Control methods for prosthetic knee joints

Abstract

This invention relates to a control method for a prosthetic knee joint, the key technical points of which include: real-time acquisition of heel pressure sensor signals. P h Foot pressure sensor signal P f and the tilt angle output by the lower leg inertial measurement unit i and angular velocity ω; when detected P f When the descent rate exceeds a preset multiple threshold of the heel pressure descent rate, the tilt angle reference is dynamically calibrated. i base And calculate the offset. i cal ;based on i cal The gait cycle is divided into six stages based on the pressure signal, and the cycle duration is calculated. T This invention constructs a gait-damping mapping knowledge base and optimizes the damping opening in the early stage of the swing through iterative learning. Finally, the adjusted U-value is applied in the early stage of the swing, while the damping remains fixed for the remaining stages. This invention primarily aims to improve the naturalness of walking and terrain adaptability of prostheses at different gait speeds.

Description

Technical Field

[0001] This invention relates to the field of lower limb prosthetic motion control technology. More specifically, this invention relates to a control method for the knee joint of a prosthetic limb. Background Technology

[0002] Currently, some prosthetic knee joint control devices employ a fixed damping parameter strategy. This strategy has limitations when the wearer's walking speed changes: when walking speed increases significantly, the flexion angle during the swing phase of the prosthetic knee joint may be insufficient, leading to shortened stride and stiff gait; when walking speed decreases significantly, excessive flexion may occur during the swing phase, affecting gait stability. These problems stem from the fact that fixed damping parameters cannot adapt to the changing biomechanical requirements of knee joint movement at different walking speeds.

[0003] The main reason for this limitation is that, in human physiological gait, the damping force required by the knee joint during the swing phase has a non-linear relationship with gait speed, and traditional methods struggle to accurately and in real-time sense changes in gait speed and dynamically adjust damping parameters accordingly. Furthermore, due to individual gait differences and the variability of walking environments, establishing a universally applicable gait speed-damping mapping relationship presents challenges. Existing methods have not yet effectively solved these problems in achieving adaptive matching of damping parameters and gait speed. Summary of the Invention

[0004] One object of the present invention is to solve at least the above-mentioned problems and to provide at least the advantages that will be described later.

[0005] Another objective of this invention is to provide a control method for the knee joint of a prosthesis, which solves the problems of unnatural gait and insufficient safety caused by the inability of traditional prostheses to adapt to changes in walking speed.

[0006] To achieve these objectives and other advantages according to the present invention, a method for controlling a prosthetic knee joint is provided, comprising:

[0007] S1. Real-time acquisition of heel pressure sensor signals P h Foot pressure sensor signal P f and the tilt angle output by the lower leg inertial measurement unit i and angular velocity ω;

[0008] S2, when detected P f rate of decline v f > k ⋅ v h At that moment, the real-time tilt angle will be... i now Defined as the new 0° reference, denoted as i base = i now ; v h for P h rate of descent k =2 is a preset multiple threshold; calculate the calibrated tilt angle offset. i cal , i cal = i - i base ;

[0009] S3, based on i cal With dual pressure signals, the gait cycle is divided into early support phase, middle support phase, late support phase, early swing phase, middle swing phase, and late swing phase.

[0010] The gait cycle duration is calculated based on the interval between the early starting points of adjacent supports. T ;

[0011] S4. Establish and utilize the gait-damping mapping knowledge base to construct the gait cycle duration. T With the damping opening in the early stage of oscillation U Mapping relationship library; detect current T The value is then queried from the velocity-damping mapping knowledge base; if a match exists... T Directly call the corresponding U Value; if unknown T Then load with T The closest known pace correspondence U As initial value;

[0012] Perform iterative learning on the unknown T:

[0013] U k+1 =U k +L(θ target - i kmax )

[0014] in U For damping opening, L For learning rate, i target The target buckling angle is 65°. i kmax For the first k Maximum buckling angle measured during the cycle;

[0015] When the buckling angle error converges to | i target - i kmax Stop iteration and store the mapping relationship when |≤5°;

[0016] S5, Apply adjusted damping opening during the initial swing phase. U The damping remains constant throughout the early, middle, and late stages of support and the later stages of oscillation.

[0017] Preferably, the control method for the prosthetic knee joint is based on i cal Based on the dual pressure signals, the gait cycle is divided into early support, middle support, late support, early oscillation, middle oscillation, and late oscillation phases, specifically:

[0018] when P h ≥ P hmax and i cal When the angle is less than -3° (heel touches the ground → full foot bears pressure to start), it is considered to be the early support phase;

[0019] when P h >50%× P hmax , P f >50%× P fmax and i cal When the rate of change is <1° / s (the stage of stable bearing pressure across the entire foot), it is determined to be the middle stage of support; P hmax For the current gait cycle P h Peak value P fmax For the current gait cycle P f Peak value;

[0020] when P h <30%× P hmax and i cal When the angle is >0° (from heel off the ground to before toes off the ground), it is considered to be the late support phase.

[0021] when P f =0 and i cal When the angle increases by ≥10° within 100ms (toes leave the ground → lower leg accelerates forward swing), it is judged as the early swing phase;

[0022] when i cal Continue to increase to the peak value and angular velocity | oh |<1° / s When the lower leg swings forward to its maximum flexion angle, it is considered to be in the middle of the swing.

[0023] when i cal Peak value detected from the current oscillation cycle i max descent value ≥ 3° and angular velocity oh< At 0 o'clock (when the lower leg decelerates and extends until the next touch of the ground), it is determined to be the late stage of the swing.

[0024] Preferably, the control method for the prosthetic knee joint performs the following optimization steps during the iterative learning process in step S4:

[0025] S41. Using 65° as the initial target buckling angle, calculate the personalized target buckling angle reference value in real time: Reference target buckling angle = adaptation coefficient × (step length / lower leg length); where, the step length is calculated through the geometric relationship between the spatial distance of the initial starting point of adjacent supports and the calibration tilt angle; the lower leg length is a preset value; the adaptation coefficient ranges from 0.4 to 0.6;

[0026] S42. When |Reference Target Buckling Angle -65°| > 10°, replace the target buckling angle in the iterative learning with the reference target buckling angle;

[0027] S43. When a slope angle β greater than 3° is detected for more than 2 seconds, in an uphill scenario, the corrected target buckling angle = current target buckling angle × (1 + 0.2 × β); in a downhill scenario, the corrected target buckling angle = the maximum value (50°, current target buckling angle × (1 - 0.1 × |β|)).

