A phased adaptive control method for rehabilitation training exoskeleton

Abstract

This invention relates to a phased adaptive control method for an exoskeleton used in rehabilitation training, belonging to the field of human-machine collaborative motion assistance technology. This method identifies gait phases based on hip joint angle and angular velocity, dividing the gait into four phases: early swing phase, late swing phase, early stance phase, and late stance phase. It then uses a polynomial curve with control points to complete basic gait planning. Finally, it adaptively adjusts the gait trajectory control points based on the swing dynamics and stance phase. This invention enables phased adaptive adjustment of gait trajectory and auxiliary torque, improving training flexibility and patient participation, effectively avoiding sensor signal interference and insufficient accuracy, and contributing to improved rehabilitation training outcomes.

Description

Technical Field

[0001] This invention relates to the field of human-machine collaborative motion assistance technology, and in particular to a phased adaptive control method for a rehabilitation training exoskeleton. Background Technology

[0002] Rehabilitation training exoskeleton robots can help patients with lower limb dysfunction to train their gait, strengthen their muscles, remodel their nerves, improve lower limb function, accelerate the rehabilitation process, and reduce the workload of rehabilitation therapists. This is of great significance for improving the efficiency and level of medical rehabilitation and alleviating the contradiction of severe shortage of rehabilitation medical supply.

[0003] Currently, rehabilitation training exoskeleton control methods are mainly divided into passive control and active control. Passive control primarily involves fixed gait planning for the exoskeleton, which drives the patient's limb movements according to a pre-set standardized gait trajectory. This method is fixed in pattern, lacks flexibility, and has low patient participation. Active control mainly relies on additional sensors such as force, electromyography (EMG), and plantar pressure to determine the patient's intentions and adjusts the training speed and joint range of motion accordingly to adapt the training mode to the patient's actual training situation. However, EMG sensor signals are susceptible to interference and have low stability. If the accuracy and sensitivity of force and plantar pressure sensors do not meet the requirements, it can easily affect the recognition and control effect. Furthermore, existing active control methods do not consider adjusting the patient's gait trajectory in stages, and cannot adaptively adjust the training program at each stage according to the patient's actual situation. Summary of the Invention

[0004] To address the aforementioned technical problems, embodiments of the present invention provide a phased adaptive control method for an exoskeleton used in rehabilitation training. This method enables active rehabilitation training without the need for additional sensors. It uses feedback current from joint motors to reflect the patient's training adaptation and performs phased adaptive control based on the training progress. The control strategy is adaptively adjusted according to the patient's intentions and state, which helps with the patient's neural remodeling.

[0005] The technical solution adopted in this invention is as follows:

[0006] A phased adaptive control method for rehabilitation training exoskeleton, comprising the following steps:

[0007] The hip joint phase angle is calculated based on the hip joint angle and angular velocity. Gait phase recognition is performed according to preset judgment conditions, and the exoskeleton gait is divided into four phases: early swing phase, late swing phase, early stance phase, and late stance phase.

[0008] Gait planning is performed based on the results of gait phase recognition. By constructing a polynomial angle curve with control points, a standardized description of the hip joint motion trajectory is achieved, providing a baseline curve and adjustable basis for phased adaptive adjustment.

[0009] The human-machine interaction state is quantified based on the deviation between the joint motor feedback current and the reference current. By adjusting the control points in the pendulum dynamic and support states differently, the hip joint angle trajectory is corrected to achieve a phased adaptive adjustment strategy.

[0010] Optionally, the initial swing phase is the stage where the left leg swings forward from the extended position of maximum hip extension until the left leg is perpendicular to the ground.

[0011] The later stage of the swing is when the left leg continues to swing forward from a position perpendicular to the ground until the heel is about to touch the ground;

[0012] The initial support phase begins with the left heel touching the ground, followed by a gradual shift of body weight from the right leg to the left leg until the left leg returns to a perpendicular position to the ground.

[0013] The later stage of the support phase involves the left leg continuing to support the body from a position perpendicular to the ground, extending backward until the hip joint reaches its maximum extension position.

[0014] Optionally, if the left leg hip joint phase angle transitions from greater than or equal to threshold A to less than threshold A, the left leg hip joint angular velocity is less than threshold B, and the right leg gait was in the early support phase at the previous moment, then the left leg gait is set to the early swing phase; where A takes values ​​ranging from... The range of values ​​for B is .

[0015] Optionally, if the change in the left leg hip joint angle does not exceed a threshold C, and the difference between the left and right leg hip joint angles is greater than a threshold D, then the left leg gait is set to the late swing phase; where C takes values ​​ranging from... The range of values ​​for D is .

[0016] Optionally, if the left leg hip joint phase angle transitions from greater than or equal to a threshold E to less than a threshold E, and the left leg hip joint angular velocity is greater than a threshold F, then the left leg gait is set to the early support phase; where E takes values ​​ranging from... The range of values ​​for F is .

[0017] Optionally, if the difference between the hip joint angles of the right and left legs does not exceed a threshold G, and the difference between the hip joint angles of the right and left legs is greater than a threshold H, then the gait of the left leg is set to late support phase; where G takes values ​​ranging from... The range of H values ​​is .

