Intelligent sports shoes based on multi-mode sensing

Abstract

The invention discloses an intelligent sneaker based on multi-mode sensing. The intelligent sneaker comprises a sneaker body, and a motion posture acquisition module, a physiological sign acquisition module, a data processing module and a power supply module which are arranged on the sneaker body, the motion posture acquisition module is used for acquiring motion posture data of a user, and the physiological sign acquisition module is used for acquiring physiological sign data of the user; the data processing module is electrically connected with the motion posture acquisition module and the physiological sign acquisition module, receives and processes the motion posture data and the physiological sign data, and then wirelessly transmits the processed data to an external terminal; and the power supply module supplies power to each power utilization module. The motion posture acquisition module and the physiological sign acquisition module are deeply integrated in a pair of shoes, a complete motion-physiological data closed loop is constructed, and accurate multi-dimensional information such as posture, heart rate and breath can be synchronously acquired without depending on external equipment.

Description

Technical Field

[0001] This invention relates to the field of smart wearable technology, and more specifically to a smart sports shoe based on multimodal sensing. Background Technology

[0002] Smart sports shoes are high-tech products that combine sensors, microprocessors, wireless communication and other technologies with traditional footwear to achieve functions such as sports monitoring, health analysis and adaptive adjustment. They are widely used in sports and fitness, health management and rehabilitation training.

[0003] Existing smart sports shoes mainly focus on single-dimensional data collection, such as step counting, running posture analysis, or foot pressure distribution monitoring. Therefore, they are used more as data loggers than to proactively protect users' health based on data. Summary of the Invention

[0004] The purpose of this invention is to provide a smart sports shoe based on multimodal sensing to solve the above-mentioned problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A smart sports shoe based on multimodal sensing includes a shoe body and a motion posture acquisition module, a physiological sign acquisition module, a data processing module and a power supply module disposed on the shoe body;

[0007] The motion posture acquisition module collects the user's motion posture data;

[0008] The physiological sign acquisition module collects the user's physiological sign data;

[0009] The data processing module is electrically connected to the motion posture acquisition module and the physiological sign acquisition module respectively, receives and processes motion posture data and physiological sign data, and then transmits the processed data wirelessly to an external terminal.

[0010] The power module supplies power to each power-consuming module.

[0011] Preferably, the motion posture acquisition module includes a first inertial measurement unit, which is disposed at the tongue and / or heel of the shoe to acquire three-dimensional acceleration and angular velocity signals of the shoe.

[0012] Preferably, the physiological sign acquisition module includes a heart rate acquisition unit and a respiration acquisition unit. The heart rate acquisition unit includes a physiological sensor disposed on the tongue of the shoe body. The physiological sensor is electrically connected to the conductive shoelaces to collect the user's heart rate or electrocardiogram signal. The respiration acquisition unit includes a pressure sensing array and a second inertial measurement unit disposed in the insole of the shoe body, and the two work together to collect the user's respiratory rate signal.

[0013] Preferably, the conductive shoelace adopts a multi-layer composite structure, which includes an outer layer made of polyester or nylon yarn, a middle layer made of spirally wound flexible conductive fibers, and an inner layer made of an insulating, skin-friendly coating, wherein the flexible conductive fibers are silver-plated copper yarn or carbon fiber composite filaments with a linear resistance ≤20Ω / m.

[0014] Preferably, the physiological sensor includes a photoplethysmography (PPG) sensor and an electrocardiogram (ECG) electrode. The PPG sensor is disposed on the inside of the shoe tongue and contacts the dorsalis pedis artery region of the user to collect signals. The ECG electrode consists of two flexible Ag / AgCl electrodes and is electrically connected to the conductive shoe strap to form an ECG signal acquisition circuit.

[0015] Preferably, the pressure sensing array is a flexible piezoresistive sensing unit array with a thickness of no more than 0.8 mm, and the sensing pressure range of a single sensing unit is 0-50 N with an accuracy of ±0.5 N.