[0028] Preferably, in the method for controlling the prosthetic knee joint, step S4 further includes adjusting the learning rate based on the rate of change of gait cycle duration:

[0029] When the standard deviation of the gait cycle duration for three consecutive cycles is less than 0.05 s, the standard learning rate L is used.

[0030] When the standard deviation is ≥0.05 seconds, the adaptive learning rate L is enabled. adaptive :L adaptive =L×[1+0.5×tanh(5×| T k - T k-1 | / T k-1 )];middle T kThe duration of the current cycle. T k-1 This is the duration of the previous cycle.

[0031] Preferably, in the method for controlling the prosthetic knee joint, step S4 further includes performing the following steps when the number of records in the gait-damping mapping knowledge base is less than 5:

[0032] For unknown gait period duration T Initial damping opening value plus safety margin: Initial damping opening = U closest ×min(1.3, 1+0.1×| T - T closest | / T closest ); U closest For the closest damping opening, T closest The closest gait cycle duration;

[0033] Set the maximum number of iterations to 10;

[0034] When the number of iterations k>5 and |65°- i kmax | When the angle is greater than 15°, reset the damping opening U. k =0.5×( U min + U max );in U min and U max These are the preset safe working range boundary values.

[0035] Preferably, in the method for controlling the prosthetic knee joint, step S4 further includes real-time monitoring of pressure signal abrupt change characteristics: if the heel pressure... P h Decrease of >80% within 20ms P hmax When the angular velocity is greater than 50° / second, the current iteration of learning is interrupted and the system switches to safety damping mode. U emergency =0.7× U min +0.3× U max , U emergency For emergency damping; after the terrain stabilizes, the damping setting should be restored to its pre-interruption state. U k Continuing the iteration, the determination of terrain stability requires that the following conditions be met simultaneously: a) Ph ≥40%× P hmax up to 500ms; b) i cal Rate of change < 2° / s and angular velocity oh c) The initial support point was detected after three consecutive sampling values ​​<10° / s.

[0036] Preferably, the control method for the prosthetic knee joint incorporates gradual damping control during the transition from the late support phase to the early swing phase.

[0037] When the calibration tilt angle is detected in the later stage of the support i cal When the angle is greater than 20°, the damping opening transitions linearly: U transition = U support +( U - U support )×( i cal -20°) / ( i swingstart -20°), where U transition For the transition period damping opening, U support The damping value is the fixed opening during the support period, and U is the damping opening during the early stage of the swing, which is the output of the current iteration. i swingstart This is the starting angle of the current cycle's early oscillation; the transition process continues until the end of the early oscillation phase, after which the complete cycle is applied. U value.

[0038] Preferably, the control method for the prosthetic knee joint includes sensor fault redundancy handling in step S1:

[0039] Real-time calculation of the effectiveness index of heel pressure signal; if it is continuous for 5 cycles P hmax <10N or signal variance <0.1N 2 If the heel pressure sensor is determined to be faulty, perform the following steps:

[0040] The vertical acceleration component a is output using an inertial measurement unit mounted proximal to the thigh prosthesis segment. z Alternative P h The ground contact threshold is set to |a z |>2g;

[0041] Simultaneously activate the heel impact sound sensor installed on the inside of the prosthetic ankle. When a sound pressure level of 200-500Hz > 80dB is detected and is consistent with az When the peak time difference is less than 50ms, the grounding event is confirmed; the backup signal grounding event triggers the reference angle reset in step S2 and the stage division in step S3.

[0042] The present invention has at least the following beneficial effects:

[0043] 1. The control method of this invention effectively improves the adaptability of prosthetic walking by sensing the wearer's gait speed changes in real time and dynamically adjusting the knee joint damping parameters. The core of this invention lies in establishing an adaptive mapping mechanism between gait cycle duration and swing phase damping opening, and introducing a closed-loop iterative learning strategy. When the wearer's gait speed changes, the system can autonomously optimize the damping opening, so that the knee joint approaches the physiological flexion trajectory in the key swing phase, significantly reducing the risk of gait disorder caused by sudden changes in gait speed. The control method of this invention overcomes the inherent limitations of fixed damping parameters in variable speed scenarios, enabling the prosthesis to maintain natural swing characteristics in different walking rhythms.

[0044] 2. For complex usage environments, this invention integrates multi-dimensional safety strategies to improve system robustness. It achieves adaptive adjustment of the learning rate through gait cycle standard deviation monitoring to avoid control instability when gait speed fluctuates drastically. It adds damping safety margin and iteration interruption protection for unknown gait speed scenarios to prevent safety hazards caused by excessive flexion. Especially in sloping terrain, it dynamically corrects the target flexion angle based on the inclination angle to enhance the knee joint support stability when walking on slopes. The control method of this invention collaboratively ensures a smooth transition from flat roads to complex terrain, reducing the wearer's need for active intervention.

[0045] 3. This invention further enhances reliability through sensor redundancy design and fine control during the motion phase. When the heel pressure sensor fails, the inertial measurement unit acceleration and heel impact acoustic signal are used to cross-verify the ground contact event, maintaining the accuracy of gait phase division. Damping gradual control is applied during the transition from the support phase to the swing phase to eliminate the mechanical impact when the joint motion state changes. The control method of this invention reduces the dependence on a single sensor, enabling the prosthesis to maintain continuous and stable damping output characteristics even in critical scenarios such as sensor malfunction or motion phase transition.

[0046] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Detailed Implementation

[0047] The present invention will be further described in detail below with reference to embodiments, so that those skilled in the art can implement it based on the description.

[0048] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0049] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0050] The present invention provides a method for controlling a prosthetic knee joint in order to achieve these objectives and other advantages according to the invention, comprising:

[0051] S1. Real-time acquisition of heel pressure sensor signals P h Foot pressure sensor signal P f and the tilt angle output by the lower leg inertial measurement unit i and angular velocity ω;

[0052] S2, when detected P f rate of decline v f > k ⋅ v h At that moment, the real-time tilt angle will be... i now Defined as the new 0° reference, denoted as i base = i now ; v h for P h rate of descent k =2 is a preset multiple threshold; calculate the calibrated tilt angle offset. i cal , i cal = i - i base ;

[0053] S3, based on i cal With dual pressure signals, the gait cycle is divided into early support phase, middle support phase, late support phase, early swing phase, middle swing phase, and late swing phase.

