CN110313918A - A kind of gait phase recognition methods and system based on plantar pressure - Google Patents

A kind of gait phase recognition methods and system based on plantar pressure Download PDF

Info

Publication number
CN110313918A
CN110313918A CN201910644477.6A CN201910644477A CN110313918A CN 110313918 A CN110313918 A CN 110313918A CN 201910644477 A CN201910644477 A CN 201910644477A CN 110313918 A CN110313918 A CN 110313918A
Authority
CN
China
Prior art keywords
gait
phase
subject
gait phase
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910644477.6A
Other languages
Chinese (zh)
Inventor
陈晓
李茂辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Quartermaster Engineering Technology Institute of Systems Engineering Academy of Military Sciences
Original Assignee
Institute of Quartermaster Engineering Technology Institute of Systems Engineering Academy of Military Sciences
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Quartermaster Engineering Technology Institute of Systems Engineering Academy of Military Sciences filed Critical Institute of Quartermaster Engineering Technology Institute of Systems Engineering Academy of Military Sciences
Priority to CN201910644477.6A priority Critical patent/CN110313918A/en
Publication of CN110313918A publication Critical patent/CN110313918A/en
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/1036—Measuring load distribution, e.g. podologic studies
    • A61B5/1038—Measuring plantar pressure during gait
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/112—Gait analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1121—Determining geometric values, e.g. centre of rotation or angular range of movement
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A61B5/6804—Garments; Clothes
    • A61B5/6807—Footwear
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/30—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Physics & Mathematics (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Molecular Biology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Dentistry (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Physiology (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Databases & Information Systems (AREA)
  • Geometry (AREA)
  • Data Mining & Analysis (AREA)
  • General Business, Economics & Management (AREA)
  • Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The present invention relates to a kind of gait phase recognition methods and system based on plantar pressure, which is characterized in that method includes the following steps: 1) obtaining several ground reaction force data of several subject's normal gaits；2) the walking movement data and joint angles data of each subject are obtained, and establish the gait phase of subject；3) according to the walking movement data of acquisition and the gait phase of foundation, hidden Markov model is constructed；4) using several ground reaction force data of several subjects as input value, hidden Markov model is solved, identifies the gait phase of each subject, the present invention can be widely applied in gait detection technique field.

