Memory alloy driving glove for hand rehabilitation, driving system and control method

Abstract

The invention discloses a memory alloy driving glove for hand rehabilitation. The memory alloy driving glove comprises a glove body, a forearm base and a traction mechanism, the glove body comprises an elastic iron sheet located on the palm side, the elastic iron sheet extends to the fingertip of each finger from the palm center and is fixed to each finger through a finger ring, and guide grooves are formed in the elastic iron sheet and the finger ring. The front arm base is fixed on the front and back surfaces of the front arm; the traction mechanism comprises memory alloy arrays and traction lines, the memory alloy arrays are arranged on the front face and the back face of the forearm base correspondingly, and the glove body is pulled by penetrating the traction lines through the guide grooves; the device further comprises a bending angle sensor and a memory alloy driver, the bending angle sensor is attached to the back face of the finger, and the memory alloy driver is used for conducting pulse heating on the memory alloy array so as to drive the finger to move. According to the invention, overshoot is suppressed while the response speed is ensured, the steady-state error is reduced, and the follow-up flexibility is improved.

Description

Technical Field

[0001] This invention relates to the fields of rehabilitation engineering and wearable robot technology, and in particular to a hand rehabilitation memory alloy driven glove, its driving system, and its control method. Background Technology

[0002] With the aging population, the number of patients with hemiplegia due to stroke and nerve damage has surged. Existing hand rehabilitation exoskeletons are mainly divided into two categories: motor-driven and pneumatic-driven. However, motor-driven structures are bulky and have high rigidity and impact, which can easily cause secondary damage to the patient's fragile spastic joints; while pneumatic drives are smooth, they are large in size, noisy, and difficult to achieve precise angle control.

[0003] In recent years, shape memory alloys (SMAs) have been introduced into the rehabilitation field due to their high power-to-weight ratio and biomimetic muscle properties. However, SMAs are typical thermally driven materials, exhibiting strong nonlinearity, severe hysteresis, and slow thermal response. Directly using traditional integer-order PID controllers often results in phenomena such as "overheating causing sudden finger bending (large overshoot)" or "constant shaking near the target position (large steady-state error)," severely hindering their practical clinical application. Summary of the Invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the prior art, the first objective of this invention is to provide a hand rehabilitation memory alloy driven glove; The second objective is to provide a drive system based on the aforementioned hand rehabilitation memory alloy drive glove; The third objective is to provide a drive control method based on the aforementioned hand rehabilitation memory alloy drive system.

[0005] Summary of the Invention: The hand rehabilitation memory alloy driven glove provided by the present invention includes: a glove body, a forearm base, and a traction mechanism; The glove body includes: an elastic iron sheet located on the palm side, the elastic iron sheet extending from the palm to the fingertip of each finger, and fixed to each finger by a finger ring, and the elastic iron sheet and the finger ring are provided with guide grooves; The forearm base is fixed to both the front and back sides of the forearm; The traction mechanism includes a shape memory alloy array and a traction line. The shape memory alloy array is set on both the front and back sides of the forearm base, and the traction line passes through the guide groove to achieve traction on the glove body. It also includes a bending angle sensor and a shape memory alloy actuator. The bending angle sensor is attached to the back of the finger, and the shape memory alloy actuator is used to pulse-heat the shape memory alloy array to drive finger movement.

[0006] Furthermore, the forearm base is provided with a fixed guide groove for the spaced memory alloy array. Multiple branches of the memory alloy are connected in parallel, and each branch is connected to the guide groove of the corresponding finger through an independent traction line to realize independent driving of each finger.

[0007] Furthermore, the guide grooves are respectively provided at the joints where the elastic iron sheet and the ring are bent.

[0008] Furthermore, the front and back sides of the forearm base are connected by elastic straps to achieve adaptive fixation.

[0009] Furthermore, the elastic sheet has a pre-bent shape, providing an initial bending moment in the flexion direction when the finger is fully extended.

[0010] Furthermore, a layered heat insulation structure is provided on the side of the forearm base that contacts the skin, and a ventilation cavity is provided opposite to the shape memory alloy array to suppress heat conduction and accelerate heat dissipation.

