Intelligent lower limb orthosis based on bimodal biofeedback and cooperative control system

Abstract

The invention relates to the technical field of intelligent medical rehabilitation instruments, in particular to an intelligent lower limb orthosis based on bimodal biofeedback and a cooperative control system.The orthosis comprises a 3D printing orthosis body, and topological optimization structural design is adopted; the embedded sensing module comprises a double-channel surface electromyographic sensor, a six-axis inertial measurement unit and a flexible piezoresistive array and is used for collecting electromyographic signals, ankle joint movement data and plantar pressure distribution in real time; the electrical stimulation execution unit is used for triggering an electrical stimulation signal of the multi-channel percutaneous electrode according to the gait phase; by fusing bimodal biofeedback and a 3D printing orthosis, accurate dynamic correction of gait abnormity of a spastic cerebral palsy patient is achieved, rehabilitation efficiency is remarkably improved, meanwhile, the gait cycle is divided into a supporting phase and a swinging phase, differentiated electrical stimulation strategies are designed for different phases, and the rehabilitation efficiency is improved. Gradient material 3D printing and a porous grid structure are adopted, so that the skin contact pressure peak value is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent medical rehabilitation devices, and particularly to an intelligent lower limb orthosis and a collaborative control system based on bimodal biofeedback. Background Art

[0002] Cerebral Palsy (CP) is the most common cause of childhood disability, manifested as movement disorders and abnormal postural development. According to the latest epidemiological survey of CP, the incidence rate of CP in China is 2.48‰, and the prevalence rate among children aged 1 - 6 years is 2.46‰. Among children aged 0 - 14 years in China, there are more than 5 million CP patients. The clinical manifestations of CP patients generally include movement disorders such as dystonia, athetoid, chorea, and ataxia. The abnormal lower limb gait of CP patients will seriously affect walking, standing, and other basic daily activities, and at the same time cause emotional stress and social exclusion, seriously affecting the quality of life of CP patients.

[0003] Spastic cerebral palsy is a common movement disorder disability, manifested as abnormal lower limb gait related to abnormal muscle strength, seriously affecting the quality of life of patients. At present, the clinical treatment methods are limited. Improving gait abnormalities requires lifelong rehabilitation exercises. Functional electrical stimulation has curative effects on abnormal muscle strength, but this rehabilitation therapy currently highly depends on equipment and venues, resulting in low patient compliance. Orthoses also have certain curative effects on improving gait abnormalities, but lack rehabilitation treatment functions. Therefore, designing and developing a 3D - printed lower limb orthosis with functional electrical stimulation treatment function and autonomous electromyogram control meets the innovative development needs of high - end medical devices and is expected to significantly improve the gait abnormalities of CP patients.

[0004] Traditional lower limb orthoses have relatively single functions and only provide mechanical support. For example, the AFO ankle - foot orthosis cannot actively intervene in muscle spasm or dynamically adapt to the gait cycle; the existing myoelectric - controlled prostheses have relatively lagged control, rely on static threshold triggering, and cannot match the dynamic changes of the gait in real - time; the adaptation degree of standardized orthoses is relatively rough. Through thermoplastic molding, the error from the patient's anatomical structure is >3mm, which is prone to cause pressure ulcers.

[0005] Therefore, in view of the above problems, the present invention proposes an intelligent lower limb orthosis and a collaborative control system based on bimodal biofeedback. Summary of the Invention

[0006] In order to overcome the problems of single - function, control lag, and rough adaptation of traditional orthoses, the present invention proposes an intelligent lower limb orthosis and a collaborative control system based on bimodal biofeedback.

