CN102138860A - Intelligentized rehabilitation training equipment for hand functions of patients suffering from cerebral injury - Google Patents

Abstract

The invention relates to hand function rehabilitation training equipment based on a spontaneous movement imagery electroencephalograph. Hardware of the instrument consists of an electroencephalograph signal acquisition and processing module, a hand force signal real-time detection sensor array, a control module based on multisource information fusion and a rehabilitation manipulator module, and software for supporting the active rehabilitation training of hand functions is arranged in a computer. The rehabilitation training equipment provides external compensation assistance or resistance hierarchically by taking movement imagery electroencephalographs of patients who suffer from cerebral injury as a starting signal and the muscle force of hands as a feedback signal in a self-adaptive mode according to the myasthenia state of the patients to assist patients who suffer from dyskinesia in performing passive, assisted and active rehabilitation training, displays the state information of the hand movement functions and the provided external power information for doctors and the patients and provides an objective and quantitative assessment method and an index for clinical rehabilitation diagnosis and treatment. The instrument is easy to use, has a friendly interface and is accepted by the doctors and the patients easily. A user can operate a system of the instrument without programming experiences, and the operating environment is not complex.

Description

A kind of intelligent brain injury patients healing hand function exercise equipment

Technical field

The present invention relates to a kind of medical supplementary instrument, particularly a kind of intelligent brain injury patients healing hand function exercise equipment of real-time brain electric control, this instrument can be with the electric control of the motion of brain injury patients imagination brain, according to patient's myasthenia state, with adaptive mode, be classified to provide external compensation power-assisted or resistance, and in real time, dynamically shows the hands functional training device of muscular strength information Xiang doctor and patient.

Background technology

Development trend from modern medical service and rehabilitation instrument and equipment, wearable technology has become the hot issue that biomedical engineering field is paid close attention to, and the various physiologic informations of human body are significant for finding and treating new disease under " accurate naturalness ".The sensor technology that can accurately measure Human Physiology information and do not influence again people's normal activity is the basis of realizing wearable technology.The development of modern micro-electro-mechanical systems (MEMS) manufacturing process allows pick off realize that microminiaturized, precision becomes possibility.The combination of MEMS manufacturing process and biomedical engineering is the inexorable trend of wearable technical development.In recent years, become the key technology of human body surface pressure detecting based on the pliable pressure sensor array of MEMS technology, this technology is mainly by from external introduction at present.The MEMS pliable pressure sensor array that is used for medical measurement that research and development have independent intellectual property right will be developed novel medical instrument and rehabilitation equipment provides necessary base for China.

Function of nervous system's information engineering is a frontier that emerges rapidly, and it has a wide range of applications in CE and rehabilitation engineering.Current, brain-computer interface and application technology thereof have become the focus of international research, and dynamically, spontaneous brain electricity at line drawing, brain behavior consciousness information levy the difficult point that technology is studied especially in real time surely.Breakthrough on this research direction not only has important academic values, also will promote the development of China's digital medical instrument industry effectively.

Follow the principle of maincenter motor control, the motion imagination is that the motor function state is exported without any tangible motion in activation in working memory.Late 1980s, motion imagination technology began to be applied to gradually rehabilitation training to the beginning of the nineties.According to mirror image neuron theory, the mirror image neuron system not only is activated when action is carried out, and also can be activated when another person finishes same action or oneself this action of the imagination when a people observes.So the mirror image neuron system just can be activated when brain injury patients was moved by the imagination and observation, thereby play the damaged effect of motor function that recovers.Discovering in recent years also confirmed this theory, the motion imagination can be improved the motor function of brain injury patients, can be used as a kind of means that activate the motion network, this therapy does not rely on patient's remaining function simultaneously, for the active rehabilitation that realizes the patient provides new method.

Multidisciplinary mixing together provides new approach, thinking and method for Medical Instruments and the rehabilitation equipment that research and development have the independent intellectual property of China.National correlation department encourages multidisciplinary expert to cooperate, and carries out comprehensive crossing research, with by integrating the advantage of each subject, merges the new thought and the new technique of each subject, promotes scientific and technical innovation.

China has extremity disabled persons 2,412 ten thousand now, and wherein the nerve damage is the modal reason that disables.Apoplexy (about 1,000 ten thousand people, annual newly-increased 4,000,000 people), cerebral trauma (annual newly-increased 100～1,200,000 people), cerebral palsy (annual newly-increased 1,000,000 people), spinal cord injury (annual newly-increased 600,000 people), poliomyelitis sequela etc. all can cause paralysis in various degree.It mainly shows as limb movement disturbance, and Activities such as daily life, work, study, amusement are all had a strong impact on.The rehabilitation of hemiplegia of limb is an international headache, and the therapy that adopts is a lot of at present, but takes notice of passive rehabilitation owing to still stopping based on the rehabilitation instrument, lacks quantized assessment of function method, and overall therapeutic effect is not satisfactory.

Summary of the invention

The object of the present invention is to provide a kind ofly based on active rehabilitation theory, adopting motion imagination brain electricity is enabling signal, and the muscular strength signal is as the intelligent brain injury patients healing hand function exercise equipment of feedback means.

For achieving the above object, the technical solution used in the present invention is: comprise eeg signal acquisition processing module, hand force signal real-time detection apparatus and rehabilitation manipulator;

Described eeg signal acquisition processing module is extracted patient's active exercise idea, in real time, at the brain wave of body, online acquisition brain injury patients, extract the characteristic information of the motion imagination, control signal as rehabilitation training inputs among the slave computer DSP, with the motion of the control of the sports consciousness information in spontaneous brain electricity signal rehabilitation manipulator;

Described hand force signal real-time detection apparatus is at hand placement sensor array, detects the motion of hand and the information of power and inputs to slave computer DSP, realizes adaptive mode, is classified to provide the external compensation power-assisted for patient's hands functional training;

Described rehabilitation manipulator comprises that the upper end has the main body mechanical framework of chute, chute internal fixation at this main body mechanical framework has the wrist supporting plate, lateral symmetry is fixed with the installing plate that has gathering sill in the lower end of main frame mechanical framework, in the gathering sill of installing plate, crossbeam is installed, the two ends of crossbeam link to each other with spring respectively, the other end of spring is fixed in the chute of main body mechanical framework, and the outfan of the cylinder that links to each other with slave computer DSP links to each other with crossbeam.

Eeg signal acquisition processing module of the present invention comprises that electrode cap, the big collection plate of preposition brain tele-release and slave computer DSP signal processing part are grouped into, three road brain wave acquisition medicated caps extract the EEG signals in head C3, C4 and Cz zone, through the big collection plate of preposition brain tele-release with other EEG signals of uV level, amplify 10,000 times through the brain wave acquisition plate, send into slave computer DSP, the DSP sample frequency is 128Hz, in slave computer after the EEG signals denoising, do the active exercise idea that the classification of task consciousness obtains the people, this moment, the healing hand function exercise equipment obtained motion actuated signal.