[0016] Preferably, the respiratory rate signal acquisition method is as follows: the rhythmic fluctuation signal of plantar pressure detected by the pressure sensing array and the upper body micro-motion resonance signal detected by the second inertial measurement unit are fused together, and the user's respiratory rate is extracted by an adaptive filtering algorithm.

[0017] Preferably, the data processing module and the power module are integrated into a detachable module, which is encapsulated in a waterproof housing and magnetically attached to the side cavity of the sole of the shoe.

[0018] Preferably, the data processing procedure of the data processing module is as follows:

[0019] S1. Calculate the vector amplitude based on the motion posture data, and perform frequency domain analysis to extract the user's step frequency. Then, determine the motion intensity level based on the root mean square value of the step frequency and the vector amplitude. In this identification process, hysteresis control logic is introduced to prevent frequent switching near the intensity threshold.

[0020] S2. The sampling rate of the physiological sign acquisition module is dynamically adjusted according to the identified exercise intensity level. When the exercise intensity is lower than the first threshold, a first sampling rate of 50Hz is used; when the exercise intensity is higher than the second threshold, a second sampling rate of 200Hz is used; when the exercise intensity is between the first threshold and the second threshold, an intermediate sampling rate of 100Hz is used. The 50Hz sampling rate is applicable to walking and below, and the 200Hz sampling rate is applicable to running and above.

[0021] S3. Using the step frequency and foot pressure rhythm information extracted from the motion posture data as reference input signals, the motion artifacts in the heart rate or electrocardiogram signal are modeled and corrected by an adaptive noise canceller.

[0022] Preferably, the method for correcting motion artifacts by the adaptive noise canceller in step S3 is as follows: based on step frequency information, the periodic time point of motion artifacts is located in the mixed physiological signal, and then a time-varying filter is constructed near the time point to extract the noise model and cancel it.

[0023] By adopting the above technical solution, the present invention has the following advantages compared with the prior art:

[0024] 1. This invention provides a smart sports shoe based on multimodal sensing. By deeply integrating the motion posture acquisition module and the physiological sign acquisition module into a single shoe, a complete motion-physiological data closed loop is constructed. It can synchronously acquire accurate multi-dimensional information such as posture, heart rate, and respiration without relying on external devices.

[0025] 2. This invention provides a smart sports shoe based on multimodal sensing. By integrating the rhythmic fluctuation of foot pressure and the upper body micro-movement resonance signal captured by inertial measurement, and extracting the breathing frequency through an adaptive filtering algorithm, it can achieve truly imperceptible and comfortable continuous breathing monitoring without the need for any additional sensors on the chest or respiratory tract. Compared with spirometers, it has a smaller comparison error.

[0026] 3. This invention provides a smart sports shoe based on multimodal sensing, which intelligently switches the sampling rate of physiological sensors according to the real-time exercise intensity. This ensures sufficient data accuracy for advanced processing during high-intensity exercise, while significantly reducing system power consumption and extending battery life during low-intensity exercise, demonstrating the system's intelligence and efficiency.

[0027] 4. This invention provides a smart sports shoe based on multimodal sensing. By using step frequency and foot pressure rhythm information provided by inertial measurement and pressure array as reference signals, the adaptive noise canceller algorithm dynamically constructs a model that is highly correlated with noise in the physiological signal and accurately cancels it. Compared with traditional filtering methods, this can greatly eliminate motion interference while retaining useful physiological signal features to the maximum extent, thereby ensuring high accuracy of R-wave detection rate even under high-intensity running conditions.

[0028] 5. This invention provides a smart sports shoe based on multimodal sensing, which adopts a multi-layer composite conductive shoelace. While ensuring excellent conductivity, it also takes into account the mechanical binding function and skin-friendly comfort of traditional shoelaces, avoiding discomfort or allergies that may be caused by direct contact of metal materials with the skin. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the overall structure of the shoe body of the present invention;

[0030] Figure 2 This is a schematic diagram of the tongue position structure of the present invention;

[0031] Figure 3 This is a schematic diagram of the insole structure of the present invention;

[0032] Figure 4 This is a schematic diagram of the conductive shoelace fabric layers of the present invention;

[0033] Figure 5 This is a flowchart of the data processing method of the present invention.

