CN104008398A - Gait classification method based on multi-sensor information fusion - Google Patents
Gait classification method based on multi-sensor information fusion Download PDFInfo
- Publication number
- CN104008398A CN104008398A CN201410257800.1A CN201410257800A CN104008398A CN 104008398 A CN104008398 A CN 104008398A CN 201410257800 A CN201410257800 A CN 201410257800A CN 104008398 A CN104008398 A CN 104008398A
- Authority
- CN
- China
- Prior art keywords
- gait
- patient
- pressure
- value
- stage
- 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.)
- Granted
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention relates to a gait classification method based on multi-sensor information fusion. The gait classification method includes the following steps that (1), plantar pressure information and ankle angle information in the walking process of multiple patients are collected; (2) according to the obtained plantar pressure information, gait stages of the patients are analyzed, the gait stages are divided into the stages of touching the ground with the feet and the stages of swinging the legs, and one gait cycle includes one stage of touching the ground with one foot and one stage of swinging the leg of the same foot of each patient; (3) a characteristic value of each gait cycle is set, so that gait characteristics of all the patients in the walking process are represented; (4), gait cluster analysis is performed on all the characteristic values of all the gait cycles of all the patients through a spectral clustering algorithm, so that the patients with the different gait characteristics are divided into different classifications. The patients are objectively classified, so that reference is provided for rehabilitation training and treatment of the patients, and a doctor can adopt different treatment modes and training intensities for the patients of the different classifications conveniently. The gait classification method can be widely used in the fields of gait analysis and medical rehabilitation.
Description
Technical field
The present invention relates to a kind of gait analysis method, particularly the Approach for Gait Classification based on multi-sensor information fusion (MSDF, Multi~sensor Data Fusion) about one.
Background technology
The change of gait feature information has reflected the variation of human physiological functions, in rehabilitation process, obtain in time accurate believable gait feature information, and for a long time gait feature is carried out to monitoring and evaluation, the Clinics and Practices of some diseases is had to important directive significance, for example rehabilitation training to paralytic's walking step state, the gait rectificative training to parkinsonism patient, the rehabilitation guide to tears of anterior cruciate ligament patient etc.
In medical clinic applications field, traditional gait analysis method mainly relies on specialist to observe patient's gait feature, or allows patient fill in meter, expects to obtain gait assessment information accurately at present.But because this method exists very large subjective factor, and the evaluation of gait feature information requires to have higher accuracy, accuracy and objectivity, has brought negative effect therefore just to later medical diagnosis on disease and the rehabilitation of patient.
Summary of the invention
For the problems referred to above, object of the present invention provides a kind of and objectively the patient of different gait features is divided into different classifications, so that doctor adopts different, the Approach for Gait Classification based on multi-sensor information fusion of recovery training method targetedly for different patients' rehabilitation.
For achieving the above object, the present invention takes following technical scheme: a kind of Approach for Gait Classification based on multi-sensor information fusion, and it comprises the following steps: plantar pressure information and ankle joint angle information when 1) gathering several patients and walking; 2) according to the plantar pressure information analysis patient's that obtains gait phase, this gait phase is divided into foot contacts to earth stage and swinging kick stage, and same pin of patient is a gait cycle through contact to earth stage and swinging kick stage of foot; 3) eigenwert of setting gait cycle, the gait feature while walking to characterize each patient; 4) adopt spectral clustering to carry out gait cluster analysis for each eigenwert in each gait cycle of each patient, so that the patient of different gait features is divided into different classifications.
Carrying out described step 1) time, adopt array pressure transducer is distributed in to patient vola, plantar pressure information while walking to gather patient, plantar pressure information comprises plantar pressure value and the pressure Two dimensional Distribution situation in vola, wherein plantar pressure value comprises vola regional force value size, and pressure comprises heel pressure, metatarsal pressure, inner side heel pressure, outside heel pressure, inner side metatarsal pressure and outside metatarsal pressure in the Two dimensional Distribution situation in vola; Ankle joint angle information while adopting Inertial Measurement Unit to gather patient and walk, ankle joint angle information comprises ankle joint angle value.
Carrying out described step 2) time, specifically comprise: 1. choose gait event threshold value, this gait event threshold value is slightly larger than the patient vola plantar pressure value sum that all pressure transducers record when unsettled; 2. contrast the plantar pressure value and the gait event threshold value that record, judge patient's gait phase; In the time that the plantar pressure value recording is greater than gait event threshold value, represent the foot event of contacting to earth has occurred, illustrate that patient enters foot and contacts to earth the stage; Otherwise, in the time that the plantar pressure value recording is less than gait event threshold value, represent the liftoff event of foot has occurred, illustrate that patient enters the swinging kick stage.
