CN110236876A - A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training - Google Patents
A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training Download PDFInfo
- Publication number
- CN110236876A CN110236876A CN201910467923.0A CN201910467923A CN110236876A CN 110236876 A CN110236876 A CN 110236876A CN 201910467923 A CN201910467923 A CN 201910467923A CN 110236876 A CN110236876 A CN 110236876A
- Authority
- CN
- China
- Prior art keywords
- wrist
- resampling
- steering engine
- forearm
- motion profile
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H1/00—Apparatus for passive exercising; Vibrating apparatus ; Chiropractic devices, e.g. body impacting devices, external devices for briefly extending or aligning unbroken bones
- A61H1/02—Stretching or bending or torsioning apparatus for exercising
- A61H1/0274—Stretching or bending or torsioning apparatus for exercising for the upper limbs
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B23/00—Exercising apparatus specially adapted for particular parts of the body
- A63B23/035—Exercising apparatus specially adapted for particular parts of the body for limbs, i.e. upper or lower limbs, e.g. simultaneously
- A63B23/12—Exercising apparatus specially adapted for particular parts of the body for limbs, i.e. upper or lower limbs, e.g. simultaneously for upper limbs or related muscles, e.g. chest, upper back or shoulder muscles
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
- A61H2201/12—Driving means
- A61H2201/1207—Driving means with electric or magnetic drive
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
- A61H2201/16—Physical interface with patient
- A61H2201/1602—Physical interface with patient kind of interface, e.g. head rest, knee support or lumbar support
- A61H2201/1635—Hand or arm, e.g. handle
- A61H2201/1638—Holding means therefor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H2230/00—Measuring physical parameters of the user
- A61H2230/08—Other bio-electrical signals
- A61H2230/085—Other bio-electrical signals used as a control parameter for the apparatus
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H2230/00—Measuring physical parameters of the user
- A61H2230/62—Posture
- A61H2230/625—Posture used as a control parameter for the apparatus
Abstract
A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training pass through the rehabilitation training of medical staff's guidance and adaptive limbs of patient.The present invention passes through teaching mode record rehabilitation training motion profile, exercise intensity identification is carried out to patient by deep learning frame again, and recognition result is embedded into the integral rehabilitation Training Control strategy based on deeply study, it exports the joint motions rate signal after adjusting and completes rehabilitation training to mechanical arm.This method is under the premise of preventing Rehabilitation training secondary injury, pass through the Vector Fusion of motor message and bio signal, the individual difference bring of hysteresis quality and bioelectrical signals that motor message can be eliminated influences, improve exercise intensity recognition accuracy, and widen the applicable surface of rehabilitation exercise track, to obtain the advantage for mitigating medical staff's burden and optimizing the Rehabilitation period, the safety and applicability of rehabilitation training are improved.
Description
Technical field
The present invention relates to rehabilitation medical field, specifically a kind of upper limb exoskeleton rehabilitation mechanical arm and utilize the upper limb dermoskeleton
The control method of bone mechanical arm progress rehabilitation training.
Background technique
Rehabilitation training is an important means of medical science of recovery therapy, mainly restores patient's suffering limb by this method of training
Normal self-care function, makes the rehabilitation of handicapped person's physiology and psychology as far as possible with trained method, reaches therapeutic effect.It is common
Rehabilitation training means have a. passive activity: carried out completely by external force, without any active contraction of muscle.External force can be by gravity, machine
Tool, other people or oneself the effect of another limbs produced by.B. actively and the activity of active power-assisted: oneself or by other people certain sides
It helps to complete the movement of limbs.Since the harmony and control ability of infant limb motion are poor, so training must be passed through
To develop the harmony of movement and improve the technical ability of functional activity.Active power-assisted activity can provide enough help, be answered with generating
Some range of motion.C. stretching activity: by increasing joint motion to joint continuous traction, it is mainly used for the pass of contracture
Section.
Rehabilitation mechanical arm is exactly to refer to the impaired paralysed patient of synkinesia function automatically or semi-automatically to complete rehabilitation instruction
Experienced electromechanical integration equipment.At present in rehabilitation medical field, rehabilitation mechanical arm achieves good using effect.Combine people
The rehabilitation mechanical arm of body engineering and intelligent control method can drive the limbs of patient to do rehabilitation exercise by the mode of wearing,
According to rehabilitation mechanical arm to the difference of the modes such as the connection of suffering limb position, support, traction of patient, rehabilitation mechanical arm can be divided into
End guiding mechanical arm and exoskeleton-type mechanical arm, end guiding mechanical arm refer to that mechanical arm drives patient's by end
Suffering limb completes rehabilitation movement, and general power source acts on the hand or wrist of patient, has easy for installation, and structure is simply special
Point is suitable for large-scale rehabilitation medical center.According to bionics principle, be that one kind is wearable sets exoskeleton-type rehabilitation mechanical arm
It is standby, by the way that mechanical arm to be worn on the arm of patient, the arm of patient is driven to complete rehabilitation movement.
Help to obtain preferable rehabilitation training effect by exoskeleton rehabilitation training mechanical arm.Patent of invention
CN106389072A proposes the method for rehabilitation of a kind of novel healing robot and virtual interacting, by introducing virtual reality skill
Art enhances human-computer interaction process, treats patient under more lively environment, can effectively promote rehabilitation training ring
Border, while also can preferably reflect rehabilitation training effect.Patent of invention CN106730638A proposes a kind of based on reciprocal force
The rehabilitation mechanical arm control method for identifying motion intention is identified that human motion is intended to by force signal, improves active training
When human-computer interaction continuity, ensured initiative rehabilitation training safety.Patent of invention CN108721058A is proposed by adopting
Collect common rehabilitation training track and store, then by choosing manually, sets corresponding rehabilitation training time, the health that real time parsing is chosen
Refreshment practices track to drive the limb motion of patient, so that rehabilitation training is completed, it is more to the optimization analysis research of track.Upper limb
Rehabilitation field is also quite paid attention in foreign countries, and international monopoly WO2018189614-A1 proposes one kind and activated by Noninvasive brain
Measurement process obtains the brain signal of user, determines that the intention of user feeds back to control system using the data of the brain signal of acquisition,
Control system synthesis processing multidimensional data information finally realizes the signal output to driver.Meanwhile in order to improve different patients
Individual difference problem, using different patients data grouping handle measure, based on bio signal intention assessment realize
Human-computer interaction has good effect to rehabilitation training effect.Japan Patent JP2018183272-A proposes that a kind of dress by VR sets
Standby acquisition human body interactive information, provides not only the good rehabilitation visual field, while providing man-machine interface, for determining patient's
Rehabilitation position and rehabilitation movement greatly improve rehabilitation environment and the rehabilitation experience of patient, meanwhile, which can be used for movement instruction
Experienced and musical instrument training.East Northwestern Polytechnical University Wang Wen etc. publishes thesis for 2017 on " Technology&Health Care "
The force control and path planning of electromagnetic induction-based
It is referred in massage robot to the dynamics control of massage robot and path planning, by adjusting the movement position of robot
And the state of massage head copes with different massage demands.
In conclusion having been achieved for preferable achievement in rehabilitation training field both at home and abroad at present, it can identify patient's
Motion intention, and it is fed back to rehabilitation mechanical arm control system, realize human-computer interaction.But to integral rehabilitation mechanical arm
For development, it still there is a problem that human body intention assessment precision is low, rehabilitation training track applicable surface is narrow, thus cannot
It is widely popularized.
Summary of the invention
To overcome human body intention assessment precision existing in the prior art is low, rehabilitation training track applicable surface is narrow to ask
The deficiency of topic, the invention proposes a kind of upper limb ectoskeleton mechanical arm and the control methods of rehabilitation training.