Described brain wave acquisition plate adopts the mode of amplifying one by one, wherein prepositionly is enlarged into 25 times, and secondary is enlarged into 40 times, three grades are enlarged into 10 times, brain electricity after the amplification is delivered to slave computer DSP by photoelectric isolating circuit and is carried out the A/D conversion, and its sample rate is 128Hz, and sampling precision is 12.

Described hand force signal real-time detection apparatus adopts the glove-type structure, and the single shaft angular transducer of the Finland SCA61T-FA1H1G of the VTI company chip of five sheet type pressure transducers that detect the finger pressure signals and three movement positions that detect hands is installed in glove, the sheet type pressure transducer is installed in middle flexor digitorum superficialis (label 1), flexor pollicis brevis (label 2,3), adductor pollicis (label 5), flexor digiti minimi brevis (label 4), the single shaft angular transducer is installed in the position of middle, the crossbeam back side, and nine pick offs are crossed serial ports through the A/D sampling module all with signal and imported host computer into.

The finger of described hand force signal real-time detection apparatus glove is connected on the crossbeam shoulder hole of rehabilitation manipulator.

The cylinder of described rehabilitation manipulator adopts medical quiet air compressor machine as pneumatic supply, the air pressure of 0-0.8Mpa is provided, check valve, air accumulator, filter, air drying lubricator, electric Proportion valve and three position five-way valve are installed on the pipeline between cylinder and the pneumatic supply, and wherein electric Proportion valve links to each other with slave computer DSP.

Described slave computer DSP also links to each other with the host computer of peripheral hardware, after brain electricity that the slave computer transmission is next and sensor signal are handled, in real time, dynamically show quantitative muscular strength information, monitoring training result Xiang doctor and patient, the force level that mechanical hand is provided when setting up training on their own by the user makes it regulate the intensity size of outside power-assisted or resistance automatically under the control of slave computer DSP.

The present invention has following technical characterstic:

1, gathers the EEG signals extraction feedback treating information of leading that a plurality of scalp electrodes obtain more,, can place different positions, extract and treat the EEG signals of dependency maximum according to different needs owing to brain injury patients is prone to the brain domain displacement;

2, according to the advanced theory of " initiatively rehabilitation ", based on hand exercise consciousness characteristic information in the spontaneous brain electricity in line drawing and recognition technology, finish a kind of limb rehabilitation training instrument of controlling by the participation of patient's spontaneous brain electricity;

3, using miniature diaphragm pressure sensor array is gathered the hand force signal, assesses its limb muscle motor function state by patient's limbs mechanical signal; Propose according to nerve-muscle theory of conduction and the meticulous motion field of hands multi-joint according to experiment, carry out the optimal spatial layout and the design of integrated framework scheme of MEMS flexible sensor array, realize the making of the sensor array of high measurement accuracy, tool " interlayer " structure.

4, with adaptive mode, by grade outside power-assisted or resistance are provided, that assisted movement dysfunction patient carries out is passive, the rehabilitation training of power-assist and aggressive mode.This system can also in real time, dynamically show hand exercise functional status information and the power-assisted that is provided or resistance information Xiang doctor and patient, for the clinical rehabilitation diagnosis and treatment provide objective, quantized assessment method and index.

5, rehabilitation training effect and the external strength information that is provided are provided medical apparatus of the present invention in real time, not only provide foundation for the doctor formulates rehabilitation training plans, and provide driving source for patient's active idea participates in limb rehabilitation training; Introducing is based on the real-time vision feedback training of virtual environment technology: release the brain training and train the clinical intervention pattern that combines with limbs, design healing hand function training guidance interface display hands movement functional status information, for the clinical rehabilitation diagnosis and treatment provide the use of objective, quantized assessment method and index system simple, friendly interface is easy to doctor and patient and accepts simultaneously.System does not require that user has the programming experience, does not provide the complicated operations environment.

Description of drawings

Fig. 1 is an overall design technique route map of the present invention;

Fig. 2 is a hardware block diagram of the present invention;

Fig. 3 is the structural representation of rehabilitation manipulator of the present invention;

Fig. 4 is the quantitative linearity relation of pressure resistance type force transducer resistance and power;

Fig. 5 is the relation of pressure transducer pressure and output;

Fig. 6 is installation site and sensor voltage output relation;

Fig. 7 is the site scattergram of arranging of flexible sensor array;

The position in kind layout viewing of Fig. 8 flexible sensor array;

Fig. 9 is the pneumatic power loop of healing hand function exercise equipment;

Figure 10 is the eeg signal acquisition circuit structure;

Figure 11 is the systems soft ware flow chart of healing hand function exercise equipment;

Figure 12 is healing hand function exercise equipment power-assisted training mode (left side) and resistance exercise pattern subprogram flow graph (right side);

Figure 13 is a software log-in interface of the present invention;

Figure 14 is software training of the present invention interface;

Figure 15 is data base querying of the present invention interface;

Figure 16 is main interface of the present invention;

Figure 17 is game training of the present invention interface.

The present invention is described in further detail below in conjunction with accompanying drawing.

The specific embodiment

The rehabilitation manipulator

Under various amyasthenic states, the rehabilitation mechanical hand should nondominant hand carries out the action of grasping and stretching, extension hands dysfunction person.By comparing for various schemes, the present invention adopts air cylinder structure simple in structure, has designed the fixed frame of rehabilitation mechanical hand, realizes the needed function of this module by the flexible push-and-pull finger of cylinder.

The hand force signal detects MEMS multisensor array and arrangement mode and encapsulation in real time

The present invention introduces grade power-assisted compensation theory, at hand placement sensor array, with the motion of detection hand and the information of power, is the key point that realizes adaptive mode, is classified to provide for patient's hands functional training the external compensation power-assisted.The present invention proposes according to nerve-muscle theory of conduction and the meticulous motion field of hands multi-joint according to experiment, carry out the optimal spatial layout and the design of integrated framework scheme of MEMS flexible sensor array, realize the making of the sensor array of high measurement accuracy, tool " interlayer " structure.

Be furnished with two kinds of pick offs at hand, a kind of is diaphragm pressure flexible MEMS pick off and sensor array, this kind pick off wide-range, high sensitivity, high accuracy, response speed are fast, the advantage that not only has generic array formula pick off, also has good pliability, can free bend even folding, under the prerequisite that does not change the movable and force method of patient, detect the size of active force between hand and the rehabilitation mechanical hand in real time.Another is the single shaft angular transducer of the Finland SCA61T-FA1H1G of VTI company chip, can detect the movement position of hand in the degree of freedom direction of hands, for the evaluation of patient's nerve-muscle function state and classification power-assisted or resistance feedback control provide foundation.