[0018] Optionally, the control point adjustment rules for the pendulum dynamic and support states are differentiated, including: in the pendulum dynamic stage, the amplitude of the forward swing is increased or decreased by adjusting the coordinates of the left control point; in the support state stage, the amplitude of the support extension is adjusted by adjusting the coordinates of the right control point.

[0019] Optionally, the judgment basis for the phased adaptive adjustment strategy is the deviation between the joint motor feedback current and the reference current, and the human-machine interaction state is quantified by the deviation coefficient.

[0020] Optionally, based on the gait phase recognition results, the complete gait cycle is divided into swing dynamics and support dynamics. Swing dynamics correspond to the forward swing phase of the lower limbs and are used to control the swing height, forward swing speed and foot landing position. Support dynamics correspond to the support and center of gravity transfer phases of the lower limbs and are used to control the support extension amplitude, center of gravity transition speed and push-off force.

[0021] Optionally, the adjustment rules take the control points in gait planning as the direct adjustment objects, and achieve smooth correction of the hip joint angle curve by changing the coordinates of the control points.

[0022] This invention provides a phased adaptive control method for exoskeleton in rehabilitation training. The technical solution provided by the embodiments of this invention brings at least the following beneficial effects:

[0023] (1) This invention does not require additional force, electromyography, plantar pressure and other sensors. It can reflect human-machine interaction force by simply using the feedback current of the joint motor, which reduces system complexity and cost and avoids control failure caused by interference from additional sensor signals and insufficient accuracy.

[0024] (2) This invention divides gait training into four phases, adaptively adjusting gait trajectory and auxiliary torque in stages to match the patient's actual training adaptation, enhance the patient's active participation, assist in neural remodeling, and optimize the rehabilitation training effect. This invention achieves real-time adaptive adjustment of the control strategy, and can dynamically match the training intensity according to the patient's intention and state, improving the flexibility and pertinence of rehabilitation training.

[0025] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 A flowchart illustrating a phased adaptive method for exoskeleton rehabilitation training;

[0028] Figure 2 This is a schematic diagram of gait phase division;

[0029] Figure 3A flowchart illustrating the process of human gait recognition;

[0030] Figure 4 A schematic diagram of the gait planning angle curve;

[0031] Figure 5 This is a schematic diagram of the angle curve adjustment. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0033] Before describing the technical solution of the present invention in detail, the technical background and technical terms involved in the technical solution will be explained first:

[0034] Exoskeleton: A wearable mechanical / mechatronic device worn on the outside of the human limb, conforming to the limb in parallel. Driven by motors, supported by structures, and powered by motors, it assists the human body in performing movements that are difficult to achieve independently. It belongs to the category of human-machine collaborative motion assistive equipment. The rehabilitation training exoskeleton of this invention is a wearable robot specifically designed for patients with lower limb dysfunction. It is used for gait training, muscle strengthening, and neural remodeling. Through preset trajectories or adaptive control, it guides / assists the patient's legs to complete a standardized gait, improving rehabilitation efficiency.

[0035] In a first aspect, the present invention provides a phased adaptive method for rehabilitation training exoskeleton. Figure 1 A flowchart illustrating a phased adaptive method for exoskeleton rehabilitation training. This method includes the following steps:

[0036] Step S101: Calculate the hip joint phase angle based on the hip joint angle and angular velocity, and perform gait phase recognition according to preset judgment conditions, dividing the exoskeleton gait into four phases: early swing phase, late swing phase, early support phase, and late support phase.

[0037] This invention targets the human-machine collaborative gait characteristics of rehabilitation training exoskeletons, finely dividing the complete gait cycle into four continuous phases: early swing phase (S1), late swing phase (S2), early stance phase (S3), and late stance phase (S4). Figure 2 Schematic diagram of gait phase division and Figure 3 The gait recognition process provides a detailed description of the complete technical solution for gait phase recognition.

[0038] 1. Physiological definition and morphological characteristics of gait phase

[0039] Taking the left leg as the core identification object, and combining the physiological laws of lower limb movement in rehabilitation training, the movement patterns and functions of the four phases are defined as follows:

[0040] Early swing phase (S1): The left leg swings forward from its maximum hip extension position until it is perpendicular to the ground. This phase corresponds to... Figure 2 The first simple sketch shows the left leg in a bent position with its head raised forward. Its main function is to overcome gravity and swing the lower limb forward to prepare for the next step of landing. This is a stage where the patient actively participates.

[0041] Late swing phase (S2): The left leg continues to swing forward from a position perpendicular to the ground until the heel is about to touch the ground. This phase corresponds to... Figure 2 The second simplified sketch shows the left leg almost perpendicular to the ground, with only the heel slightly raised. The movement speed gradually slows down, and the swing amplitude needs to be controlled to avoid impact in order to land smoothly.

[0042] Early Support Phase (S3): After the left heel touches the ground, the body's center of gravity gradually shifts from the right leg to the left leg until the left leg returns to a perpendicular position to the ground. This phase corresponds to... Figure 2 The third simplified pose in the middle is when the left leg steps forward and lands to support the body weight, which is a key stage where the exoskeleton provides auxiliary support.