Description

A kind of gait phase recognition methods and system based on plantar pressure
Technical field
The present invention relates to a kind of gait phase recognition methods and system based on plantar pressure, belong to gait detection technique In field.
Background technique
With biped robot, the development of artificial limb and for the rehabilitation training demand of the wounded, online step State phase recognition methods seem extremely important.There are many signals for judging gait phase, in general, adopts when studying gait phase Signal is mainly surface muscle electric signal, but the bio signal of human body is faint, unstable, therefore acquisition, the place of signal It manages highly difficult.
It is in the prior art the motion information based on upper and lower extremities for the identification of gait phase, passes through angle information processing system System obtains shoulder joint and kneed angular speed to judge gait phase, but the sensing device that this method needs is more, letter Number processing is also more difficult.Two levels that the prior art also uses one layer of decision tree classifier to combine with one layer of linear classifier point Class method distinguishes static and dynamic using decision tree classifier first, then carries out linear classification identification, this method to it respectively Main purpose be the lower limb motion mode based on gait phase identification, be more focused on and motor pattern identified, for How to judge that gait phase research is less.Currently, due to the simplification and practicability that measure foot pressure, researcher has been set Shoe style sensor is counted out, and has proposed several systems for detecting gait phase, including for wireless gait analysis and in fact When the shoes that feed back integrate sensing system and gait phase detection system etc., however, these systems by gait phase be detected as from The event of dissipating is inaccurate in actual walking pattern movement.
Although from the foregoing, it will be observed that the prior art never studied in face of the identification of gait phase by Tongfang, and making phase The contribution answered, but the method for the prior art has very defect, therefore, how the detection gait phase of continuously smooth, and can The problem of detection recognizes the data of abnormal gait, is urgent need to resolve in this field.
Summary of the invention
In view of the above-mentioned problems, capableing of the detection gait phase of continuously smooth the object of the present invention is to provide one kind and can examine Survey the gait phase recognition methods and system based on plantar pressure for recognizing abnormal gait data.
To achieve the above object, the present invention takes following technical scheme: a kind of gait phase identification based on plantar pressure Method, which comprises the following steps: 1) obtain several ground reaction force data of several subject's normal gaits； 2) the walking movement data and joint angles data of each subject are obtained, and establish the gait phase of subject；3) basis obtains The gait phase of the walking movement data and foundation that take constructs hidden Markov model；4) by several ground of several subjects Reaction force data solve hidden Markov model, identify the gait phase of each subject as input value.
Further, the detailed process of the step 1) are as follows: the 1.1) difference on pressure insole from heel to forefoot Pressure sensor is respectively set in position；1.2) several ground for obtaining several subject's normal gaits respectively by pressure insole are anti- Ground reaction force data are approximately Gaussian Profile by force data:
Wherein, p (y/ μ, ∑) is the probability of observation y appearance in the case of given characteristic parameter μ and ∑, y, μ and ∑ difference For the signal vector of GRF data, mean vector and covariance matrix；N is the quantity of pressure sensor.
Further, the detailed process of the step 2) are as follows: 2.1) when obtaining the ground reaction force data of subject, obtain Take the walking movement data of corresponding subject and the joint angles data of lower limb；2.2) according to the walking movement data of acquisition and Joint angles data establish the gait phase of subject.
Further, the gait phase of the subject includes starting the phase of contacting to earth, the load-bearing reaction phase, intergrade of standing, standing The terminal phase swings early period and shaking peroid, wherein starts preceding 1~5% that the phase of contacting to earth is the entire gait cycle of subject, load-bearing is anti- Answering the phase is the 6~10% of the entire gait cycle of subject, and intergrade of standing is the 11~35% of the entire gait cycle of subject, is stood The vertical terminal phase is the 36~55% of the entire gait cycle of subject, swing that early period is the entire gait cycle of subject 56~ 65%, shaking peroid is the 66~100% of the entire gait cycle of subject.
Further, the detailed process of the step 3) are as follows: 3.1) according to the walking movement data of acquisition and the gait of foundation Phase determines the pressure characteristic parameter of each ground reaction force data；3.2) it based on determining pressure characteristic parameter, constructs hidden Markov model.