Upper limb ectoskeleton mechanical arm proposed by the present invention includes pitman shaft, linear motor, straight-line electric rack, large arm outside
Plate, damper connecting shaft, stepper motor, stepper motor frame, stepping motor gear, large arm interior plate, forearm interior plate, outside forearm
Side plate, wrist outer ring, wrist locating snap ring, wrist inside, wrist gear, steering engine link block, steering engine bracket, steering engine and hand are drawn
Bar.Wherein:
The wrist locating snap ring is inlaid in the card slot of wrist inner ring external peripheral surface upper end.The wrist gear set exists
The external peripheral surface of the wrist inner ring.The cantilever end for being fixed on the forearm interior plate of the wrist inner ring side is hinged on the inside of large arm
Plate;One end of forearm outer panel is fixed on the wrist inner ring other side, and stepper motor is fixed on the forearm by stepper motor frame
The outer surface of outer panel;The cantilever end of the forearm outer panel and one end of large arm outer panel are hinged.Stepping motor gear installation
It is engaged on the output shaft of stepper motor, and with the wrist gear.Damper connecting shaft is mounted on the forearm outer panel,
And close to the cantilever end of the forearm outer panel.One end of pitman shaft is packed into the damper connecting shaft, the other end and straight-line electric
Machine is connected;The linear motor is mounted on the outer surface of the large arm outer panel other end.Wrist exterior ring cap is mounted in described
The external peripheral surface of wrist gear, and make to be fitted close therebetween.Stop with the wrist upper surface of the wrist gear
The lower end surface of rotating ring is bonded, and the lower end surface of the wrist gear is concordant with the end face of wrist inner ring.The steering engine passes through steering engine bracket
It is mounted on steering engine link block on the mounting plate of the wrist exterior ring cap, and the steering engine axis for making the steering engine and the axis on the mounting plate
Hole is rotatably assorted.Hand pull rod is hinged on the mounting plate of the wrist outer ring.Straight-line electric rack and large arm outer panel are connected.Greatly
Arm outer panel and forearm outer panel are hinged, and large arm interior plate and forearm interior plate are hinged.On the outside of two panels large arm support plate and large arm
Plate and large arm interior plate are connected, while two panels forearm support plate and forearm interior plate and forearm outer panel being connected.Mechanical wrist
There are two freedom degree, the swings and twisting of corresponding wrist in joint.Stepping motor gear is engaged with wrist gear.
The internal diameter of the wrist inner ring is 120~150mm.The external peripheral surface of the wrist inner ring one end has the recessed of annular
Slot, for being embedded in wrist locating snap ring.There is the annular snap-gauge of radially protruding in the wrist inner ring the other end face, it is described for placing
Wrist gear, after the wrist gear set is in the wrist inner ring, under the upper surface of the annular slab and the wrist gear
End face fitting;The outer end face of the annular snap-gauge is evenly equipped with axial mounting plate, draws for installing the steering engine bracket and hand
Bar.
The internal diameter of the wrist outer ring is identical as the outer diameter of the wrist inner ring, and makes to be rotatably assorted therebetween.
The outer diameter of the wrist outer ring is identical as the internal diameter of the wrist gear.One end face of the wrist outer ring is evenly equipped with a pair of axial
The mounting plate of protrusion.There is the opening of wrist gear between a pair of mounting plate.There is connecting hole in the semi-circumference of the wrist outer ring,
And it is located at the center line of the connecting hole on the plane of symmetry of the semi-circumference.In the pair of mounting plate, have on a mounting plate
The mounting hole of steering engine bracket and the mounting hole of steering engine axis have the pin hole for installing the hand pull rod on another mounting plate.
The internal diameter of the wrist gear is identical as the wrist outside diameter of inner ring, and makes to be rotatably assorted therebetween.?
The circumferential surface of the wrist gear half is the flank of tooth engaged with stepping motor gear.In the circumferential surface of the wrist gear
On be machined with location hole corresponding with the connecting hole on the wrist outer ring, the center of the location hole and the circumferential direction of the flank of tooth are right
Angle between face is referred to as 180 °.
The large arm outer panel is lath-shaped.There is bearing hole respectively at the both ends of the large arm outer panel, is respectively used to hinged
Forearm outer panel and straight-line electric rack.The mounting hole of large arm support plate is machined on the large arm outer panel.
There is the bearing hole for hinged forearm interior plate in one end of the large arm interior plate.It is processed on the large arm interior plate
There is the mounting hole of large arm support plate.
There is the through-hole of connection stepper motor frame and wrist inner ring in one end of the forearm outer panel.In the forearm outer panel
There is the mounting hole of damper connecting shaft at middle part, is rotatably assorted between the mounting hole and the damper connecting shaft.On the outside of the forearm
The other end of plate has the axis hole connecting with the large arm outer panel.The installation of forearm support plate is machined on the forearm outer panel
Hole.
One end of the forearm interior plate has and the even through-hole that connect of wrist inner ring.The other end of the forearm interior plate has
The axis hole being connect with the large arm interior plate.The mounting hole of forearm support plate is machined on the forearm interior plate.
The hand pull rod is made of two connecting plates and a pull rod.There is sliding slot on described two connecting plates, it is described
The pin shaft at pull rod both ends is respectively charged into the sliding slot on described two connecting plates, and can be slided along the sliding slot.Described in one
There is the axis hole with steering engine shaft key groove cooperation in connecting plate inner end end;Outer surface in another connecting plate inner end end point
There is pin shaft, which is packed into the pin hole on one mounting plate of the wrist outer ring.
One surface of the steering engine link block is the plane with the mounting plate surface engagement of the wrist outer ring, another
There is the step-like convex block chimeric with the steering engine bracket at the half of one long side on surface.Have on the steering engine link block
The installation strut of two steering engine brackets, and the installation strut is made to be located at another long side side of the steering engine link block
Side surface both ends;There is threaded hole respectively in the end face of each installation strut.It is described to state the company of being distributed on steering engine link block
Connect the mounting hole of wrist outer ring.
The detailed process proposed by the present invention for carrying out refreshment white silk using the upper limb ectoskeleton mechanical arm is:
Step 1: the original motion trajectory point of each joint kinematic parameter of rehabilitation mechanical arm is generated
Set rehabilitation exercise motion and recovery period.Medical staff's demonstration rehabilitation exercise motion.Rehabilitation mechanical arm is remembered in real time
Record and save the original motion trajectory point in each joint when current presentation movement.
The kinematic parameter in each joint is angle position and the angular speed of motor.
Step 2: patient dresses ectoskeleton mechanical arm, and disposes EGC sensor in the finger of patient.It will by computer
The control model of the linear motor of ectoskeleton mechanical arm, stepper motor and steering engine is adjusted to mode position.
Step 3: the predetermined movement intensity of rehabilitation training is determined.
The exercise intensity is divided into high-intensitive, middle intensity and low-intensity, respectively corresponds the different phase of rehabilitation: low
Intensity movements correspond to rehabilitation initial stage patient, and patient's capacity of will in this stage is weaker;Middle intensity kinesis corresponds to convalescence trouble
Person, the patient in this stage has certain capacity of will, but has not yet fully recovered;High-intensity exercise corresponds to rehabilitation later period patient, this stage
Patient has been provided with capacity of will, but there is still a need for reinforce muscle recovery.
Step 4: the motion profile point after obtaining resampling.
Computer reads the original motion trajectory point, and is carried out by track of the Lan Suosi resampling methods to reading slotting
It mends, to obtain the point of the motion profile after resampling.Tracing point number=original tracing point of motion profile point after the resampling
Number × 2 forms resampling motion profile point.Save the resampling motion profile point.