Carrying out described step 3) time, the eigenwert of gait cycle comprises foot contact to earth eigenwert and the eigenwert in swinging kick stage in stage: 1. the contact to earth eigenwert in stage of foot comprises pressure characteristic value and temporal characteristics value, and wherein pressure characteristic value comprises: H/M: the ratio of heel pressure and metatarsal pressure; HM/HL: the ratio of inner side heel pressure and outside heel pressure; MM/ML: the ratio of inner side metatarsal pressure and outside metatarsal pressure; Temporal characteristics value comprises: HD/SD: the ratio of the shared vola of heel contact time contact time; MD/SD: the ratio of the shared vola of metatarsal contact time contact time; 2. the eigenwert in swinging kick stage comprises angle character value, this angle character value is designated as SW/ST: the contact to earth ratio of ankle joint angle average in stage of ankle joint angle average and the foot in swinging kick stage, and wherein ankle joint angle average is that the ankle joint angle value recording is averaged; Gait feature while adopting H/M, HM/HL, MM/ML, HD/SD, MD/SD and SW/ST to describe patient to walk, distinguishes different patients according to each patient's different gait feature.
Carrying out described step 4) time, common N step sample when gathering several patients and walking, wherein N is positive integer, each step of patient is a gait cycle, carry out spectral clustering for each eigenwert in gait cycle, it comprises the following steps: 1. adopt Euclidean distance formula to calculate any two step sample x in N step sample
ik, x
jkbetween distance:
wherein, x
ik, x
jkbe proper vector, the dimension of p representation feature vector, k=1 represents to get one dimension, i=1,2 ..., N, j=1,2 ..., N; 2. N is walked to each step sample in sample and be set as a classification; 3. two nearest classifications are aggregated into a new classification, remaining classification is other classifications; 4. the new classification that adopts mean distance formula between class to calculate to obtain 3. from step mean distance respectively and between the class of other classifications:
wherein, C
prepresent the new classification that step is obtained in 3., n
prepresent C
pin the number of samples that comprises, C
qrepresent any one classification in other classifications, n
qrepresent C
qin the number of samples that comprises; If 5. step only has a new classification in 4., and without other classifications, enters next step; Otherwise, get back to step 3.; 6. according to step 1.~implementation 5., draw cluster pedigree chart; 7. according to drawn cluster pedigree chart, determine cluster number, it is different classes of that the patient that this cluster number is different gait features is divided into.
The present invention is owing to taking above technical scheme, it has the following advantages: 1, the present invention is by analyzing different patients' plantar pressure information and the similarity degree of ankle joint angle information, utilize spectral clustering that patient similar gait feature is divided into a class, lack the deficiency of objective classification evaluation method to make up gait, reach objectively patient is classified, thereby for Rehabilitation training for treatment provides reference, so that doctor can take different therapeutic modalities and training strength to different classes of patient.2, the present invention adopts multi-sensor information fusion technology, not only gather patient's plantar pressure information, also gathered paralytic's ankle joint angle information simultaneously, gait feature while more adopting H/M, HM/HL, MM/ML, HD/SD, MD/SD and SW/ST to describe patient to walk, so that the feature comprehensively when objectively analyzing patient and walk.3, the present invention adopts spectral clustering to analyze gained gait feature data, and compared to traditional clustering algorithm, the globally optimal solution that spectral clustering can be restrained in sample space arbitrarily, has ensured to obtain the legitimate result to patient's gait classification.In view of above advantage, the present invention can be widely used in gait analysis and medical rehabilitation field.
Brief description of the drawings
Fig. 1 is the stationary value schematic diagram that pressure transducer is exported in the time of zero load
Fig. 2 is that gait phase is analyzed schematic diagram
Fig. 3 is that H/M cluster result is analyzed schematic diagram
Fig. 4 is that MM/ML cluster result is analyzed schematic diagram
Fig. 5 is that HM/HL cluster result is analyzed schematic diagram
Fig. 6 is that HD/SD cluster result is analyzed schematic diagram
Fig. 7 is that SW/ST cluster result is analyzed schematic diagram
Fig. 8 is that MD/SD cluster result is analyzed schematic diagram
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in detail.