[0054] The gait cycle duration is calculated based on the interval between the early starting points of adjacent supports. T ;

[0055] S4. Establish and utilize the gait-damping mapping knowledge base to construct the gait cycle duration. TWith the damping opening in the early stage of oscillation U Mapping relationship library; detect current T The value is then queried from the velocity-damping mapping knowledge base; if a match exists... T Directly call the corresponding U Value; if unknown T Then load with T The closest known pace correspondence U As initial value;

[0056] Perform iterative learning on the unknown T:

[0057] U k+1 =U k +L(θ target - i kmax )

[0058] in U For damping opening, L For learning rate, i target The target buckling angle is 65°. i kmax For the first k Maximum buckling angle measured during the cycle;

[0059] When the buckling angle error converges to | i target - i kmax Stop iteration and store the mapping relationship when |≤5°;

[0060] S5, Apply adjusted damping opening during the initial swing phase. U The damping remains constant throughout the early, middle, and late stages of support and the later stages of oscillation.

[0061] In the above technical solution, the gait cycle duration refers to the time interval between two consecutive heel touch-the-ground events, reflecting the walking speed; the damping opening is the flow regulation parameter of the hydraulic valve or magnetorheological fluid device, and the damping decreases when the opening increases; the tilt angle offset is the angle between the lower leg axis and the vertical direction, and the installation error is eliminated after dynamic benchmark calibration; iterative learning is an adaptive process that iteratively optimizes the damping parameters based on historical gait data.

[0062] In the above technical solution, heel and ball of foot pressure sensors are installed in the lower leg segment of the prosthesis, and an inertial measurement unit is integrated synchronously. When the system detects that the rate of decrease in ball of foot pressure significantly exceeds the rate of decrease in heel pressure (determined by a preset multiple threshold), the current tilt angle is immediately set as the new reference angle, and the real-time offset is calculated accordingly. Based on this offset and the change characteristics of the dual-channel pressure signals, the gait is decomposed into six continuous stages, and the gait cycle duration is calculated by the interval between the starting points of adjacent support phases.

[0063] In the above technical solution, the system has a built-in gait-damping mapping knowledge base, storing the verified correspondence between gait cycle duration and damping opening in the early swing phase. For known gait speeds, the matching damping value is directly called; for unknown gait speeds, the damping opening closest to the historical value is loaded as the initial value, and iterative learning is initiated: based on the deviation between the measured maximum buckling angle and the target buckling angle during the current swing phase, the damping opening is gradually corrected according to the learning rate until the buckling angle deviation stably converges to a physiologically reasonable range over multiple consecutive cycles. Finally, the optimized damping value is applied only in the early swing phase, while the preset fixed damping is maintained in other phases.

[0064] Existing technologies employ preset multiple damping modes (such as slow / medium / fast modes), requiring manual switching or relying on coarse acceleration thresholds for triggering. Their limitations include: inability to differentiate individual gait differences, with different wearers potentially exhibiting over-flexion or under-flexion at the same damping level; lag in parameter switching during variable-speed walking, resulting in discontinuous gait; and a lack of continuous self-optimization capabilities, requiring repeated manual calibration.

[0065] This technical solution eliminates individual installation differences through dynamic reference angle calibration, accurately quantifies gait speed using gait cycle duration, and achieves autonomous progressive optimization of damping parameters through closed-loop iterative learning, fundamentally overcoming the mechanical response defects of fixed gear switching.

[0066] This technical solution significantly improves the prosthesis's adaptability to changes in the wearer's walking rhythm through closed-loop control of gait speed sensing and damping adjustment. In scenarios with fluctuating gait speed, the knee joint swing flexion angle automatically approaches the physiological range of motion, eliminating gait stiffness or instability caused by damping mismatch. The iterative learning mechanism ensures that the system completes parameter optimization within several gait cycles, reducing the need for manual intervention. The phased damping strategy balances swing flexibility and support stability, making the prosthesis's movement trajectory closer to a natural gait.

[0067] The control method for the prosthetic knee joint in the above technical solution includes:

[0068] Signal acquisition and dynamic calibration:

[0069] The prosthetic lower leg assembly integrates a piezoresistive thin-film pressure sensor and a nine-axis inertial measurement unit. Heel pressure signal. P h and foot pressure signalsP f Real-time data is acquired at a sampling rate of 100Hz, and mechanical vibration noise is eliminated by Kalman filtering. The inertial measurement unit calculates the tilt angle using a quaternion fusion algorithm. i and angular velocity oh The output frequency is 200Hz. During dynamic calibration:

[0070] calculate P f rate of descent v f =Δ P f / Δ t and P h rate of descent v h =Δ P h / Δ t (Time window Δ) t =100ms);

[0071] When satisfied v f >2 v h and v h When the speed is >5 N / s, lock the tilt angle at that moment. i now As a new benchmark i base ;

[0072] Output calibration offset i cal = i - i base This eliminates the tilt angle deviation during prosthesis installation.

[0073] Gait cycle segmentation and duration calculation:

[0074] based on i cal Constructing a six-stage state machine using dual pressure signals:

[0075] Supporting the initial starting point: When P h ≥0.9 P hmax and i cal A timestamp is marked when the temperature is less than -3°. t 1 ;

[0076] Support in the medium term: Dual-path pressure > 50% peak value and slope angle change rate |d ical / d t |<1° / s;

[0077] Early stage of swing: P f =0 and i cal Increase by ≥10° within 100ms.

[0078] Interval between the starting points of adjacent support lines T = t 2- t 1 represents the gait cycle duration, reflecting real-time walking speed.

[0079] Damped mapping and iterative learning

[0080] The pace-damping knowledge base uses a hash table to store key-value pairs. T i , U i Initially, three sets of reference mappings are preset (damping opening U refers to the flow regulation parameter of the hydraulic valve, U=0 indicates that the damping is completely closed (maximum damping force 150 N·m·s / rad), U=1 indicates that the damping is completely open (minimum damping force 20 N·m·s / rad), and Usupport=0.2 corresponds to a fixed damping force of 100 N·m·s / rad during the support period).

[0081] T =1.5s→U=0.8 (slow speed)

[0082] T =1.0s→U=0.5 (medium speed)

[0083] T =0.6s→U=0.3 (Fast)

[0084] unknown pace T :

[0085] Load closest T closest corresponding U closest As initial U 0;

[0086] Perform iterative learning:

[0087] U k+1 = U k +λ(65°- i kmax ) ( l =0.1)

[0088] in i kmax Indicates the first k Within each gait cycle swing phase, the tilt angle offset is calibrated. i cal Measured maximum value

[0089] When three consecutive cycles are satisfied, store the new mapping. T , U k ).