[0007] The technical solution of the present invention is as follows: The intelligent lower limb orthosis based on bimodal biofeedback includes:

[0008] 3D printed orthosis body, which is modeled based on the patient's lower limb MRI / CT data, adopts topological optimization structure design, and its material adopts gradient density filling technology, with high-rigidity PLA as the outer layer and flexible TPU as the inner layer. The pressure distribution is optimized by finite element analysis, so that the pressure distribution uniformity CV ≤ 8%;

[0009] Embedded sensing module, including dual-channel surface electromyography sensor, six-axis inertial measurement unit and flexible piezoresistive array, for real-time acquisition of electromyographic signals, ankle joint motion data and plantar pressure distribution;

[0010] The electrical stimulation execution unit adopts a multi-channel transcutaneous electrode and is used to trigger the electrical stimulation signal of the multi-channel transcutaneous electrode according to the gait phase.

[0011] Preferably, the dual-channel surface electromyography sensor is attached to the tibialis anterior muscle and the lateral head of the gastrocnemius muscle, with a bandwidth of 20-450Hz. The sensor adopts a bipolar differential measurement method and combines a reference ground electrode to suppress environmental noise. During the gait cycle, the EMG signal of TA is used to detect foot drop, while the EMG signal of GL is used to identify gastrocnemius spasm.

[0012] Preferably, the sampling rate of the six-axis inertial measurement unit is 200 Hz, which measures the angular velocity and acceleration of the ankle joint in real time, fuses multi-axis data through the Madgwick filtering algorithm, and calculates the three-dimensional posture angle of the ankle joint with an accuracy of ±1°.

[0013] Preferably, the resolution of the flexible piezoresistive array is 1.4 sens / cm 2 , embedded in the sole of the foot to map the trajectory of the pressure center in real time, the sensor array adopts the piezoresistive principle, with a dynamic range of 0-1000kPa and a linear error of <0.01%F·S to detect tiny changes in pressure distribution.

[0014] Preferably, the frequency of the electrical stimulation execution unit is 1-100 Hz and the pulse width is 50-400 μs. The unit supports independent control of multi-channel transcutaneous electrodes. Based on the gait phase, the system dynamically adjusts the stimulation mode:

[0015] Stance phase: high-frequency stimulation of 50 Hz and 300 μs pulse width inhibits gastrocnemius muscle spasm;

[0016] Swing phase: 30Hz, low-frequency stimulation with a pulse width of 200μs enhances the contraction of the tibialis anterior muscle and corrects foot drop;

[0017] The stimulation intensity is linearly mapped to the EMG signal amplitude, so that the parameters can be adjusted in real time through closed-loop feedback.

[0018] Preferably, the inner contact surface of the 3D printed orthosis adopts a porous TPU grid structure with a pore diameter of 1 mm and a porosity of 70%.

[0019] Preferably, the device further includes a low-power power management module, which adopts an energy recovery circuit and a sleep mode, with a standby current < 10 μA, a full battery life of up to 52 hours, and an IP67 waterproof rating. The sleep strategy is as follows:

[0020] No motion signal for 10 minutes: Enter the low-power mode;

[0021] The IMU detects slight movement: Wake up instantly.

[0022] Preferably, the cooperative control system based on bimodal biofeedback includes:

[0023] A signal preprocessing module, which is responsible for real-time noise reduction and feature extraction of the original sEMG signal. First, wavelet transform is used to eliminate motion artifacts and high-frequency noise, and then the original electromyogram signal is converted into an amplitude envelope line through the RMS envelope extraction algorithm, retaining the low-frequency components of 0 - 20 Hz to reflect the muscle activation intensity. At the same time, the module integrates a 50 Hz notch filter to eliminate power frequency interference;

[0024] A phase recognition module, which detects the gait cycle based on IMU data and divides it into a stance phase and a swing phase;

[0025] A dual-threshold trigger mechanism that activates electrical stimulation when the EMG amplitude > 100 μV and the COP displacement > 5 cm.

[0026] Preferably, the system further includes an adaptive current regulation module, which dynamically adjusts the electrical stimulation output current by real-time analyzing the sEMG signal intensity, and adopts a linear mapping algorithm to make the stimulation intensity proportional to the degree of abnormal muscle activation.