[0043] Late Support Phase (S4): The left leg continues to support the body from a position perpendicular to the ground, extending backward until the hip joint reaches its maximum extension position. This phase corresponds to... Figure 2 The fourth simplified gait posture is with the left leg extended backward, accumulating potential energy for the next swing and propelling the body forward. This is the final stage of the gait cycle.

[0044] The left and right legs have an alternating and coordinated phase. When the left leg is in the swing phase, the right leg is in the support phase, and vice versa, thus forming a complete walking gait cycle.

[0045] 2. Data Acquisition and Core Feature Calculation

[0046] To achieve accurate phase recognition, the core feature quantities of hip joint movement are first collected and calculated. The specific process is as follows:

[0047] Angle acquisition: The exoskeleton joint encoder acquires the angle data of the left and right hip joints in real time, which is recorded as the left leg hip joint angle. Right leg hip joint angle The sampling period is consistent with the control period T, with a value range of 10ms to 25ms, to ensure that the data update frequency matches the real-time requirements of rehabilitation training.

[0048] Angular velocity calculation: Based on the difference in hip joint angles between adjacent control cycles, the angular velocities of the left and right hip joints are calculated using the difference method. The calculation formula is as follows:

[0049] ;

[0050] ;

[0051] in , These are the angular velocities of the hip joints of the left and right legs at time t, respectively. They are used to characterize the speed and trend of lower limb movement and are key indicators for judging the swing / support state.

[0052] Phase angle calculation: Based on the two-dimensional motion state of the hip joint angle and angular velocity, calculate the phase angle of the hip joint for the left and right legs respectively. , The calculation formula is as follows:

[0053] ;

[0054] ;

[0055] in This is a four-quadrant arctangent function with an output range of [-π, π], which can be mapped to the [0, 2π] interval as needed to fully characterize the phase period. The essence of the phase angle calculation above is to map angles and angular velocities into phase coordinates, used to characterize the relative position of the hip joint movement within the gait cycle, providing a quantitative basis for phase interval division.

[0056] 3. Phase determination logic and constraints

[0057] Combination Figure 3 The gait recognition flowchart uses the left leg as the core recognition object and achieves precise phase switching through multi-condition logic judgment. The triggering conditions and constraint logic of each phase are as follows:

[0058] (1) Early stage of oscillation S1

[0059] The trigger condition requires the simultaneous fulfillment of the following three constraints for the left leg gait to be... Setting it to 1 ensures the accuracy and coordination of phase switching:

[0060] Constraint 1: ;

[0061] The transition of the left leg hip joint phase angle from greater than or equal to threshold A to less than threshold A indicates that the motion phase has shifted from the late support phase to the early swing phase; the range of A is... .

[0062] Constraint 2: ;

[0063] Left leg hip joint angular velocity A value less than threshold B indicates that the left leg swing speed has entered a low-speed range, marking the start of the forward swing motion; where the value of B ranges from [value missing]. .

[0064] Constraint 3: ;

[0065] At the previous moment, the right leg was in the early support phase (rGait(t-1)=3), ensuring the alternation and coordination of the left and right legs to avoid gait disorder.

[0066] If all the above constraints (constraints 1 to 3) are met, the left leg gait is set to the early swing phase; otherwise, the gait of the previous frame is kept unchanged to prevent misjudgment caused by noise interference.

[0067] (2) S2 in the later stage of the oscillation

[0068] The trigger condition requires that the following two constraints be met simultaneously: left leg gait. Setting it to 2 indicates that the left leg's forward swing is entering the finishing phase:

[0069] Constraint 4: ;

[0070] The change in the left leg hip joint angle does not exceed the threshold C, indicating that the swing amplitude of the left leg tends to stabilize and the speed gradually decreases; the value of C ranges from [value missing]. .

[0071] Constraint 5: ;

[0072] A difference in hip joint angles between the left and right legs greater than a threshold D indicates that the left leg is still in a forward swing posture while the right leg is in a support posture, consistent with the gait alternation pattern; the value of D ranges from [value missing]. .

[0073] If all the above constraints (constraints 4 to 5) are met, then the left leg gait is set to the late swing phase; otherwise, the previous gait remains unchanged.

[0074] (3) Support the early stage S3

[0075] The trigger condition requires that the following two constraints be met simultaneously: left leg gait. Setting it to 3 indicates that the left leg is switching from swinging to supporting:

[0076] Constraint 6: ;

[0077] The transition of the left leg hip joint phase angle from greater than or equal to a threshold E to less than a threshold E indicates that the motion phase has shifted from the late swing phase to the early support phase; the value of E ranges from [value missing]. .

[0078] Constraint 7: ;

[0079] A left leg hip joint angular velocity greater than the threshold F characterizes the velocity characteristics of the left leg upon landing, distinguishing it from a slow swinging state; the value of F ranges from [value missing]. .

[0080] If all the above constraints (constraints 6 to 7) are met, then the left leg gait is set to the early support phase; otherwise, the previous gait remains unchanged.