[0090] Phased damping implementation

[0091] The hydraulic proportional valve switches the damping opening according to the gait phase: Support phase (early / mid / late stage): Maintains a fixed opening U support =0.2 (high damping); early stage of oscillation: applied learned output value U∈[0.3,0.8]; mid-to-late stage of oscillation: fixed opening Uswing=0.7 (low damping). The phase switching command is issued within 5ms of the boundary event trigger, and the PID controller ensures that the opening tracks the target value within 20ms (error 0.02±0.02).

[0092] In the above technical solution, dynamic benchmark calibration addresses the issue caused by prosthesis tilt during wear. i Measurement error (e.g., 5° deviation between inward and outward rotation); cycle duration T More resistant to stride variation disturbances than acceleration amplitude, accurately quantifying stride speed; iterative learning enables damping parameters to converge within 3-5 steps under unknown stride speed; six-stage damping distribution balances support stability (high damping to prevent collapse) and oscillation efficiency (dynamic damping to adapt to stride speed).

[0093] In another technical solution, the control method for the prosthetic knee joint is based on i cal Based on the dual pressure signals, the gait cycle is divided into early support, middle support, late support, early oscillation, middle oscillation, and late oscillation phases, specifically:

[0094] when P h ≥ P hmax and i cal When the angle is less than -3° (heel touches the ground → full foot bears pressure to start), it is considered to be the early support phase;

[0095] when P h >50%× P hmax , P f >50%× P fmax and ical When the rate of change is <1° / s (the stage of stable bearing pressure across the entire foot), it is determined to be the middle stage of support; P hmax For the current gait cycle P h Peak value P fmax For the current gait cycle P f Peak value;

[0096] when P h <30%× P hmax and i cal When the angle is >0° (from heel off the ground to before toes off the ground), it is considered to be the late support phase.

[0097] when P f =0 and i cal When the angle increases by ≥10° within 100ms (toes leave the ground → lower leg accelerates forward swing), it is judged as the early swing phase;

[0098] when i cal Continue to increase to the peak value and angular velocity | oh |<1° / s When the lower leg swings forward to its maximum flexion angle, it is considered to be in the middle of the swing.

[0099] when i cal Peak value detected from the current oscillation cycle i max descent value ≥ 3° and angular velocity oh< At time 0 (when the lower leg decelerates and extends until the next contact with the ground), it is determined to be the late swing phase. In the above technical solution, the dynamic threshold of the pressure peak is a percentage criterion based on the maximum pressure value measured by the sensor within the current gait cycle, eliminating the influence of individual differences in stepping force; the angle change rate is the rate of change of the tilt angle offset per unit time, used to identify the steady state of motion; phase boundary events, such as heel contact and toe lift-off, are key actions that mark the transition of gait phases.

[0100] The above technical solution designs a multi-condition joint criterion by simultaneously analyzing three types of signals: heel pressure, ball of the foot pressure, and post-calibration tilt angle offset.

[0101] Support for early detection: It is necessary to simultaneously meet the condition that the heel pressure reaches the peak of the cycle and the lower leg is in a backward tilted position to ensure accurate ground contact event recognition;

[0102] Supporting mid-term assessment: Both pressure streams exceed half-peak values ​​and the rate of change of tilt angle approaches zero, filtering out body sway interference;

[0103] Supporting the later stage of judgment: the pressure on the heel drops significantly to the low threshold range and the lower leg turns forward, accurately capturing the moment when the heel leaves the ground;

[0104] Early swing detection: When the foot pressure reaches zero and the tilt angle increases rapidly, it marks the starting point of the lower leg's accelerated forward swing.

[0105] Mid-cycle oscillation determination: When the tilt angle increases to its peak and the angular velocity approaches zero, the maximum buckling position is locked.

[0106] Late-stage swing assessment: The tilt angle decreases from its peak and is accompanied by an extension angular velocity, predicting the posture for the next touchdown.

[0107] Existing technologies use single-signal threshold segmentation (e.g., if the pressure is greater than a fixed value, it is judged as the support period). Its limitations are: differences in the sensitivity of pressure sensors can easily lead to missed ground contact events (e.g., light ground contact scenarios); it does not combine the angle change rate, so it cannot distinguish between static bearing (mid-support period) and dynamic transition (late-support period); the division of the swing period depends on the absolute value of the angle, ignoring the differences in individual buckling range.

[0108] In this technical solution, the robustness of phase boundary recognition is significantly improved by pressure-angle dual signal coupling and dynamic threshold design, especially adapting to gait abnormalities (such as limping) and complex terrain (such as soft ground).

[0109] In this technical solution, the multi-dimensional signal fusion mechanism significantly improves the accuracy of gait phase segmentation, provides a reliable timing reference for damping parameter switching, supports mid-term steady-state criteria to avoid minor swaying that could trigger damping adjustments, and ensures stability during the load-bearing period; rapid identification in the early swing phase ensures timely reduction of damping and releases the knee joint's flexion degree of freedom; extension prediction in the later swing phase optimizes the ground contact preparation posture; and overall, it achieves a high degree of synchronization between the prosthesis's movement rhythm and the human gait, reducing energy loss caused by sudden changes in joint movement.

[0110] In the above technical solution, gait phase segmentation includes:

[0111] Signal synchronization and preprocessing:

[0112] The prosthetic control system simultaneously acquires three signals at a sampling rate of 200Hz: heel pressure. P h Foot pressure P f Calibrate tilt angle offset i cal The pressure signal is low-pass filtered at 10Hz to eliminate impact oscillations, and the tilt angle signal is smoothed using a fourth-order Butterworth filter (cutoff frequency 5Hz). Three types of derived parameters are calculated in real time: peak pressure.P hmax =max( P h )and P fmax =max( P f (Within the current period); Rate of change of tilt angle oh θ = dth cal / d t (Calculation using the central difference method); Pressure drop rate v h =- dP h / dt , v f =- dP f / dt (100ms window movement).

[0113] Six-stage partitioning logic implementation

[0114] Support for early judgment: When the following conditions are met simultaneously: P h ≥0.9 P hmax ; i cal < -3°; v h >5N / s; Detect heel contact event and mark the time stamp. t heel-strike This condition ensures that light-touch interference is eliminated.