[0011] Correspondingly, the provided hand rehabilitation memory alloy drive system, based on the hand rehabilitation memory alloy drive glove, includes: a host computer, a data acquisition card, a voltage to PWM module, a power drive module, a memory alloy driver, and a bending angle sensor; The host computer has a built-in control algorithm that calculates control quantities in real time and sends control voltages through the analog output interface of the data acquisition card. After being mapped into a duty cycle signal by the voltage-to-PWM module, the voltage-to-PWM module triggers the power drive module to pulse-heat the memory alloy driver, thereby driving finger movement. At the same time, the bending sensor monitors the finger joint angle in real time, converts it into a voltage signal, and feeds it back to the host computer through the analog input interface of the data acquisition card, forming a physical closed loop.

[0012] Correspondingly, the provided hand rehabilitation memory alloy drive control method, based on a hand rehabilitation memory alloy drive system, includes the following steps: S1. Obtain the desired angle θ1 and the measured angle θ2 fed back by the bending sensor to obtain the error signal e(t), e(t) = θ1 - θ2; S2. Input the error signal into the fractional-order PI control algorithm to calculate the initial control quantity u(t); the fractional-order integral operator uses the impulse response invariance method for discretization approximation to obtain the discrete transfer function to form a discrete controller; S3. Based on the sign of the error signal e(t) and its rate of change, determine whether the current stage is flexion or extension under hysteresis logic, and then select the target execution object between the memory alloy arrays of the palm and the back of the hand, or perform cooperative action. S4. The initial control quantity is limited in amplitude by combining current limiting protection, and the control quantity is converted into a duty cycle signal by voltage to PWM module mapping. The signal is then applied to the execution target selected in S3 by power drive module to heat the shape memory alloy.

[0013] In S2, the discretization sampling period Ts is set to 1ms, and the fitting order N is set to 5 times to meet the embedded real-time requirements while preserving the frequency domain characteristics of the fractional-order operator. The hysteresis logic in S3 is specifically configured as follows: the error signal serves as both the input to the controller and the criterion for the switching logic. When the error exceeds the positive threshold, the memory alloy array drive of the palm is locked; when the error exceeds the negative threshold, the drive of the memory alloy array on the back of the hand is switched. Within the threshold range, the current operating state is maintained or bidirectional fine-tuning is performed to eliminate frequent switching caused by signal jitter.

[0014] Furthermore, S4 employs a linear mapping relationship. Based on the allocation result of S3, the power drive module energizes only the selected single branch memory alloy array, or in the cooperative operation mode, proportionally allocates the duty cycle to the two branch memory alloy arrays to achieve alternating or coordinated driving.

[0015] Beneficial effects: Compared with the prior art, the significant advantages of this invention are: through the dual traction transmission chain, a repeatable initial bending moment can be provided under fully straightened conditions, thereby reducing the starting "dead zone" and backlash difference, and improving the transmission stiffness and buckling start consistency; based on the designed training device, a fractional-order PI is introduced and the impulse response invariance method is used to discretize the continuous domain fractional-order integral operator, achieving low computational load and excellent tracking performance under embedded sampling period constraints, manifested in overshoot suppression, oscillation reduction and steady-state error reduction, the system achieves comprehensive improvement in starting response, compliant follow-up and thermal safety, and enhanced clinical usability and operational reliability. Attached Figure Description

[0016] Figure 1 This is a diagram showing the result of driving the glove according to the present invention; Figure 2 This is a schematic diagram of the dual traction method of the present invention; Figure 3 This is a diagram of the driving system of the present invention; Figure 4 This is the fractional-order control logic diagram of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] like Figure 1 and 2 The hand rehabilitation memory alloy driven glove shown includes: glove body 1, forearm base 2 and traction mechanism 3; The glove body 1 includes: an elastic metal plate 101 located on the palm side, extending from the palm to the fingertip of each finger, and fixed to each finger by a finger ring 102; guide grooves 103 are provided on the elastic metal plate 101 and the finger ring 102; the guide grooves 103 are respectively provided at the joints of the elastic metal plate 101 and the finger ring 102 corresponding to bending. The elastic metal plate 101 has a pre-bent shape, providing an initial bending torque in the flexion direction when the fingers are fully extended.