[0034] Figure label:

[0035] 1. Shoe body; 11. Cavity; 2. Motion posture acquisition module; 21. First inertial measurement unit; 3. Physiological characteristic acquisition module; 31. Heart rate acquisition unit; 311. Photoplethysmography (PPG) sensor; 312. Electrocardiogram (ECG) electrode; 32. Respiration acquisition unit; 321. Pressure sensor array; 322. Second inertial measurement unit; 4. Data processing module; 5. Power supply module; 6. Conductive shoelaces; 61. Outer layer; 62. Middle layer; 63. Lining; 7. Detachable module. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0037] Example

[0038] Please refer to Figures 1 to 5 As shown, this invention discloses a smart sports shoe based on multimodal sensing, which includes a shoe body 1 and a motion posture acquisition module 2, a physiological characteristic acquisition module 3, a data processing module 4, and a power supply module 5 disposed on the shoe body 1. The motion posture acquisition module 2 collects the user's motion posture data, and the physiological characteristic acquisition module 3 collects the user's physiological characteristic data. The data processing module 4 is electrically connected to the motion posture acquisition module 2 and the physiological characteristic acquisition module 3 respectively, receives and processes the motion posture data and physiological characteristic data, and then transmits the processed data wirelessly to an external terminal. The power supply module 5 supplies power to each power module.

[0039] The motion posture acquisition module 2 includes a first inertial measurement unit 21, which is located at the tongue of the shoe body 1 to acquire the three-dimensional acceleration and angular velocity signals of the shoe body 1. In this embodiment, the first inertial measurement unit 21 uses a six-axis sensor of model MPU6050, which integrates a three-axis accelerometer and a three-axis gyroscope, and can accurately acquire the three-dimensional acceleration and angular velocity signals of the shoe body 1. The first inertial measurement unit 21 can also be located at the heel of the shoe body 1 to acquire the three-dimensional acceleration and angular velocity signals of the shoe body 1.

[0040] To ensure the accuracy and stability of the acquired signals, the first inertial measurement unit 21 is fixed to the tongue of the shoe body 1 with waterproof adhesive. The tongue is positioned close to the instep, resulting in minimal relative displacement with the foot during movement. This effectively reduces the impact of sensor wobbling on signal acquisition, ensuring that the acquired three-dimensional acceleration and angular velocity signals accurately reflect the user's movement posture. For example, during running, the first inertial measurement unit 21 can detect acceleration changes during foot landing and swinging phases, thus providing raw data for motion analysis.

[0041] The physiological sign acquisition module 3 includes a heart rate acquisition unit 31 and a respiration acquisition unit 32. The heart rate acquisition unit 31 includes a physiological sensor located on the tongue of the shoe body 1. The physiological sensor is electrically connected to the conductive shoelace 6 to collect the user's heart rate or electrocardiogram signal. The respiration acquisition unit 32 includes a pressure sensor array 321 and a second inertial measurement unit 322 located in the insole of the shoe body. The two work together to collect the user's respiratory rate signal.

[0042] The physiological sensors include a photoplethysmography (PPG) sensor 311 and electrocardiogram (ECG) electrodes 312. The PPG sensor, an integrated sensor (model MAX30102), is located on the inside of the shoe tongue 1, in close contact with the dorsalis pedis artery region. This sensor emits light signals of a specific wavelength and receives the light signals reflected by blood in the blood vessels. It collects pulse wave signals based on changes in the intensity of the light signals and then calculates heart rate data. The ECG electrodes 312 are two flexible Ag / AgCl electrodes, fabricated on a flexible substrate on the inside of the shoe tongue using a screen printing process. The two electrodes are 5 cm apart and located on either side of the dorsalis pedis artery. The flexible Ag / AgCl electrodes have good biocompatibility and conductivity, allowing them to adhere closely to the skin and collect ECG signals.