[0115] Supporting the mid-term switching condition requires continuously satisfying two steady states: P h >0.5 P hmax ; P f >0.5P fmax ; ∣ω θ ∣ <1° / s (lasting ≥300ms); tilt angle change rate threshold filters out minor body swaying.

[0116] Early-stage boundary capture during swing: satisfies the rapid takeoff characteristic: P f =0 (lasting ≥50ms); Δθ cal ≥10° (t∈[t0, t0+100ms]) and ω>30° / s; where Δ i cal = ical ( t 0 +100ms) - i cal ( t 0 ).

[0117] Anti-interference mechanism design

[0118] Peak value dynamically updated: when P h >1.1 P hmax Reset P hmax = P h To prevent misjudgments caused by the failure of historical peak values. Even better, when... P h >1.1 P hmax And if the duration is greater than 20ms, reset. P hmax = P h This avoids triggering updates due to momentary interference.

[0119] State maintenance constraints: Stable maintenance conditions must be maintained for ≥300ms to enter the support phase; avoid phase oscillations caused by transient signal jumps.

[0120] Cross-cycle verification: The end of the oscillation phase needs to be detected. P h ≥0.85 P hmax Only then is it permissible to switch to the early support phase; the influence of abnormal gait such as single-leg hopping is excluded.

[0121] This technical solution improves the accuracy of heel strike detection: the peak pressure ratio threshold (90%) combined with negative tilt angle conditions results in a false negative rate of <0.5% (actual measured data); it improves the anti-interference performance during the support phase: dual pressure thresholds + tilt angle steady-state maintenance requirements filter out more than 85% of body swaying; and it ensures real-time toe-off performance: pressure returning to zero combined with a 10° tilt angle change within 100ms results in a detection delay of ≤15ms.

[0122] In another technical solution, the control method for the prosthetic knee joint performs the following optimization steps during the iterative learning process in step S4:

[0123] S41. Using 65° as the initial target flexion angle (65° is based on the average maximum flexion angle of the knee joint during the swing phase when a healthy adult walks on flat ground), calculate the personalized target flexion angle baseline value in real time: Baseline target flexion angle = Fit coefficient × (Step length / Lower leg length); where, step length is calculated through the geometric relationship between the spatial distance of adjacent support pre-start points and the calibration tilt angle; lower leg length is a preset value, and step length... S = L ×tan( i cal) × r ,in, L Calf length, i cal represents the calibration tilt angle offset, a coefficient. r The value is 1.2-1.5, determined based on gait statistics of healthy individuals; the fit coefficient ranges from 0.4 to 0.6.

[0124] S42. When |Reference Target Buckling Angle -65°| > 10°, replace the target buckling angle in the iterative learning with the reference target buckling angle;

[0125] S43, the slope angle β is calculated by fusing the pitch and roll angles of the thigh segment inertial measurement unit, using a complementary filtering algorithm (cutoff frequency 10Hz), and the formula is as follows: ,in a x , a y , a z The acceleration components of the thigh inertial unit are triaxial.

[0126] When a slope angle β greater than 3° is detected for more than 2 seconds, in an uphill scenario, the corrected target buckling angle = current target buckling angle × (1 + 0.2 × β); in a downhill scenario, the corrected target buckling angle = the maximum value (50°, current target buckling angle × (1 - 0.1 × |β|)).

[0127] In the above technical solution, the step length reference value is the horizontal projection distance between the contact points of adjacent heels, reflecting the individual stride characteristics; the adaptation coefficient is the proportional factor that relates step length to lower leg length, reflecting the personalized biomechanical needs; the slope duration angle is the inclination angle of the slope where the prosthesis is located, and it is necessary to continuously detect the time exceeding the threshold to filter out instantaneous interference.

[0128] In the above technical solution, the personalized target buckling angle is calculated in real time during the gait cycle:

[0129] Individualized baseline calculation: The actual step length is calculated through spatial geometric relationships, combined with the preset lower leg length, and a baseline target angle is generated according to the adaptation coefficient ratio (baseline value = coefficient × step length / leg length). The adaptation coefficient is set within a reasonable range based on biomechanical research.

[0130] Benchmark correction mechanism: When the deviation between the benchmark target angle and the initial reference value exceeds the allowable range, the target angle in the iterative learning is replaced;

[0131] Dynamic slope compensation: When a continuous slope is detected, the target angle is increased proportionally uphill to enhance push-off assist, while the target angle is decreased downhill with a safety lower limit to prevent excessive buckling. The slope compensation amount changes linearly with the incline angle to ensure a smooth transition.

[0132] Existing technologies rely on preset multi-target angle modes (such as flat / climbing modes), which have limitations: they depend on manual mode switching and cannot respond in real time to sudden changes in slope; the target angles within a mode remain fixed values, failing to consider dynamic changes in stride length (such as transitioning from brisk to slow walking); and downhill safety restrictions use global high damping, sacrificing gait naturalness. This technical solution achieves individual adaptation through stride-driven baseline angle calculation, combined with online correction of slope-sensitive target angles, maximizing freedom of movement while ensuring safety.

[0133] In the above technical solution, the benchmark target angle design associated with stride length accurately matches individual gait characteristics, solving the problem of stride limitation or excessive flexion caused by a uniform target value. The slope compensation mechanism dynamically optimizes biomechanical requirements: increasing the target angle uphill to enhance propulsion, and limiting the lower limit of the target angle downhill to prevent instability, while avoiding a stiff gait with globally high damping. The system automatically maintains the rationality of the target angle in scenarios of stride changes and slope transitions, reducing the frequency of manual recalibration.

[0134] In another technical solution, the method for controlling the prosthetic knee joint further includes step S4 of adjusting the learning rate based on the rate of change of gait cycle duration.

[0135] When the standard deviation of the gait cycle duration for three consecutive cycles is less than 0.05 s, the standard learning rate L is used.

[0136] When the standard deviation is ≥0.05 seconds, the adaptive learning rate L is enabled. adaptive :L adaptive =L×[1+0.5×tanh(5×| T k - T k-1 | / T k-1 )];middle T k The duration of the current cycle. T k-1 This is the duration of the previous cycle.

[0137] In the above technical solution, the degree of gait speed fluctuation is the statistical discreteness of the duration of continuous gait cycles, reflecting the stability of walking rhythm; the adaptive learning rate is the parameter correction intensity that is dynamically adjusted with the rate of change of gait speed; the smoothing constraint function is a nonlinear transformation mechanism that limits the amplitude of learning rate change and prevents abrupt changes.