[0027] Preferably, the phase recognition module detects the gait cycle based on the real-time data of a six-axis IMU. This module uses the Perry gait cycle model to divide the gait into 8 sub-phases, accurately identifies key events through the angular velocity peak detection algorithm, and calculates the three-dimensional attitude angle of the ankle joint in combination with the Madgwick quaternion filter.

[0028] Advantages of the present invention:

[0029] 1. By integrating bimodal biofeedback (sEMG + IMU) and 3D printed orthosis, precise dynamic correction of gait abnormalities in spastic cerebral palsy patients is achieved, the accuracy of gait phase recognition is improved, the stimulation delay is reduced, and thus the rehabilitation efficiency is significantly enhanced.

[0030] 2. Divide the gait cycle into the stance phase and the swing phase, and design differentiated electrical stimulation strategies for different phases.

[0031] 3. Adopt gradient material 3D printing and porous grid structure to reduce the peak value of skin contact pressure, and there is no redness or swelling after continuous wearing for 8 hours, taking into account both mechanical support and comfort.

[0032] 4. Adaptive electrical stimulation adjustment and low-power energy recovery enable patients to perform home rehabilitation away from the hospital environment, solving the pain point of poor compliance in traditional treatments. Brief Description of the Drawings

[0033] Figure 1 The figure shows a schematic diagram of the working process of the present invention;

[0034] Figure 2 The figure shows a schematic diagram of the technical route of the present invention. Detailed Description of the Invention

[0035] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] Please refer to Figure 1 - Figure 2 , the present invention provides an embodiment: an intelligent lower limb orthosis based on dual-modal biofeedback, including:

[0037] A 3D printed orthosis body, which is personalized modeled based on the patient's lower limb MRI / CT data, designed with a topology optimization structure, and its material adopts a gradient density filling technology. The outer layer is made of high-rigidity PLA to provide stable support, and the inner layer is made of flexible TPU to enhance wearing comfort, and the pressure distribution is optimized through finite element analysis to make the uniformity of the pressure distribution CV ≤ 8%;

[0038] An embedded sensing module, including a dual-channel surface electromyography sensor, a six-axis inertial measurement unit, and a flexible piezoresistive array, for real-time acquisition of electromyography signals, ankle joint movement data, and plantar pressure distribution;

[0039] An electrical stimulation execution unit, which uses multi-channel percutaneous electrodes to trigger electrical stimulation signals of the multi-channel percutaneous electrodes according to the gait phase.

[0040] Preferably, the dual-channel surface electromyography sensor is attached to the tibialis anterior muscle and the lateral head of the gastrocnemius muscle, has a bandwidth of 20-450Hz, and can accurately capture muscle electrical signals. The sensor adopts a bipolar differential measurement method and combines a reference ground electrode to suppress environmental noise. During the gait cycle, the EMG signal of TA is used to detect foot drop, and the EMG signal of GL is used to identify gastrocnemius spasm.

[0041] Preferably, the sampling rate of the six-axis inertial measurement unit is 200 Hz, which measures the angular velocity and acceleration of the ankle joint in real time, fuses multi-axis data through the Madgwick filtering algorithm, and calculates the three-dimensional posture angle of the ankle joint with an accuracy of ±1°.

[0042] Preferably, the resolution of the flexible piezoresistive array is 1.4 sens / cm 2 , embedded in the sole of the foot to map the trajectory of the pressure center in real time, the sensor array adopts the piezoresistive principle, with a dynamic range of 0-1000kPa and a linear error of <0.01%F·S to detect tiny changes in pressure distribution.

[0043] Preferably, the frequency of the electrical stimulation execution unit is 1-100 Hz and the pulse width is 50-400 μs. The unit supports independent control of multi-channel transcutaneous electrodes. Based on the gait phase, the system dynamically adjusts the stimulation mode:

[0044] Stance phase: high-frequency stimulation (50 Hz, pulse width 300 μs) inhibits gastrocnemius spasm;

[0045] Swing phase: low-frequency stimulation (30 Hz, pulse width 200 μs) enhances the contraction of the tibialis anterior muscle and corrects foot drop;

[0046] The stimulation intensity is linearly mapped to the EMG signal amplitude, so that the parameters can be adjusted in real time through closed-loop feedback.