Approach for Gait Classification based on multi-sensor information fusion provided by the invention comprises the following steps:
1) plantar pressure information and ankle joint angle information while gathering several patients and walk;
In the present embodiment, plantar pressure information and ankle joint angle information while adopting multi-sensor fusion technology collection patient to walk, be distributed in patient vola by array pressure transducer, plantar pressure information while walking to gather patient, plantar pressure information comprises plantar pressure value and the pressure Two dimensional Distribution situation in vola, wherein plantar pressure value comprises vola regional force value size, and pressure comprises heel pressure, metatarsal pressure, inner side heel pressure, outside heel pressure, inner side metatarsal pressure and outside metatarsal pressure in the Two dimensional Distribution situation in vola.Ankle joint angle information while adopting Inertial Measurement Unit (IMU, Inertial Measurement Unit are called for short inertial navigation) to gather patient and walk, ankle joint angle information comprises ankle joint angle value.
2) according to the plantar pressure information analysis patient's who obtains gait phase, specifically comprise:
1. choose gait event threshold value, this gait event threshold value is slightly larger than the patient vola plantar pressure value sum that all pressure transducers record when unsettled;
Due to the design problem of pressure transducer in prior art, be distributed in the pressure transducer in patient vola when unsettled in vola output be not zero, a but stationary value.As shown in Figure 1, horizontal ordinate represents the time, plantar pressure value when ordinate represents that vola is unsettled, as can be seen from the figure stationary value is 850N, this stationary value is the plantar pressure value of vola when unsettled, now can choose 1.1 times of the vola plantar pressure value that pressure transducer records when unsettled as gait event threshold value.
2. contrast the plantar pressure value and the gait event threshold value that record, judge patient's gait phase, this gait phase is divided into foot and contacts to earth stage and swinging kick stage;
In the time that the plantar pressure value recording is greater than gait event threshold value, represent the foot event (Heel-Strike, HS) of contacting to earth has occurred, illustrate that patient enters foot and contacts to earth the stage.Otherwise, in the time that the plantar pressure value recording is less than gait event threshold value, represent the liftoff event of foot (Toe-off, TO) has occurred, illustrate that patient enters the swinging kick stage.Contact to earth event and the liftoff event of foot of the foot of same pin of patient forms a complete gait cycle, and same pin of patient forms a complete gait cycle through contact to earth stage and swinging kick stage of foot.
As shown in Figure 2, horizontal ordinate represents the time, plantar pressure value when ordinate represents that patient walks.In figure, go forward circumference point of each crest value is the foot event of contacting to earth, and the round dot at rear is the liftoff event of foot.Crest between every adjacent two round dots represents that foot contacts to earth the stage, and the trough between every adjacent two round dots represents the swinging kick stage.Contact to earth stage and swinging kick stage of adjacent foot determined a gait cycle.
3) eigenwert of setting gait cycle, the gait feature while walking to characterize each patient, this eigenwert refers to the conclusion and the sign that use low volume data to carry out the rule that in a gait cycle, mass data presents;
Comprise that as gait cycle foot contacts to earth stage and swinging kick stage, the eigenwert of gait cycle also comprises foot contact to earth eigenwert and the eigenwert in swinging kick stage in stage:
1. the contact to earth eigenwert in stage of foot comprises pressure characteristic value and temporal characteristics value, and wherein pressure characteristic value comprises:
H/M: the ratio of heel pressure (Heel-Pressure) and metatarsal pressure (Meta-Pressure);
HM/HL: the ratio of inner side heel pressure (Heel Medial-Pressure) and outside heel pressure (Heel Lateral-Pressure);
MM/ML: the ratio of inner side metatarsal pressure (Meta Medial-Pressure) and outside metatarsal pressure (Meta Lateral-Pressure).
Temporal characteristics value comprises:
HD/SD: the ratio of heel contact time (Heel-Duration) shared vola contact time (Stance-Duration);
MD/SD: the ratio of metatarsal contact time (Meat-Duration) shared vola contact time (Stance-Duration).
2. the eigenwert in swinging kick stage comprises angle character value, this angle character value is designated as SW/ST, be the contact to earth ratio of ankle joint angle average (Mean-Stance-Angle) in stage of ankle joint angle average (Mean-Swing-Angle) and the foot in swinging kick stage, wherein ankle joint angle average is that the ankle joint angle value recording is averaged.
Gait feature while adopting H/M, HM/HL, MM/ML, HD/SD, MD/SD and SW/ST to describe patient to walk, feature when gait feature represents that patient walks, as turned up in drop foot or foot, wherein drop foot mainly characterizes with H/M, turns up mainly and characterize etc. with MM/ML in foot.Because each patient's gait feature is different, therefore can distinguish different patients by gait feature, and then patient similar gait feature is divided into a class.
