CN109528203A - A kind of interactive patients with cerebral apoplexy gait training and evaluating system based on Multi-source Information Fusion - Google Patents

Abstract

The invention discloses a kind of interactive patients with cerebral apoplexy gait training and evaluating system based on Multi-source Information Fusion, structure include pose acquisition module, plantar pressure acquisition module, tired information acquisition module, wireless communication module, Data Analysis Services module, intelligent cloud platform, the AR helmet and mobile phone terminal；The system is directed to the patients with cerebral apoplexy with certain walking ability, is based on augmented reality, can intelligently be supplied to the interactive rehabilitation training game of patient's high-immersion, effectively improve its gait rehabilitation training effect；Obtain multidimensional gait information when Rehabilitation training in real time by gait acquisition module, after Data Analysis Services module, treated information and date by wireless module is transferred to mobile phone terminal, gait evaluation result height visually Real-time Feedback to doctor and patient；It is monitored based on muscular fatigue, guarantees the safety of Rehabilitation training process；Intelligent cloud platform can also realize automatically tracking, evaluate and test and pushing for gait condition.

Description

A kind of interactive patients with cerebral apoplexy gait training and evaluation and test based on Multi-source Information Fusion System

Technical field

The present invention relates to body gait training and evaluation and test fields, and in particular to a kind of interactive mode based on Multi-source Information Fusion Patients with cerebral apoplexy gait training and evaluating system.

Background technique

Gait is the behavioural characteristic of mankind's walking, is related to the factors such as behavioural habits, occupation, education, age and gender, equally Also it is influenced by a variety of diseases；The control of gait is sufficiently complex, including central command, balance and coordinated control, is related to The cooperative motion in lower limb each joint and muscle, at the same it is also related with the posture of upper limb and trunk；The imbalance of any link all may Walking and gait are influenced, and exception is also possible to by compensatory or cover；Gait evaluation and test is to study the inspection method of walking rule, purport By biomethanics and kinematics means, the key link and influence factor of abnormal gait are disclosed, to instruct Rehabilitation Assessment and treatment, it helps clinical diagnosis, curative effect evaluation and mechanism study etc..

Since nervous centralis is impaired, gait and normal person can make a big difference patients with cerebral apoplexy, pass through detection brain soldier The gait information of middle patient is compared with normal person's standard gait, can help to the state of an illness shape for accurately obtaining patients with cerebral apoplexy Condition overcomes conventional method to only rely on to visually observe and can not accurately provide patients with cerebral apoplexy condition in entire rehabilitation training The problem of；In addition to this, patients with cerebral apoplexy is more susceptible to fatigue compared with normal person, in addition the central nervous system function of most of patients Impaired, in walking, brain in patients cannot obtain the feedback information in relation to muscle activity situation in time, with adding for degree of fatigue Weight, the tension of muscle can significantly rise so as to cause spasm, the serious consequences such as pull, and easily cause the secondary damage of human muscle, Therefore, the tired information of on-line real-time measuremen patient muscle is very crucial.

The patent of Publication No. CN103505219 describes a kind of body gait evaluating system and method, passes through three three It ties up acceleration transducer and acquires data, and terminal is transmitted data to by wireless transmission and is analyzed, on integrative medicine Tinetti evaluation and test table provides evaluation result, but system can only acquire three-dimensional acceleration signal, and parameter is limited, and there is no patient's steps The feedback mechanism of state rehabilitation training function and muscular fatigue information, while being also difficult to realize the rehabilitation instruction of long-term follow record patient Practice effect.

Summary of the invention

In view of the above problems, the invention discloses a kind of, the interactive patients with cerebral apoplexy gait based on Multi-source Information Fusion is instructed Experienced and evaluating system, the system obtain the gait and pose of patient by Multi-source Information Fusion, are carried out based on the AR helmet virtually existing Real gait training obtains more comprehensive multi-source multidimensional gait information while patients with cerebral apoplexy training, utilizes cloud platform Automatic judgment, tracking and push are carried out to patient's gait condition, while muscular fatigue feedback mechanism is added, increases cerebral apoplexy trouble The safety of person's rehabilitation training system.