[0019] The forearm base 2 is fixed to both sides of the forearm and connected by elastic straps 202 to achieve adaptive fixation. The forearm base 2 is provided with a fixing guide groove 201 for the spaced memory alloy array 301. Multiple branches of the memory alloy are connected in parallel, and each branch is connected to the guide groove 103 of the corresponding finger through an independent traction wire 302 to achieve independent driving of each finger. The memory alloys located in the palm and back of the hand are the first and second actuation branches, respectively.

[0020] The traction mechanism 3 includes a shape memory alloy array 301 and a traction line 302. The shape memory alloy array 301 is respectively set on the front and back sides of the forearm base 2, and the traction line 302 passes through the guide groove 103 to achieve traction on the glove body 1. It also includes a bending angle sensor and a memory alloy driver. The bending angle sensor is attached to the back of the finger, and the memory alloy driver is used to pulse heat the memory alloy array 301, thereby driving the finger movement.

[0021] The forearm base 2 has a layered heat insulation structure on the side in contact with the skin and a ventilation cavity opposite to the shape memory alloy array 301 to suppress heat conduction and accelerate heat dissipation.

[0022] The bending angle sensor is a flexible resistive sensor that can change its resistance characteristics as the finger joint bends, enabling real-time monitoring of the joint bending angle.

[0023] like Figure 3 The hand rehabilitation memory alloy drive system shown constructs a hardware closed-loop system of "host computer - data acquisition - execution - feedback", including: host computer, data acquisition card, voltage to PWM module, power drive module, memory alloy driver and bending angle sensor; The host computer has a built-in control algorithm that calculates control quantities in real time using MATLAB / Simulink. The control voltage is sent through the analog output interface of the data acquisition card. After being mapped into a duty cycle signal by the voltage-to-PWM module, the power drive module is triggered to pulse-heat the memory alloy driver, thereby driving finger movement. At the same time, the bending sensor monitors the finger joint angle in real time, converts it into a voltage signal, and feeds it back to the host computer through the analog input interface of the data acquisition card, forming a physical closed loop.

[0024] like Figure 4 As shown, the control method of the above-mentioned hand rehabilitation memory alloy drive system includes the following steps: S1. Obtain the desired angle θ1 and the measured angle θ2 fed back by the bending sensor to obtain the error signal e(t), e(t) = θ1 - θ2; S2. Input the error signal into the fractional-order PI control algorithm to calculate the initial control quantity u(t); the fractional-order integral operator is discretized and approximated using the impulse response invariance method to obtain the discrete transfer function to form a discrete controller.

[0025] S3. Based on the sign and rate of change of the error signal e(t), the system determines whether the current state is flexion or extension under hysteresis logic, thereby selecting the target execution object between the memory alloy arrays on the palm and back of the hand, or performing coordinated action. The specific configuration of the hysteresis logic is as follows: the error signal serves as both the input to the controller and the criterion for the switching logic. When the error exceeds the positive threshold, the drive of the memory alloy array on the palm is locked; when the error exceeds the negative threshold, the drive of the memory alloy array on the back of the hand is switched. Within the threshold range, the current action state is maintained or bidirectional fine-tuning is performed to eliminate frequent switching caused by signal jitter. S4. The initial control quantity is amplitude-limited by current limiting protection, and the control quantity is converted into a duty cycle signal through a voltage-to-PWM module. This signal is then applied to the execution target selected in S3 by the power drive module to heat the shape memory alloy. A linear mapping relationship is used; the power drive module, based on the allocation result in S3, either powers only the selected single branch of the shape memory alloy array, or, in cooperative operation mode, proportionally allocates the duty cycle to both branches of the shape memory alloy array to achieve alternating or coordinated driving.