Eeg signal acquisition, processing module

This module is the important core of " autonomous intelligence control ", also is the key link that realizes " initiatively rehabilitation " theory.Introduce the participation of patient's active exercise idea, in real time, at the brain of body, online acquisition brain injury patients, extract the characteristic information of the motion imagination, control signal as rehabilitation training, motion with mechanical hand in the control of the sports consciousness information in the spontaneous brain electricity signal rehabilitation system, not only help the functional rehabilitation of limbs nerve-musculature, and help the functional rehabilitation of brain injury patients cerebral nerve function and corresponding nerve conduction path thereof.

This module at first utilizes a plurality of scalp electrodes to be used for eeg signal acquisition, sample frequency is 128Hz, and with the EEG signals of gathering extract, amplification and Filtering Processing, after amplifying signal converts digital signal to by 12 A/D, carry out EEG signals pretreatment, feature extraction, with classification.

Because phenomenons such as compensatory displacement can take place in the brain injury patients functional areas, this device is used the preconditioning technique of common space pattern, realizes the selection for brain injury patients optimum electrode position, and the online noise reduction process of finishing various pseudo-difference signals, noise.By research brain-computer interface technology and signal processing technology, in the online BCI system, use the AR model to extract the spectrum signature of special frequency channel, directly do classification with linear classifier then, good classification effect and stable is fit to be applied to online BCI system and realization easily.So the present invention adopts the feature extracting method of AR model as motion imagination EEG signals, the Fisher discriminant is realized the ONLINE RECOGNITION to hand sports consciousness task, nicety of grading＞90% as sorting technique.

Control module based on Multi-source Information Fusion

Feedback control circuit carries out the Multi-source Information Fusion of brain electricity, force signal feature etc., finishes the chirokinesthetic ACTIVE CONTROL of rehabilitation machinery based on brain electricity, pressure, and regulates control based on the power-assisted of hand force signal or the self adaptation of resistance size.

Software of the present invention is by constituting with lower module;

People-machine control display interface not only can in real time, dynamically show quantitative muscular strength information, monitoring training result Xiang doctor and patient, but also can set up when training force level of being provided of mechanical hand on their own by the user, make it under the control of microprocessor, can regulate the intensity size of outside power-assisted or resistance automatically.

Administration module is used for training process and experimenter's information are managed, and comprises establishment, inquiry, modification, the deletion of trainer's information, and the selection of training program, and the storage of the desired index of feedback training scheme and form and data is set;

The data base is used to store training program, trainer's the information and the process storage of training thereof;

The feedback training module, mechanical hand is provided in the time of can setting up training on their own by the user in this module power-assisted or resistance grade make it can regulate the intensity size of power-assisted or resistance automatically under the control of microprocessor;

Display module is analyzed five parts by result's printing, the demonstration of brain electricity, brain electricity analytical, muscular strength demonstration, muscular strength and is formed, and is used for monitoring in real time and shows the rehabilitation of patients physical training condition.

The overall design technique route map of system of the present invention as shown in Figure 1, based on the real-time monitoring of spontaneous brain electricity and hand with force signal, is core with hand exercise consciousness characteristic information in the spontaneous brain electricity at line drawing and recognition technology, the detection of hand force signal and the adaptive control technology of analytical technology and fusion multi-source information, serves as to detect and execution module to comprise miniature thin-film pressure sensor unit and graded external power-assisted/resistance unit at interior mechanical hand, constitutes the autonomous intelligence hand rehabilitation system based on brain-computer interface, fusion multi-source information.The present invention will be according to the advanced theory of " initiatively rehabilitation ", finish a kind of by patient's spontaneous brain electricity participate in control, can according to patient's myasthenia state, with adaptive mode, graduation the hand rehabilitation medical supplemental training device of external strength is provided.

The hardware and the software of system are constructed as follows:

The hardware of system constitutes

Healing hand function exercise equipment system hardware block diagram is seen Fig. 2.It is by three modules that the present invention can be divided into by task, is respectively to start module, feedback control module and man machine interface.

Start module and comprise that three roads electricity collection medicated cap, the big collection plate of preposition brain tele-release and slave computer DSP signal processing part are grouped into, be responsible for extracting the active consciousness of human hand movement, the enabling signal of healing hand function exercise equipment is provided.Patient wear three road brain wave acquisition medicated caps extract the EEG signals of head C3, C4 and Cz, again with other EEG signals of uV level, amplify 10,000 times through homemade brain wave acquisition plate, send into slave computer.In slave computer, after the EEG signals denoising, do the classification of task consciousness, can obtain people's active exercise idea, this moment, the healing hand function exercise equipment obtained motion actuated signal.

Feedback control module comprises that hand machinery, hand MEMS flexible sensor array and slave computer feedback control partly form.Be responsible for according to patient's myasthenia state, with adaptive mode, providing outside power-assisted with classifying.After obtaining enabling signal, according to the sensor array of hand, obtain the state of real-time motion of hand and power, according to different rehabilitation modalities, the control hand provides the size variation of power-assisted.The signal that sensor array is collected is transferred to host computer simultaneously, as the foundation of rehabilitation evaluation.

Man machine interface is to be presented on the PC that healing hand function exercise equipment software is housed, after brain electricity that the slave computer transmission is next and sensor signal are handled, in real time, dynamically show quantitative muscular strength information, monitoring training result Xiang doctor and patient, but also can set up when training force level of being provided of mechanical hand on their own by the user, make it under the control of microprocessor, can regulate the intensity size of outside power-assisted or resistance automatically.Rehabilitation training effect and the external strength information that is provided are provided in real time, not only provide foundation, and provide driving source for patient's active idea participates in limb rehabilitation training for the doctor formulates rehabilitation training plans.Introducing is released the brain training and is combined the clinical intervention pattern with the limbs training based on the real-time vision feedback training of virtual environment technology, design healing hand function training guidance interface.Show hands movement functional status information, for the clinical rehabilitation diagnosis and treatment provide objective, quantized assessment method and index

1.1 tool design of rehabilitation machinery mobile phone and power plant module thereof

Carry out the design of frame for movement according to product expectation function and index,, can make some simple extensions and return the action of holding by nondominant hand, adopted the scheme of cylinder control in order to reach hands dysfunction person under various amyasthenic states.Because the stroke patient hand can not move for a long time, hand may produce certain distortion, adopts cylinder structure simple in structure, realizes the function that we need by the flexible push-and-pull finger of cylinder.

As shown in Figure 3, the rehabilitation manipulator comprises that the upper end has the main body mechanical framework 1 of chute 7, chute 7 internal fixation at this main body mechanical framework 1 have wrist supporting plate 8, be fixed with the installing plate 9 that has gathering sill 3 in the lower end of main frame mechanical framework 1 lateral symmetry, in the gathering sill 3 of installing plate 9, crossbeam 2 is installed, the two ends of crossbeam 2 link to each other with spring 5 respectively, the other end of spring 5 is fixed in the chute 7 of main body mechanical framework 1, and the outfan of the cylinder 4 that links to each other with slave computer DSP links to each other with crossbeam 2.