[0081] (4) Supporting the later stage S4

[0082] The trigger condition requires that the following two constraints be met simultaneously: left leg gait. Setting it to 4 indicates that the left leg enters the extension support phase:

[0083] Constraint 8: ;

[0084] The difference in hip joint angle between the right and left legs does not exceed the threshold G, indicating that the angle difference between the left and right legs is in the late support phase; the value of G ranges from [value missing]. .

[0085] Constraint 9: ;

[0086] If the difference in hip joint angle between the right and left legs is greater than the threshold H, non-support late-stage states with excessively small angle differences are excluded; the value range of H is... .

[0087] If all the above constraints (constraints 8 to 9) are met, then the left leg gait is set to late support phase; otherwise, the previous gait remains unchanged.

[0088] (5) Gait maintenance logic

[0089] If the current moment does not meet any of the above phase switching conditions, the gait phase of the previous moment remains unchanged. The "hysteresis hold" mechanism avoids phase jumps caused by transient noise, fluctuations in patient exertion, or sensor errors, thereby improving recognition stability.

[0090] 4. Overall recognition process and real-time performance guarantee

[0091] Combination Figure 3 The flowchart shows the complete gait phase recognition process:

[0092] Process 1: After the system (i.e., the phased adaptive control system for rehabilitation training exoskeleton) is started, it enters the control loop, first collecting the current angle of the left leg hip joint. Right leg hip joint angle .

[0093] Step 2: Calculate the angular velocities of the left and right hip joints based on angle data. , Thus, the motion speed characteristics are obtained.

[0094] Step 3: Calculate the phase angles of the left and right hip joints based on angle and angular velocity. , The phase coordinate characteristics are obtained.

[0095] Step 4: Sequentially perform conditional checks for the early swing phase, late swing phase, early support phase, and late support phase, and update the left leg gait phase based on the check results. .

[0096] Step 5: If none of the switching conditions are met, then the current phase remains unchanged.

[0097] Process 6: Enter the next control cycle and repeat the above steps to achieve continuous and real-time gait phase recognition.

[0098] Through the above data acquisition, feature calculation and multi-condition constraint logic, the system can accurately identify the four phases of the gait of the rehabilitation training exoskeleton, providing a reliable phase basis for subsequent gait planning and phased adaptive control, while ensuring the real-time recognition and anti-interference ability.

[0099] Step S102: Based on the results of gait phase recognition, gait planning is performed. By constructing a polynomial angle curve with control points, a standardized description of the hip joint motion trajectory is achieved, providing a baseline curve and adjustable basis for phased adaptive adjustment.

[0100] After completing gait phase recognition, the control system enters the gait planning stage based on the identified four phases: early swing phase, late swing phase, early stance phase, and late stance phase. Gait planning is the core intermediate link connecting phase recognition and phased adaptive adjustment. Its function is to generate a smooth, continuous joint angle trajectory that conforms to the normal walking pattern of the human body based on the motion laws and physiological characteristics of different phases, providing a baseline curve and adjustable foundation for subsequent phased adaptive adjustment. This step achieves a standardized description of the hip joint motion trajectory by constructing a polynomial angle curve with control points, while reserving control point adjustment interfaces so that the subsequent adaptive control stage can make precise and stable dynamic corrections to the gait trajectory based on the patient's human-computer interaction status.

[0101] 1. Formula for gait planning angle curve

[0102] This invention employs a Bernstein polynomial (Bezier curve) with control points as the gait planning angle curve to achieve a standardized description of the hip joint movement trajectory. By precisely controlling the curvature direction, degree of curvature, and amplitude of the trajectory curve through control points, the gait trajectory satisfies both the requirements of smooth and continuous movement and possesses an online adjustable adaptive basis. The gait planning angle curve of this invention uses the following formula:

[0103] ;

[0104] The meanings of each parameter are as follows:

[0105] θ(t): The target angle of the hip joint at time t, which is the dependent variable of the curve and represents the target posture of the hip joint at the current gait moment.

[0106] t: Normalized time parameter, with a value range of [0,1], representing the relative progress within the gait cycle. t=0 corresponds to the start of the gait cycle, and t=1 corresponds to the end of the gait cycle.

[0107] n: Polynomial order (i.e., the number of control points minus one). The larger the order n, the higher the accuracy of the curve's fitting control of gait movement, and the more delicate the matching of human physiological gait characteristics.

[0108] The binomial coefficient (combination number) is calculated using the following formula: This determines the weight distribution of the influence of each control point on the shape of the training curve, ensuring that the curve transitions smoothly and without abrupt changes between each gait phase.

[0109] Bernstein basis functions assign weights to each control point that vary with time t, making the curve smooth through the sequence of control points.

[0110] : Coordinates P of the i-th control point i (x i ,y i The gait trajectory is the core adjustment unit. Its coordinate position directly controls the bending direction and degree of the training curve, and it is the direct object of subsequent phased adaptive adjustment.