It is described that the motion profile point detailed process after resampling is obtained by Lan Suosi resampling methods are as follows:
The first step determines that the corresponding Lan Suosi window function weight L (x) of each tracing point is in corresponding original motion trajectory point
Wherein, sinc is sinc function, and x is the time value of original motion trajectory point, and b is algorithm hyper parameter, takes 3.
Second step, Lan Suosi resampling.
Lan Suosi resampling is carried out to original motion trajectory point.Increase new track among each original motion trajectory point
Point.Angle position S (the x of newly-increased tracing point*)
Wherein, j is summation algorithm parameter, x*For the time value for increasing motion profile point newly, sjFor in the initial trace of j time
The numerical value of the angle position of point.
Step 5: the angle position of the resampling motion profile point is exported to linear motor, stepper motor and steering engine:
Computer reads the motion profile of resampling, exports resampling campaign rail in real time according to current exercise intensity selection
Mark point, and the angle position of the resampling motion profile point is exported to linear motor, stepper motor and steering engine respectively.
In selection in real time output resampling motion profile point:
When exercise intensity is high, the third resampling motion profile after previous resampling motion profile point is selected
Point, even if being spaced two resampling motion profile points between selected two adjacent resampling motion profile points.
When exercise intensity is middle, second resampling motion profile after previous resampling motion profile point is selected
Point, even if being spaced a resampling motion profile point between selected two adjacent resampling motion profile points.
When exercise intensity is low, the next resampling campaign rail adjacent with previous resampling motion profile point is selected
Mark point.
Computer output position controls signal, and receives signal feedback and each encoder from EGC sensor
Feedback angle position signal and angular velocity signal;Computer carries out data prediction to received above-mentioned each signal.Electrocardio passes
Sensor is that analog quantity turns digital quantity, sample rate 1k.When the sampling number N of the EGC sensor reaches 3000, will handle
At data import exercise intensity identification model.
The data prediction is characteristic value, the time domain approximate entropy that temporal criterion difference is extracted to collected electrocardiosignal
The characteristic value of characteristic value, the characteristic value of root mean square frequency domain and frequency domain criteria difference;From collected linear motor encoder feedback angle
It is opposite that track is extracted respectively in the encoder feedback angle signal of degree signal, stepper motor encoder feedback angle signal and steering engine
The characteristic value of speed difference.
Each characteristic value extracted is constituted into multimode vector, after being standardized, imports exercise intensity perception
Model obtains current exercise intensity.It is that each characteristic value is in parallel that each characteristic value, which constitutes multimode vector,;When in parallel, make the heart
Each characteristic value extracted in electric signal is located at first four.
The temporal criterion difference calculation method SD are as follows:
Wherein xiFor the signal value of corresponding electrocardiosignal time series, N is the sampled point number of EGC sensor, and μ is the heart
The average value of electric transducer signal.
The time domain approximate entropy are as follows:
For the one-dimensional electrocardio discrete signal a (1) that constant duration sampling obtains, a (2) ..., a (N) are reconstructed into three
Dimensional vector A (1), A (2) ..., A (N-m+1), i.e. A (i)=[a (i), a (i+1) ..., a (i+m-1)], statistics is as 1≤i≤N-m
When+1, meet the reconstruct vector number of the following conditions:
Wherein d [A (i), A (k)] is vector distance, is defined as each dimension absolute difference maximum one in two reconstruct vectors
Item numerical value, k are the sequential parameters of algorithm, and value range is [1, N-m+1], including k=i.The current reconstruct dimension quantity of state of definitionIt is obtained by the difference of the quantity of state and current dimension quantity of state of high reconstruct dimension
To approximate entropy (ApEn): ApEn=φm(r)-φm+1(r)
The m is algorithm hyper parameter.M=3.
The calculation method of the frequency domain character value are as follows:
Wherein P is the frequency domain amplitude of the spectrogram obtained by collected electrocardiosignal by Fast Fourier Transform (FFT), and f is
The frequency of spectrogram, MSF are square frequency, and FC is gravity frequency, and VF is frequency variance, and RMSF is root mean square frequency, and RVF is frequency
Rate standard deviation.
The linear motor angle position, stepper motor angle position and steering engine angle position are with the signaling point of resampling
On the basis of, according to the angle signal of the linear motor, stepper motor and steering engine encoder feedback calculate separately the linear motor,
The angular speed of stepper motor and steering engine.
The speed difference be resampling motion profile point angular speed respectively with linear motor angular speed, stepper motor angle
The difference of speed and steering engine angular speed.
The multimode vector data is standardized as S-core method.
The exercise intensity sensor model is made of deep neural network.
The exercise intensity sensor model is optimized using Adams gradient method;Optimization object is cross entropy, optimization
Data be multiple and different experiment testers actuator angle signals and electrocardiosignal for being acquired in multiple rehabilitation training.
Step 6: computer chooses next motion profile point according to the exercise intensity that exercise intensity sensor model obtains,
And signal is controlled to linear motor, stepper motor and steering engine output position respectively, acquire electrocardio feedback signal and encoder feedback
Signal imports exercise intensity sensor model after pretreatment.
Repeat to choose next motion profile point described in this step -- output position control signal -- acquisition electrocardio feedback letter
Number and encoder feedback signal -- import exercise intensity sensor model process, until complete setting the rehabilitation exercise period.
Compared with prior art, the beneficial effect that the present invention obtains is:
The present invention passes through teaching mould by medical staff's guidance and the recovery training method of adaptive limbs of patient, this method
Formula records rehabilitation training motion profile, then carries out exercise intensity identification to patient by deep learning frame, and by recognition result
Be embedded into based on deeply study integral rehabilitation Training Control strategy in, export adjusting after joint motions rate signal to
Mechanical arm completes rehabilitation training.This method can be believed under the premise of preventing Rehabilitation training secondary injury by movement
Vector Fusion number with bio signal, can eliminate the hysteresis quality of motor message and the individual difference bring of bioelectrical signals
It influences, improves exercise intensity recognition accuracy, and widen the applicable surface of rehabilitation exercise track, mitigate medical staff to obtain
Burden and the advantage in optimization Rehabilitation period, improve the safety and applicability of rehabilitation training.
A kind of control program based on the present invention provides the frame by deep learning assists medical staff to complete the complete of patient
The whole rehabilitation training course for the treatment of.Its characteristic is to carry out finishing man-machine interaction process using deeply learning algorithm, wherein human-computer interaction
Realization mainly by sensor detect patient motor message and electromyography signal and feed back to control system realize and control system
System driving rehabilitation mechanical arm drives wearer's limbs to carry out rehabilitation exercise.In some embodiments, collected limbs of patient
Motor message and collected limbs of patient bioelectrical signals are merged by characteristic vector pickup and multi-C vector, then use depth
Neural fusion limb motion intensity perception, obtains exercise intensity.Thereafter, the human-computer interaction signal Yu mechanical arm of feedback upload
The motor movement signal of sensor feedback imports control system, and control system determines the motor fortune of future time point according to exercise intensity
Dynamic variable, and joint motor driver is passed to, manipulator motion is driven, human body limb is driven to complete people finally by bandage
Machine interactive process.