Technical solution of the invention is as follows.

The patients with cerebral apoplexy training of the Multi-source Information Fusion and gait evaluating system include pose acquisition module, vola Pressure acquisition module, tired information acquisition module, wireless communication module, Data Analysis Services module, intelligent cloud platform, the AR helmet And mobile phone terminal.

The pose acquisition module is a nine axis attitude transducers, which includes three axis accelerometer, three axis tops Spiral shell instrument and three axis magnetometric sensors, the parameters such as instantaneous position, speed, acceleration for the movement of comprehensive acquisition body gait, The position of acquisition is located at the waist of patient, can extract in this way human body position and posture and improve body gait index evaluation and test Accuracy.

The plantar pressure acquisition module is encapsulated in by plantar pressure sensor using people's foot as in the film substrate of model, root According to the composition of vola bone, the sensor block made by force sensitive ink is distributed in the foot such as five big phalanx points, calcaneum, metatarsal The main force part in bottom, place the presser sensor patch of large area by the other positions in vola acquire slight plantar pressure and its Variation, each plantar pressure sensor block is by the progress data transmitting of internal thin wire, by operational amplifier by each sensor The pressure data of acquisition is fused into pressure value, is changed according to left and right plantar pressure to judge lifting and landing for left and right foot.

The fatigue information acquisition module is by surface myoelectric sensor, flesh sound sensor and oximetry sensor three One sensor array of class sensor integration, it is compact-sized, it can conveniently be attached to human body acquisition position, the acquisition position of the module At the row's myenteron and musculus soleus that walking is easiest to fatigue.

The wireless communication module is for the wireless communication between modules.

The Data Analysis Services module is that the data for obtaining the acquisition module are handled, and is passed from nine axis postures Position and posture information when extracting human locomotion in the data of sensor acquisition, when obtaining walking from plantar pressure sensor Double-legged alternate variation, the posture when walking of patient can be intuitively shown by the AR helmet, can also be logical in user terminal Cross and virtual threshold height be set to increase gait training difficulty, if patient's foot-up height reach virtual door sill height and smoothly across It crosses, scores, otherwise do not score, shown the practical posture of patient by virtual reality technology, and walking is real-time Patient is fed back to, training effect is enhanced；Simultaneously in order to enhance trained interest, virtual game training is divided into two people and interacts confrontation Training, medical worker guide the gait rehabilitation training both of which of auxiliary patient；Interact the virtual scene of dual training are as follows: two people The footrace in same racing track will appear virtual threshold obstacle, successfully cross over since same starting point beginning during footrace Obstacle can obtain extra bonus point, while can also be by remaining where one is to virtual obstacles are arranged on the racing track of opponent, but simultaneously can The time of the user is consumed, the used time is shorter, and person's bonus point is more, and trained evaluation result is finally provided according to comprehensive score；It was training Cheng Zhong, the walking obtained by Position and attitude sensor are by the length of time that plantar pressure sensor detects foot bottom stress Whether whether the variation of uniform and pressure value is normal for no consistent, pressure value distribution；Tinetti is evaluated and tested in table on integrative medicine Determine whether the gait information of object to be evaluated is normal, and result is intuitively showed into user by mobile phone terminal；It is described Tinetti evaluation and test table in index include: bipod relative position, the size of the difference of bipod stress time, plantar pressure variation Uniformly whether, whether paces are uniform.

The intelligent cloud platform uses NoSql distributed data base and Hadoop computing unit；NoSql distributed data Library is connected with management terminal and Hadoop computing unit, for storing the gait information data and evaluation result of patients with cerebral apoplexy, It is stored simultaneously plus after user tag for distinguishing user during storage for data；NoSql distributed data base with Hadoop uses clustering to design, and provides redundancy backup equipment, guarantees the reliability of system；Using isolated forest (Isolation Forest) carrys out the abnormal data in detection data, achievees the purpose that reduce data and improves data accuracy, The isolated forest algorithm is to isolate abnormal point under the strategy of random division, achievees the purpose that separating abnormality data；Intelligence Energy cloud platform can realize automatically tracking, evaluate and test and pushing for patient's gait condition.

