threshold by 0.90.

[0303] • Moderate training frequency (3-5 times per week) and moderate intensity: maintain standard risk threshold.

[0304] • Low training frequency (less than 3 times per week) and low intensity: appropriately increase the risk threshold by 5%, that is, multiply the threshold by 1.05.

[0305] In practical applications, the above factors can be combined to calculate an individual's risk threshold, resulting in a threshold that better reflects the individual. For example, for a runner who weighs 80kg, is 45 years old, has 5 years of running experience, normal bone density, and a high training frequency, the following comprehensive risk assessment formula can be applied:

[0306] Individualized risk threshold = Standard risk threshold * 0.95 (age-adjusted) * 1.10 (exercise history-adjusted) * 0.95 (weight-adjusted) * 1.05 (bone health-adjusted) * 0.90 (training intensity-adjusted)

[0307] The risk assessment model above integrates various individual factors through a product approach to obtain a more personalized risk threshold. This final individualized threshold will be used to determine whether the runner's fatigue injury value (Dtotal) has reached a dangerous level, thus enabling a scientific risk assessment.

[0308] To achieve more scientific risk management, runners can be dynamically monitored to continuously track their fatigue injury levels. By regularly calculating Dtotal and updating risk assessment results, trends in fatigue accumulation can be identified, allowing for early prevention of injuries. We use the following two steps for dynamic monitoring: First, by collecting real-time data on pressure insoles and leg sleeves, we estimate tibial force during running and update fatigue injury levels within the gait cycle via a feedback system. Second, by utilizing the Dtotal trend for each training cycle, we predict the future rate of fatigue injury accumulation and changes in risk level. If fatigue injury values ​​gradually increase and approach high-risk or extremely high-risk thresholds, we recommend taking preventative measures in advance.

[0309] S2 uses a prediction model to extract features and assign weights to the predicted ankle joint net force and muscle force, thereby estimating the tibial load of the monitored person's lower limb.

[0310] Example 20:

[0311] The wearable monitoring device for lower limb movement load based on a flexible pressure sensor, the main technical contents of which are described in any one of Examples 11 to 19, and further, the time series data prediction method includes, but is not limited to, Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Bidirectional Long Short-Term Memory (BiLSTM).

[0312] Example 21:

[0313] See Figures 1 to 3 The main technical contents of the wearable monitoring device for lower limb movement load based on flexible pressure sensors include:

[0314] The real-time lower limb tibial load monitoring device consists of the following components: a flexible sensing insole, a flexible sensing leg sleeve, hardware circuitry (which may include IMU sensors), and related software algorithms and a user interface, as shown in the attached document. Figure 1 As shown.

[0315] A wide-range, high-sensitivity flexible tactile sensor was selected, and flexible sensing insoles and leg sleeves were developed based on human tibial load analysis and muscle group force characteristics. The flexible tactile sensors are distributed, and their distribution map will be designed based on muscle group distribution or foot pressure points, and according to the detection requirements (see attached diagram). Figure 2 - Appendix Figure 3 Flexible tactile sensors can be distributed in various ways, including but not limited to array-based and random distributions. Their measurement principles include, but are not limited to, resistive, piezoelectric, capacitive, triboelectric, and ionoelectric methods. Signals detected by flexible sensing insoles are used to predict parameters such as ground reaction force and ankle joint net torque. Flexible sensing leg sleeves are used to predict muscle force, and then an algorithmic model is used to estimate the tibial load on the lower limb.

[0316] The detailed principle of predicting parameters such as ground reaction force and ankle joint net torque using signals detected by flexible sensing insoles is as follows: Starting from the Newton-Euler equations, a physical model can be constructed that outputs from the flexible sensing insole regarding the relationship between ankle joint net force and torque. The force and torque defined at the proximal end of the foot can be expressed as:

[0317] F a =m f (a f -g)-F grf (1)

[0318]

[0319] Ankle joint net torque and τ grf r a,grf F grf Related, and the output of the pressure insole can be calculated down to these three parameters. First, the coordinates r of the point in the foot coordinate system are... F Switch to global coordinate system coordinates r G The point of application of the ground force, r a,grf Since the plantar pressure center (CoP) can be represented, the overall CoP of the plantar surface in the local coordinate system can be expressed by the weighted average method as follows:

[0320]

[0321] In actual modeling, to simplify mechanical calculations, only the vertical ground reaction force is considered, while the horizontal component is ignored. It can be approximated as follows:

[0322]

[0323] The reaction torque can be estimated from the force obtained from the pressure insole. and its point of action The position is calculated relative to the reference point r. ref The torque (usually chosen as the origin of the ankle or foot coordinate system) is expressed as follows:

[0324]

[0325] Substituting equations (3), (4), and (5) into equations (1) and (2) will establish a biomechanical model relationship between the pressure insole output and the net force and torque of the ankle joint.