[0043] The conductive shoelace 6 employs a multi-layered composite structure. The outer layer 61 is woven from polyester yarn, possessing excellent abrasion resistance and flexibility to meet the daily needs of shoelace use. The middle layer 62 consists of spirally wound flexible conductive fibers, using silver-plated copper yarn with a linear resistance ≤15Ω / m. This conductive fiber not only exhibits excellent conductivity but also good flexibility, allowing it to bend and deform with the shoelace without affecting its conductivity. The inner layer 63 is an insulating, skin-friendly coating made of medical-grade silicone material, preventing discomfort caused by direct contact between the conductive fibers and the skin, while also providing insulation protection. The conductive shoelace 6 is electrically connected to the electrocardiogram electrode 312 via conductive silver paste, forming a complete electrocardiogram signal acquisition circuit.

[0044] The pressure sensor array 321 employs a flexible piezoresistive sensor unit array, fabricated using flexible printed circuit board technology, with an overall thickness of only 0.6mm, meeting the requirements for a thin and lightweight insole. This array consists of 8×8 sensor units distributed in the forefoot area of ​​the insole. Each sensor unit can sense pressure ranging from 0-50N with an accuracy of ±0.3N, enabling precise detection of pressure changes in different areas of the sole. The pressure sensor array 321 is bonded to the inside of the insole with waterproof adhesive, deforming synchronously with the insole to ensure constant close contact with the sole during exercise.

[0045] The second inertial measurement unit 322 uses the same MPU6050 sensor as the first inertial measurement unit 21, and is fixed to the arch of the insole. This position is relatively stable during movement and can effectively collect the micro-motion resonance signals transmitted from the upper body to the sole of the foot through the lower limbs.

[0046] In this way, smart athletic shoes can monitor user information from the dimensions of physiological signs and movement posture data. However, in this process, strong mechanical interference can easily contaminate the weak physiological signals collected, resulting in motion artifacts. Specifically, motion artifacts originate from the relative displacement between the sensor and the skin caused by actions such as running and jumping, changes in blood flow due to muscle compression, and vibrations of the shoe itself. The amplitude of these interference signals is often much larger than that of physiological signals such as electrocardiogram (ECG) and photoplethysmography (PPG), resulting in an extremely low signal-to-noise ratio of the collected signals, making it impossible to stably and accurately extract key physiological parameters such as heart rate and respiration during exercise.

[0047] Therefore, in a preferred embodiment, the present invention proposes a method for acquiring respiratory frequency signals by combining fusion detection and algorithm extraction: the pressure sensor array 321 detects the rhythmic fluctuation signal of plantar pressure, the second inertial measurement unit 322 detects the upper body micro-movement resonance signal, and after transmitting these two signals to the data processing module 4, the signals are denoised and fused by an adaptive filtering algorithm to finally extract the user's respiratory frequency.

[0048] Specifically:

[0049] First, bandpass filtering is applied to the pressure array signal 321 to extract the rhythmic fluctuation component. Respiration causes slight fluctuations in the body's center of gravity, resulting in low-frequency fluctuations in plantar pressure synchronized with respiration (0.1–0.5 Hz, corresponding to a respiratory rate of 12–30 breaths / minute). However, this fluctuation is susceptible to gait impacts (high-frequency noise, 1–5 Hz) and weight changes (baseline drift). Therefore, high-pass filtering (cutoff frequency 0.05 Hz) or a sliding window detrending algorithm is needed to remove slowly changing baselines (such as weight changes and continuous postural shifts).

[0050] Secondly, the Z-axis acceleration signal from the second inertial measurement unit 322 of the heel is filtered in the 0.2-0.5Hz band to extract the upper body resonance component. The upper body micro-movement resonance signal driven by the rise and fall of the chest / abdomen during breathing is detected as low-frequency vibration (same as the breathing frequency) by the second inertial measurement unit 322, but it is easily affected by limb swaying (high-frequency noise, 2-10Hz) and posture adjustment. Therefore, a bandpass filter of 0.2-0.5Hz is needed to initially retain the breathing frequency range and filter out high-frequency noise.