Advantage of the invention: the gait training system uses nine axis attitude transducers and plantar pressure sensor Overall Acquisition The gait information of patients with cerebral apoplexy, by the AR helmet by 3-D image Real-time Feedback to patient, and have one and it is double Reality-virtualizing game training function enhances training effect；Gait detection passes through the nine collected multi-source informations of axis attitude transducer It is merged, wants more accurate compared to acceleration transducer acquisition gait information was depended merely in the past；And introduce the inspection of human muscle's fatigue The feedback mechanism of survey improves safety when patients with cerebral apoplexy gait training；The number that intelligent cloud platform can obtain terminal According to being saved, the data based on preservation carry out collected data by MapReduce distribution off-line calculation model Deep layer is excavated, and provides the pathological model to match with patient's gait analysis result, and make the stronger rehabilitation training of specific aim Plan.

Detailed description of the invention

Fig. 1 is interactive patients with cerebral apoplexy gait training and evaluating system structure chart based on Multi-source Information Fusion.

Fig. 2 is interactive patients with cerebral apoplexy gait training and evaluating system block diagram based on Multi-source Information Fusion.

Fig. 3 is interactive patients with cerebral apoplexy gait training and evaluating system working principle diagram based on Multi-source Information Fusion.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, technical solution in the embodiment of the present invention and application method into Row system, it is complete, be explicitly described.

As shown in Figure 1 interactive patients with cerebral apoplexy gait training and evaluating system based on Multi-source Information Fusion includes position Appearance acquisition module (1), plantar pressure acquisition module (2), tired information acquisition module (3), wireless communication module (4), data point Analyse processing module (5), intelligent cloud platform (6), the AR helmet (7) and mobile phone terminal (8).

Nine axis attitude transducers as shown in Figure 1 are fixed on measured's waist, and acquisition is using sensor position as coordinate in real time OriginXYZThe acceleration signal of axis,XYZThe angle signal and direction signal of axis；Number is sent the signal to using wireless transmission According to processing terminal, speed, step pitch and height when human locomotion are obtained by integral calculation.

Pressure resistance type flexibility plantar pressure sensor as shown in Figure 1, uses that pressure drag curve characteristic is good, good tolerance Plantar pressure sensor, resistance and pressure are in power function relationship, and resistance-reciprocal and pressure are in linear approximate relationship, resistance and pressure Force characteristic curve repeatability is good, and functional material moisture-proof thermal friction is good, and packaging technology is stablized, and has good protective effect； Its foot type and gait correcting function use myoelectric sensor, for monitoring muscle condition in real time, by measuring muscle fibers contract Fluctuating range, to judge the physiological datas such as movements of parts of the body state and muscular fatigue situation, electronic circuit and wireless Module realizes the wireless transmission reception of data acquisition control, data buffer storage, data.

Tired information acquisition module as shown in Figure 1 is used by surface myoelectric sensor, flesh sound sensor and blood oxygen concentration The sensor array design of three kinds of sensor integrations of sensor is to rely on nonconducting composite and flexible material, is integrated with nine Surface myoelectric sensor electrode, three flesh sound sensor probes and an oximetry sensor probe, collection point is according to people The muscle layout and vascular distribution of body lower limb are installed.

Tired information acquisition module as shown in Figure 1, the position that myoelectric sensor electrode is placed are in contact with belly of muscle, myoelectricity Electrode selects that potential stabilization, favorable reproducibility, internal resistance be low, electrode of high sensitivity, extracts human muscle by non-invasive methods The electric signal of skin surface, while two-pass DINSAR F, L electrode and all the way reference electrode R are used, to improve the accurate of sampled signal Degree；Flesh sound acquisition terminal is distributed in around electromyographic electrode, for detecting the vibration signal of belly of muscle, is directly pasted by piezoelectric ceramic piece In skin surface, so that faint piezoelectric voltage, i.e. muscle signals are acquired, convenient and high sensitivity；The survey of oximetry value Measuring principle is to emit feux rouges and infrared light respectively by red-light LED and infrared light LED, by above after tissue and blood vessel Photoelectric detector receives reflected light, then by photoelectric conversion, has converted optical signals to current signal, and amplified output, The variation of analysis current signal obtains blood oxygen saturation, expression formula are as follows:

Sao2=C(Hbo2)/(C(Hbo2)+C(Hb))*100%

The acquisition terminal of blood oxygen saturation is fixed on sensor array outermost close at ankle, and blood vessel is most intensive herein, thus Convenient for signal acquisition, processing and accurate analysis；The Position Design of terminal is acquired all using optimal acquisition position, the number of acquisition Also most accurate according to value, all acquisition modes all use hurtless measure to acquire, and highly integrated sensor array dresses detection device It is more convenient.