Further, the detailed process of the step 3.1) are as follows: 3.1.1) according to the walking movement data of acquisition, it marks respectively The gait phase of each ground reaction force data, wherein if including in the gait phase of each ground reaction force data Mean vector μ and covariance matrix ∑ value；3.1.2 all mean vector μ and association in each ground reaction force data) are calculated Conditional probability of the variance matrix ∑ in marked gait phase；3.1.3) the mean vector μ of alternative condition maximum probability and association side Pressure characteristic parameter of the poor matrix ∑ as corresponding ground reaction force data:
Wherein, p is conditional probability；ytFor the observation of time t；qtFor the state of time t；p(yt|qt=i) it is time t In the state of event ytThe probability of generation；I is contact to earth successively indicating gait phase by 1~6 phase, load-bearing reaction phase, standing Intergrade, standing terminal phase swing one of them in early period and shaking peroid.
Further, the hidden Markov model are as follows:
Wherein, p (qt| it y) is posterior probability；p(y|qt) it is the probability that a certain event y occurs under given hidden state；p (qt) it is the probability for giving hidden state；P (y) is the probability that event y occurs；T is total experimental period；Parameter alpha (qt) and β (qt) It is defined as follows:
α(qt)=p (y0..., yt|qt)
β(qt)=p (yt+1..., yT|qt)
The α and β of each time step are obtained by following recursion equations:
Further, when solving hidden Markov model in the step 4), the hidden state transfer matrix of normal gait is in Now it is three diagonal shapes:
Wherein, diagonal entry indicate in gait phase from transition probability, other nonzero elements indicate adjacent gait phase Transition probability between position.
Further, the walking movement data of the subject are obtained using video camera, the joint angle of subject's lower limb Degree is obtained according to using encoder and dipmeter.
A kind of gait phase identifying system based on plantar pressure characterized by comprising ground reaction force data obtain Modulus block, for obtaining several ground reaction force data of several subject's normal gaits；Gait phase establishes module, is used for The walking movement data and joint angles data of each subject are obtained, and establish the gait phase of subject；Model construction mould Block, for constructing hidden Markov model according to the walking movement data of acquisition and the gait phase of foundation；Gait phase identification Module, for solving hidden Markov model, identification using several ground reaction force data of several subjects as input value The gait phase of each subject.
The invention adopts the above technical scheme, which has the following advantages: if the 1, present invention passes through several subjects' Dry GRF data, are approximately Gaussian Profile, and extract pressure characteristic parameter, establish Hidden Markov Model, if will acquire Dry GRF data solve Hidden Markov Model as input value, can identify, realize to the gait phase of subject To the accurate judgement of gait phase in movement.2, the present invention in subject walking movement data and joint angles data acquisition It is simple with processing, and using the pressure insole for being provided with several pressure sensors, the GRF data of subject are continuously put down It detected slidingly, be able to detect identification abnormal gait data, can be widely applied in gait detection technique field.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention；
Fig. 2 is the gait phase schematic diagram in the present invention；
Fig. 3 is the Gaussian distribution curve figure in the present invention at pressure sensor seven；
Fig. 4 is the Gaussian distribution curve figure of the GRF data of pressure sensor one and pressure sensor seven in the present invention；
Fig. 5 is the transition schematic diagram in the present invention between gait phase；
Fig. 6 is the element schematic diagram of the transition matrix of three gait phases in the present invention.
Specific embodiment
Come to carry out detailed description to the present invention below in conjunction with attached drawing.It should be appreciated, however, that attached drawing has been provided only more Understand the present invention well, they should not be interpreted as limitation of the present invention.
Hidden Markov model (HMM) is a kind of statistical model, suitable for modeling to continuous data.In form, HMM It is defined as the random process of a dual insertion, with unobservable undeniable process (it is hiding), still It can only be observed by generating another group of random process of observation sequence.This means that the state of data generating procedure behind is hidden Hiding, it can be inferred by observing.
Therefore, as shown in Figure 1, the gait phase recognition methods provided by the invention based on plantar pressure, including following step It is rapid:
1) seven positions on insole from heel to forefoot are respectively set pressure sensor and constitute pressure insole, tool Body are as follows:
There are six hidden states, i.e. six gait phases in gait motion, as shown in Fig. 2, circled representative pressure passes Sensor successively uses one to seven to indicate from heel to forefoot, can pass through seven pressure on pressure insole for convenience of studying Continuous ground reaction force (GRF) data that sensor issues are observed.