[0051] Finally, the above-mentioned rhythmic fluctuation components and upper body resonance components are input into a normalized least mean square adaptive filter for fusion, and the filter output is the estimated respiratory rate signal.

[0052] In one fusion example, the respiratory component consistent with the P signal (rhythmic fluctuation component) is first extracted from the M signal (upper body resonance component). That is, the preprocessed P signal is used as the expected value, and the weight of the filter is adjusted by NLMS to make the output close to the expected value, so as to obtain the enhanced signal y1(n). Secondly, the respiratory component consistent with M is extracted from P in the same way to obtain the enhanced signal y2(n). Finally, the two enhanced signals y1(n) and y2(n) are fused, and the weights are dynamically allocated based on the signal-to-noise ratio (SNR) of the two.

[0053] Calculate the real-time SNR (signal power / noise power, noise power is estimated by non-breathing band energy) of y1(n) and y2(n); weighting formula: w1=SNR1 / (SNR1+SNR2), w2=1-w1; fuse the signals, Sfusion(n)=w1·y1(n)+w2·y2(n).

[0054] Perform a short-time Fourier transform (STFT) on Sfusion(n) and find the power peak in the 0.1–0.5 Hz frequency band, which is the respiratory frequency signal.

[0055] In this embodiment, the data processing module 4 uses an STM32L476RG low-power microcontroller as its core processing chip. This chip has powerful data processing capabilities and low power consumption, making it suitable for long-term battery life requirements. Wireless transmission uses a Bluetooth 5.0 module, which supports low-power data transmission and has a transmission distance of up to 10m, enabling stable transmission of processed data to external terminals such as mobile phones and tablets.

[0056] The power module 5 uses a 500mAh flexible lithium polymer battery, which is characterized by good flexibility, small size and high energy density, and can adapt to the structural requirements of the shoe body.

[0057] The data processing module 4 and the power supply module 5 are integrated into a detachable module 7. The detachable module 7 uses a waterproof enclosure with an IP67 rating. The enclosure is made of ABS engineering plastic and has good waterproof, dustproof and impact resistance.

[0058] The sole of the shoe body 1 has a cavity 11 on its side that matches the detachable module 7. The cavity 11 contains magnetic contacts, and the detachable module 7 is magnetically attached to it. This magnetic connection not only facilitates the removal and installation of the module but also ensures a stable and reliable electrical connection between the module and the shoe body. The power module is electrically connected to each power-consuming module via wires, providing them with operating voltage. The detachable module also features a Micro-USB charging port for convenient battery charging.

[0059] The data processing module's data processing procedure is as follows:

[0060] S1. Calculate the vector amplitude based on motion posture data and perform frequency domain analysis to extract the user's step frequency. Then, determine the motion intensity level based on the root mean square value of the step frequency and vector amplitude. Hysteresis control logic is introduced during this recognition process to prevent frequent switching near the intensity threshold. Specifically:

[0061] Based on the three-dimensional acceleration and angular velocity signals acquired by the motion posture acquisition module 2, their vector amplitudes are calculated. A Fast Fourier Transform (FFT) is performed on the vector amplitudes, and frequency domain analysis is conducted to extract the user's step frequency. The motion intensity level is determined based on the step frequency and the root mean square value of the vector amplitude. The following settings are defined:

[0062] A cadence of ≤80 steps / minute and a root mean square value of vector amplitude ≤0.5g are considered light exercise.

[0063] A cadence of 80-150 steps per minute and a root mean square value of vector amplitude of 0.5-1.5g constitute moderate exercise.

[0064] A step frequency of ≥150 steps / minute and a root mean square value of vector amplitude ≥1.5g are considered high-intensity exercises.

[0065] Hysteresis control logic is introduced during the recognition process, and the following settings are made:

[0066] The threshold for switching from light to moderate exercise is 85 steps / minute, and the threshold for the root mean square value of the vector amplitude is 0.6g.

[0067] The threshold for switching from moderate to light exercise is 75 steps / minute and the threshold for the root mean square value of vector amplitude is 0.4g.