[0138] The above technical solution involves real-time monitoring of the fluctuation characteristics of continuous gait cycle duration: Stability criterion: When the fluctuation of gait cycle duration is below a set threshold, a standard learning rate is maintained to ensure the accuracy of parameter fine-tuning; Fluctuation response mechanism: When significant fluctuations in gait speed are detected, an adaptive learning rate is generated based on the relative rate of change of adjacent cycle durations. This learning rate scales smoothly with the amplitude of gait speed changes; the larger the rate of change, the higher the learning rate increase, but the maximum amplification factor is constrained by a nonlinear function; Smooth transition design: The adaptive learning rate automatically decays to the standard value when the gait speed returns to stability, avoiding abrupt changes in damping parameters.

[0139] Existing technologies employ a constant learning rate iterative strategy, which has limitations: the learning rate is decoupled from gait state, resulting in low optimization efficiency during fast-to-slow speed transitions; a fixed learning rate causes knee flexion angle oscillations during sudden gait changes, requiring multiple falls for recovery before convergence; and manually preset learning rates cannot adapt to individual gait differences. This technical solution achieves a balance between optimization efficiency and stability by adaptively adjusting the learning rate driven by gait fluctuations, overcoming the rigid constraints of the fixed learning rate strategy.

[0140] In another technical solution, the method for controlling the prosthetic knee joint, step S4 further includes performing the following steps when the number of records in the gait-damping mapping knowledge base is less than 5:

[0141] For unknown gait period duration T Initial damping opening value plus safety margin: Initial damping opening = U closest ×min(1.3, 1+0.1×| T - T closest | / T closest ); U closest For the closest damping opening, T closest The closest gait cycle duration;

[0142] Set the maximum number of iterations to 10;

[0143] When the number of iterations k>5 and |65°- i kmax | When the angle is greater than 15°, reset the damping opening U. k =0.5×( U min+ U max );in U min and U max These are the preset safe working range boundary values: Umin=0.3 (corresponding to a damping force of 80 N·m·s / rad) and Umax=0.8 (corresponding to a damping force of 30 N·m·s / rad), determined based on knee joint mechanical strength tests to ensure that the damping force is within the safe range of 20-150 N·m·s / rad.

[0144] In the above technical solution, the damping safety margin is a protective amplification of the initial damping value to prevent the risk of underdamping; the iterative circuit breaker mechanism is a safety strategy to interrupt and reset parameters when the learning process is abnormal; the safe working range is the damping opening boundary value to ensure the stability of knee joint movement.

[0145] In the above technical solution, when the system detects insufficient gait-damping mapping records, triple protection is activated:

[0146] Initial damping enhancement: Based on the damping value corresponding to the closest historical step speed, the initial value is amplified according to the actual step speed difference, but the maximum increase threshold is limited to avoid being overly conservative.

[0147] Learning process monitoring: Continuously track the buckling angle error during the iteration process. If it exceeds the safety tolerance and more than half of the iterations have been completed, immediately interrupt the learning and reset the damping to the preset safety range midpoint.

[0148] Hard resource constraints: Set a maximum limit on the number of iterations, forcibly terminate unconverged processes, and release computing resources.

[0149] Existing technologies employ a globally conservative damping strategy (e.g., uniformly setting all unknown gait speeds to high damping), which has limitations: sacrificing gait naturalness for safety results in stiffness during the swing phase; the lack of an iterative process monitoring mechanism leads to the risk of mechanical damage due to abnormal learning; and rigid resource allocation and inefficient iteration slow down system response. This technical solution maximizes the degrees of freedom of movement while ensuring safety through a collaborative design of neighbor-value scaling, circuit breaker reset, and resource constraints.

[0150] In the above technical solution, the safety margin design provides robust initial damping for scenarios with unknown walking speeds, preventing the risk of knee joint loss of control. The circuit breaker mechanism promptly interrupts the parameter divergence process, resetting to a safe range to eliminate the risk of mechanical overload. Iteration number constraints optimize the allocation of computational resources, ensuring the system's real-time response capability. These three elements work together to ensure a balance between learning security and efficiency in scenarios with sparse knowledge bases.

[0151] In another technical solution, the method for controlling the prosthetic knee joint, step S4 further includes real-time monitoring of pressure signal abrupt change characteristics: if the heel pressure...P h Decrease of >80% within 20ms P hmax When the angular velocity is greater than 50° / second, the current iteration of learning is interrupted and the system switches to safety damping mode. U emergency =0.7× U min +0.3× U max , U emergency For emergency damping; after the terrain stabilizes, the damping setting should be restored to its pre-interruption state. U k Continuing the iteration, the determination of terrain stability requires that the following conditions be met simultaneously: a) P h ≥40%× P hmax up to 500ms; b) i cal Rate of change < 2° / s and angular velocity oh c) The initial support point was detected after three consecutive sampling values ​​<10° / s.

[0152] The above technical solution includes: pressure mutation characteristics, which indicate a signal pattern of a significant decrease in heel pressure within a very short time, predicting a sudden change in the support surface; emergency damping opening, which refers to a compromise damping value that balances braking requirements and motion continuity; and terrain steady-state criteria, which refers to the multi-dimensional recovery conditions that integrate pressure recovery, angle changes, and phase events.

[0153] In this technical solution, the coupling abrupt change characteristics of heel pressure and angular velocity are monitored in real time:

[0154] Real-time instability detection: When a sharp decrease in heel pressure is detected accompanied by rapid knee extension, it is immediately identified as a risk of instability.

[0155] Safety mode switching: interrupts the learning process and switches to the preset emergency damping opening, which is between high-damping braking and motion freedom;

[0156] Intelligent recovery mechanism: Learning can only resume when three conditions are met simultaneously: heel pressure is continuously restored to a safe range, joint movement is stable, and a new gait cycle starting point is detected to ensure the terrain is truly stable.

[0157] Existing technologies employ safety freeze strategies (such as locking joints after triggering), which have limitations: relying solely on angle thresholds to determine instability, they may miss slip risks caused by sudden pressure changes; adaptive learning is completely disabled in emergency situations, requiring a system restart to recover; and recovery conditions are singular (such as only reaching a pressure threshold), leading to incorrect recovery in swaying terrain. This technical solution improves instability detection sensitivity through pressure-angular velocity coupling recognition, designs a compromise emergency damping to ensure basic motion capabilities, and establishes multi-dimensional recovery criteria to achieve seamless continuation of the learning process.