2) six GRF data of six subject's normal gaits are obtained respectively by pressure insole, wherein subject is not There is the subject of any known obstacle, normal gait is walking speed in normal gait velocity interval (4~5km/h), and is walked Scanning frequency degree will not influence the gait of GRF amplitude, include the GRF data of seven pressure sensors acquisition, tool in every GRF data Body are as follows:
For the mean value and variance for obtaining Gaussian Profile, need to collect from six subjects (age: 28.3 scholars 1.03, often Position subject 6 step) normal gait 36 GRF data.And subject is required not have any known obstacle, and with normal walking Speed 40~50m of walking on level land.Statistics indicate that walking speed is not when walking speed is in normal gait velocity interval The size of GRF can be significantly affected, and weight is a main factor, i.e. the size of GRF data increase proportional to weight.Cause This, the size of GRF data is by weight normalized, and the time span of GRF data is by stride Percentage Criterion, stride percentage It is distinguished by heel contact.
As shown in Figure 3, it is shown that the Gaussian Profile under a stride at big toe, the heavy line in bottom plane are GRF number According to average value, the upper and lower part dotted line in bottom plane be with the standard deviation of average value 1.96 (Gaussian Profile 95% confidence interval).Since the freedom degree of GRF data is very big, t distribution is approximately Gaussian Profile.Gray line shows each stride hundred Divide the Gaussian Profile of ratio.
It is approximately Gauss point by GRF data since the GRF data of seven pressure sensors on pressure insole are measured simultaneously 36 GRF data in total of cloth, 6 subjects follow following Gaussian Profile:
Wherein, p (y/ μ, ∑) is the probability of observation y appearance in the case of given characteristic parameter μ and ∑, y, μ and ∑ difference For the signal vector of GRF data, mean vector and covariance matrix, n are the quantity of pressure sensor.In this case, by It is 7, n=7 in the observation number in given time, ∑ is one 7 × 7 symmetrical matrix.
3) the walking movement data and joint angles data of each subject are obtained, and establish the gait phase of subject, Specifically:
3.1) when obtaining the GRF data of subject, the walking movement data of corresponding subject are obtained by video camera, and The joint angles data of corresponding subject's lower limb (hip, knee and ankle) are obtained by encoder and dipmeter.
3.2) according to the walking movement data of acquisition and joint angles data, the gait phase of subject is established, wherein Gait phase includes starting the phase of contacting to earth (IC), load-bearing reaction phase (LR), standing intergrade (MS), the standing terminal phase (TS), swing Early period (PS) and shaking peroid (SW), start the phase of contacting to earth be the entire gait cycle of subject preceding 1~5%, load-bearing react the phase be by The 6~10% of the entire gait cycle of examination person, standing intergrade are the 11~35% of the entire gait cycle of subject, standing terminal phase It is the 36~55% of the entire gait cycle of subject, swings 56~65% that early period is the entire gait cycle of subject, shaking peroid It is the 66~100% of the entire gait cycle of subject.
By taking pressure sensor one and seven as an example, as shown in Figure 4, it is shown that the pressure sensor one and seven in pressure insole Normal GRF frequency band (solid line and fine dotted line) and the normal GRF data marked by gait phase (thick dashed line).For convenience of research, adopt It successively indicates to start the phase of contacting to earth with 1~6, the load-bearing reaction phase, intergrade of standing, the standing terminal phase, swing early period and shaking peroid.
4) according to the walking movement data of acquisition and the gait phase of foundation, the pressure characteristic ginseng of every GRF data is determined Number, specifically:
4.1) according to the walking movement data of acquisition, the gait phase of every GRF data is marked respectively.
It include several mean vector μ and covariance matrix ∑ value in the gait phase that one marks, for example, starting to contact to earth Interim includes 5 different mean vector μ and covariance matrix ∑, because 1~5% normal GRF data has been used to be used to divide Class.
4.2) all mean vector μ and covariance matrix ∑ are calculated in every GRF data in the item of marked gait phase Part probability.
4.3) the mean vector μ of alternative condition maximum probability and covariance matrix ∑ are special as the pressure of corresponding GRF data Levy parameter.
The value of mean vector μ and covariance matrix ∑ needs to classify by gait phase, for find mean vector μ and The most probable conditional probability of covariance matrix ∑ is calculating all mean vector μ and covariance matrix ∑ in marked gait After the conditional probability of phase, pressure of the mean vector μ and covariance matrix ∑ of alternative condition maximum probability as corresponding GRF data Power characteristic parameter, is shown below:
Wherein, p is conditional probability；ytFor the observation of time t；qtFor the state of time t；p(yt|qt=i) it is time t In the state of event ytThe probability of generation；I is contact to earth successively indicating gait phase by 1~6 phase, load-bearing reaction phase, standing Intergrade, standing terminal phase swing one of them in early period and shaking peroid.
5) based on determining pressure characteristic parameter, hidden Markov model is constructed, specifically:
Based on determining pressure characteristic parameter, the equation by Bayes rule guidance is constructed hidden Markov model It is as follows:
Wherein, p (qt| it y) is posterior probability；p(y|qt) it is that a certain event y (i.e. observation) occurs under given hidden state Probability；p(qt) it is the probability for giving hidden state；P (y) is the probability that event y occurs；T is total experimental period；Parameter alpha (qt) With β (qt) it is defined as follows and states shown in formula (4) and (5):
α(qt)=p (y0..., yt|qt) (4)
β(qt)=p (yt+1..., yT|qt) (5)
The α and β of each time step can be obtained by following recursion equations:
6) using observation several GRF data of data sequence, that is, several subjects as input value, hidden Markov mould is solved Type, output generate the probability distribution of each subject's hidden state sequence, and according to the general of each subject's hidden state sequence Rate distribution, identifies the gait phase of each subject, specifically:
Hidden state (transition or itself transition between i.e. adjacent gait phase) may infer that as posterior probability p (qt| y).The reasoning problems of HMM are related to observe data sequence as input, and export the probability distribution for generating base state.Due to Dependence between hidden state, this problem are substantially complicated, but can easily pass through simple recurrence side Journey solves.
Due to there is six kinds of hidden states in hidden Markov model, transfer matrix has six rows and six column.In normal gait, Gait phase be successively rendered as shown in Figure 5 in solid line, therefore, the hidden state transfer matrix of normal gait should be rendered as (hidden state transfer matrix indicates the sequence between hidden state, that is, gait phase to three diagonal shapes shown in following formula, in other words, walks The probability that state phase 1 is transformed to gait phase 5 is 0, but the probability for being transformed to gait phase 2 is very big, is a12, which is HMM Constantly iterative solution calculates gained):
Wherein, diagonal entry indicate in gait phase from transition probability, other nonzero elements (a12, a21, a23, a32 Deng) indicate transition probability between adjacent gait phase.Diagonal entry is almost one, because number state conversion is mostly From conversion.In addition, 0 in formula (8) indicates do not have transition between non-adjacent gait phase.
As shown in fig. 6, depicting the three gait phase load-bearing detected reaction phases, stand intergrade and standing terminal Phase is expressed as diagonal term (a from conversion22, a33And a44), the conversion between them is expressed as aI+1, iAnd aI, i+1(a23, a32With a34, a43)。
The a of normal gait34And a43It is almost nil, because the state conversion from SW to IC is very prominent due to heel strike So.By taking pressure sensor one and seven as an example, normal GRF data are from the GRF data for constructing normal GRF frequency band.All GRF Data are in 95% confidence interval of normal GRF frequency band.Gait phase, that is, posterior probability, by formula (3), according to extraction Pressure characteristic parameter, the GRF data for inputting acquisition are estimated.
Gait phase successively occurs with correctly sequence, and the transfer matrix of estimation is as shown in table 1 below.It is pair from transition probability Angle element is almost 1, and state conversion occurs over just between adjacent states.Therefore, it has three diagonal forms as shown in formula (8) Shape.
Table 1: the transfer matrix of estimation
IC LR MS TS PS SW
IC 0.9824 0.0176 0.0000 0.0000 0.0000 0.0000
LR 0.0005 0.9704 0.0287 0.0000 0.0000 0.0000
MS 0.0000 0.0074 0.9807 0.0114 0.0000 0.0000
TS 0.0000 0.0000 0.0043 0.8923 0.1054 0.0000
PS 0.0000 0.0000 0.0000 0.0210 0.9275 0.0506
SW 0.0000 0.0000 0.0000 0.0000 0.0000 0.9998
Based on the above-mentioned gait phase recognition methods based on plantar pressure, the present invention also provides a kind of based on plantar pressure Gait phase identifying system, comprising:
Ground reaction force data acquisition module, for obtaining several ground reaction forces of several subject's normal gaits Data；Gait phase establishes module, for obtaining the walking movement data and joint angles data of each subject, and establish by The gait phase of examination person；Model construction module, for constructing hidden according to the walking movement data of acquisition and the gait phase of foundation Markov model；Gait phase identification module, for using several ground reaction force data of several subjects as input Value solves hidden Markov model, identifies the gait phase of each subject.
The various embodiments described above are merely to illustrate the present invention, wherein the structure of each component, connection type and manufacture craft etc. are all It can be varied, all equivalents and improvement carried out based on the technical solution of the present invention should not exclude Except protection scope of the present invention.