Pose acquisition module, plantar pressure acquisition module and tired information acquisition module as shown in Figure 1 passes through wireless Communication module is connected with Data Analysis Services module, and Data Analysis Services module, mobile phone terminal, the AR helmet are also by wireless communication Module connection.

Linear acceleration in the acquired information of pose acquisition module as shown in Figure 2 carries out Kalman filtering, then with Europe Angle is drawn to be converted according to Euler matrivx:

WhereinXFor roll angle,YFor pitch angle,ZFor yaw angle, the linear acceleration perpendicular to the ground of human locomotion process is calculatedaccz, speed when human locomotion, human height；It will be measured by plantar pressure sensor All values it is cumulative after, obtain lower limb to ground exerted forces:, foot-up height；Root According to the accuracy of two kinds of sensors, respectively multiplied by being added after corresponding weight value, i.e.,, in conjunction with doctor Whether bipod relative position, the size of the difference of bipod stress time, plantar pressure variation are equal in Tinetti evaluation and test table on Even, the paces evaluation and test information such as whether uniform determines whether the gait information of object to be evaluated is normal, show that body gait index is commented Estimate as a result, and evaluation result is presented to doctor and user by the human-computer interaction interface of mobile phone terminal.

Intelligent cloud platform as shown in Figure 2, including management terminal and cloud control platform；Management terminal include hospital terminal and User terminal two parts are realized using C# language Programming, and the data of cloud platform are obtained by display terminal, and doctor can obtain The gait data of patients with cerebral apoplexy and matched pathological model, quickly analyze the state of an illness of patient；Meanwhile patient can also It is quickly got in touch with doctor using user terminal, seeks advice from oneself physical condition and oneself state of an illness rehabilitation situation；Data interaction Functional module is realized by the way of wireless transmission with acquisition terminal online, completes its data interaction, by ICP/IP protocol and The interaction of cloud control platform realization data；Data analysis uses BDP data analysis tool, is supplied to one analysis directions choosing of user It trims vegetables for cooking list, allows the autonomous selection analysis direction of user, management terminal is established by data interaction functional module and cloud control platform to be connected It connects, resulting pertinent model information is calculated to local data base mould by Hadoop computing unit in caching downloading cloud control platform Block；The data of acquisition match comparing with the data of long-range cloud platform, realize the analysis matching of pathological symptom；Cloud control is flat Platform includes NoSql distributed data base and Hadoop unit module；The data interaction of NoSql distributed data base and management terminal Functional module is connected, for storing the body detection data of patient, the diagnosis and therapy recording and its data operation result of doctor；By original Longitudinal Extension (scale up) on a single computer is changed into (scale out) extending transversely on computer cluster, I.e. a kind of big data processing technique NoSql(Not Only SQL being easily programmed in cluster environment and execution efficiency is high) distribution Formula database, can be the closely related data organization of content at a kind of structure abundant, and is shown and stored, so as to It is accessed as a unit (unit), i.e. polymerization mechanism (aggregate)；Meanwhile NoSql distributed data base is open source Database can be run in the cluster, have " non-mode " data processing function, that is, not have to modification structure definition in advance, can also Freely to add field；When running on cluster, number of nodes required for acquisition data is needed to be preferably minimized；If database In include clearly paradigmatic structure, and which data clearly polymerize, these data can be placed in a node, while also have master From copy function, there is very powerful redundancy ability；The Hadoop unit module is one by Apache fund club The distributed system infrastructure of exploitation, can reliable, efficient, telescopic mode realize data processing；Hadoop is depended on Community service, cost is relatively low, anyone can use；By extending to historical data tradition ODS, store from data The patient's pathological examination analyzed and excavated in warehouse, calls for applications and inquires；We are by patient history's body number of user Hadoop is written according to batch；Then the map () and reduce () function of MapReduce distribution off-line calculation model are used, it is right User behavior classification model construction is realized and is labelled according to the historical data and illness inspection result of patient to every kind of pathology, simultaneously History gait data when according to known disease incidence constructs the basic pathology model of related pathologies；Each gait data hair Changing carries out preliminary matching by the model of foundation；For the scene that reply data volume is huge, this system is using isolated forest (Isolation Forest) carrys out the abnormal data in detection data, achievees the purpose that reduce data and improves data accuracy.