The finger of glove is connected on the crossbeam, and the shoulder hole of crossbeam is in order fixedly to tie up the sticky cloth of glove, to use the screw in compression sticky cloth.The crossbeam two ends are movable in gathering sill, and spring connects spring crossbeam junction point for one section and connects other end connecting bolt mounting groove.The junction point that connects cylinder is arranged in the middle of crossbeam, and cylinder is done stretching motion, and the power of healing hand function exercise equipment is provided.

Spring selects tension spring (pulling force is less than 250 Ns, and big footpath is less than 15 millimeters, tensile elongation 80 millimeter), exchanges left-right symmetric for a few money springs that reach the optional different pulling force of better rehabilitation efficacy in addition.Less than 100 millimeters, therefore as long as placement location is suitable, sitting crouches all can use whole palm apart from the baseplane.But contacting with groove at the volley, spring can send slight sound.The pick-and-place motion of whole device imitation staff is mainly done passive resistive exercise according to the elastic force that spring provides, to reach better rehabilitation efficacy.

1.2 the controllability with mechanical hand that provides of moving power plant module classification power-assisted of rehabilitation machinery tinea manus or resistance realizes.Concrete pneumatic circuit as shown in Figure 9.The cylinder 4 of rehabilitation manipulator adopts medical quiet air compressor machine as pneumatic supply, the air pressure of 0-0.8Mpa is provided, check valve, air accumulator, filter, air drying lubricator, electric Proportion valve and three position five-way valve are installed on the pipeline between cylinder 4 and the pneumatic supply, and wherein electric Proportion valve links to each other with slave computer DSP.

Adopt medical quiet air compressor machine as pneumatic supply, the air pressure of 0-0.8Mpa is provided,, after the dedusting oil removing, offer cylinder through voltage stabilizing.Three position five-way valve control gas flow paths realizes the control of the direction of motion of cylinder.When high-potential voltage is provided, push away before the cylinder, hand is done stretching; When low-potential voltage is provided, contract behind the cylinder, hand is done and is rolled up.Electric Proportion valve can be controlled air pressure change linearly according to the variation of voltage, and the air pressure of the corresponding 0.005-0.5Mpa of the voltage of 0-10V is controlled the size of outside power-assisted scalably.All qigong control element buyings are from SMC company.

The rehabilitation mechanical hand should nondominant hand carries out the action of grasping and stretching, extension.By comparing for various schemes, the present invention adopts air cylinder structure, realizes function by control push-and-pull finger.

1.3MEMS multisensor array and arrangement mode thereof and encapsulation

The finger pressure signal extracts by the sheet type pressure transducer, demarcate by test of many times and obtain accurately finger pressure and to hold the power corresponding relation early stage, DSP A/D sampling module reads signal and calculates hand grip value by processing, imports host computer into by serial ports and further handles with signal and show.

The present invention introduces grade power-assisted compensation theory, at hand placement sensor array, with the motion of detection hand and the information of power, is the key point that realizes adaptive mode, is classified to provide for patient's hands functional training the external compensation power-assisted.The present invention proposes according to nerve-muscle theory of conduction and the meticulous motion field of hands multi-joint according to experiment, carry out the optimal spatial layout and the design of integrated framework scheme of MEMS flexible sensor array, realize the making of the sensor array of high measurement accuracy, tool " interlayer " structure.

Be furnished with two kinds of pick offs at hand, a kind of is diaphragm pressure flexible MEMS pick off and sensor array, this kind pick off wide-range, high sensitivity, high accuracy, response speed are fast, the advantage that not only has generic array formula pick off, also has good pliability, can free bend even folding, under the prerequisite that does not change the movable and force method of patient, detect the size of active force between hand and the rehabilitation mechanical hand in real time.Another is the single shaft angular transducer of the Finland SCA61T-FA1H1G of VTI company chip, can detect the movement position of hand in the degree of freedom direction of hands, for the evaluation of patient's nerve-muscle function state and classification power-assisted or resistance feedback control provide foundation.

Pressure transducer adopts piezoresistive transducer FSR (Force Sensing Resistor), sheet type, little (the long 38.1mm of size, pick off Validity Test section diameter 7.6mm), sensitivity is higher, the linearity is good, and resistance is linear substantially during greater than 100g at pressure as shown in Figure 4, and the stress test scope is 0-10Kg.In fingerstall, fingerstall is fixed in rehabilitation machinery on hand with sensor package, the good accuracy that contacts with test that has guaranteed pick off and finger.The relation of pressure transducer pressure and output as shown in Figure 5.

Utilize angular transducer measurement articulations digitorum manus angle to change and judge the hand exercise situation.Angular transducer adopts the single shaft angular transducer of the Finland SCA61T-FA1H1G of VTI company chip, the 0-5v analog signal interface, and measuring range is ± 90 °, voltage: output services mode.Pressure sensor dimensions is 10.48mm*11.31mm*5.08mm.

Installation site and sensor voltage output relation output voltage when the pick off level is 2.5V as shown in Figure 6, and the inclination angle output voltage that changes can produce corresponding linear change.Native system rationally is fixed in angular transducer on the healing hand function exercise equipment, determines that according to output voltage values the articulations digitorum manus angle changes and then judgement hand exercise situation.When hand is held with a firm grip or grip when not enough, the hand stop motion, the pressure transducer output voltage values remains unchanged, and identifies the healing hand function situation with this.

By experiment, compare according to testee's hand grasping, and in conjunction with self-test assessment.5 zones of the influenced maximum of palm when determining static grasping.These 5 regional centralized are at several positions such as middle flexor digitorum superficialis, flexor pollicis brevis, adductor pollicis, flexor digiti minimi breviss.According to experimental result, 1 point in 5 zones, 2 and 3 areal pressures are the most obvious, and 5 areal pressures take second place, and 4 areal pressures are minimum.The back of the hand has three influenced maximums of regional power, as shown in Figure 7.

Choose these 8 zones as test point, utilize pick off to extract force value, by calculating the voltage output signal of pressure, the hand grip size when reflecting static grasping with this.The grip size has substantial connection with the muscular strength signal, and when grip increased, the electrical signal amplitude of muscular strength and frequency all can increase.Arrange as shown in Figure 8 in the position in kind of pick off.

Pressure sensor package is fitted in the flexible glove, the position of each pick off is corresponding to 8 top set of regions mid points, glove can adopt the scalable gloves for game of elasticity, material is mainly: polyester fiber, nylon, Lycra, glove are designed to bilayer, sensor package is fixed in the middle of the interlayer, the outer polyester fiber abrasionproof that uses, durable, internal layer is a nylon, Lycra, contact soft with hands, good springiness and ventilative is tested tested personnel and is with the measurement glove, and glove are fixed on the crossbeam of putting into the rehabilitation training machinery, hand electric pressure signal and antebrachial muscle force signal are extracted in experiment, judge the hand rehabilitation situation.