[0068] The threshold for switching from moderate to high-intensity exercise is 155 steps / minute, and the threshold for the root mean square value of the vector amplitude is 1.6g.

[0069] The threshold for switching from high-intensity exercise to moderate-intensity exercise is 145 steps / minute, and the threshold for the root mean square value of vector amplitude is 1.4g, to prevent frequent switching near the intensity threshold.

[0070] S2. The sampling rate of the physiological sign acquisition module 4 is dynamically adjusted according to the identified exercise intensity level. When the exercise intensity is below the first threshold (mild exercise), a first sampling rate of 50Hz is used; when the exercise intensity is above the second threshold (high-intensity exercise), a second sampling rate of 200Hz is used; when the exercise intensity is between the first and second thresholds (moderate exercise), an intermediate sampling rate of 100Hz is used. The 50Hz sampling rate is suitable for walking and below, meeting the basic physiological sign monitoring needs; the 200Hz sampling rate is suitable for running and above, accurately capturing rapidly changing physiological signals.

[0071] S3. Motion Artifact Correction: Using cadence and foot pressure rhythm information extracted from motion posture data as reference input signals, an adaptive noise canceller is used to model and correct motion artifacts in heart rate or ECG signals. Specifically, the correction method involves locating the periodic time points of motion artifacts in the mixed physiological signals based on cadence information. The period corresponding to the cadence is T = 60 / cadence. A time-varying filter is constructed within 50ms before and after the start of each period. The bandwidth of this filter is dynamically adjusted with the exercise intensity: 1-5Hz for light exercise, 1-10Hz for moderate exercise, and 1-15Hz for high-intensity exercise. The noise model is extracted through the time-varying filter and then canceled out from the mixed physiological signals to correct motion artifacts.

[0072] In this example, the adaptive noise canceller uses the least mean square or recursive least square algorithm to dynamically update the filter coefficients.

[0073] When exercising, the modules of the smart sports shoe in this embodiment work together according to the following process:

[0074] Start the device: The user inserts the detachable module 7 into the side cavity 11 of the shoe sole, and the power module 5 starts to supply power to each power module, and the device starts up and enters the working state.

[0075] Data acquisition: The first inertial measurement unit 21 of the motion posture acquisition module 2 acquires the three-dimensional acceleration and angular velocity signals of the shoe body 1; in the physiological signs acquisition module 3, the PPG sensor 311 and electrocardiogram electrode 312 of the heart rate acquisition unit 31 acquire heart rate or electrocardiogram signals in cooperation with the conductive shoelace 6, and the pressure sensor array 321 and the second inertial measurement unit 322 of the breathing acquisition unit 32 acquire the rhythmic fluctuation signal of the sole pressure and the upper body micro-motion resonance signal, respectively.

[0076] Data processing: The data processing module 4 receives the data transmitted by each acquisition module, first identifies the exercise intensity level, then dynamically adjusts the sampling rate of the physiological sign acquisition module 3 according to the exercise intensity level, and finally corrects the motion artifacts in the heart rate or electrocardiogram signal through an adaptive noise canceller, and extracts the respiratory frequency by fusing the signals from the pressure sensor array 321 and the second inertial measurement unit 322.

[0077] Data transmission: The data processing module 4 transmits the processed motion posture data (step frequency, exercise intensity level, etc.) and physiological sign data (heart rate, electrocardiogram signal, respiratory rate, etc.) to an external terminal via Bluetooth 5.0. Users can view the relevant data through the APP on the terminal to achieve real-time monitoring of their exercise and health status.

[0078] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart sports shoe based on multimodal sensing, characterized in that: It includes a shoe body and a motion posture acquisition module, a physiological sign acquisition module, a data processing module, and a power supply module installed on the shoe body; The motion posture acquisition module collects the user's motion posture data; The physiological sign acquisition module collects the user's physiological sign data; The data processing module is electrically connected to the motion posture acquisition module and the physiological sign acquisition module respectively, receives and processes motion posture data and physiological sign data, and then transmits the processed data wirelessly to an external terminal. The power module supplies power to each power-consuming module.