[0158] In the aforementioned technical solution, the pressure-motion coupling mechanism accurately captures the precursors of sudden instability, giving the wearer critical reaction time. The compromise emergency damping design strikes a balance between fall prevention and mobility maintenance, avoiding secondary risks caused by joint locking. Multi-dimensional recovery criteria ensure that the learning process automatically resumes after the terrain is fully stable, maintaining the continuity of adaptive control and comprehensively improving the prosthesis's active protection capability in the face of sudden terrain changes.

[0159] In another technical solution, the control method for the prosthetic knee joint adds damping gradual control during the transition phase from the late support stage to the early swing stage:

[0160] When the calibration tilt angle is detected in the later stage of the support i cal When the angle is greater than 20°, the damping opening transitions linearly: U transition = U support +( U - U support )×( i cal -20°) / ( i swingstart -20°), where U transition For the transition period damping opening, U support The damping value is the fixed opening during the support period, and U is the damping opening during the early stage of the swing, which is the output of the current iteration. i swingstart This is the starting angle of the current cycle's early oscillation; the transition process continues until the end of the early oscillation phase, after which the complete cycle is applied. U value.

[0161] In the above technical solution, the transition starting angle refers to the critical point of the tilt angle that triggers the linear change of damping; the fixed damping during the support period refers to the high damping value used in the weight-bearing stage of the prosthesis; and the target damping during the swing period refers to the ideal damping value in the early stage of the swing period optimized by iterative learning.

[0162] In the above technical solution, when the offset of the later support tilt angle exceeds the transition start angle:

[0163] Linear transition trigger: Calculate the transition value between the fixed damping during the support period and the target damping during the swing period based on the proportion of the current tilt angle in the transition interval;

[0164] Dynamic gradual control: As the tilt angle increases, the damping value continuously transitions from high damping during the support phase to low damping during the swing phase, with the transition rate matching the joint movement speed;

[0165] Seamless switching guarantee: The transition continues until the end of the early stage of the swing, after which the target damping value is fully applied.

[0166] Existing technologies employ phase boundary step switching, which has limitations: sudden damping changes can cause knee joint impact noise and accelerate mechanical wear; sudden release of push-off energy can cause unnatural forward leg swing, affecting gait smoothness; and it does not consider the changing transition time requirements due to differences in gait speed. This technical solution achieves dynamic synchronization between damping force and joint movement through an angle-proportion-driven linear transition mechanism.

[0167] In the above technical solution, the tilt angle ratio control allows the damping force to change continuously and gradually with joint flexion, completely eliminating the mechanical impact during the transition between movement phases. The push-off kinetic energy is released smoothly through the damping gradient, improving the efficiency of the lower leg forward swing and gait continuity. The transition duration adapts to the joint movement speed, ensuring natural switching at different speeds. Overall, it reduces prosthetic movement noise and the wearer's metabolic burden.

[0168] In another technical solution, the control method for the prosthetic knee joint adds sensor fault redundancy processing in S1:

[0169] Real-time calculation of the effectiveness index of heel pressure signal; if it is continuous for 5 cycles P hmax <10N or signal variance <0.1N 2 If the heel pressure sensor is determined to be faulty, perform the following steps:

[0170] The vertical acceleration component a is output using an inertial measurement unit mounted proximal to the thigh prosthesis segment. z Alternative P h The ground contact threshold is set to |a z |>2g;

[0171] Simultaneously activate the heel impact sound sensor installed on the inside of the prosthetic ankle. When a sound pressure level of 200-500Hz > 80dB is detected and is consistent with a z When the peak time difference is less than 50ms, the grounding event is confirmed; the backup signal grounding event triggers the reference angle reset in step S2 and the stage division in step S3.

[0172] In the above technical solution, the signal validity index refers to the combination of features that quantify the reliability of the pressure signal (such as peak level and fluctuation intensity).

[0173] Vertical impact acceleration is the transient acceleration component experienced by the prosthesis along its axis, reflecting the intensity of the impact upon contact with the ground;

[0174] Voiceprint event synchronization represents the spatiotemporal consistency verification of different physical quantity sensors detecting the same event.

[0175] In the above technical solution, the reliability of the heel pressure signal is evaluated in real time. When abnormal signal characteristics are detected for multiple consecutive cycles: multi-source signal switching - the vertical acceleration signal of the thigh segment inertial unit is used to replace the heel pressure, and the ground contact event is determined by the impact acceleration threshold; voiceprint collaborative verification - the heel impact characteristics captured by the acoustic sensor at the ankle are analyzed synchronously, and ground contact is confirmed only when the sound pressure characteristics are consistent with the acceleration peak in time and space; lossless functional migration - the ground contact event of the backup sensor triggers the reference angle reset and gait phase division of the original process to ensure the continuity of system function.

[0176] Existing technologies employ a single backup sensing channel (e.g., accelerometer only), which has limitations: the acceleration threshold is easily affected by external impacts during walking (e.g., bumps), leading to false ground contact; there is no multi-source signal cross-verification mechanism, resulting in a false judgment rate exceeding safety tolerance in fault scenarios; and the backup signal only maintains basic motion, losing the ability to finely segment gait phases. This technical solution achieves highly reliable ground contact recognition and full functional maintenance in fault scenarios through a spatiotemporal coupling verification mechanism of acoustic and accelerometer dual physical quantities.

[0177] In this technical solution, signal validity monitoring enables early warning of sensor failures, preventing erroneous data from contaminating the control system. The acoustic-accelerometer spatiotemporal coupling verification mechanism eliminates interference from walking, ensuring the accuracy of ground contact event determination. The lossless functional migration design ensures that gait segmentation and damping adjustment remain fully supported even when the sensor fails, significantly improving the system's fault tolerance and continuous operating time.

[0178] The number of devices and processing scale described herein are for the purpose of simplifying the description of the invention. Applications, modifications, and variations of the invention will be readily apparent to those skilled in the art.

[0179] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details.