4) adopt spectral clustering to carry out gait cluster analysis for each eigenwert in each gait cycle of each patient, so that the patient of different gait features is divided into different classifications.
In the present embodiment, the common N step sample when gathering several patients and walking, wherein N is positive integer, and each step sample of patient is a gait cycle, carries out spectral clustering for each eigenwert in each gait cycle, and it comprises the following steps:
1. adopt Euclidean distance formula to calculate any two step sample x in N step sample
ik, x
jkbetween distance:
Wherein, x
ik, x
jkbe proper vector, the dimension of p representation feature vector, k=1 represents to get one dimension, i=1,2 ..., N, j=1,2 ..., N;
2. N is walked to each step sample in sample and be set as a classification;
3. two nearest classifications are aggregated into a new classification, remaining classification is other classifications;
4. the new classification that adopts mean distance formula between class to calculate to obtain 3. from step mean distance respectively and between the class of other classifications:
Wherein, C
prepresent the new classification that step is obtained in 3., n
prepresent C
pin the number of samples that comprises, C
qrepresent any one classification in other classifications, n
qrepresent C
qin the number of samples that comprises;
If 5. step only has a new classification in 4., and without other classifications, enters next step; Otherwise, get back to step 3.;
6. according to step 1.~implementation 5., draw cluster pedigree chart, this cluster pedigree chart is the conventional form of expression that those skilled in the art adopt spectral clustering deal with data acquired results, therefore no longer describe in detail;
7. according to drawn cluster pedigree chart, determine cluster number, it is different classes of that the patient that this cluster number is different gait features is divided into.
Because reasonably cluster result feature is more obvious, therefore those skilled in the art pass through drawn cluster pedigree chart, can directly draw cluster number.
In sum, gather the every step of patient as a sample, and each sample is made as to a classification, two the most similar samples are polymerized to a new classification, again the similarity between new classification basis and other classifications of polymerization is carried out to polymerization again afterwards, constantly loop cluster computing, along with weakening of similarity between of all categories, last all categories is all polymerized to a large classification, thereby obtain a cluster pedigree chart being polymerized according to similarity size, the patient of different gait features can be divided into different classifications by observing cluster pedigree chart, so that doctor adopts different for different patients' rehabilitation, recovery training method targetedly.
Collect four paralytic A, paralytic B, paralytic C and paralytic D and carry out normal level walking, obtain altogether 75 step effective samples, wherein the effective sample of paralytic A, C, D is respectively 20 steps, and the effective sample of paralytic A is 1~20, the effective sample of paralytic C is 36~55, the effective sample of paralytic D is 56~75; Paralytic B effective sample is 15 steps, and its effective sample is 21~35.Adopt spectral clustering in this 75 step effective sample, six eigenwerts (H/M, HM/HL, MM/ML, HD/SD, MD/SD and SW/ST) of each step are carried out gait cluster analysis:
As shown in Figure 3, horizontal ordinate is sample number, and ordinate is between class distance.Wherein, effective sample 11, effective sample 13 and effective sample 14 are outlier, intuitively observe figure and find, outlier in the end flocks together and just can be judged to be profit group point with large class, illustrates that these independent data points are far away apart from other data points.Reasonably gait cluster result should be defined as two classes, and effective sample 1~35 is a class, and effective sample 36~75 is another kind of.Drop foot situation when paralytic C, paralytic D walking is more serious, the two is classified as to a class, and paralytic A, paralytic B walking time does not have drop foot phenomenon, the two is classified as another kind of, thereby just paralytic can be divided into two classes according to gait cluster result.
As shown in Figure 4, horizontal ordinate is sample number, and ordinate is between class distance.Wherein, effective sample 36, effective sample 54 are outlier, so reasonably gait cluster result is still two large classes.In figure, effective sample 1~20 major part from paralytic A is a class by gait cluster, after removing two outlier effective samples 36 and effective sample 54, all effective samples 37~53 from paralytic C are also a class by gait cluster, are another kind of from the effective sample of all the other two paralytic B, D by gait cluster.Again by the analysis concrete to its eigenwert, can obtain conclusion paralytic A, C walking time, strephenopodia phenomenon is more serious afterwards, and the strephenopodia phenomenon of paralytic B, D is lighter.Because the effective sample 11,12,13 from paralytic A is classification 2 by gait cluster, the strephenopodia phenomenon outline of the paralytic A that can also reach a conclusion is better than paralytic C.