2. The smart sports shoe based on multimodal sensing as described in claim 1, characterized in that: The motion posture acquisition module includes a first inertial measurement unit, which is located at the tongue and / or heel of the shoe to acquire three-dimensional acceleration and angular velocity signals of the shoe.

3. The smart sports shoe based on multimodal sensing as described in claim 1, characterized in that: The physiological sign acquisition module includes a heart rate acquisition unit and a respiration acquisition unit. The heart rate acquisition unit includes a physiological sensor located on the tongue of the shoe body. The physiological sensor is electrically connected to the conductive shoelaces to collect the user's heart rate or electrocardiogram signal. The respiration acquisition unit includes a pressure sensor array and a second inertial measurement unit located in the insole of the shoe body. The two work together to collect the user's respiratory rate signal.

4. A smart sports shoe based on multimodal sensing as described in claim 3, characterized in that: The conductive shoelace adopts a multi-layer composite structure, which includes an outer layer made of polyester or nylon yarn, a middle layer made of spirally wound flexible conductive fibers, and an inner layer made of an insulating skin-friendly coating. The flexible conductive fibers are silver-plated copper yarn or carbon fiber composite filaments with a linear resistance ≤20Ω / m.

5. A smart sports shoe based on multimodal sensing as described in claim 3, characterized in that: The physiological sensor includes a photoplethysmography (PPG) sensor and an electrocardiogram (ECG) electrode. The PPG sensor is disposed on the inside of the shoe tongue and contacts the dorsalis pedis artery region of the user to collect signals. The ECG electrode consists of two flexible Ag / AgCl electrodes and is electrically connected to the conductive shoe strap to form an ECG signal acquisition circuit.

6. A smart sports shoe based on multimodal sensing as described in claim 3, characterized in that: The pressure sensing array is a flexible piezoresistive sensing unit array with a thickness of no more than 0.8 mm, and the sensing pressure range of a single sensing unit is 0-50 N with an accuracy of ±0.5 N.

7. A smart sports shoe based on multimodal sensing as described in claim 3, characterized in that: The method for acquiring the respiratory rate signal is as follows: the rhythmic fluctuation signal of the plantar pressure detected by the pressure sensing array is fused with the upper body micro-motion resonance signal detected by the second inertial measurement unit, and the user's respiratory rate is extracted through an adaptive filtering algorithm.

8. A smart sports shoe based on multimodal sensing as described in claim 1, characterized in that: The data processing module and the power module are integrated into a detachable module, which is encapsulated in a waterproof housing and magnetically attached to the side cavity of the sole of the shoe.

9. A smart sports shoe based on multimodal sensing as described in claim 3, characterized in that: The data processing procedure of the data processing module is as follows: S1. Calculate the vector amplitude based on the motion posture data, and perform frequency domain analysis to extract the user's step frequency. Then, determine the motion intensity level based on the root mean square value of the step frequency and the vector amplitude. In the recognition process, hysteresis control logic is introduced to prevent frequent switching near the intensity threshold. S2. The sampling rate of the physiological sign acquisition module is dynamically adjusted according to the identified exercise intensity level. When the exercise intensity is lower than the first threshold, a first sampling rate of 50Hz is used; when the exercise intensity is higher than the second threshold, a second sampling rate of 200Hz is used; when the exercise intensity is between the first threshold and the second threshold, an intermediate sampling rate of 100Hz is used. The 50Hz sampling rate is applicable to walking and below, and the 200Hz sampling rate is applicable to running and above. S3. Using the step frequency and foot pressure rhythm information extracted from the motion posture data as reference input signals, the motion artifacts in the heart rate or electrocardiogram signal are modeled and corrected by an adaptive noise canceller.

10. A smart sports shoe based on multimodal sensing as described in claim 9, characterized in that: The method for correcting motion artifacts using the adaptive noise canceller described in step S3 is as follows: based on step frequency information, the periodic time points of motion artifacts are located in the mixed physiological signals, and then a time-varying filter is constructed near these time points to extract the noise model and cancel it.

Images