[0111] This curve uses control point Pi as the core adjustment carrier. By changing the coordinates of the control point, the curve shape can be adjusted globally without refitting the entire curve, providing a simple and efficient adjustment interface for subsequent phased adaptive adjustments.

[0112] 2. Physical meaning and regulation basis of control points

[0113] The control point Pi(xi,yi) is the core adjustment unit of the gait planning angle curve. Its coordinate position directly determines the bending direction, bending degree and movement amplitude of the curve: the horizontal axis xi controls the time / phase distribution within the gait phase interval, affecting the movement rhythm and gait sequence; the vertical axis yi controls the amplitude of the hip joint angle under the corresponding phase, affecting movement parameters such as swing height and extension amplitude.

[0114] To achieve adaptive trajectory adjustment, this invention is based on the feedback current of the joint motor. Calculate the deviation coefficient The human-computer interaction state is converted into the basis for adjusting control points. The formula for calculating the deviation coefficient is as follows:

[0115] ;

[0116] The meanings of each parameter are as follows:

[0117] The real-time acquisition of the feedback current of the hip joint drive motor is positively correlated with the interaction force between the exoskeleton and the patient's limb; the greater the interaction force, the greater I.

[0118] The pre-calibrated reference current for the ideal collaborative state represents the motor current required to drive the exoskeleton under standard gait trajectory without additional human-machine interaction force. It is determined by the mechanical structure of the exoskeleton, transmission efficiency, and energy consumption during normal gait.

[0119] The current deviation coefficient is a relative deviation value that can be positive or negative. It directly reflects the degree and direction of human-computer interaction deviating from the ideal coordination.

[0120] If k>0 (i.e. I>I0), it indicates that the human-computer interaction force is too large, the patient's muscle strength is insufficient, and the exoskeleton is in a passive assisted state of "dragging the patient". The training angle needs to be reduced to reduce the exercise load.

[0121] If k≤0 (i.e. I≤I0), it indicates that the human-computer interaction force is moderate / too small, the patient has sufficient muscle strength or high tolerance, and can actively participate in gait movements. The training angle needs to be increased to strengthen the rehabilitation stimulus.

[0122] The larger |k| is, the more significant the difference between the current state and the ideal cooperative state, and the greater the required trajectory adjustment. The smaller |k| is, the closer the state is to the ideal, the more moderate the adjustment, and the less overcorrection is needed.

[0123] In the aforementioned gait planning, the hip joint is used as the controlled object, and a polynomial function is employed to construct the basic trajectory curve of the joint angle changing over time. This curve not only ensures a smooth transition between each gait phase, avoiding abrupt changes, shocks, or jerks in joint movement, but also allows for global control of the entire curve's shape through a small number of control points, providing a simple and reliable adjustment method for adaptive adjustment. Control points are the key adjustment units of the gait trajectory; their coordinate positions directly determine the bending direction, degree of bending, and amplitude of the angle curve. Different gait phases correspond to different curve shapes and control point constraints. During the swing phase, the angle curve mainly reflects the speed and position changes of the lower limb's forward swing, and control points are used to adjust the swing height, swing speed, and foot landing position. During the support phase, the angle curve mainly reflects the weight transfer, support extension, and push-off process, and control points are used to adjust the support stiffness, joint extension amplitude, and center of gravity transition speed.

[0124] The gait planning method employed in this invention updates the target joint angle by selecting the curve parameters of the corresponding segment based on the currently identified gait phase within each control cycle. After generating the basic trajectory, the planned angle curve is output to subsequent control stages as the initial reference trajectory for phased adaptive adjustment. When the patient's muscle strength, movement intention, or human-machine interaction force changes, the adaptive adjustment module only needs to modify the coordinates of the control points under the corresponding gait phase. This allows for online correction of the gait trajectory without disrupting the continuity and smoothness of the curve, thus achieving a complete closed-loop control from gait phase recognition to baseline trajectory generation and real-time adaptive correction. Therefore, gait planning not only inherits the phased information provided by gait phase recognition, ensuring that the trajectory conforms to the normal gait sequence and movement patterns, but also provides an adjustable and optimizable basic object for the phased adaptive adjustment strategy. This makes the entire control method interdependent and progressively advances in the three stages of recognition, planning, and adjustment, ultimately achieving safe, compliant, and personalized phased adaptive control of the rehabilitation training exoskeleton.

[0125] Step S103: Quantify the human-machine interaction state based on the deviation between the joint motor feedback current and the reference current, and execute the correction of the hip joint angle trajectory through the differentiated control point adjustment rules of the pendulum dynamic and the support state to achieve a phased adaptive adjustment strategy.

[0126] The phased adaptive adjustment strategy is the core closed-loop component of the rehabilitation training exoskeleton control method. It follows the phase division of gait phase recognition, and is based on the Bernstein polynomial curve and control point mechanism of gait planning. It uses the joint motor feedback current as the only sensing signal, without relying on additional force, electromyography or plantar pressure sensors, to achieve precise and dynamic adjustment of the hip joint trajectory under different phase states. Ultimately, it matches the patient's real-time muscle strength status and training tolerance, ensuring the safety, personalization and effectiveness of rehabilitation training.