1.4 EEG signals extraction module

As Figure 10 is the overall construction drawing of eeg collection system.

Just can extract the information relevant because 3 lead the brain electricity,, be respectively applied for and amplify 3 EEG signals of leading so this acquisition system has three paths with treatment.Acquisition system is divided into analog-and digital-two parts on constituting.(1) main amplification and the impedance detection that realizes EEG signals of simulation part; (2) numerical portion is mainly realized the AD conversion of signal, and the pretreatment of signal is also passed to upper PC.Because the brain electricity is very faint signal (uV rank), thereby interference to external world is very responsive; And numerical portion is because its frequency height can produce a lot of interference.In order to remove this interference, be necessary digital and analog circuit is partly isolated.Isolate and to be divided into two parts, the one, the isolation of power supply (i.e. numeral and simulation partly be not ground) altogether realizes by the insulating power supply module; The 2nd, the isolation of signal realizes by the photoelectricity coupling.

Because the amplitude of brain electricity so will show brain wave patterns, need be amplified them about 10,000 times between 10-100uV.So big amplification obviously can not once realize, therefore adopts the mode of amplifying one by one.We adopt three grades of amplifications in this system, and the amplification of design is 10,000 times.Wherein preposition being enlarged into about 25 times, secondary is enlarged into about 40 times, and three grades are enlarged into about 10 times.Because the influence of electrode polarization voltage, offset voltage and some other direct current factor; carry out to make amplifier saturated when big multiple amplifies; signal in entering DSP the time waist through holding circuit; brain electricity after amplifying is at last delivered to DSP by photoelectric isolating circuit and is carried out the A/D conversion; its sample rate is 128Hz, and sampling precision is 12.

Whole system adopts the USB power supply, thereby convenient and safe.USB can provide the electric current of 5V voltage and 500mA, and general power is 2.5W, can satisfy the power consumption requirement of system fully.

Realize that the big main difficulty of brain tele-release is to remove various interference when high-gain is amplified, promptly guarantee good signal-to-noise; And the prestage amplification is the core that achieves this end.

1.5 multiple signals processing module

Signal processing module is to be realized by the fixed-point dsp TMS320F2812DSP and the peripheral circuit thereof based on the TMS320C2xx kernel of TI company.Though analogue signal has been passed through the analog filtering processing before entering digital signal processing module, removed the power frequency interference, but by the amplification of 10000 times in amplifier, there are some interference in addition in the signal of being gathered into by A/D and have some physiology artifacts (for example: electrocardio, eye electricity, myoelectricity or the like).False judgment in feedback procedure, to occur in order reducing, need to carry out pretreatment the signal of gathering.In DSP inside, we have carried out the low-pass filtering of 0-30Hz to signal, have further removed the interference of power frequency and other signals.Treated signal can obtain the wish of people's active exercise, as the enabling signal of rehabilitation training instrument by the processing of classifying of AR model method.

After obtaining enabling signal, gather machinery on hand power and the signal of displacement transducer, can judge hand power and movable information, required power-assisted or resistance are provided as required.

Software constitutes

Realized mainly that based on the software of the healing hand function exercise equipment of motion imagination brain electricity dynamic demonstration, eigenvalue extractions, the Fisher linear classification of the collection of brain electricity and processing, brain electricity, the dynamic demonstration and the muscular strength of muscular strength feed back the rehabilitation training module, idiographic flow is seen Figure 11.After DSP powers on, pass through serial communication, obtain user's various information with host computer, rehabilitation modality is set, and after initialization finished, AD started working, gather brain telecommunications, do pretreatment after, pass to host computer, after host computer is done classification of task,, then enter default rehabilitation modality as obtaining enabling signal, the value of AD pick-up transducers array then, handle the back as feedback, control the air pressure of pneumatic circuit adaptively, thereby control the size of extraneous power-assisted.Give host computer with various transfer of data simultaneously, know this time to move to terminal point, begin to wash once circulation, judge the brain electricity, do classification.

2.1DSP in software section

Main communication function, the content of data acquisition process function and feedback control function three bulks of realizing in DSP.

2.1.1 serial ports (SCI) data communication module

By finishing the transmission of user characteristics and parameter with communicating by letter of host computer, the selection of training mode, and the transmission of brain electricity and pressure sensor signal are so that set up customer data base at host computer.The transmission of data is mainly finished by the SCI module of F2812, and its pressure sensor data with the collection of ADC sampling module is uploaded to PC and shows, monitors patient's rehabilitation situation in real time for the doctor, simultaneously the hand rehabilitation state is carried out the trace detection.

The transfer of data of F2812 and PC relies on RS-232 interface, and this is by Electronic Industries Association (Electronic Industries Association, the asynchronous transmission standard interface of EIA) being formulated.Usually RS-232 interface occurs with 9 pins (DB-9) or the kenel of 25 pins (DB-25), has two groups of RS-232 interface on the general personal computer, is called COM1 and COM2.

The baud rate that serial communication adopts is 9600, the no parity check position, and 8 bit data positions, 1 position of rest, the data format that is provided with DSP SCI module is identical.The reception area data are the sampled value of 6 next ADC passages of DSP transmission.Wherein first passage (ADCINA0) inserts the sine wave of 2Hz, second passage (ADCINA1) ground connection, and remaining channel is unsettled, and sampled value is at random.

2.1.2 data acquisition and D/A voltage output module

The sampling work of brain electricity and pressure transducer numerical value is mainly finished by the ADC sampling module.The data acquisition module of EEG signals mainly is to be finished by the sub-thread of data acquisition, sample frequency is 128Hz, the real-time pretreatment module of EEG signals is done further pretreatment to the digital signal that collects host computer and is handled, and eliminates the interference that is produced in transmission course.

Per 0.1 second once sampling of sensor settings, promptly sample rate is made as 10Hz.F2812 has 3 CPU intervalometers, can be used for realizing easily timing sampling.

The AD sampling module of F2812 has 12 precision, that is to say that error should be reduced to 0.1%, but in fact the sampled value of its AD and theoretical value have bigger error, and maximum can reach 9%.This obviously can not satisfy accurate Testing requirement, but by adding calibration circuit and improving on sampling routine, can control the transformation result error about 0.1%, thereby meet design requirement.

2.1.3 providing of classification power-assisted adaptively

D/A is responsible for digital signal being changed into analogue signal, the external connection electrical apparatus proportioning valve.Gather hand MEMS flexible sensor signal in real time, judge the power and the displacement characteristic information of hand, adaptive providing is minimum power-assisted and the maximum resistance that staff can move.