The present invention is able to record joint motor motor message, realizes and acquires rehabilitation exercise track, and and controller combination, storage
Deposit a variety of rehabilitation tracks.Different rehabilitation exercises is chosen in practical rehabilitation training link to cooperate different patients and different treatments
Journey, wearable motion profile guarantee the accuracy of rehabilitation exercise close to true rehabilitation exercise track.The acquisition of this recovery training method
Limbs of patient kinematics signal and bioelectrical signals during rehabilitation exercise, it is available compared with sports-like using kinematics signal
Intensity recognition accuracy, while the characteristics of have both the shorter recognition time of acquisition biological signals, people when prominent patient motion
Machine interactive process, it is ensured that the rehabilitation exercise of patient is safe and man-machine integration.Finally, the electromyography signal recorded in rehabilitation exercise
It is also convenient for observation muscle to have an effect state, facilitates medical staff to the subsequent analysis of Rehabilitation Outcome.As shown in fig. 7, passing through
Acquire bioelectrical signals and kinematics signal of the patient in rehabilitation exercise experiment, the fusion vector of composition, using t-SNE algorithm
Realize dimension reduction and visualization, the point separation of corresponding different tag intensities is relatively opened, and the point of same label is at a distance of relatively close, it can be seen that this
Invention is preferable to human motion intensity recognition effect, can obtain 95% or more discrimination, acquires data and recognition time is low
In 0.2s.
Figure of description
Fig. 1 is hardware platform structural schematic diagram of the invention.
Fig. 2 is that the wrist joint of Fig. 1 splits structure chart.
Fig. 3 is the wrist joint twisting freedom degree structure chart of Fig. 1.
Fig. 4 is the structural schematic diagram of damper connecting shaft.
Fig. 5 is the structural schematic diagram of wrist outer ring.
Fig. 6 is the structural schematic diagram of locating snap ring.
Fig. 7 is the structural schematic diagram of wrist inner ring.
Fig. 8 is the structural schematic diagram of wrist gear.
Fig. 9 is the structural schematic diagram of steering engine link block.
Figure 10 is the structural schematic diagram of steering engine bracket.
Figure 11 is the structural schematic diagram of hand pull rod.
Figure 12 is carpal top view.
Figure 13 is the sectional view of Figure 12.
Figure 14 is elbow joint partial schematic diagram.
Figure 15 is exercise intensity identification model operation schematic diagram
Figure 16 is to feature vector t-SNE Two dimensional Distribution schematic diagram
In figure: 1. pitman shafts;2. linear motor;3. straight-line electric rack;4. connecting plate;5. large arm outer panel;6. damper
Connecting shaft;7. stepper motor;8. stepper motor frame;9. stepper motor axle sleeve;10. stepping motor gear;11. large arm interior plate;
12. large arm support plate;13. forearm interior plate;14. forearm support plate;15. forearm outer panel;16. wrist outer ring;17. wrist stops
Rotating ring;18. on the inside of wrist;19. wrist gear;20. steering engine link block;21. steering engine bracket;22. steering engine;23. hand pull rod.
Specific embodiment
The present embodiment is a kind of wearable upper limb exoskeleton rehabilitation mechanical arm of Three Degree Of Freedom, including pitman shaft 1, straight line
Motor 2, straight-line electric rack 3, connecting plate 4, large arm outer panel 5, damper connecting shaft 6, stepper motor 7, stepper motor frame 8, step
Into motor shaft sleeve 9, stepping motor gear 10, large arm interior plate 11, large arm support plate 12, forearm interior plate 13, forearm support plate
14, forearm outer panel 15, wrist outer ring 16, wrist locating snap ring 17, wrist inner ring 18, wrist gear 19, steering engine link block 20, rudder
Machine support 21, steering engine 22 and hand pull rod 23.There are two the large arm support plate 12 and forearm support plate 11 is each.Wherein:
The wrist locating snap ring 17 is inlaid in the card slot of 18 external peripheral surface upper end of wrist inner ring.The wrist gear 19
It is sleeved on the external peripheral surface of the wrist inner ring.The cantilever end for being fixed on the forearm interior plate 13 of the wrist inner ring side is hinged with
Large arm interior plate 11;One end of forearm outer panel 15 is fixed on the wrist inner ring other side, and stepper motor 7 passes through stepper motor frame
8 are fixed on the outer surface of the forearm outer panel 15;The cantilever end of the forearm outer panel and one end of large arm outer panel 5 are hinged.
Stepping motor gear 10 is mounted on the output shaft of stepper motor by retarder, and is engaged with the wrist gear 19.Damping
Device connecting shaft 6 is mounted on the forearm outer panel 15, and close to the cantilever end of the forearm outer panel.One end of pitman shaft 1 fills
Enter in the damper connecting shaft, the other end and linear motor 2 are connected;The linear motor 2 is mounted on the large arm outer panel 5
On the outer surface of the other end.Wrist outer ring 16 is sleeved on the external peripheral surface of the wrist gear 19, and makes tight therebetween
Close fit.The upper surface of the wrist gear is bonded with the lower end surface of the wrist locating snap ring 17, the lower end of the wrist gear
Face is concordant with the end face of wrist inner ring 18.The steering engine 22 is mounted on outside the wrist by steering engine bracket 21 and steering engine link block 20
On the mounting plate of ring set, and the steering engine axis of the steering engine is made to be rotatably assorted with the axis hole on the mounting plate.Hand pull rod 23 is hinged
On the mounting plate of the wrist outer ring 16.
The wrist inner ring 18 is annular shape, and internal diameter is 120~150mm.The external peripheral surface of the wrist inner ring one end
There is the groove of annular, for being embedded in wrist locating snap ring 17.There is the annular snap-gauge of radially protruding in the wrist inner ring the other end face,
For placing the wrist gear 19, after the wrist gear 19 is sleeved in the wrist inner ring, the upper surface of the annular slab
It is bonded with the lower end surface of the wrist gear 19;The outer end face of the annular snap-gauge is evenly equipped with axial mounting plate, for installing
The steering engine bracket 21 and hand pull rod 23.
The wrist outer ring 16 is annular shape.The internal diameter of the wrist outer ring is identical as the outer diameter of the wrist inner ring 18, and
Make to be rotatably assorted therebetween.The outer diameter of the wrist outer ring is identical as the internal diameter of the wrist gear 19.Outside the wrist
One end face of ring is evenly equipped with a pair of axially projecting mounting plate.By the wrist outer ring excision between a pair of mounting plate, formed
The opening of wrist gear.There is connecting hole in the semi-circumference of the wrist outer ring, and the center line of the connecting hole is made to be located at described half
On the plane of symmetry of circumference.In the pair of mounting plate, there are the mounting hole of steering engine bracket 21 and the peace of steering engine axis on a mounting plate
Hole is filled, has the pin hole for installing the hand pull rod on another mounting plate.
The wrist locating snap ring 17 is split ring.The outer diameter of the wrist locating snap ring is bigger than the outer diameter of the wrist inner ring 18
2mm。
The wrist gear 19 is annular shape.The internal diameter of the wrist gear is identical as 18 outer diameter of wrist inner ring, and makes
Therebetween it can be rotatably assorted.It is the tooth engaged with stepping motor gear 10 in the circumferential surface of the wrist gear half
Face.Location hole corresponding with the connecting hole on the wrist outer ring, the positioning are machined on the circumferential surface of the wrist gear
Angle between the center in hole and the circumferential plane of symmetry of the flank of tooth is 180 °.
The large arm outer panel 5 is lath-shaped.There is bearing hole respectively at the both ends of the large arm outer panel, is respectively used to hinged
Forearm outer panel 15 and straight-line electric rack 3.The mounting hole of large arm support plate 12 is machined on the large arm outer panel.
11 lath-shaped of large arm interior plate.There is the axis for hinged forearm interior plate 13 in one end of the large arm interior plate
Bearing bore.The mounting hole of large arm support plate 12 is machined on the large arm interior plate.