[0026] In S2: ; In the formula: K p This is the proportional gain coefficient, which determines the system's response speed to the current error. In this system, it is set to 8000. K i This is the integral gain coefficient. This parameter is used to eliminate the steady-state error of the system and ensure that the finger can ultimately stay precisely at the set angle. In this system, it is set to 30. D -μ Fractional integral operator, where μ is the order of integration (0 < μ < 1).

[0027] Fractional integral operator S -μ Discrete approximation is performed using the impulse invariance method to obtain the discrete transfer function, which is then used to form a discrete controller. -μTreating it as an analog filter, the impulse response h(t) is obtained using the inverse Laplace transform. The unit impulse response h(t) is sampled at equal intervals with a sampling period Ts to obtain discrete sampling points h(nTs) of length N, n=0,1,…,N−1. The discrete sampled signal h(nTs) is fitted using the Prony function to obtain the discrete transfer function of the fractional integral operator. ; Discrete fractional integral element is defined as G μ (z)≈H(z).

[0028] In MATLAB / Simulink, calling `irid_fod(α, Ts, N)` directly calculates the coefficients of the discrete transfer function. i},{α i} (where α is the fractional order, Ts is the sampling period, and N is the order of the fit), and this is filled into the DiscreteTransfer Fcn module to implement Gμ(z). Therefore, the discrete implementation of FOPI in this study is: ; To balance real-time performance and phase-preserving characteristics, μ is selected from 0.4 to 0.9, preferably from 0.6 to 0.85, and further preferably from 0.73; the sampling period Ts is set to 1 ms; and the fitting order N is set to 5. The coefficients of the discrete fractional integral element G(z) are rounded to three decimal places, resulting in: ; The discrete integral transfer function G is obtained above. μ (z) is then substituted into the discrete expression of the fractional-order PI controller. In each sampling period Ts, the host computer calculates the current error... e ( k The current control quantity is calculated using the discrete algorithm described above. u ( k To prevent the calculated control quantity from exceeding physical limits, the control quantity is then subjected to saturation limiting and finally linearly mapped to a PWM duty cycle signal of 0-100%.

[0029] After the system is powered on, it enters soft-start mode, reads the current value of the bending sensor as the initial position, and loads the target angle. In the initial stage of the flexion movement (i.e., when the error is relatively large), the allocation logic unit will recognize this state, enabling the finger to respond quickly to the start command.

[0030] As the SMA continues to heat up, the finger angle gradually approaches the target value. During this process, the fractional-order PI controller adjusts the PWM duty cycle in real time according to the rate of change of the error. If the error decreases too quickly, the controller will automatically reduce the duty cycle to prevent overshoot; If a switch from buckling to extension (i.e., error sign flipping) is required, the allocation logic unit will immediately cut off the power supply to the first execution branch (buckling SMA) and determine whether to enable the second execution branch (extension SMA) based on the hysteresis threshold. This logic-based switching mechanism avoids repeated oscillations near zero error.

[0031] When the measured angle enters the allowable error band of the target angle (e.g., ±2°), the system enters steady-state hold mode. At this time, the controller outputs a small hold duty cycle, which is used only to compensate for the heat loss of the SMA to the environment, so that the finger is stably held in a specific posture.

Claims

1. A hand rehabilitation memory alloy driven glove, characterized in that, include: The glove body (1), the forearm base (2), and the traction mechanism (3); The glove body (1) includes: an elastic iron sheet (101) located on the palm side, the elastic iron sheet (101) extending from the palm to the fingertip of each finger and fixed to each finger through a finger ring (102), and a guide groove (103) provided on the elastic iron sheet (101) and the finger ring (102). The forearm base (2) is fixed to both sides of the forearm; The traction mechanism (3) includes a shape memory alloy array (301) and a traction line (302). The shape memory alloy array (301) is set on both the front and back sides of the forearm base (2), and the traction line (302) passes through the guide groove (103) to achieve traction on the glove body (1). It also includes a bending angle sensor and a memory alloy actuator, the bending angle sensor being fitted to the back of the finger and the memory alloy actuator being used to pulse heat the memory alloy array (301) to drive finger movement.