As shown in figure 12, be the subroutine flow chart under power-assisted rehabilitation modality and the resistance rehabilitation modality.Under the power-assisted rehabilitation modality, at first initialization, pick off zeroing, the value of reading angular pick off is done average and is handled, if find in the 1s, the value of pick off does not change, and promptly hand is static, then increases D/A output voltage Δ V according to setting progression automatically, thereby output pressure increases Δ P, and power-assisted increases Δ F, then the judgement that circulates and enter sensor values next time, enter circulation next time, up to producing hand motion, move to terminal point, loop ends.

Resistance mode is just in time opposite, and the value of reading angular pick off is done average and handled, if find that the value of pick off does not change in the 1s, promptly hand is static, then reduce D/A output voltage Δ V according to setting progression automatically, thereby output pressure reduces Δ P, and resistance reduces Δ F, then the judgement that circulates and enter sensor values next time, enter circulation next time, up to producing hand motion, move to terminal point, loop ends.

This kind mode does not have the standard that unified hand can motoricity, and the power of needs varies with each individual because staff can move, this kind method is being sought all the time, that point that hand can just move allows the patient can use up own maximum effort all the time, reaches maximum rehabilitation training effect.

2.2 the software section in the host computer

2.2.1 development platform and developing instrument

Operating system: Windows XP

Development platform: Microsoft Visual Studio.NET 2008

Development language: C#

Data base: Microsoft SQL Server 2005

As Figure 13, the software log-in interface comprises data base administration, and eigenvalue extraction and processing etc. are landed module and are used to guarantee safety of user data.

Software section in the host computer is mainly by constituting with lower module:

2.2.2 EEG feature extraction and sort module

Autoregression model based on the burg algorithm is used in the feature extraction of EEG signals, be called for short AR (auto-regressive) model.

No matter x (n) is deterministic signal or stochastic signal, input u (n) and output x (n) are always there is following relation:

$x\left(n\right)=-\underset{k=1}{\overset{p}{\mathrm{\Σ}}}{a}_{k}x(n-k)+\underset{k=0}{\overset{q}{\mathrm{\Σ}}}{b}_{k}u(n-k)---\left(1\right)$

Promptly

$x\left(n\right)=\underset{k=0}{\overset{\∞}{\mathrm{\Σ}}}h\left(k\right)u(n-k)---\left(2\right)$

Wherein h (k) is the ssystem transfer function that u (n) excitation produces signal x (n).a _kWith b _kCoefficient for respective items.

Transform is got on last two formula both sides respectively, and supposition b ₀=1, can get its transfer function and be:

$H\left(z\right)=\frac{X\left(z\right)}{U\left(z\right)}=\frac{1+\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}{b}_k{z}^{-\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="HDA0000043226210000021.png,HDA0000043226210000031.png"\; state="\{\{state\}\}">k</figure-callout>}}}{1+\underset{k=1}{\overset{p}{\mathrm{\&Sigma;}}}{a}_k{z}^{-k}}=\frac{B\left(z\right)}{A\left(z\right)}---\left(3\right)$

Wherein X (z), U (z) are respectively the Z territory expression-form of x (n), u (n), and

Suppose that u (n) is that an average is zero, variance is σ ²White noise sequence, by the theory of stochastic signal, make z=e in the following formula by linear system ^JwBe Fourier transformation as can be known, the power spectrum of output sequence x (n) is:

$P_x\left(e^{\mathrm{jw}}\right)=\frac{{\mathrm{\&sigma;}}^{2}B\left(e^{\mathrm{jw}}\right)B^{*}\left(e^{\mathrm{jw}}\right)}{A\left(e^{\mathrm{jw}}\right)A^{*}\left(e^{\mathrm{jw}}\right)}=\frac{{\mathrm{\&sigma;}}^2{\left|B\left(e^{\mathrm{jw}}\right)\right|}^2}{{\left|A\left(e^{\mathrm{jw}}\right)\right|}^2}---\left(4\right)$

σ wherein ²Be the variance of white noise u (n), B ^*(e ^Jw), A ^*(e ^Jw) represent B (e respectively ^Jw), A (e ^Jw) conjugation.B (e ^Jw), A (e ^Jw) expression B (e ^Jw), A (e ^Jw) mould.

If the variances sigma of excitation white noise ², and the parameter a of model ₁, a ₂..., a _p, b ₁, b ₂..., b _qKnown, can calculate the power spectrum of output sequence x (n) so by following formula.In formula (1), if b ₁, b ₂..., b _qBe zero entirely, then signal, transfer function, power spectrum, promptly formula (1), (3), (4) can be write as respectively:

$x\left(n\right)=-\underset{k=1}{\overset{p}{\mathrm{\&Sigma;}}}{a}_{k}x(n-k)+u\left(n\right)---\left(5\right)$

$H\left(z\right)=\frac{1}{A\left(z\right)}=\frac{1}{1+\underset{k=1}{\overset{p}{\mathrm{\&Sigma;}}}{a}_k{z}^{-k}}---\left(6\right)$

$P_x\left(e^{\mathrm{jw}}\right)==\frac{{\mathrm{\&sigma;}}^2}{{|1+\underset{k=1}{\overset{p}{\mathrm{\&Sigma;}}}{a}_k{e}^{-\mathrm{jwk}}|}^2}---\left(7\right)$

The model that formula (5), (6), (7) provide is called autoregression model.Wherein p is the exponent number of model, and a is the AR model coefficient, and we select p=6, so comprise 6 coefficient a undetermined in this model ₁, a ₂... a ₆, this group coefficient is provided by burg algorithm recursion.The recursion step of Burg algorithm following (wherein m represents the AR model order):

1) by initial condition M=1,

Obtain by (8) formula again

[(wherein, f represents forward prediction, and b indicates back forecast)]

${\hat{k}}_m=\frac{-2\underset{n=m}{\overset{N-1}{\mathrm{\&Sigma;}}}{e}_{m-1}^f\left(n\right)e_{m-1}^{b^{*}}(n-1)}{\underset{n=m}{\overset{N-1}{\mathrm{\&Sigma;}}}{\left|e_{m-1}^f\left(n\right)\right|}^2+\underset{n=m}{\overset{N-1}{\mathrm{\&Sigma;}}}{\left|e_{m-1}^b(n-1)\right|}^2},m=\mathrm{1,2},...,p---\left(8\right)$

E wherein _m(n) expression is according to extrapolate the constantly predictive value of m time of n, and subscript f represents to predict forward that b represents to predict backward.