There is the through-hole of connection stepper motor frame and wrist inner ring in one end of the forearm outer panel 15.In the forearm outer panel
Middle part have the mounting hole of damper connecting shaft 6, be rotatably assorted between the mounting hole and the damper connecting shaft.Outside the forearm
The other end of side plate has the axis hole connecting with the large arm outer panel.Forearm support plate 14 is machined on the forearm outer panel
Mounting hole.
One end of the forearm interior plate 13 has and the even through-hole that connect of wrist inner ring.The other end of the forearm interior plate
There is the axis hole connecting with the large arm interior plate.The mounting hole of forearm support plate 14 is machined on the forearm interior plate.
The shape of the hand pull rod 23 is H-shaped, is made of two connecting plates and a pull rod.On described two connecting plates
There is sliding slot, the pin shaft at the pull rod both ends is respectively charged into the sliding slot on described two connecting plates, and can be sliding along the sliding slot
It is dynamic.There is the axis hole with steering engine shaft key groove cooperation in connecting plate inner end end;At another connecting plate inner end end
Outer surface at head has pin shaft, which is packed into the pin hole on 16 1 mounting plates of the wrist outer ring.
The steering engine link block 20 is rectangular plate-like, and a surface of the steering engine link block is the peace with the wrist outer ring
The plane of loading board surface engagement has the platform chimeric with the steering engine bracket 21 at the half of one long side on another surface
Scalariform convex block.There are two the installation struts of the steering engine bracket on the steering engine link block, and are located at the installation strut
The both ends of the side surface of another long side side of steering engine link block;There is screw thread respectively in the end face of each installation strut
Hole.It is described to state the mounting hole that connection wrist outer ring is distributed on steering engine link block 20.
The steering engine bracket 21 is rectangle deckle board.Have and the steering engine link block on a surface of the steering engine bracket one end
20 chimeric convex blocks, the convex block are step-like.
The hand pull rod 23 is made of two pieces of connecting plates and a pull rod, and shape is H-shaped.On two pieces of connecting plates
There is sliding slot, the pin shaft at the pull rod both ends is respectively placed in each sliding slot, and can be slided along the sliding slot.In two pieces of connecting plates,
There is the axis hole being fixedly connected with the steering engine axis in the inner end end of one piece of connecting plate, and the inner end end of another piece of mounting plate has and institute
State the pin shaft of the pin hole cooperation on one mounting plate of wrist outer ring.
The pitman shaft 1 is the output shaft of linear motor 2, and the damping equipped with deep groove ball bearing is protruded into 1 one end of coupling rod
In the through-hole of device connecting shaft 6, constitutes damper connecting shaft 6 and forearm outer panel 15 by deep groove ball bearing GB276-94 and rotate
It is secondary.
The linear motor 2 is connected in straight-line electric rack 3, and straight-line electric rack 3 and large arm outer panel 5 are solid by screw thread
Even.Large arm outer panel 5 is hinged by deep groove ball bearing GB276-94 with forearm outer panel 15, on the inside of large arm interior plate 11 and forearm
Plate 13 is hinged.Two panels large arm support plate 12 and large arm outer panel 5 and large arm interior plate 11 are connected by screw thread, meanwhile, by two panels
Forearm support plate 14 and forearm interior plate 13 and forearm outer panel 15 are connected by screw thread.There are two freely for mechanical carpal joint
Degree, the swing and twisting of corresponding wrist.7 output shaft of stepper motor is connected by interference fit and stepper motor axle sleeve 9,
Stepping motor gear 10 is engaged with wrist gear 19.
Specific recovery training method realizes that steps are as follows:
Step 1: according to conditions of patients, doctor sets rehabilitation exercise motion and recovery period.Medical staff dresses rehabilitation machines
Tool arm, rehabilitation exercise motion of demonstrating.The mode for adjusting the linear motor 2, stepper motor 7 and steering engine 22 simultaneously is current-mode
Formula.Rehabilitation mechanical arm records the kinematic parameter in each joint when current presentation movement in real time, and generates original motion trajectory point;According to
The original motion trajectory point is saved according to time sequencing.
The kinematic parameter in each joint is angle position and the angular speed of motor.
Step 2: patient dresses ectoskeleton mechanical arm.
Patient dresses ectoskeleton mechanical arm, and disposes EGC sensor in the finger of patient.By computer by ectoskeleton
The control model of the linear motor of mechanical arm, stepper motor and steering engine is adjusted to mode position.
EGC sensor selects finger-clipped Heat Rate Clamp, can choose folder and is worn on any one finger.
Step 3: rehabilitation exercise motion and recovery period are set according to doctor, determine the predetermined movement intensity of rehabilitation training.
The exercise intensity is divided into high-intensitive, middle intensity and low-intensity, respectively corresponds the different phase of rehabilitation: low-intensity movement pair
Rehabilitation initial stage patient is answered, patient's capacity of will in this stage is weaker;Middle intensity kinesis corresponds to reconvalescent, this stage
Patient has certain capacity of will, but has not yet fully recovered;High-intensity exercise corresponds to rehabilitation later period patient, and the patient in this stage has been provided with
Capacity of will, but there is still a need for reinforce muscle recovery.
Step 4: computer reads the original motion trajectory point, and by Lan Suosi resampling methods to the rail of reading
Mark carries out interpolation, to obtain the point of the motion profile after resampling.The tracing point number of motion profile point after the resampling=
Former track points × 2, form resampling motion profile point.Save the resampling motion profile point.The Lan Suosi resampling is calculated
Method detailed process are as follows:
The first step determines that the corresponding Lan Suosi window function weight L (x) of each tracing point is in corresponding original motion trajectory point
Wherein, sinc is sinc function, and x is the time value of original motion trajectory point, and b is algorithm hyper parameter, takes 3.
Second step, Lan Suosi resampling.
Lan Suosi resampling is carried out to original motion trajectory point.Increase new track among each original motion trajectory point
Point.Angle position S (the x of newly-increased tracing point*)
Wherein, j is summation algorithm parameter, x*For the time value for increasing motion profile point newly, sjFor in the initial trace of j time
The numerical value of the angle position of point.
Step 5: computer reads the motion profile of resampling, and according to current exercise intensity selection, output is adopted again in real time
Sample motion profile point, and the angle position of the resampling motion profile point is exported to linear motor, stepper motor and steering engine respectively.
In selection in real time output resampling motion profile point:
When exercise intensity is high, the third resampling motion profile after previous resampling motion profile point is selected
Point, even if being spaced two resampling motion profile points between selected two adjacent resampling motion profile points.
When exercise intensity is middle, second resampling motion profile after previous resampling motion profile point is selected
Point, even if being spaced a resampling motion profile point between selected two adjacent resampling motion profile points.
When exercise intensity is low, the next resampling campaign rail adjacent with previous resampling motion profile point is selected
Mark point.
Computer output position controls signal, and receives signal feedback and each encoder from EGC sensor
Feedback angle position signal and angular velocity signal;Computer carries out data prediction to received above-mentioned each signal.Electrocardio passes
Sensor is that analog quantity turns digital quantity, sample rate 1k.When the sampling number N of the EGC sensor reaches 3000, will handle
At data import exercise intensity identification model.
The data prediction is to be mentioned using general Visual Studio2017 software to collected electrocardiosignal
Take the feature of the characteristic value of temporal criterion difference, the characteristic value of time domain approximate entropy, the characteristic value of root mean square frequency domain and frequency domain criteria difference
Value;From the volume of collected linear motor encoder feedback angle signal, stepper motor encoder feedback angle signal and steering engine
The characteristic value of track relative speed difference is extracted in code device feedback angle signal respectively.
Each characteristic value extracted is constituted into multimode vector, after being standardized, imports exercise intensity perception
Model obtains current exercise intensity.It is that each characteristic value is in parallel that each characteristic value, which constitutes multimode vector,;When in parallel, make the heart
Each characteristic value extracted in electric signal is located at first four.As shown in figure 15.