[0127] 1. Core sensing basis: Quantitative definition of current deviation coefficient

[0128] The core criterion for the phased adaptive adjustment strategy is the deviation between the joint motor feedback current and the reference current. The human-computer interaction state is quantified by the deviation coefficient k, which has been described in detail above and will not be repeated here.

[0129] 2. Phase-based adjustment basis: Gait phase and curve segment division

[0130] Figure 4 This is a schematic diagram of the gait planning angle curve. Figure 5 This is a schematic diagram of the angle curve adjustment. Based on the gait phase recognition results, the complete gait cycle is divided into a swing dynamic (early swing phase lGait=1, late swing phase lGait=2) and a support phase (early support phase lGait=3, late support phase lGait=4), corresponding to... Figure 4 , Figure 5 The left segment (t∈[0,1]s) and the right segment (t∈[1,2]s) of the mid-hip angle curve:

[0131] Swing dynamics (left side segment): This corresponds to the forward swing phase of the lower limb. The core is to control the swing height, forward swing speed, and foot landing position, which directly affects the patient's comfort and safety when lifting their leg.

[0132] Support position (right side): This corresponds to the lower limb support and weight transfer phase. The key is to control the support extension range, the speed of weight transfer, and the force of the push-off, which directly affects the patient's standing stability and weight-bearing tolerance.

[0133] The physiological characteristics of movement differ significantly in different phase segments, therefore, differentiated control point adjustment rules are required to ensure that the adjustment movements conform to the timing of human gait.

[0134] 3. Differentiated control point adjustment rules

[0135] The adjustment rules directly adjust the Bernstein polynomial control points Pi(xi,yi) in gait planning, achieving smooth correction of the hip joint angle curve by changing the coordinates of these control points without refitting the entire curve. Combined with... Figure 4 (Control point adjustment direction) and Figure 5 (Multi-curve comparison), the adjustment logic under different phases is as follows:

[0136] (1) Adjustment rules for dynamic swing (lGait=1 / 2, left segment)

[0137] The oscillation corresponds to the left region of the curve, with control point Pi located in the rising segment (early stage of the oscillation) and the peak segment (late stage of the oscillation):

[0138] Scenario 1: Reduce the training angle (k>0, excessive interaction force)

[0139] Objective: To cause the left curve to contract inward, reducing the peak angle and amplitude of the hip joint's forward swing.

[0140] Adjust direction: Move the control point Pi(xi,yi) to the lower right (corresponding to...) Figure 4 (Left "reduced" arrow)

[0141] The horizontal axis, xi, increases (moves to the right), causing the peak moment of the forward swing to shift backward, thus slowing down the swing rhythm.

[0142] The vertical axis, yi, decreases (moves downwards), directly reducing the peak angle of the forward swing and decreasing the swing height.

[0143] Curve effect: such as Figure 5 The contraction pattern of the θ2 curve relative to the θ1 curve, with a significantly reduced forward swing amplitude, alleviates the burden on the patient's leg lifting.

[0144] Scenario 2: Increase the training angle (k≤0, moderate / small interaction force)

[0145] Objective: To expand the left curve outward and increase the peak angle and amplitude of the hip joint's forward swing.

[0146] Adjust direction: Move the control point Pi(xi,yi) to the upper left (corresponding to...) Figure 4 (Left "increase" arrow)

[0147] The horizontal axis xi decreases (shifts to the left), causing the peak moment of the forward swing to occur earlier and accelerating the swing rhythm.

[0148] The vertical axis, yi, increases (moves upwards), directly increasing the peak angle of the forward swing and thus increasing the swing height.

[0149] Curve effect: such as Figure 5 Compared to the θ1 curve, the θ3 curve has a more outward-convex baseline shape and an increased forward swing amplitude, which enhances the stimulation of muscle contraction in the patient's lower limbs.

[0150] (2) Adjustment rules for support state (lGait=3 / 4, right segment)

[0151] The support state corresponds to the area on the right side of the curve, with control point Pi located in the descending segment (early support phase) and the final segment (late support phase) of the curve:

[0152] Scenario 1: Reduce the training angle (k>0, excessive interaction force)

[0153] Objective: To cause the right curve to contract inward, reducing the hip joint support extension angle and the speed of weight transfer.

[0154] Adjust direction: Move the control point Pi(xi,yi) to the lower left (corresponding to...) Figure 4 (Right "reduction" arrow)

[0155] The horizontal axis, xi, decreases (shifts to the left), causing the peak moment of support extension to arrive earlier and slowing down the pace of center of gravity shift.

[0156] The vertical axis, yi, decreases (moves downwards), directly reducing the peak angle of the support extension and decreasing the load amplitude.

[0157] Curve effect: The curve on the right side contracts inward relative to the baseline shape, reducing the support extension angle and alleviating the weight-bearing pressure on the patient's lower limbs.

[0158] Scenario 2: Increase the training angle (k≤0, moderate / small interaction force)

[0159] Objective: To expand the right curve outward, thereby increasing the hip joint's support extension angle and the force of weight transfer.