[0326] The detailed principle of predicting muscle force using a flexible sensing leg sleeve is as follows: The muscle deformation measured by the flexible sensing ring essentially reflects the surface expansion stress caused by muscle contraction. The surface stress of the muscle in the cross-section mainly originates from the uniformly distributed σ along the radial direction. xx Therefore, the muscle force can be obtained as the integral of the longitudinal stress on the cross section:

[0327]

[0328] Substituting into the viscoelastic constitutive equation, we get:

[0329]

[0330] Assuming surface stress σ surf (t) and internal strain ε xx (t) satisfies the spatial weighting relation:

[0331]

[0332] Where G(x,y,t) is the spacetime Green's function, characterizing the delay and attenuation of stress transmission. Finally, combining the above two formulas, the total muscle force can be expressed as the spacetime integral of the epidermal muscle deformation:

[0333]

[0334] Combining continuum mechanics and the assumption of transverse isotropy, stress-strain and viscoelastic constitutive relations were established, and the total force of the muscle was expressed as the spatiotemporal integral of the surface muscle deformation.

[0335] In addition to physical models, algorithmic models can be established between the output of the pressure insole and the net ankle torque, and between the output of the sensor leg sleeve and muscle force. Specifically, data is simultaneously collected from the pressure insole and laboratory equipment (motion capture system, force table). The data collected from the laboratory equipment is input into Opensim to calculate the net ankle torque as the output, and the output of the pressure insole is used as the input to establish a fitting model between the pressure insole output and the net ankle torque. Similarly, data is simultaneously collected from the sensor leg sleeve and an IsoMed 2000 (Germany) isokinetic muscle strength machine. The muscle torque collected by the isokinetic muscle strength machine is used as the output, and the sensor leg sleeve is used as the input to establish a fitting model between the sensor leg sleeve output and muscle force.

[0336] The hardware circuitry includes, but is not limited to, external interface circuitry, operational amplifiers, microcontroller units (MCUs), power supplies, serial ports, wireless communication modules (e.g., Bluetooth, ZigBee, Wi-Fi), and IMU modules, enabling data acquisition, analysis, and transmission. Flexible printed circuit boards can also be used for the hardware circuitry. The software algorithm for tibial load estimation is a pre-trained prediction model, which is loaded into the application. This training model can be a combination of various time-series data prediction methods and attention mechanisms, including but not limited to convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and bidirectional long short-term memory networks (BiLSTMs), performing feature extraction and weight assignment on the data to ultimately achieve tibial load prediction.

[0337] The user interface refers to an APP installed on a smart terminal, which can view the lower limb movement status and real-time bone load curve of the monitored person in real time, and can also provide lower limb injury warnings for the monitored person.

Claims

1. A wearable monitoring device for lower limb motion load based on a flexible pressure sensor, characterized in that, include: Flexible sensing leg sleeve, flexible sensing insole, data processing module, tibial load prediction module; Flexible tactile sensors are arranged on both the flexible sensing leg sleeve and the flexible sensing insole. The flexible sensing leg sleeve uses a flexible tactile sensor to collect muscle movement signals of the leg muscles of the monitored person; The flexible sensing insole uses a flexible tactile sensor to collect force signals on the foot of the monitored person. The data processing module analyzes the collected muscle movement signals of the leg muscles to obtain the predicted muscle force. The data processing module analyzes the collected force signals on the foot to obtain the predicted net force of the ankle joint. The tibial load prediction module estimates the tibial load of the monitored person's lower limb based on the predicted net ankle force and muscle force, and obtains a real-time bone load curve.

2. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The steps for analyzing the collected muscle movement signals of the leg muscles to obtain the predicted muscle force are as follows: A1 constructs a three-dimensional coordinate system and calculates the surface stress of the muscle in the cross-section, as shown below: In the formula, t represents time; F1(t) represents the surface stress of the muscle in the cross-section; σ xx (t) represents the longitudinal stress on the cross-section; A represents the cross-sectional area of ​​the muscle. Substituting a2, representing the surface stress of the muscle in the cross-section, into the viscoelastic constitutive equation, we get the following: In the formula, τ σ τ ε These are the stress relaxation and delay time constants, respectively; F is the muscle force. υ ⊥ These represent the equilibrium elastic modulus along the fiber direction, the transient transverse elastic modulus, and Poisson's ratio, respectively; the x-axis is defined along the muscle's long axis, and the horizontal plane is defined as the yz plane, ε... xx (t) represents the internal strain along the x-axis; For internal strain ε xx The differential of (t); σ yy Let σ be the stress along the y-axis. zz The stress is along the z-axis. a3 sets the internal strain ε xx (t) and surface stress σ surf (t) satisfies the spatial weighting relation as shown below: In the formula, τ represents time; ε xx (x,y,t) represents the internal strain along the x-direction at position (x,y) at time t; σ surf (x,y,τ) represents the surface stress of the muscle at position (x,y) at time τ; G(x,y,t-τ) is the spacetime Green's function; The predicted muscle force is calculated using a4, as shown below: In the formula, F(t) is the predicted muscle force; K1(x,y) and K2(x,y) are both spatial weight coefficients; τ1 and τ2 are both characteristic time constants.

3. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The steps for analyzing the collected foot force signals to obtain the predicted ankle joint net force are as follows: b1. Establish a foot coordinate system and determine the point of application r of the ground force using a weighted average method. a,grf As shown below: In the formula, i is the sensor unit index; N is the total number of sensors; r i p represents the position coordinates of the i-th sensor; i Let A be the pressure value at the i-th sensor; i Let be the sensing area at the i-th sensor; Let be the center of pressure on the sole of the foot; T is the homogeneous coordinate transformation matrix corresponding to the transformation from the foot coordinate system to the global coordinate system; b2 Calculate the ground reaction force F grf As shown below: b3 Calculate the ground reaction moment τ grf As shown below: In the formula, r ref Use as a reference point; b4 calculates the net force and net torque at the ankle joint, as shown below: F a =m f (a f -g)-F grf (8) In the formula, F a Net force at the ankle joint; τ a The net torque of the ankle joint; m f For foot mass; a f F is the acceleration of the foot's center of mass. grf The ground reaction force is g; g is the acceleration due to gravity; t is time; I f ω is the moment of inertia of the foot's center of mass. f The foot angular velocity; r a,grf The point of application of the ground force; r a The position of the center of gravity; τ grf This is the ground reaction torque.

4. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The wearable monitoring device for lower limb exercise load also includes a data transmission module and a user interface; The data transmission module is used for signal transmission between the flexible tactile sensor, the data processing module, and the tibial load prediction module arranged on the flexible sensing leg sleeve and the flexible sensing insole. The signal transmission methods include Bluetooth transmission, ZigBee transmission, Wi-Fi transmission, and wired transmission. The user interface is integrated into the smart terminal; The user interface is used to display the lower limb movement status and real-time bone load curve of the monitored person; The lower limb movement states include standing, running, and walking; When the load on the tibia of the monitored person's lower limb exceeds the preset fatigue threshold, a lower limb injury warning is displayed on the user interface.

5. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The flexible sensing leg sleeve and flexible sensing insole each integrate one or more flexible tactile sensors. When multiple flexible tactile sensors are integrated, these flexible tactile sensors are arranged in a distributed manner; The distributed layout includes, but is not limited to, array-based distribution and random distribution; The integration method includes, but is not limited to, physical bonding, knitting processes, and integrated fabrication of sensors and fabric substrates.

6. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The flexible tactile sensors include resistive sensors, piezoelectric sensors, capacitive sensors, triboelectric sensors, and ionoelectric sensors.

7. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The data processing module stores an ankle joint net force algorithm model; The steps for constructing the ankle joint net force algorithm model are as follows: c1 simultaneously acquires foot force signals detected by flexible sensor insoles and data detected by laboratory equipment; The laboratory equipment includes a motion capture system and a force measuring table; c2 inputs the data from laboratory equipment testing into Opensim to calculate the net ankle joint torque; c3 uses the foot force signal detected by the flexible sensor insole as input and the ankle joint net torque as output to establish an ankle joint net force algorithm model.

8. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The data processing module stores a muscle strength algorithm model; The steps for constructing the muscle strength algorithm model are as follows: d1 synchronously acquires the muscle movement signals of the leg muscles detected by the flexible sensor leg sleeve and the muscle torque detected by the isokinetic muscle strength machine; d2 uses the muscle movement signal of the leg muscles as input and the muscle torque as output to establish a muscle force algorithm model.

9. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 1, characterized in that, The steps for estimating the tibial load of the monitored person's lower limb are as follows: s1 constructs a prediction model that combines time series data prediction methods with an attention mechanism; S2 uses a prediction model to extract features and assign weights to the predicted ankle joint net force and muscle force, thereby estimating the tibial load of the monitored person's lower limb.

10. The wearable monitoring device for lower limb motion load based on a flexible pressure sensor according to claim 9, characterized in that, The time series data prediction methods include, but are not limited to, convolutional neural networks, long short-term memory networks, and bidirectional long short-term memory networks.

Images