[0047] Preferably, the inner contact surface of the 3D printed orthosis body adopts a porous TPU grid structure with a pore size of 1 mm and a porosity of 70%.

[0048] Preferably, the device also includes a low-power power management module, using an energy recovery circuit and a sleep mode, with a standby current of <10μA, a full-power endurance of 52 hours, and a waterproof rating of IP67, wherein the sleep strategy is:

[0049] No motion signal for 10 minutes: enter low power mode;

[0050] IMU detects slight movement: instant wake-up.

[0051] Further, the usage of the intelligent lower limb orthosis is described in detail:

[0052] Before use, the patient needs to undergo lower limb MRI / CT scans at a hospital or rehabilitation center to obtain accurate three-dimensional anatomical data of the limb. Technicians use professional software to reconstruct a digital model and generate an orthosis structure that is lightweight and optimized in mechanical properties through a topology optimization algorithm. 3D printing uses a dual-nozzle FDM technology. The outer layer uses rigid PLA material to provide support, and the inner layer uses flexible TPU (porous grid structure, pore diameter 1 mm) to ensure a comfortable fit. After receiving the orthosis, the patient should first try it on, check whether the pressure distribution is uniform (verified by a plantar pressure-sensitive film), and have the rehabilitation physician adjust the tightness of the straps to ensure there are no local pressure points. A small amount of medical silicone lubricant can be applied to the inner side of the orthosis to reduce friction. Before wearing it every day, the skin and the contact surface of the orthosis need to be cleaned.

[0053] Before formal use, sensor calibration is required:

[0054] The sEMG electrodes are attached to the tibialis anterior muscle and the gastrocnemius muscle. Conductive gel needs to be applied to ensure that the signal impedance < 5 kΩ, and the real-time electromyogram is observed through the mobile phone APP, and the electrode position is adjusted until the signal amplitude is stable.

[0055] The IMU module is located at the ankle joint and needs to be statically calibrated. At this time, keep the lower limb stationary for 10 seconds to eliminate the zero offset, and then walk slowly at a speed of 1 m / s to dynamically calibrate the gait parameters.

[0056] The plantar pressure array needs to be tested by standard stepping on a flat ground to ensure that the detection error of the COP (center of pressure) trajectory < 0.5 cm. After calibration is completed, long-press the power button of the orthosis for 3 seconds to start the system. The Bluetooth is automatically connected to the mobile phone APP, and the real-time gait data is displayed on the mobile phone APP. The gait data includes step frequency, step length, and ankle joint angle.

[0057] The system provides three preset training modes, which can be switched through the APP:

[0058] Spasm inhibition mode: Suitable for equinovarus caused by gastrocnemius spasm. When the IMU detects that the dorsiflexion angular velocity of the ankle joint > 50° / s (characteristic of the stance phase), a 50 Hz high-frequency electrical stimulation is automatically triggered to inhibit abnormal muscle contraction. The patient needs to perform 10 minutes of walking training every day under the guidance of a physician and gradually increase the slope to strengthen the effect.

[0059] Foot drop correction mode: For weakness of the tibialis anterior muscle. During the swing phase, the system identifies the heel lift event through the IMU and applies a 30 Hz electrical stimulation with a 50 ms delay to increase the height of the toe off the ground. Marking lines can be laid on the ground during training to gradually shorten the double support time of the gait cycle.

[0060] Adaptive learning mode: The system records the patient's historical gait data based on the LSTM network and automatically optimizes the stimulation parameters, such as frequency and pulse width. The patient needs to complete a standardized 6-minute walking test 3 times a week, and the system generates a rehabilitation progress report based on this.

[0061] After each training, the APP automatically generates a multi-dimensional evaluation report, which includes step length symmetry, energy consumption (estimated VO2), and the number of spasm events, etc. The patient can adjust the stimulation intensity through the APP.