Claims (10)

1. a kind of gait phase recognition methods based on plantar pressure, which comprises the following steps:
1) several ground reaction force data of several subject's normal gaits are obtained；
2) the walking movement data and joint angles data of each subject are obtained, and establish the gait phase of subject；
3) according to the walking movement data of acquisition and the gait phase of foundation, hidden Markov model is constructed；
4) using several ground reaction force data of several subjects as input value, hidden Markov model is solved, identification is every The gait phase of one subject.
2. a kind of gait phase recognition methods based on plantar pressure as described in claim 1, which is characterized in that the step 1) detailed process are as follows:
1.1) pressure sensor is respectively set in the different location on pressure insole from heel to forefoot；
1.2) several ground reaction force data of several subject's normal gaits are obtained respectively by pressure insole, ground is anti- Force data is approximately Gaussian Profile:
Wherein, p (y/ μ, ∑) is the probability of observation y appearance in the case of given characteristic parameter μ and ∑, and y, μ and ∑ are respectively GRF Signal vector, mean vector and the covariance matrix of data；N is the quantity of pressure sensor.
3. a kind of gait phase recognition methods based on plantar pressure as described in claim 1, which is characterized in that the step 2) detailed process are as follows:
2.1) when obtaining the ground reaction force data of subject, the walking movement data and lower limb for corresponding to subject are obtained Joint angles data；
2.2) according to the walking movement data of acquisition and joint angles data, the gait phase of subject is established.
4. a kind of gait phase recognition methods based on plantar pressure as claimed in claim 3, which is characterized in that described tested The gait phase of person includes starting the phase of contacting to earth, the load-bearing reaction phase, intergrade of standing, the standing terminal phase, swinging early period and shaking peroid, Wherein, start preceding 1~5% that the phase of contacting to earth is the entire gait cycle of subject, load-bearing reacts the phase for the entire gait cycle of subject 6~10%, stand intergrade be the entire gait cycle of subject 11~35%, the standing terminal phase be the entire gait of subject The 36~55% of period swing 56~65% that early period is the entire gait cycle of subject, and shaking peroid is the entire gait of subject The 66~100% of period.
5. a kind of gait phase recognition methods based on plantar pressure as claimed in claim 4, which is characterized in that the step 3) detailed process are as follows:
3.1) according to the walking movement data of acquisition and the gait phase of foundation, the pressure of each ground reaction force data is determined Characteristic parameter；
3.2) based on determining pressure characteristic parameter, hidden Markov model is constructed.
6. a kind of gait phase recognition methods based on plantar pressure as claimed in claim 5, which is characterized in that the step 3.1) detailed process are as follows:
3.1.1) according to the walking movement data of acquisition, the gait phase of each ground reaction force data is marked respectively, wherein It include several mean vector μ and covariance matrix ∑ value in the gait phase of each ground reaction force data；
3.1.2 all mean vector μ and covariance matrix ∑) are calculated in each ground reaction force data in marked gait phase The conditional probability of position；
3.1.3) the mean vector μ of alternative condition maximum probability and covariance matrix ∑ are as corresponding ground reaction force data Pressure characteristic parameter:
Wherein, p is conditional probability；ytFor the observation of time t；qtFor the state of time t；p(yt|qt=i) be time t state Lower event ytThe probability of generation；I be by 1~6 successively indicate gait phase contact to earth the phase, load-bearing react the phase, stand intergrade, The standing terminal phase swings one of them in early period and shaking peroid.
7. a kind of gait phase recognition methods based on plantar pressure as claimed in claim 5, which is characterized in that the hidden horse Er Kefu model are as follows:
Wherein, p (qt| it y) is posterior probability；p(y|qt) it is the probability that a certain event y occurs under given hidden state；p(qt) be The probability of given hidden state；P (y) is the probability that event y occurs；T is total experimental period；Parameter alpha (qt) and β (qt) definition It is as follows:
α(qt)=p (y0..., yt|qt)
β(qt)=p (yt+1..., yT|qt)
The α and β of each time step are obtained by following recursion equations:
8. a kind of gait phase recognition methods based on plantar pressure as claimed in claim 5, which is characterized in that the step 4) when solving hidden Markov model in, the hidden state transfer matrix of normal gait is rendered as three diagonal shapes:
Wherein, diagonal entry indicate in gait phase from transition probability, other nonzero elements indicate adjacent gait phase it Between transition probability.
9. a kind of gait phase recognition methods based on plantar pressure as claimed in any one of claims 1 to 8, feature exist In the walking movement data of the subject are obtained using video camera, and the joint angles data of subject's lower limb are using volume Code device and dipmeter obtain.
10. a kind of gait phase identifying system based on plantar pressure characterized by comprising
Ground reaction force data acquisition module, for obtaining several ground reaction force numbers of several subject's normal gaits According to；
Gait phase establishes module, for obtaining the walking movement data and joint angles data of each subject, and establish by The gait phase of examination person；
Model construction module, for constructing hidden Markov mould according to the walking movement data of acquisition and the gait phase of foundation Type；
Gait phase identification module, for solving hidden using several ground reaction force data of several subjects as input value Markov model identifies the gait phase of each subject.
CN201910644477.6A 2019-07-17 2019-07-17 A kind of gait phase recognition methods and system based on plantar pressure Pending CN110313918A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910644477.6A CN110313918A (en) 2019-07-17 2019-07-17 A kind of gait phase recognition methods and system based on plantar pressure