[0160] Adjust direction: Move the control point Pi(xi,yi) to the upper right (corresponding to...) Figure 4 (Right-side "Enlarge" arrow)

[0161] The horizontal axis, xi, increases (shifts to the right), causing the peak moment of support extension to be delayed, thus prolonging the center of gravity transfer time.

[0162] The vertical axis, yi, increases (moves upward), directly raising the peak angle of the support extension and enhancing the push-off force.

[0163] Curve effect: The curve on the right side protrudes outward relative to the baseline shape, which increases the range of support and stretching, and strengthens the involvement of the patient's core muscles and extensor muscles.

[0164] 4. Dynamic update formula for control point coordinates: Segmented recursion achieves smooth adjustment.

[0165] To ensure the continuity and smoothness of trajectory adjustment and avoid abrupt changes in joint movement, this invention employs a piecewise recursive formula to achieve real-time updates of control point coordinates, strictly distinguishing between the adjustment directions in pendulum dynamics and support states. The formula is as follows:

[0166] (1) x-axis Update formula

[0167] ;

[0168] The meanings of each parameter are as follows:

[0169] The horizontal coordinate of the control point in the previous control cycle ensures the continuity of adjustment.

[0170] k: Current deviation coefficient, which determines the adjustment range and direction (k>0 means decreasing the angle, k≤0 means increasing the angle).

[0171] Δx: The preset step size for adjusting the horizontal coordinate, a positive constant whose sign is determined by the gait phase;

[0172] Dynamic state: +kΔx → The horizontal coordinate increases / decreases with the sign of k (corresponding to the lower right / upper left movement); Support state: -kΔx → The horizontal coordinate decreases / increases with the sign of k (corresponding to the lower left / upper right movement).

[0173] (2) Vertical axis Update formula

[0174] ;

[0175] : The ordinate of the control point in the previous control cycle;

[0176] Δy: The preset adjustment step size for the ordinate, which is a positive constant;

[0177] When k>0 (decreasing angle): -kΔy → decreases the ordinate (moves downward); when k≤0 (increasing angle): -kΔy is equivalent to +|k|Δy → increases the ordinate (moves upward).

[0178] (3) Control cycle and real-time guarantee

[0179] The control point update frequency is consistent with the control period T (T∈[10ms,25ms]), and is completed within each control period:

[0180] Acquire the current gait phase lGait(t).

[0181] Collect the real-time motor current I and calculate the deviation coefficient k.

[0182] Based on the phase and the sign of k, the new control point coordinates Pi(xi,yi) are calculated by substituting them into the update formula.

[0183] Based on the new control points, the target hip angle is recalculated using the Bernstein polynomial formula.

[0184] The target angle is output to the motor drive module to complete real-time control.

[0185] The aforementioned phased adaptive adjustment strategy, following gait phase recognition, strictly adheres to the temporal division of the four gait phases, precisely binding adjustment actions to the swing and support phases. This ensures the adjustment logic conforms to the physiological laws of human gait, avoids cross-phase erroneous adjustments, and guarantees the integrity and stability of the gait temporal sequence. Using the control point Pi as the core adjustment interface, the abstract human-computer interaction force state is quantified into specific coordinate adjustment quantities through a deviation coefficient k. Without modifying the order and basis functions of the Bernstein polynomial, trajectory correction is achieved simply by updating the control point, perfectly connecting to the gait planning stage. The updated control point directly generates a new target angle curve, providing precise instructions for the motor drive and realizing a complete closed loop of "perception-decision-execution." Through phased and directional smooth adjustments, it avoids sports injuries caused by excessive interaction forces and can enhance training intensity when the patient's tolerance is sufficient, ultimately achieving personalized, adaptive rehabilitation training to aid in neural remodeling and functional recovery.

[0186] The method described above in this invention is implemented by a phased adaptive control system for rehabilitation training exoskeleton. This system is a core hardware and software suite for realizing gait phase recognition, gait planning, and phased adaptive control, and specifically includes the following components:

[0187] 1. Hardware level

[0188] Exoskeleton body: including the joint mechanical structure of the hip and knee joints and drive motors, used to perform motion assistance.

[0189] Sensor module: Includes a joint encoder for real-time acquisition of the left leg hip joint angle. Right leg hip joint angle .

[0190] Controller: Typically an embedded microcontroller or industrial computer, responsible for data processing, algorithm calculation and control command output.

[0191] Drive module: Receives commands from the controller and drives the joint motors to complete gait movements.

[0192] 2. Software level

[0193] The controller in the phased adaptive control system of the rehabilitation training exoskeleton is the hardware core and execution carrier of the entire control system. The data acquisition and preprocessing module, gait phase recognition module, gait planning module, and adaptive adjustment module are software functional units running on the controller. Together, they form a complete closed loop of "perception-decision-control".

[0194] Data acquisition and preprocessing module: responsible for reading encoder data and calculating angular velocity and phase angle.

[0195] Gait phase recognition module: namely the four-phase judgment logic described in this invention, which outputs the current gait phase based on the collected feature data.

[0196] Gait planning module: Generates basic gait trajectory curves based on the identified phases.