[0062] A collaborative control system based on bimodal biofeedback includes:

[0063] A signal preprocessing module, which is responsible for real-time noise reduction and feature extraction of the original sEMG signal. First, wavelet transform is used to eliminate motion artifacts and high-frequency noise, and then the original electromyogram signal is converted into an amplitude envelope line through the RMS envelope extraction algorithm. The low-frequency components of 0 - 20Hz are retained to reflect the muscle activation intensity. At the same time, the module integrates a 50Hz notch filter to eliminate power frequency interference;

[0064] A phase recognition module, which detects the gait cycle based on IMU data and divides it into the stance phase and the swing phase;

[0065] A dual-threshold triggering mechanism, which activates electrical stimulation when the EMG amplitude > 100μV and the COP displacement > 5cm.

[0066] Preferably, the system further includes an adaptive current regulation module, which dynamically adjusts the electrical stimulation output current by real-time analyzing the sEMG signal intensity. The linear mapping algorithm is used to make the stimulation intensity proportional to the degree of abnormal muscle activation.

[0067] Preferably, the phase recognition module detects the gait cycle based on the real-time data of a six-axis IMU. This module uses the Perry gait cycle model to divide the gait into 8 sub-phases, accurately identifies key events through the angular velocity peak detection algorithm, and combines the Madgwick quaternion filter to calculate the three-dimensional attitude angle of the ankle joint.

[0068] Example two: Spasm inhibition mode

[0069] Applicable scenario: The patient's gastrocnemius spasm causes varus foot.

[0070] Control strategy:

[0071] Stance phase detection: When the IMU detects that the dorsiflexion angular velocity of the ankle joint > 50° / s, trigger the GL electrical stimulation (frequency 50Hz, pulse width 300μs).

[0072] Spasm inhibition: Reduce the sensitivity of muscle spindles through alternating frequency stimulation (5Hz carrier + 50Hz modulation).

[0073] Example 3: Foot Drop Correction Mode

[0074] Applicable Scenario: Weak anterior tibial muscle causes foot drop during the swing phase.

[0075] Control Strategy:

[0076] Swing Phase Trigger: After the IMU detects the heel-off event, a 50-ms delay is applied to the TA for electrical stimulation (frequency 30 Hz, pulse width 200 μs).

[0077] Dynamic Adjustment: Adjust the stimulation intensity according to the toe-off height (IMU Z-axis acceleration) to ensure that the foot clearance distance > 2 cm.

[0078] Example 4: Personalized Adaptation Verification

[0079] Test Data: After 3 cases of cerebral palsy patients (GMFCS level II) wore this orthosis:

[0080] Gait Symmetry: The step length difference decreased from (15.2 ± 3.1) cm to (4.3 ± 1.5) cm (P < 0.01).

[0081] Energy Consumption: The oxygen consumption (VO2) during walking decreased by 18%.

[0082] Furthermore, the technical effects of the present invention are described in detail:

[0083] Explanation of the Precise Control of the System:

[0084] Gait Phase Recognition Accuracy: 98.7% (vs. 85.2% for the traditional threshold method).

[0085] Electrical Stimulation Delay: Support phase (75 ± 12 ms), swing phase (82 ± 15 ms).

[0086] Autonomous Surface Electromyogram Control: By attaching a wearable sensor film to the muscle surface, the characteristic electromyogram signals generated during specific muscle activities are used to trigger the regulation of the downstream electronic component switches.

[0087] Signal Transmission Time: The signal transmission time of the wired sensor is less than 50 ms, and the signal transmission time of the wireless sensor is less than 200 ms;

[0088] Electromyogram Signal Amplitude: Between 0.01 mV and 0.1 mV;

[0089] Electromyogram Signal Amplification Factor: Not less than about 10,000 times;

[0090] Gain of the Electromyogram Signal Amplifier: Not less than 100 dB;

[0091] Peak value of electromyographic signal: not less than 5000uV;

[0092] EMG signal collection interference frequency: no higher than 30Hz;

[0093] EMG signal acquisition parameters: 16 channels (compatible with patch electrodes), sampling rate not less than 2000Hz, bandpass filter range is 20Hz ~ 500Hz, notch filter filters out 50Hz power frequency noise. Bipolar differential measurement method and reference ground electrode are used to reduce noise interference as much as possible, improve common mode suppression capability, and can collect EMG data with medical grade accuracy in real time.