Gait evaluating system working principle based on Multi-source Information Fusion as shown in Figure 3, here with patients with cerebral apoplexy Its concrete operating principle is illustrated for gait training: when patient starts gait training, pose acquisition module and plantar pressure Acquisition module starts to acquire the gait data of patient, is handled by Data Analysis Services module, will processing by wireless transmission Information and data afterwards is transferred to mobile phone terminal, is post-processed based on mobile phone terminal to data, then be wirelessly transferred and be sent to the AR helmet, User is trained according to the virtual game of the AR helmet, and user can also switch scene of game by mobile phone terminal or setting game is difficult Degree discriminates whether the requirement for reaching threshold crossing obstacle based on the step pitch of acquisition and lift leg height；Meanwhile user can also be in hand Machine terminal selects double training mode, and when two users carry out interaction dual training, the gait data of two users can be simultaneously It is sent to Data Analysis Services module, is uploaded to mobile phone terminal after treatment, then is uniformly sent to the AR helmet of two users, The virtual scene that two users see is identical, but distributes in different racing tracks, and user can be by way of remaining where one is to other side Racing track is placed obstacles；During training, gait evaluation result is provided according to Tinetti evaluating standard, mobile phone terminal into Row display；At the same time, the muscular fatigue information at tired information acquisition module acquisition patient's training position, after data processing Analysis obtains muscular fatigue evaluation result, in this, as the value of feedback of Rehabilitation training；All collected data can pass through Mobile phone terminal is uploaded to cloud platform, carries out storage and big data analysis in cloud platform, provides the evaluation and test knot of rehabilitation training for patient Fruit, intelligent cloud platform can also realize automatically tracking, evaluate and test and pushing for patient's gait condition.

Specific embodiment

Plantar pressure shoes are worn in patients with cerebral apoplexy before gait training, on foot, lower limb bind fatigue detecting module, waist system Nine axis attitude transducer modules wear the AR helmet on head；Open equipment power supply, allow evaluated object normal stand for a period of time, After evaluation and test terminal carries out data calibration, then trained instruction is issued by data analysing terminal；At this point, can be selected by mobile phone terminal Trained complexity and training mode, trained complexity have high, medium and low three kinds, and training mode has single player mode and double People's mode；After the completion of selecting training mode, patient initially enters gait rehabilitation training: under single training mode, patient passes through The image of AR helmet feedback carries out gait rehabilitation training, and caregiver and doctor can check gait evaluation and test knot by mobile phone terminal The fatigue conditions of fruit and patient；Under double training mode, two users can carry out interaction dual training and medical worker draws Lead the gait rehabilitation training of auxiliary patient；If muscular fatigue occurs in patient, training system can be automatically stopped, and simultaneously in mobile phone Terminal is shown；Gait data and fatigue conditions can be uploaded to cloud platform in real time, will after the data analysis and process in cloud Patient's gait condition is automatically pushed to user；After training, power supply is closed, equipment is removed from subject, placed In home, it is finally completed entire training and test process.

Above-described embodiment is only to clearly demonstrate citing of the invention, and be not the restriction to embodiment；For institute For the those of ordinary skill in category field, other various forms of variations or change can also be made on the basis of the above description It is dynamic；There is no necessity and possibility to exhaust all the enbodiments；And obvious variation extended from this or change It is dynamic to be still in the protection scope of this invention.