2) by initial condition

Getting m=1[asks

] time e ₁ ^f(n), e ₁ ^b(n), k ₁And the AR estimates of parameters of this moment

3) by

When (9), (10) formula is tried to achieve m=2 respectively

Estimate by (8) formula again

$e_m^f\left(n\right)=e_{m-1}^f\left(n\right)+k_m{e}_{m-1}^b(n-1),m=\mathrm{1,2},...,p---\left(9\right)$

$e_m^b\left(n\right)=e_{m-1}^b(n-1)+k_m^{*}{e}_{m-1}^f\left(n\right),m=\mathrm{1,2},...,p---\left(10\right)$

4) according to [Levinson] recurrence relation of (11a), (11b) and (12) formula, when obtaining m=2

And

${\hat{a}}_m\left(k\right)={\hat{a}}_{m-1}\left(k\right)+{\hat{k}}_m{\hat{a}}_{m-1}^{*}(m-k),k=\mathrm{1,2},...,m-1---\left(11a\right)$

${\hat{a}}_m\left(m\right)={\hat{k}}_m,k=m---\left(11b\right)$

${\widehat{\mathrm{\&rho;}}}_m=(1-{\left|{\hat{k}}_m\right|}^2){\widehat{\mathrm{\&rho;}}}_{m-1}---\left(12\right)$

5) repeat said process,, can obtain the parameter [the AR parameter when having obtained all orders] of p rank AR model up to m=p.

Each leads and can calculate 6 parameters, because system gathers three simultaneously and leads the brain electricity, program needs parallel processing three to lead EEG signals, obtains 18 parameters altogether, so characterize the vectorial a=(a that the EEG signals feature of each period is exactly one ten octuple ₁, a ₂... a ₁₈).

The classification of brain electrical feature comprises two steps, i.e. training process and use.In ensuing process, with x=(x ₁, x ₂..., x ₁₈) replacement a=(a ₁, a ₂... a ₁₈), training process and use are respectively described below:

1) training process

The main purpose of training process is to determine the classification weight vector and the classification thresholds of particular patient.

In training process, computer calls the graphical cues interface, as shown in figure 14, provides prompting images such as " left side ", " right side ", " plus sige " at random, the patient will make the motion imaginations such as " moving left hand ", " the moving right hand ", " loosening ", utilizes multithreading to receive EEG signals simultaneously.Training process is handled the EEG signals that obtains after finishing.

Requirement to the prompting figure is: prompting figure constant duration occurs at random, and always pointing out number of times is 4*N, and " left side ", " right side ", " plus sige " occurrence number are respectively N, N, 2*N.We select N=30, and promptly total prompting figure is 120 width of cloth, and " left side ", " right side ", " plus sige " are respectively 30,30,60;

Data procedures is: lead parallel 120 sections of being divided into equal length of EEG signals with three, each section is carried out eigenvalue extract, obtaining 120 forms is x=(x ₁, x ₂..., x ₁₈) characteristic vector, the characteristic vector that corresponding " left side ", " right side " prompting figure are obtained is divided into one group, amounts to 60, the characteristic vector of corresponding " plus sige " prompting figure is divided into one group, amount to 60, utilize this two stack features value to carry out the Fisher classification, weight vector a=(a obtains classifying ₁, a ₂... a ₁₈) and classification thresholds.The weight vector of will classifying and classification thresholds deposit the data base in, in order to calling.

2) use

In the use, need handle in real time EEG signals, 5 seconds data of every collection are carried out a sub-eigenvalue and are extracted, and obtain characteristic vector x=(x ₁, x ₂..., x ₁₈), call this user weight vector a=(a that in training process, obtains classifying ₁, a ₂... a ₁₈), calculate following formula

y＝a ^Tx＝a ₁x ₁+a ₂x ₂+......+a ₁₈x ₁₈

Compare by the size of this value, can draw the classification results of " moving " and " motionless " with classification thresholds.

Figure 14, training prompting interface needs to import the picture interval in this interface, set and change frame time, clicks " beginning " afterwards, enters screen mode toggle, can occur hint image at random in the middle of the screen, the motion imagination that the prompting user is correlated with.Software writes down the order of prompting figure simultaneously, is used for sample classification; Utilize multithreading to receive EEG signals and deposit, call EEG signals after training process finishes automatically and carry out that eigenvalue extracts and classification, and will obtain classifying weight vector and classification thresholds deposit the data base in.

2.2.3 feedback training module

This module is according to patient's myasthenia state, and the scheme that is used for selecting according to the trainer is carried out active to the trainer and resumed training.At the patient of different degree of injury, we can be provided with different rehabilitation modalities to instrument, assistant mode and resistance mode (flow chart as shown in figure 12).Assistant mode is primarily aimed at the myasthenia of 0-3 level, and patient's myasthenia completely maybe can only carry out non-loaded light exercise.Resistance mode is at the myasthenia state more than 4 grades, mainly helps to recover the hand muscle strength.

Under assistant mode, gather hand MEMS flexible sensor signal in real time, judge the power and the displacement characteristic information of hand, adaptive providing is the minimum power-assisted that staff can move.If find that hand does not move, then increase power-assisted according to setting progression automatically, know to produce hand motion.Resistance mode is just in time opposite, as according to sensor signal, finds that hand can't move, and then progressively reduces external drag, up to moving, provides the maximum resistance that staff can move this moment.

Feedback training can farthest be transferred patient's active property of participation, helps the foundation once more and the muscle recovery of patient's nervous pathway.Force transducer and displacement transducer are placed in the position of hand and cylinder, obtain the strength that hand initiatively participates in and the state of hand exercise in real time.According to different situations, classification is provided, the scalable hand exercise makes it to obtain corresponding power-assisted or resistance.

2.2.4 brain electricity, muscular strength show

It is the Active electroencephalogram (EEG) of making the patient who accepts feedback training that brain electricity shows, is that the EEG signals waveform dynamic real-time ground with the person of undergoing training is presented on the display, and native system has been realized that electroencephalogram does not dynamically have and traced.

The collection of muscular strength information and processing mainly provide user's hands function status information, and these information can be used for the design of healing hand function training program, also can be used as training program Evaluation on effect parameter simultaneously.

As Figure 16, the mirror image neuron system not only is activated when action is carried out, and also can be activated when another person finishes same action or oneself this action of the imagination when a people observes.In this interface, if select " motion imagination task " pattern, the fingerprint type in the upper right corner can or stop according to the classification campaign of brain electrical feature signal; If " movement observation task " pattern of selection, the fingerprint type can be finished a series of contraction and stretching according to predefined program.So the mirror image neuron system just can be activated when patients with cerebral apoplexy moved by the imagination and observation, thereby play the damaged effect of motor function that recovers.Comprise three of EEG signals display interfaces in the interface, one of electromyographic signal display interface, one of hand pressure signal display interface; The control information panel can carry out task choosing, the selection of signal display interface; Status signal shows each wave band energy value of selected EEG signals.