2. The hand rehabilitation memory alloy driven glove according to claim 1, characterized in that: The forearm base (2) is provided with a fixed guide groove (201) for the memory alloy array (301) to be spaced apart. Multiple branches of the memory alloy are connected in parallel, and each branch is connected to the guide groove (103) of the corresponding finger through an independent traction line (302) to realize the independent driving of each finger.

3. The hand rehabilitation memory alloy driven glove according to claim 1, characterized in that: The guide groove (103) is respectively provided at the joints of the elastic iron sheet (101) and the ring (102) where they are bent.

4. The hand rehabilitation memory alloy driven glove according to claim 1, characterized in that: The front and back sides of the forearm base (2) are connected by elastic straps (202) to achieve adaptive fixation.

5. The hand rehabilitation memory alloy driven glove according to claim 1, characterized in that: The elastic sheet (101) has a pre-bent shape, providing an initial bending moment in the flexion direction when the finger is fully extended.

6. The hand rehabilitation memory alloy driven glove according to claim 1, characterized in that: The forearm base (2) is provided with a layered heat insulation structure on the side in contact with the skin and a ventilation cavity is provided opposite to the shape memory alloy array (301) to suppress heat conduction and accelerate heat dissipation.

7. A hand rehabilitation memory alloy drive system, characterized in that, The hand rehabilitation memory alloy driven glove according to claim 1 includes: a host computer, a data acquisition card, a voltage to PWM module, a power drive module, a memory alloy driver, and a bending angle sensor. The host computer has a built-in control algorithm that calculates control quantities in real time and sends control voltages through the analog output interface of the data acquisition card. After being mapped into a duty cycle signal by the voltage-to-PWM module, the voltage-to-PWM module triggers the power drive module to pulse-heat the memory alloy driver, thereby driving finger movement. At the same time, the bending sensor monitors the finger joint angle in real time, converts it into a voltage signal, and feeds it back to the host computer through the analog input interface of the data acquisition card, forming a physical closed loop.

8. A method for driving and controlling a memory alloy for hand rehabilitation, characterized in that, The hand rehabilitation memory alloy drive system according to claim 7 includes the following steps: S1. Obtain the desired angle θ1 and the measured angle θ2 fed back by the bending sensor to obtain the error signal e(t), e(t) = θ1 - θ2; S2. Input the error signal into the fractional-order PI control algorithm to calculate the initial control quantity u(t); the fractional-order integral operator uses the impulse response invariance method for discretization approximation to obtain the discrete transfer function to form a discrete controller; S3. Based on the sign of the error signal e(t) and its rate of change, determine whether the current stage is flexion or extension under hysteresis logic, and then select the target execution object between the memory alloy arrays of the palm and the back of the hand, or perform cooperative action. S4. The initial control quantity is limited in amplitude by combining current limiting protection, and the control quantity is converted into a duty cycle signal by voltage to PWM module mapping. The signal is then applied to the execution target selected in S3 by power drive module to heat the shape memory alloy.

9. The hand rehabilitation memory alloy drive control method according to claim 8, characterized in that: In S2, the discretization sampling period Ts is set to 1ms, and the fitting order N is set to 5, so as to meet the embedded real-time requirements while preserving the frequency domain characteristics of the fractional-order operator; the hysteresis logic in S3 is specifically configured as follows: the error signal serves as both the input of the controller and the criterion for the switching logic. When the error exceeds the positive threshold, the memory alloy array drive of the palm is locked; when the error exceeds the negative threshold, the drive of the memory alloy array on the back of the hand is switched. Maintain the current operating state or perform bidirectional fine-tuning within the threshold range to eliminate frequent switching caused by signal jitter.

10. The hand rehabilitation memory alloy drive control method according to claim 8, characterized in that: S4 employs a linear mapping relationship. Based on the allocation result of S3, the power drive module energizes only the selected single branch memory alloy array, or in the cooperative operation mode, it proportionally allocates the duty cycle to the two branch memory alloy arrays to achieve alternating or coordinated driving.

Images