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910644477.6A CN110313918A (en) 2019-07-17 2019-07-17 A kind of gait phase recognition methods and system based on plantar pressure

Publications (1)

Publication Number Publication Date
CN110313918A true CN110313918A (en) 2019-10-11

Family

ID=68123824

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910644477.6A Pending CN110313918A (en) 2019-07-17 2019-07-17 A kind of gait phase recognition methods and system based on plantar pressure

Country Status (1)

Country Link
CN (1) CN110313918A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110638449A (en) * 2019-09-30 2020-01-03 福州大学 Muscle quantitative analysis method based on mechanical work
CN110786862A (en) * 2019-11-05 2020-02-14 哈工大机器人湖州国际创新研究院 Gait cycle identification method based on torque and angle feedback fusion
CN111528852A (en) * 2020-05-08 2020-08-14 广州市懒洋洋慢运动用品有限公司 Gait detection method and system based on plantar pressure
CN113658707A (en) * 2021-08-26 2021-11-16 华南理工大学 Foot varus angle detection modeling method and system

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1808414A (en) * 2004-12-06 2006-07-26 索尼株式会社 Method and apparatus for learning data, method and apparatus for recognizing data, method and apparatus for generating data and computer program
CN101984454A (en) * 2010-11-19 2011-03-09 杭州电子科技大学 Multi-source multi-characteristic information fusion method based on data drive
US20110313714A1 (en) * 2007-11-28 2011-12-22 Erez Lieberman Determining Postural Stability
US20140088729A1 (en) * 2005-03-31 2014-03-27 Massachusetts Institute Of Technology Powered Ankle-Foot Prosthesis
CN104021573A (en) * 2014-06-13 2014-09-03 哈尔滨工程大学 Human movement classification and identification method based on joint pose angles
CN106887115A (en) * 2017-01-20 2017-06-23 安徽大学 Old people falling monitoring device and falling risk assessment method
CN108836346A (en) * 2018-04-16 2018-11-20 大连理工大学 A kind of Human Body Gait Analysis method and system based on inertial sensor