2.2.5 DBM

At first, the user need import essential information, as numbering, name, sex, age, the scheme of undergoing training, the number of times of undergoing training, contact address etc.Then, physician guidance user accepts the EEG signals training process before the rehabilitation training, in the training process, the user need finish corresponding motion imagination task according to prompting, system receives the EEG signals by the dsp chip transmission in real time, carry out feature extraction, the parameter that obtains classifying deposits this user's categorical data in data base; In the rehabilitation training process, use the EEG signals training parameter that has stored among the data base, motion imagination brain electric information to the user carries out real-time feature extraction and classification, the output category result, drive the motion of hand, receive rehabilitation machinery Micro-force sensor signal on hand simultaneously, measure the hand muscular strength, real-time Transmission is given main frame, and determining of rehabilitation scheme and providing of real-time force are provided.

Database module stores user basic information, brain electrical characteristic values, sorting parameter, little force information, feedback training information etc., by structured creation, access and inquiry to desired data in case management module and the feedback treating module, realize the unification of case system, management efficiently, be convenient to design comparatively perfect rehabilitation training scheme and carry out the assessment of healing hand function situation.

As Figure 15, database management module is used for essential information, brain electric information, myoelectric information and the muscular strength information etc. of managing patient, the relevant information that this user can be checked in input user account or name.Can and pass to DSP with it with brain electricity classification weight vector, be used to move and imagine the classification of brain electricity; Myoelectricity and muscular strength information can be accessed, be used for rehabilitation scheme design and rehabilitation assessment.

2.2.6 rehabilitation assessment and interest and appeal game module

The assessment of grip is one of quantitatively more common measurement of muscle function.For acute paralytic, can write down the recovery of grip, be that initial upper limb recovers one of evaluation with the sensitivity of accurately predicting the functional rehabilitation in future.It also illustrates the upper extremity exercise function damage to apoplexy, and using the grip measurement is to estimate preferably.Estimate chronic paralytic's " reliability " " validity " and " sensitivity " with grip, and grip and the data that influences the upper extremity function injuring relation are confirmed all.We equally with the autonomous grip of the maximum of strong limb as a comparison.

In the feedback training link, utilize trivial games to strengthen the interest of training.As shown in figure 17, this recreation has the grade selector button for fishing trivial games, top, interface, selects the fish of different brackets, represents different resistance information.After the recreation beginning, main window can provide the information that fish is risen to the bait, and the prompting user begins firmly, and the fingerprint type in the upper left corner also can changing to the range degree according to the user.The messagewindow on right side, interface is divided into game information and rehabilitation information, and the rehabilitation messagewindow comprises resistance information, muscular strength information, and power-assisted information, hand pressure information etc., and rehabilitation situation graded.This interface can increase the interest of rehabilitation training, has reasonably to instruct rehabilitation training.

Claims (7)

1. an intelligent brain injury patients healing hand function exercise equipment is characterized in that: comprise eeg signal acquisition processing module, hand force signal real-time detection apparatus and rehabilitation manipulator;

Described eeg signal acquisition processing module is extracted patient's active exercise idea, in real time, at the brain wave of body, online acquisition brain injury patients, extract the characteristic information of the motion imagination, control signal as rehabilitation training inputs among the slave computer DSP, with the motion of the control of the sports consciousness information in spontaneous brain electricity signal rehabilitation manipulator;

Described hand force signal real-time detection apparatus is at hand placement sensor array, detects the motion of hand and the information of power and inputs to slave computer DSP, realizes adaptive mode, is classified to provide the external compensation power-assisted for patient's hands functional training;

Described rehabilitation manipulator comprises that the upper end has the main body mechanical framework (1) of chute (7), chute (7) internal fixation at this main body mechanical framework (1) has wrist supporting plate (8), lower end lateral symmetry at main frame mechanical framework (1) is fixed with the installing plate (9) that has gathering sill (3), in the gathering sill (3) of installing plate (9), crossbeam (2) is installed, the two ends of crossbeam (2) link to each other with spring (5) respectively, the other end of spring (5) is fixed in the chute (7) of main body mechanical framework (1), and the outfan of the cylinder (4) that links to each other with slave computer DSP links to each other with crossbeam (2).

2. intelligent brain injury patients healing hand function exercise equipment according to claim 1, it is characterized in that: described eeg signal acquisition processing module comprises electrode cap, big collection plate of preposition brain tele-release and slave computer DSP signal processing part are grouped into, three road brain wave acquisition medicated caps extract head C3, the EEG signals in C4 and Cz zone, through the big collection plate of preposition brain tele-release with other EEG signals of uV level, amplify 10,000 times through the brain wave acquisition plate, send into slave computer DSP, the DSP sample frequency is 128Hz, in slave computer after the EEG signals denoising, do the active exercise idea that the classification of task consciousness obtains the people, this moment, the healing hand function exercise equipment obtained motion actuated signal.

3. intelligent brain injury patients healing hand function exercise equipment according to claim 2, it is characterized in that: described brain wave acquisition plate adopts the mode of amplifying one by one, wherein prepositionly be enlarged into 25 times, secondary is enlarged into 40 times, three grades are enlarged into 10 times, brain electricity after the amplification is delivered to slave computer DSP by photoelectric isolating circuit and is carried out the A/D conversion, and its sample rate is 128Hz, and sampling precision is 12.

4. intelligent brain injury patients healing hand function exercise equipment according to claim 1, it is characterized in that: described hand force signal real-time detection apparatus adopts the glove-type structure, and the single shaft angular transducer of the Finland SCA61T-FA1H1G of the VTI company chip of five sheet type pressure transducers that detect the finger pressure signals and three movement positions that detect hands is installed in glove, the sheet type pressure transducer is installed in middle flexor digitorum superficialis (label 1), flexor pollicis brevis (label 2,3), adductor pollicis (label 5), flexor digiti minimi brevis (label 4), the single shaft angular transducer is installed in the position of middle, the crossbeam back side, and nine pick offs are crossed serial ports through the A/D sampling module all with signal and imported host computer into.

5. intelligent brain injury patients healing hand function exercise equipment according to claim 1 is characterized in that: the finger of described hand force signal real-time detection apparatus glove is connected on the crossbeam shoulder hole of rehabilitation manipulator.

6. intelligent brain injury patients healing hand function exercise equipment according to claim 1, it is characterized in that: the cylinder of described rehabilitation manipulator (4) adopts medical quiet air compressor machine as pneumatic supply, the air pressure of 0-0.8Mpa is provided, check valve, air accumulator, filter, air drying lubricator, electric Proportion valve and three position five-way valve are installed on the pipeline between cylinder (4) and the pneumatic supply, and wherein electric Proportion valve links to each other with slave computer DSP.

7. intelligent brain injury patients healing hand function exercise equipment according to claim 1, it is characterized in that: described slave computer DSP also links to each other with the host computer of peripheral hardware, after brain electricity that the slave computer transmission is next and sensor signal are handled, in real time, dynamically show quantitative muscular strength information, monitoring training result Xiang doctor and patient, the force level that mechanical hand is provided when setting up training on their own by the user makes it regulate the intensity size of outside power-assisted or resistance automatically under the control of slave computer DSP.

Images