Claims (3)

1. the invention discloses a kind of interactive patients with cerebral apoplexy gait training and evaluating system based on Multi-source Information Fusion, packet It includes: pose acquisition module, plantar pressure acquisition module, tired information acquisition module, wireless communication module, Data Analysis Services mould Block, mobile phone terminal, the AR helmet and intelligent cloud platform；Wherein the pose acquisition module and plantar pressure acquisition module are as trouble The acquisition device of person's gait detection, is located at waist and vola；Tired information acquisition module uses sensor array design, packet Surface myoelectric sensor, flesh sound sensor and oximetry sensor are included, acquisition position is located at lower limb, adds by sleeve winding With fixation；The Data Analysis Services module by collected data carry out Kalman filtering, pose data to Eulerian angles into Row Euler's transformation accurately calculates human locomotion acceleration, passes through integral calculation walking speed of acceleration to all directions Degree, quadratic integral calculate gait displacement, and provide the evaluation result of patient's gait automatically based on standard gait index；It will test flesh The data of meat fatigue carry out windowed function processing, and the calculating of median frequency and frequency of average power is carried out to every segment data, according to The change rate of this two indexs determines the muscular fatigue stage；It is flat that gait data and fatigue data are finally uploaded to the intelligent cloud Platform, intelligent cloud platform can realize automatically tracking, evaluate and test and pushing for patient's gait condition.

2. according to pose acquisition module, plantar pressure acquisition module described in right 1 and tired information acquisition module, feature exists In: it is based on above-mentioned multi-sensor information fusion, the accurate gait information for obtaining patients with cerebral apoplexy and tired information, by wirelessly passing Contribute's primary data is sent to Data Analysis Services module, data and information and is uniformly processed, and will be located based on wireless communication module The multi-source multidimensional data transfer managed post-processes through mobile phone terminal to mobile phone terminal, re-sends to the AR helmet, be based on virtual reality technology Gait 3-D image result height visually real-time exhibition and is fed back into patient, patient can be by having designed on the AR helmet High-immersion virtual reality game carries out rehabilitation training, effectively improves the enthusiasm of patient's gait rehabilitation training；The data Analysis and processing module can receive simultaneously and handle two paths data, realize two user interaction dual trainings and medical matters people The gait rehabilitation training of member guidance auxiliary patient；Position and attitude sensor obtains walking, when plantar pressure sensor obtains walking Plantar nervous arch evaluates and tests table in conjunction with medically Tinetti, provides the gait evaluation result of subject automatically；Wherein Index in Tinetti evaluation and test table include: bipod relative position, the time difference of bipod stress, plantar pressure variation it is whether equal Whether even, paces are uniform；Tired information acquisition module obtains lower limb muscles fatigue characteristic, and muscular fatigue rank is fed back to simultaneously Mobile phone terminal and the AR helmet, if reaching the default grade of muscular fatigue, training system is automatically stopped, and patient is prompted to stop Breath.

3. including management terminal and cloud control platform according to cloud platform described in right 1；Management terminal includes hospital terminal and use Family terminal two parts are realized using C# language Programming, for showing the data of cloud platform；The data of cloud platform, which are analyzed, to be used BDP data analysis tool, management terminal establish connection by data interaction function and cloud control platform, and downloading is counted by Hadoop The resulting model of unit is calculated to local data base；The data of acquisition match comparing with the data of cloud platform, realize pathology The analysis of symptom matches；Cloud control platform includes NoSql distributed data base and Hadoop unit module；The NoSql points Cloth database is connected with the data interaction functional module of management terminal, for storing the body detection data of patient, doctor Diagnosis and therapy recording and data operation result；The Hadoop depends on community service, and cost is relatively low, by the patient history of user Hadoop, map () and reduce () function based on MapReduce distribution off-line model, to user's row is written in batch data It for classification model construction, realizes and is labelled according to the history body data and illness inspection result of patient to every kind of pathology, while root History gait data when according to known disease incidence constructs the basic pathology model of related pathologies and borrows when each gait changes Above-mentioned basic pathology model is helped to be matched；For the scene that reply data volume is huge, this system is using isolated forest (Isolation Forest) carrys out the abnormal data in detection data, achievees the purpose that reduce data and improves data accuracy.