Claims

1. A method for controlling the knee joint of a prosthetic limb, characterized in that, include: S1. Real-time acquisition of heel pressure sensor signal P h Foot pressure sensor signal P f The tilt angle θ and angular velocity ω output by the lower leg inertial measurement unit; S2, when P is detected f rate of descent v f >k·v h At that moment, the real-time tilt angle θ now Defined as the new 0° reference, denoted as θ base =θ now ;v h For P h The rate of descent, k=2 is a preset multiple threshold; Calculate the calibrated tilt angle offset θ cal θ cal =θ-θ base ; S3, based on θ cal With dual pressure signals, the gait cycle is divided into early support phase, middle support phase, late support phase, early swing phase, middle swing phase, and late swing phase. The gait cycle duration T is calculated based on the interval between the initial starting points of adjacent supports. S4. Establish and call the gait-damping mapping knowledge base to construct a mapping relationship library between gait cycle duration T and swing early damping opening U; detect the current T value and query the gait-damping mapping knowledge base. If a matching T exists, directly call the corresponding U value; if the T is unknown, load the known gait corresponding U that is closest to T as the initial value. Perform iterative learning on the unknown T: U k+1 =U k +L(θ target -θ kmax ) Where U is the damping opening, L is the learning rate, and θ target For a target buckling angle of 65°, θ kmax The measured maximum buckling angle in the k-th cycle; When the buckling angle error converges to |θ target -θ kmax Stop iteration and store the mapping relationship when |≤5°; S5. Apply the adjusted damping opening U during the early stage of oscillation; maintain fixed damping during the early, middle, and late stages of support and oscillation. Based on θ cal Based on the dual pressure signals, the gait cycle is divided into early support, middle support, late support, early oscillation, middle oscillation, and late oscillation phases, specifically: When P h ≥P hmax And θ cal When the temperature is less than -3°, it is considered to be in the early stage of support. When P h >50%×P hmax P f >50%×P fmax And θ cal When the rate of change is <1° / s, it is judged as a medium-term support level; P hmax P in the current gait cycle h Peak value, P fmax P in the current gait cycle f Peak value; When P h <30%×P hmax And θ cal When the angle is greater than 0°, it is considered to be in the later stage of support. When P f =0 and θ cal An increase of ≥10° within 100ms is considered the early stage of the oscillation. When θ cal When the angular velocity increases continuously to the peak value and |ω| < 1° / s, it is determined to be in the middle stage of the oscillation; When θ cal The peak value θ detected from the current oscillation period max When the descent value is ≥3° and the angular velocity ω <0, it is determined to be the late stage of the oscillation.

2. The method for controlling the prosthetic knee joint as described in claim 1, characterized in that, The following optimization steps are performed during the iterative learning process in step S4: S41. Using 65° as the initial target buckling angle, calculate the personalized target buckling angle reference value in real time: Reference target buckling angle = adaptation coefficient × (step length / lower leg length); where, the step length is calculated through the geometric relationship between the spatial distance of the initial starting point of adjacent supports and the calibration tilt angle; the lower leg length is a preset value; the adaptation coefficient ranges from 0.4 to 0.6; S42. When |benchmark target buckling angle -65°| > 10°, replace the target buckling angle in the iterative learning with the benchmark target buckling angle; S43. When a slope angle β greater than 3° is detected for more than 2 seconds, in an uphill scenario, the corrected target buckling angle = the current target buckling angle × (1 + 0.2 × β); in a downhill scenario, the corrected target buckling angle = the maximum value (50°, current target buckling angle × (1 - 0.1 × |β|)).

3. The method for controlling the prosthetic knee joint as described in claim 1, characterized in that, Step S4 also includes adjusting the learning rate based on the rate of change of gait cycle duration: When the standard deviation of the gait cycle duration for three consecutive cycles is less than 0.05 s, the standard learning rate L is used. When the standard deviation ≥ 0.05 seconds, enable the adaptive learning rate L adaptive : L adaptive = L × [1 + 0.5 × tanh(5 × |T k - T k-1 | / T k-1 )]; where T k is the current cycle duration, and T k-1 is the previous cycle duration.

4. The method for controlling the prosthetic knee joint as described in claim 3, characterized in that, Step S4 also includes performing the following steps when the number of records in the gait-damping mapping knowledge base is less than 5: Add a safety margin to the initial damping opening value for the unknown gait period duration T: initial damping opening =U closest ×min(1.3, 1+0.1×|TT) closest | / T closest );U closest For the closest damping opening, T closest The closest gait cycle duration; Set the maximum number of iterations to 10; When the iteration number k>5 and |65°-θ kmax When the angle is greater than 15°, reset the damping opening U. k =0.5×(U min +U max ); where U min with U max These are the preset safe working range boundary values.

5. The method for controlling the prosthetic knee joint as described in claim 4, characterized in that, Step S4 also includes real-time monitoring of pressure signal abrupt change characteristics: if the heel pressure P h Decrease of >80% within 20ms × P hmax When the angular velocity is greater than 50° / second, interrupt the current iteration of learning and switch to the safety damping mode: U emergency =0.7×U min +0.3×U max U emergency For emergency damping opening; Once the terrain stabilizes, restore the U-shaped path as it was before the interruption. k Continuing the iteration, the determination of terrain stability requires that the following conditions be met simultaneously: a) P h ≥40%×P hmax up to 500ms; b)θ cal c) The rate of change is <2° / s and the angular velocity ω is <10° / s for three consecutive samplings; c) The early support starting point was detected.

6. The method for controlling the prosthetic knee joint as described in claim 1, characterized in that, During the transition from the later stage of support to the early stage of oscillation, damping gradual control is added: When the calibration tilt angle θ is detected in the later stage of the support cal When the angle is greater than 20°, the damping opening transitions linearly: U transition =U support +(UU support )×(θ cal -20°) / (θ swingstart -20°), where U transition For the damping opening during the transition period, U support Let U be the fixed damping opening during the support period, and θ be the initial damping opening of the swing phase output by the current iteration. swingstart This is the starting angle of the current cycle's early oscillation; the transition process continues until the end of the early oscillation phase, after which the full U value is applied.

7. The method for controlling the prosthetic knee joint as described in claim 1, characterized in that, Add sensor fault redundancy handling in S1: Real-time calculation of the effectiveness index of heel pressure signal; if P for 5 consecutive cycles hmax <10N or signal variance <0.1N 2 If the heel pressure sensor is determined to be faulty, perform the following steps: The vertical acceleration component a is output using an inertial measurement unit mounted proximal to the thigh prosthesis segment. z Replace P h The ground contact threshold is set to |a z |>2g; Simultaneously activate the heel impact sound sensor installed on the inside of the prosthetic ankle. When a sound pressure level of 200-500Hz > 80dB is detected and is consistent with a z When the peak time difference is less than 50ms, the grounding event is confirmed; the backup signal grounding event triggers the reference angle reset in step S2 and the stage division in step S3.