[0094] Explanation for the improvement of comfort:

[0095] Peak skin contact pressure: conventional orthosis (125 kPa) → this design (68 kPa).

[0096] The incidence of redness and swelling after continuous wearing for 8 hours: 0% (vs. 37% for traditional products).

[0097] Pressure and shear force signal data are collected and identified in real time:

[0098] Recognition accuracy: not less than 0.05% F·S;

[0099] Identification error: no more than 0.01% F·S;

[0100] Others: Control the on or off of the functional electrical stimulation function of the orthosis through autonomous surface electromyography.

[0101] Functional electrical stimulation matches the gait cycle: A gait cycle during walking involves the movement of multiple or multiple groups of lower limb muscle groups. Through a wearable sensor film attached to the muscle surface, the characteristic electromyographic signals generated during specific muscle activity are captured and analyzed to identify the wearer's gait stage. The frequency and amplitude of the gait cycle can be accurately identified through the characteristic electromyographic signals emitted by two groups of muscles in different stages of the gait cycle. The characteristic electromyographic signals are connected to the functional electrical stimulation therapy device, which performs electrical stimulation therapy at a specific stage of the gait cycle, thereby specifically improving specific muscle groups with abnormal muscle strength.

[0102] Response time: The duration of functional electrical stimulation in response to gait electromyographic signals is less than 100ms.

[0103] Detectable gait parameter range:

[0104] Step frequency: Normal speed should not be less than 110 steps / min, slow speed should not be less than 70 steps / min;

[0105] Stride length: The vertical distance between the two heels is 0 to 15 cm.

[0106] Therapeutic effect of gait abnormality improvement:

[0107] Gait analysis: According to the gait analysis results of the patient, the items with significant statistical differences from normal children (step length, step width, walking speed, step frequency, stride length, gait cycle, double support time), after wearing the lower limb orthosis, the values of these items in this patient showed significant statistical differences before and after treatment (P<0.05).

[0108] Through the above steps, by integrating dual-modal biofeedback and 3D printed orthosis, precise dynamic correction of gait abnormalities in spastic cerebral palsy patients is achieved, thus significantly improving the rehabilitation efficiency. At the same time, the gait cycle is divided into the stance phase and the swing phase, and differential electrical stimulation strategies are designed for different phases. The 3D printing with gradient materials and the porous grid structure are used to reduce the peak value of skin contact pressure, so as to solve the problems of single function, control lag and rough adaptability of traditional orthoses.

Claims

1. An intelligent lower limb orthosis based on bimodal biofeedback, characterized in that, It includes: A 3D printed orthosis body, which is modeled based on the patient's lower limb MRI / CT data, designed with a topology optimization structure, using a gradient density filling technology for its material, with a high-rigidity PLA on the outer layer and a flexible TPU on the inner layer, and optimizing the pressure distribution through finite element analysis to make the pressure distribution uniformity CV ≤ 8%; An embedded sensing module, including a dual-channel surface electromyography sensor, a six-axis inertial measurement unit, and a flexible piezoresistive array, for real-time acquisition of electromyography signals, ankle joint movement data, and plantar pressure distribution; An electrical stimulation execution unit, which uses multi-channel percutaneous electrodes to trigger electrical stimulation signals of the multi-channel percutaneous electrodes according to the gait phase.

2. The intelligent lower limb orthosis based on dual-modal biofeedback according to claim 1, wherein: The dual-channel surface electromyography sensor is attached to the tibialis anterior muscle and the lateral head of the gastrocnemius muscle, with a bandwidth of 20 - 450 Hz. The sensor uses a bipolar differential measurement method and combines a reference ground electrode to suppress environmental noise. In the gait cycle, the EMG signal of the TA is used to detect foot drop, while the EMG signal of the GL is used to identify gastrocnemius muscle spasm.