The temporal criterion difference calculation method SD are as follows:
Wherein xiFor the signal value of corresponding electrocardiosignal time series, N is the sampled point number of EGC sensor, and μ is the heart
The average value of electric transducer signal.
The time domain approximate entropy are as follows:
For the one-dimensional electrocardio discrete signal a (1) that constant duration sampling obtains, a (2) ..., a (N) are reconstructed into three
Dimensional vector A (1), A (2) ..., A (N-m+1), i.e. A (i)=[a (i), a (i+1) ..., a (i+m-1)], statistics is as 1≤i≤N-m
When+1, meet the reconstruct vector number of the following conditions:
Wherein d [A (i), A (k)] is vector distance, is defined as each dimension absolute difference maximum one in two reconstruct vectors
Item numerical value, k are the sequential parameters of algorithm, and value range is [1, N-m+1], including k=i.The current reconstruct dimension quantity of state of definitionIt is obtained by the difference of the quantity of state and current dimension quantity of state of high reconstruct dimension
To approximate entropy (ApEn): ApEn=φm(r)-φm+1(r)
The m is algorithm hyper parameter.M=3.
The calculation method of the frequency domain character value are as follows:
Wherein P is the frequency domain amplitude of the spectrogram obtained by collected electrocardiosignal by Fast Fourier Transform (FFT), and f is
The frequency of spectrogram, MSF are square frequency, and FC is gravity frequency, and VF is frequency variance, and RMSF is root mean square frequency, and RVF is frequency
Rate standard deviation.
The linear motor angle position, stepper motor angle position and steering engine angle position are with the signaling point of resampling
On the basis of, according to the angle signal of the linear motor, stepper motor and steering engine encoder feedback calculate separately the linear motor,
The angular speed of stepper motor and steering engine.
The speed difference be resampling motion profile point angular speed respectively with linear motor angular speed, stepper motor angle
The difference of speed and steering engine angular speed.
The multimode vector data is standardized as S-core method.
The exercise intensity sensor model is made of deep neural network.The present embodiment constructs three-layer neural network as sense
Perception model: two layers close to signal input layer is first two layers, and preceding two layers of activation primitive selects relu;The last layer activation primitive makes
Use softmax.
The exercise intensity sensor model is optimized using Adams gradient method;Optimization object is cross entropy, optimization
Data be multiple and different experiment testers actuator angle signals and electrocardiosignal for being acquired in multiple rehabilitation training.This reality
It applies example and chooses each 300 groups multimode vector training datas with tag along sort under every kind of exercise intensity, perceived for exercise intensity
Model optimization.Separately take the lower 300 groups of multimode vectors of every kind of exercise intensity as model measurement data.Exercise intensity perception after optimization
Model reaches 99.0% in the identification of multimode vector training data, and the identification under test data can reach 95.3%.Structure
At the visual image that is obtained in T-SNE method dimensionality reduction of multimode vector.As shown in figure 16.
Step 6: computer chooses next motion profile point according to the exercise intensity that exercise intensity sensor model obtains,
And signal is controlled to linear motor, stepper motor and steering engine output position respectively, acquire electrocardio feedback signal and encoder feedback
Signal imports exercise intensity sensor model after pretreatment.
Repeat to choose next motion profile point described in this step -- output position control signal -- acquisition electrocardio feedback letter
Number and encoder feedback signal -- import exercise intensity sensor model process, until complete setting the rehabilitation exercise period.
Claims (10)
1. a kind of upper limb ectoskeleton mechanical arm, which is characterized in that on the outside of pitman shaft, linear motor, straight-line electric rack, large arm
Plate, damper connecting shaft, stepper motor, stepper motor frame, stepping motor gear, large arm interior plate, forearm interior plate, outside forearm
Side plate, wrist outer ring, wrist locating snap ring, wrist inside, wrist gear, steering engine link block, steering engine bracket, steering engine and hand are drawn
Bar;Wherein:
The wrist locating snap ring is inlaid in the card slot of wrist inner ring external peripheral surface upper end;The wrist gear set is in the wrist
The external peripheral surface of portion's inner ring;The cantilever end for being fixed on the forearm interior plate of the wrist inner ring side is hinged with large arm interior plate;
One end of forearm outer panel is fixed on the wrist inner ring other side, and stepper motor is fixed on the outside of the forearm by stepper motor frame
The outer surface of plate;The cantilever end of the forearm outer panel and one end of large arm outer panel are hinged;Stepping motor gear is mounted on step
It is engaged on the output shaft of motor, and with the wrist gear;Damper connecting shaft is mounted on the forearm outer panel, and is leaned on
The cantilever end of the nearly forearm outer panel;One end of pitman shaft is packed into the damper connecting shaft, and the other end and linear motor are solid
Even;The linear motor is mounted on the outer surface of the large arm outer panel other end;Wrist exterior ring cap is mounted in the wrist
The external peripheral surface of gear, and make to be fitted close therebetween;The upper surface of the wrist gear and the wrist locating snap ring
Lower end surface fitting, the lower end surface of the wrist gear is concordant with the end face of wrist inner ring;The steering engine passes through steering engine bracket and rudder
Machine link block is mounted on the mounting plate of the wrist exterior ring cap, and turns the steering engine axis of the steering engine and the axis hole on the mounting plate
Dynamic cooperation;Hand pull rod is hinged on the mounting plate of the wrist outer ring;Straight-line electric rack and large arm outer panel are connected;Outside large arm
Side plate and forearm outer panel are hinged, and large arm interior plate and forearm interior plate are hinged;Two panels large arm support plate and large arm outer panel and
Large arm interior plate is connected, while two panels forearm support plate and forearm interior plate and forearm outer panel being connected;Mechanical carpal joint
There are two freedom degree, the swing and twisting of corresponding wrist;Stepping motor gear is engaged with wrist gear.
2. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that the internal diameter of the wrist inner ring be 120~
150mm;The external peripheral surface of the wrist inner ring one end has the groove of annular, for being embedded in wrist locating snap ring;In the wrist inner ring
The other end face has the annular snap-gauge of radially protruding, for placing the wrist gear, when the wrist gear set is in the wrist
After in inner ring, the upper surface of the annular slab is bonded with the lower end surface of the wrist gear;The outer end face of the annular snap-gauge is uniformly distributed
There is axial mounting plate, for installing the steering engine bracket and hand pull rod.
3. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that the internal diameter and the wrist of the wrist outer ring
The outer diameter of inner ring is identical, and makes to be rotatably assorted therebetween;The internal diameter of the outer diameter of the wrist outer ring and the wrist gear
It is identical;One end face of the wrist outer ring is evenly equipped with a pair of axially projecting mounting plate;There is wrist between a pair of mounting plate
The opening of gear;There is connecting hole in the semi-circumference of the wrist outer ring, and the center line of the connecting hole is made to be located at the semi-circumference
The plane of symmetry on;In the pair of mounting plate, there are the mounting hole of steering engine bracket and the mounting hole of steering engine axis on a mounting plate, separately
There is the pin hole for installing the hand pull rod on one mounting plate.
4. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that the internal diameter and the wrist of the wrist gear
Outside diameter of inner ring is identical, and makes to be rotatably assorted therebetween;It is and stepping in the circumferential surface of the wrist gear half
The flank of tooth of motor gear engagement;It is machined on the circumferential surface of the wrist gear corresponding with the connecting hole on the wrist outer ring
Location hole, the angle between the center of the location hole and the circumferential plane of symmetry of the flank of tooth is 180 °.
5. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that the large arm outer panel is lath-shaped;At this
There is bearing hole at the both ends of large arm outer panel respectively, are respectively used to hinged forearm outer panel and straight-line electric rack;On the outside of the large arm
The mounting hole of large arm support plate is machined on plate;
There is the bearing hole for hinged forearm interior plate in one end of the large arm interior plate;It is machined on the large arm interior plate big
The mounting hole of arm support plate;
There is the through-hole of connection stepper motor frame and wrist inner ring in one end of the forearm outer panel;At the middle part of the forearm outer panel
There is the mounting hole of damper connecting shaft, is rotatably assorted between the mounting hole and the damper connecting shaft;The forearm outer panel
The other end has the axis hole connecting with the large arm outer panel;The mounting hole of forearm support plate is machined on the forearm outer panel;
One end of the forearm interior plate has and the even through-hole that connect of wrist inner ring;The other end of the forearm interior plate has and institute
State the axis hole of large arm interior plate connection;The mounting hole of forearm support plate is machined on the forearm interior plate.
6. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that the hand pull rod is by two connecting plates and one
A pull rod composition;There is sliding slot on described two connecting plates, the pin shaft at the pull rod both ends is respectively charged into described two connecting plates
On sliding slot in, and can be slided along the sliding slot;Have in connecting plate inner end end and cooperates with the steering engine shaft key groove
Axis hole;Outer surface in another connecting plate inner end end point has pin shaft, which is packed into wrist outer ring one peace
In pin hole in loading board.
7. upper limb ectoskeleton mechanical arm as described in claim 1, which is characterized in that a surface of the steering engine link block be with
The plane of the mounting plate surface engagement of the wrist outer ring has and the rudder at the half of one long side on another surface
The chimeric step-like convex block of machine support;There are two the installation struts of the steering engine bracket on the steering engine link block, and make the peace
Dress strut is located at the both ends of the side surface of another long side side of steering engine link block;At the end of each installation strut
There is threaded hole in face respectively;It is described to state the mounting hole that connection wrist outer ring is distributed on steering engine link block.
8. a kind of carry out the method that refreshment is practiced using upper limb ectoskeleton mechanical arm described in claim 1, which is characterized in that specific step
Suddenly it is:
Step 1: the original motion trajectory point of each joint kinematic parameter of rehabilitation mechanical arm is generated: setting rehabilitation exercise motion and health
The multiple period;Medical staff's demonstration rehabilitation exercise motion;Rehabilitation mechanical arm records and saves each joint when current presentation movement in real time
Original motion trajectory point;
The kinematic parameter in each joint is angle position and the angular speed of motor;
Step 2: patient dresses ectoskeleton mechanical arm: patient dresses ectoskeleton mechanical arm and passes in the finger of patient placement electrocardio
Sensor;The control model of the linear motor of ectoskeleton mechanical arm, stepper motor and steering engine is adjusted to position mould by computer
Formula;
Step 3: determine the predetermined movement intensity of rehabilitation training: the exercise intensity is divided into high-intensitive, middle intensity and low strong
Degree, respectively correspond the different phase of rehabilitation: low-intensity moves corresponding rehabilitation initial stage patient, and the patient in this stage is autonomous
Ability is weaker;Middle intensity kinesis corresponds to reconvalescent, and the patient in this stage has certain capacity of will, but has not yet fully recovered;It is high-strength
Degree moves corresponding rehabilitation later period patient, and the patient in this stage has been provided with capacity of will, but there is still a need for reinforce muscle recovery;
Step 4: the motion profile point after obtaining resampling:
Computer reads the original motion trajectory point, and carries out interpolation by track of the Lan Suosi resampling methods to reading,
To obtain the point of the motion profile after resampling;The tracing point number of motion profile point after the resampling=original track points ×
2, form resampling motion profile point;Save the resampling motion profile point;
Step 5: the angle position of the resampling motion profile point is exported to linear motor, stepper motor and steering engine: calculating machine-readable
The motion profile for taking resampling exports resampling motion profile point according to current exercise intensity selection in real time, and respectively to straight
Line motor, stepper motor and steering engine export the angle position of the resampling motion profile point;
In selection in real time output resampling motion profile point:
When exercise intensity is high, the third resampling motion profile point after previous resampling motion profile point is selected,
Even if being spaced two resampling motion profile points between selected two adjacent resampling motion profile points;
When exercise intensity is middle, second resampling motion profile point after previous resampling motion profile point is selected,
Even if being spaced a resampling motion profile point between selected two adjacent resampling motion profile points;
When exercise intensity is low, the next resampling motion profile adjacent with previous resampling motion profile point is selected
Point;
Computer output position controls signal, and signal feedback and each encoder of the reception from EGC sensor is anti-
Present angle position signal and angular velocity signal;Computer carries out data prediction to received above-mentioned each signal;EGC sensor
Turn digital quantity, sample rate 1k for analog quantity;When the sampling number N of the EGC sensor reaches 3000, processing is completed
Data import exercise intensity identification model;
The data prediction is characteristic value, the feature of time domain approximate entropy that temporal criterion difference is extracted to collected electrocardiosignal
Value, the characteristic value of the characteristic value of root mean square frequency domain and frequency domain criteria difference;Believe from collected linear motor encoder feedback angle
Number, extract track relative velocity respectively in the encoder feedback angle signal of stepper motor encoder feedback angle signal and steering engine
The characteristic value of difference;
Each characteristic value extracted is constituted into multimode vector, after being standardized, imports exercise intensity sensor model,
Obtain current exercise intensity;It is that each characteristic value is in parallel that each characteristic value, which constitutes multimode vector,;When in parallel, make electrocardiosignal
Each characteristic value of middle extraction is located at first four;
The exercise intensity sensor model is optimized using Adams gradient method;Optimization object is cross entropy, the number of optimization
According to the actuator angle signal and electrocardiosignal acquired in multiple rehabilitation training for multiple and different experiment testers;
Step 6: computer chooses next motion profile point according to the exercise intensity that exercise intensity sensor model obtains, and divides
Signal is not controlled to linear motor, stepper motor and steering engine output position, acquires electrocardio feedback signal and encoder feedback signal,
Exercise intensity sensor model is imported after pretreatment;
Repeat to choose next motion profile point described in this step -- output position controls signal -- acquisition electrocardio feedback signal and
Encoder feedback signal -- the process of exercise intensity sensor model is imported, until completing the rehabilitation exercise period of setting.
9. upper limb ectoskeleton mechanical arm as claimed in claim 8 carries out the method that refreshment is practiced, which is characterized in that described to pass through blue rope
This resampling methods obtains the motion profile point detailed process after resampling are as follows:
The first step determines that the corresponding Lan Suosi window function weight L (x) of each tracing point is in corresponding original motion trajectory point
Wherein, sinc is sinc function, and x is the time value of original motion trajectory point, and b is algorithm hyper parameter, takes 3;
Second step, Lan Suosi resampling;
Lan Suosi resampling is carried out to original motion trajectory point;Increase new tracing point among each original motion trajectory point;
Angle position S (the x of newly-increased tracing point*)
Wherein, j is summation algorithm parameter, x*For the time value for increasing motion profile point newly, sjFor in the initial trace of j time point
The numerical value of angle position.