[0197] Adaptive adjustment module: Adjusts gait trajectory and auxiliary torque in stages based on motor feedback current.

[0198] In summary, this invention constructs a complete closed-loop control system around a rehabilitation training exoskeleton, encompassing "gait phase recognition—gait planning—phased adaptive adjustment." Its core is to achieve personalized, safe, and efficient lower limb rehabilitation training solely through feedback current from joint motors and gait phase information, without relying on additional force / electromyography / plantar pressure sensors. This invention relies solely on joint motor feedback current, eliminating the need for additional sensors, thus reducing system complexity and cost while improving reliability. Based on four-phase gait recognition, phased control is achieved, conforming to the physiological laws of human gait and avoiding cross-phase misadjustments. Using Bernstein polynomial control points as the adjustment interface, smooth and efficient trajectory correction is achieved, ensuring exercise safety and comfort. The human-machine interaction state is quantified through a current deviation coefficient, dynamically adjusting training intensity to both prevent sports injuries and enhance rehabilitation effects.

[0199] In a second aspect, the present invention provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described rehabilitation training exoskeleton phased adaptive control method.

[0200] The present invention may also provide a storage medium, which may be a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the above-described rehabilitation training exoskeleton phased adaptive control method.

[0201] The computer-readable storage medium provided by this invention may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0202] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A phased adaptive control method for a rehabilitation training exoskeleton, characterized in that, The method includes the following steps: The hip joint phase angle is calculated based on the hip joint angle and angular velocity. Gait phase recognition is performed according to preset judgment conditions, and the exoskeleton gait is divided into four phases: early swing phase, late swing phase, early stance phase, and late stance phase. Gait planning is performed based on the results of gait phase recognition. By constructing a polynomial angle curve with control points, a standardized description of the hip joint motion trajectory is achieved, providing a baseline curve and adjustable basis for phased adaptive adjustment. The human-machine interaction state is quantified based on the deviation between the joint motor feedback current and the reference current. By adjusting the control points in the pendulum dynamic and support states differently, the hip joint angle trajectory is corrected to achieve a phased adaptive adjustment strategy.

2. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1, characterized in that, The initial phase of the swing is when the left leg starts from the extended position of the hip joint at its maximum extension and swings forward until the left leg is perpendicular to the ground; The later stage of the swing is when the left leg continues to swing forward from a position perpendicular to the ground until the heel is about to touch the ground; The initial support phase begins with the left heel touching the ground, followed by a gradual shift of body weight from the right leg to the left leg until the left leg returns to a perpendicular position to the ground. The later stage of the support phase involves the left leg continuing to support the body from a position perpendicular to the ground, extending backward until the hip joint reaches its maximum extension position.

3. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1 or 2, characterized in that, If the left leg hip joint phase angle transitions from greater than or equal to threshold A to less than threshold A, the left leg hip joint angular velocity is less than threshold B, and the right leg gait was in the early support phase at the previous moment, then the left leg gait is set to the early swing phase; where A takes values ​​ranging from... The range of values ​​for B is .

4. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1 or 2, characterized in that, If the change in the left leg hip joint angle does not exceed threshold C, and the difference between the left and right leg hip joint angles is greater than threshold D, then the left leg gait is set to the late swing phase; where C ranges from [value missing]. The range of values ​​for D is .

5. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1, characterized in that, If the left leg hip joint phase angle transitions from greater than or equal to threshold E to less than threshold E, and the left leg hip joint angular velocity is greater than threshold F, then the left leg gait is set to the early support phase; where E takes values ​​ranging from... The range of values ​​for F is .

6. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1, characterized in that, If the difference between the hip angles of the right and left legs does not exceed the threshold G, and the difference between the hip angles of the right and left legs is greater than the threshold H, then the gait of the left leg is set to late support phase; where G takes values ​​ranging from... The range of H values ​​is .

7. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1, characterized in that, The control point adjustment rules are differentiated between the pendulum dynamic and the support state. In the pendulum dynamic stage, the forward swing amplitude is increased or decreased by adjusting the coordinates of the left control point. In the support state stage, the support extension amplitude is adjusted by adjusting the coordinates of the right control point.

8. The staged adaptive control method for rehabilitation training exoskeleton according to claim 1, characterized in that, The basis for the phased adaptive adjustment strategy is the deviation between the joint motor feedback current and the reference current, and the human-computer interaction state is quantified by the deviation coefficient.

9. The staged adaptive control method for rehabilitation training exoskeleton according to claim 8, characterized in that, Based on the gait phase recognition results, the complete gait cycle is divided into swing dynamics and support dynamics. Swing dynamics correspond to the forward swing phase of the lower limbs and are used to control the swing height, forward swing speed and foot landing position. Support dynamics correspond to the support and center of gravity transfer phases of the lower limbs and are used to control the support extension amplitude, center of gravity transition speed and push-off force.

10. The staged adaptive control method for rehabilitation training exoskeleton according to claim 9, characterized in that, The adjustment rules take the control points in gait planning as the direct adjustment objects, and achieve smooth correction of the hip joint angle curve by changing the coordinates of the control points.

Images