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1808414A (en) * 2004-12-06 2006-07-26 索尼株式会社 Method and apparatus for learning data, method and apparatus for recognizing data, method and apparatus for generating data and computer program
US20140088729A1 (en) * 2005-03-31 2014-03-27 Massachusetts Institute Of Technology Powered Ankle-Foot Prosthesis
US20110313714A1 (en) * 2007-11-28 2011-12-22 Erez Lieberman Determining Postural Stability
CN101984454A (en) * 2010-11-19 2011-03-09 杭州电子科技大学 Multi-source multi-characteristic information fusion method based on data drive
CN104021573A (en) * 2014-06-13 2014-09-03 哈尔滨工程大学 Human movement classification and identification method based on joint pose angles
CN106887115A (en) * 2017-01-20 2017-06-23 安徽大学 Old people falling monitoring device and falling risk assessment method
CN108836346A (en) * 2018-04-16 2018-11-20 大连理工大学 A kind of Human Body Gait Analysis method and system based on inertial sensor

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
M. M. DE ROSSI 等: "《Gait Segmentation Using Bipedal Foot Pressure Patterns》", 《THE FOURTH IEEE RAS/EMBS INTERNATIONAL CONFERENCE ON BIOMEDICAL ROBOTICS AND BIOMECHATRONICS》 *
张向刚等: "《一种基于隐马尔科夫模型的步态识别算法》", 《计算机科学》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110638449A (en) * 2019-09-30 2020-01-03 福州大学 Muscle quantitative analysis method based on mechanical work
CN110786862A (en) * 2019-11-05 2020-02-14 哈工大机器人湖州国际创新研究院 Gait cycle identification method based on torque and angle feedback fusion
CN111528852A (en) * 2020-05-08 2020-08-14 广州市懒洋洋慢运动用品有限公司 Gait detection method and system based on plantar pressure
CN111528852B (en) * 2020-05-08 2023-07-11 广州市懒洋洋慢运动用品有限公司 Gait detection method and system based on plantar pressure
CN113658707A (en) * 2021-08-26 2021-11-16 华南理工大学 Foot varus angle detection modeling method and system

Similar Documents

Publication Publication Date Title
CN110313918A (en) A kind of gait phase recognition methods and system based on plantar pressure
US11660024B2 (en) Method and system for analyzing human gait
US10314520B2 (en) System and method for characterizing biomechanical activity
Zhang et al. Accurate ambulatory gait analysis in walking and running using machine learning models
Wang et al. Recognizing human daily activities from accelerometer signal
Zhou et al. Validation of an IMU gait analysis algorithm for gait monitoring in daily life situations
Rast et al. Systematic review on the application of wearable inertial sensors to quantify everyday life motor activity in people with mobility impairments
Dubois et al. A gait analysis method based on a depth camera for fall prevention
Song et al. Speed estimation from a tri-axial accelerometer using neural networks
Jeong et al. Classification of three types of walking activities regarding stairs using plantar pressure sensors
Li et al. Pre-impact fall detection based on a modified zero moment point criterion using data from Kinect sensors
CN107330249A (en) A kind of Parkinsonian symptoms area of computer aided method of discrimination based on KINECT skeleton datas
CN106295544A (en) A kind of unchanged view angle gait recognition method based on Kinect
Papageorgiou et al. Hidden markov modeling of human normal gait using laser range finder for a mobility assistance robot
Brahms et al. Stride length determination during overground running using a single foot-mounted inertial measurement unit
Ladetto et al. Human walking analysis assisted by DGPS
CN112115923A (en) Multichannel time sequence gait analysis algorithm based on direct feature extraction
CN115346272A (en) Real-time tumble detection method based on depth image sequence
KR20210046121A (en) Apparatus and method for identify patients with parkinson's disease and patients with podarthritis by performing neural network analysis by various detection information
Barralon et al. Classification of daily physical activities from a single kinematic sensor
Scheltinga et al. Estimating 3D ground reaction forces in running using three inertial measurement units
Majumder et al. A wireless IoT system towards gait detection in stroke patients
KR102194313B1 (en) Apparatus and method for identifying individuals by performing neural network analysis for various detection information
KR102302719B1 (en) Apparatus and method for classification of gait type by performing neural network analysis for various detection information
CN111539507B (en) Recovery movement speed calculation model parameter identification method based on particle swarm optimization algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20191011

RJ01 Rejection of invention patent application after publication