10. upper limb ectoskeleton mechanical arm as claimed in claim 8 carries out the method that refreshment is practiced, which is characterized in that
The temporal criterion difference calculation method SD are as follows:
Wherein xiFor the signal value of corresponding electrocardiosignal time series, N is the sampled point number of EGC sensor, and μ is electrocardio sensing
The average value of device signal;
The time domain approximate entropy are as follows:
For one-dimensional electrocardio discrete signal a (1), a (2) ..., a (N) that constant duration sampling obtains, be reconstructed into it is three-dimensional to
It measures A (1), A (2) ..., A (N-m+1), i.e. A (i)=[a (i), a (i+1) ..., a (i+m-1)], statistics is as 1≤i≤N-m+1
When, meet the reconstruct vector number of the following conditions:
Wherein d [A (i), A (k)] is vector distance, is defined as each maximum item number of dimension absolute difference in two reconstruct vectors
Value, k are the sequential parameters of algorithm, and value range is [1, N-m+1], including k=i;The current reconstruct dimension quantity of state of definitionIt is obtained by the difference of the quantity of state and current dimension quantity of state of high reconstruct dimension
To approximate entropy (ApEn): ApEn=φm(r)-φm+1(r)
The m is algorithm hyper parameter;M=3;
The calculation method of the frequency domain character value are as follows:
Wherein P is the frequency domain amplitude of the spectrogram obtained by collected electrocardiosignal by Fast Fourier Transform (FFT), and f is frequency spectrum
The frequency of figure, MSF are square frequency, and FC is gravity frequency, and VF is frequency variance, and RMSF is root mean square frequency, and RVF is frequency mark
It is quasi- poor;
The linear motor angle position, stepper motor angle position and steering engine angle position are using the signaling point of resampling as base
Standard calculates separately the linear motor, stepping according to the angle signal of the linear motor, stepper motor and steering engine encoder feedback
The angular speed of motor and steering engine;
The speed difference be resampling motion profile point angular speed respectively with linear motor angular speed, stepper motor angular speed
With the difference of steering engine angular speed;
The multimode vector data is standardized as S-core method;
The exercise intensity sensor model is made of deep neural network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910467923.0A CN110236876A (en) | 2019-05-31 | 2019-05-31 | A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910467923.0A CN110236876A (en) | 2019-05-31 | 2019-05-31 | A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110236876A true CN110236876A (en) | 2019-09-17 |
Family
ID=67885628
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910467923.0A Pending CN110236876A (en) | 2019-05-31 | 2019-05-31 | A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110236876A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111228743A (en) * | 2020-02-27 | 2020-06-05 | 利辛县儒康医药有限公司 | Joint movement auxiliary device |
CN111281741A (en) * | 2020-02-26 | 2020-06-16 | 南京邮电大学 | Reconfigurable exoskeleton upper limb rehabilitation robot for different body types |
CN111759662A (en) * | 2020-07-03 | 2020-10-13 | 浙江工业大学 | Arm auxiliary device |
CN111816309A (en) * | 2020-07-13 | 2020-10-23 | 国家康复辅具研究中心 | Rehabilitation training prescription self-adaptive recommendation method and system based on deep reinforcement learning |
CN112022619A (en) * | 2020-09-07 | 2020-12-04 | 西北工业大学 | Multi-mode information fusion sensing system of upper limb rehabilitation robot |
CN113878612A (en) * | 2021-09-23 | 2022-01-04 | 北京邮电大学 | Dual-mode driving joint for rehabilitation robot |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1364755B1 (en) * | 2002-05-22 | 2009-11-04 | Organisation Intergouvernementale Dite Agence Spatiale Europeenne | Exoskeleton for a human arm, especially for spatial applications |
CN205459684U (en) * | 2016-01-11 | 2016-08-17 | 昆明理工大学 | Rehabilitation training robotic arm |
CN205795993U (en) * | 2016-04-18 | 2016-12-14 | 常熟理工学院 | Healing robot |
CN108582155A (en) * | 2018-04-16 | 2018-09-28 | 西北工业大学 | It is a kind of that there is the mechanical arm for pressing down function of quivering |
-
2019
- 2019-05-31 CN CN201910467923.0A patent/CN110236876A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1364755B1 (en) * | 2002-05-22 | 2009-11-04 | Organisation Intergouvernementale Dite Agence Spatiale Europeenne | Exoskeleton for a human arm, especially for spatial applications |
CN205459684U (en) * | 2016-01-11 | 2016-08-17 | 昆明理工大学 | Rehabilitation training robotic arm |
CN205795993U (en) * | 2016-04-18 | 2016-12-14 | 常熟理工学院 | Healing robot |
CN108582155A (en) * | 2018-04-16 | 2018-09-28 | 西北工业大学 | It is a kind of that there is the mechanical arm for pressing down function of quivering |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111281741A (en) * | 2020-02-26 | 2020-06-16 | 南京邮电大学 | Reconfigurable exoskeleton upper limb rehabilitation robot for different body types |
CN111228743A (en) * | 2020-02-27 | 2020-06-05 | 利辛县儒康医药有限公司 | Joint movement auxiliary device |
CN111228743B (en) * | 2020-02-27 | 2021-08-13 | 鹤壁市人民医院 | Joint movement auxiliary device |
CN111759662A (en) * | 2020-07-03 | 2020-10-13 | 浙江工业大学 | Arm auxiliary device |
CN111816309A (en) * | 2020-07-13 | 2020-10-23 | 国家康复辅具研究中心 | Rehabilitation training prescription self-adaptive recommendation method and system based on deep reinforcement learning |
CN112022619A (en) * | 2020-09-07 | 2020-12-04 | 西北工业大学 | Multi-mode information fusion sensing system of upper limb rehabilitation robot |
CN113878612A (en) * | 2021-09-23 | 2022-01-04 | 北京邮电大学 | Dual-mode driving joint for rehabilitation robot |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110236876A (en) | A kind of upper limb ectoskeleton mechanical arm and the control method of rehabilitation training | |
CN111816309B (en) | Rehabilitation training prescription self-adaptive recommendation method and system based on deep reinforcement learning | |
Zannatha et al. | Development of a system based on 3D vision, interactive virtual environments, ergonometric signals and a humanoid for stroke rehabilitation | |
CN102058466B (en) | Chinese massage robot | |
JP5105508B2 (en) | Methods and instruments for rehabilitation and training | |
CN109091819A (en) | Upper limb rehabilitation robot control system | |
CN110742775A (en) | Upper limb active and passive rehabilitation training robot system based on force feedback technology | |
CN204207994U (en) | Two upper limb coordination exercise training and testing equipment | |
CN109288651A (en) | Personalized upper-limbs rehabilitation training robot system and its recovery training method | |
CN114366556B (en) | Multimode training control system and method for lower limb rehabilitation | |
CN104873356B (en) | A kind of device and method of the auxiliary upper limb rehabilitation based on true feeling technology | |
CN201033143Y (en) | Intelligent limb guiding healing and training therapeutic machine | |
CN110533763B (en) | Intelligent orthopedic external fixation system based on cloud platform | |
Sramka et al. | Combined methods of rehabilitation of patients after stroke: virtual reality and traditional approach | |
CN106974795A (en) | A kind of drive lacking upper limb rehabilitation robot control system | |
CN106621207A (en) | Training method for recovery robot and recovery robot | |
CN111067543A (en) | Man-machine interaction system of horizontal stepping type rehabilitation training robot | |
CN110393894A (en) | Preceding shoulder joint training adapter, rehabilitation training equipment and its Training Control method | |
CN102600031A (en) | Treatment and rehabilitation system | |
CN206566167U (en) | Healing robot | |
Hu et al. | A massage robot based on Chinese massage therapy | |
CN106691478A (en) | Sub-site hand function rehabilitation evaluation method and device | |
CN106020472A (en) | Brain computer interface system on basis of motor imageries of different uplifting amplitudes of lower limbs | |
Wang et al. | Interval estimation of motion intensity variation using the improved inception-V3 model | |
CN114669028A (en) | Interaction system and method for collaborative training robot based on motion rehabilitation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190917 |