3. The intelligent lower limb orthosis based on dual-modal biofeedback according to claim 1, characterized in that: The sampling rate of the six-axis inertial measurement unit is 200 Hz, which real-time measures the angular velocity and acceleration of the ankle joint, fuses multi-axis data through the Madgwick filtering algorithm, and calculates the three-dimensional attitude angle of the ankle joint with an accuracy of ±1°.

4. The intelligent lower limb orthosis based on dual-modal biofeedback according to claim 1, wherein: The resolution of the flexible piezoresistive array is 1.4 sens / cm 2 , which is embedded in the sole of the foot to map the pressure center trajectory in real time. The sensor array uses the piezoresistive principle, with a dynamic range of 0 - 1000 kPa and a linear error of < 0.01% F·S, to detect tiny changes in pressure distribution.

5. The intelligent lower limb orthosis based on dual-modal biofeedback according to claim 1, characterized in that: The frequency of the electrical stimulation execution unit is 1 - 100 Hz, and the pulse width is 50 - 400 μs. This unit supports independent control of multi-channel percutaneous electrodes. Based on the gait phase, the system dynamically adjusts the stimulation mode: Support phase: High-frequency stimulation of 50 Hz and a pulse width of 300 μs suppresses gastrocnemius muscle spasm; Swing phase: Low-frequency stimulation of 30 Hz and a pulse width of 200 μs enhances the contraction of the tibialis anterior muscle and corrects foot drop; The stimulation intensity is linearly mapped to the EMG signal amplitude, so as to adjust the parameters in real time through closed-loop feedback.

6. The intelligent lower limb orthosis based on dual-modal biofeedback according to claim 1, characterized in that: The inner contact surface of the 3D printed orthosis body adopts a porous TPU grid structure with a pore diameter of 1 mm and a porosity of 70%.

7. The intelligent lower limb orthosis based on bimodal biofeedback according to claim 1, characterized in that: The device also includes a low-power power management module, which uses an energy recovery circuit and a sleep mode, with a standby current < 10 μA and a full-charge battery life of up to 52 hours, and an IP67 waterproof rating. The sleep strategy is as follows: No movement signal for 10 minutes: Enter the low-power mode; The IMU detects slight movement: Wake up instantaneously.

8. A collaborative control system based on bimodal biofeedback, which uses the intelligent lower limb orthosis based on bimodal biofeedback described in claims 1-5, is characterized in that, It includes: A signal preprocessing module, which is responsible for real-time noise reduction and feature extraction of the original sEMG signal. First, wavelet transform is used to eliminate motion artifacts and high-frequency noise, and then the original electromyography signal is converted into an amplitude envelope through the RMS envelope extraction algorithm, retaining the low-frequency components of 0 - 20 Hz to reflect the muscle activation intensity. At the same time, the module integrates a 50 Hz notch filter to eliminate power frequency interference; A phase recognition module, which detects the gait cycle based on IMU data and divides it into a support phase and a swing phase; A dual-threshold triggering mechanism, which activates electrical stimulation when the EMG amplitude > 100 μV and the COP displacement > 5 cm.

9. The collaborative control system based on bimodal biofeedback according to claim 8, characterized in that: The system also includes an adaptive current regulation module, which dynamically adjusts the electrical stimulation output current by real-time analyzing the sEMG signal intensity, and uses a linear mapping algorithm to make the stimulation intensity proportional to the degree of abnormal muscle activation.

10. The collaborative control system based on bimodal biofeedback according to claim 8, wherein: The phase recognition module detects the gait cycle based on the real-time data of the six-axis IMU. This module uses the Perry gait cycle model to divide the gait into 8 sub-phases, accurately identifies key events through the angular velocity peak detection algorithm, and calculates the three-dimensional attitude angle of the ankle joint by combining the Madgwick quaternion filter.

Images