Can find out by observing Fig. 5~Fig. 8, all the other four eigenwert HM/HL, HD/SD, MD/SD and SW/ST do not have obvious cluster feature, and these four eigenwerts can not be distinguished four paralytics' kinematics and dynamic characteristic in more detail in other words.By the eigenwert of above-mentioned four people's walking step states is carried out after cluster analysis, kinematics and the dynamic characteristic that can find four paralytic's walking step states still have difference clearly, such as drop foot degree, in the degree etc. of turning up, this provides significant reference information to patient's clinical medicine rehabilitation training from now on.
The various embodiments described above are only for illustrating the present invention, and wherein the implementation of each step can change to some extent, and every equivalents of carrying out on the basis of technical solution of the present invention and improvement, all should not get rid of outside protection scope of the present invention.
Claims (7)
1. the Approach for Gait Classification based on multi-sensor information fusion, it comprises the following steps:
1) plantar pressure information and ankle joint angle information while gathering several patients and walk;
2) according to the plantar pressure information analysis patient's that obtains gait phase, this gait phase is divided into foot contacts to earth stage and swinging kick stage, and same pin of patient is a gait cycle through contact to earth stage and swinging kick stage of foot;
3) eigenwert of setting gait cycle, the gait feature while walking to characterize each patient;
4) adopt spectral clustering to carry out gait cluster analysis for each eigenwert in each gait cycle of each patient, so that the patient of different gait features is divided into different classifications.
2. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 1, it is characterized in that: carrying out described step 1) time, adopt array pressure transducer is distributed in to patient vola, plantar pressure information while walking to gather patient, plantar pressure information comprises plantar pressure value and the pressure Two dimensional Distribution situation in vola, wherein plantar pressure value comprises vola regional force value size, pressure comprises heel pressure in the Two dimensional Distribution situation in vola, metatarsal pressure, inner side heel pressure, outside heel pressure, inner side metatarsal pressure and outside metatarsal pressure, ankle joint angle information while adopting Inertial Measurement Unit to gather patient and walk, ankle joint angle information comprises ankle joint angle value.
3. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 1, is characterized in that: carrying out described step 2) time, specifically comprise:
1. choose gait event threshold value, this gait event threshold value is slightly larger than the patient vola plantar pressure value sum that all pressure transducers record when unsettled;
2. contrast the plantar pressure value and the gait event threshold value that record, judge patient's gait phase;
In the time that the plantar pressure value recording is greater than gait event threshold value, represent the foot event of contacting to earth has occurred, illustrate that patient enters foot and contacts to earth the stage; Otherwise, in the time that the plantar pressure value recording is less than gait event threshold value, represent the liftoff event of foot has occurred, illustrate that patient enters the swinging kick stage.
4. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 2, is characterized in that: carrying out described step 2) time, specifically comprise:
1. choose gait event threshold value, this gait event threshold value is slightly larger than the patient vola plantar pressure value sum that all pressure transducers record when unsettled;
2. contrast the plantar pressure value and the gait event threshold value that record, judge patient's gait phase;
In the time that the plantar pressure value recording is greater than gait event threshold value, represent the foot event of contacting to earth has occurred, illustrate that patient enters foot and contacts to earth the stage; Otherwise, in the time that the plantar pressure value recording is less than gait event threshold value, represent the liftoff event of foot has occurred, illustrate that patient enters the swinging kick stage.
5. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 1 or 2 or 3 or 4, it is characterized in that: carrying out described step 3) time, the eigenwert of gait cycle comprises foot contact to earth eigenwert and the eigenwert in swinging kick stage in stage:
1. the contact to earth eigenwert in stage of foot comprises pressure characteristic value and temporal characteristics value, and wherein pressure characteristic value comprises:
H/M: the ratio of heel pressure and metatarsal pressure;
HM/HL: the ratio of inner side heel pressure and outside heel pressure;
MM/ML: the ratio of inner side metatarsal pressure and outside metatarsal pressure;
Temporal characteristics value comprises:
HD/SD: the ratio of the shared vola of heel contact time contact time;
MD/SD: the ratio of the shared vola of metatarsal contact time contact time;
2. the eigenwert in swinging kick stage comprises angle character value, this angle character value is designated as SW/ST: the contact to earth ratio of ankle joint angle average in stage of ankle joint angle average and the foot in swinging kick stage, and wherein ankle joint angle average is that the ankle joint angle value recording is averaged; Gait feature while adopting H/M, HM/HL, MM/ML, HD/SD, MD/SD and SW/ST to describe patient to walk, distinguishes different patients according to each patient's different gait feature.
6. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 1 or 2 or 3 or 4, it is characterized in that: carrying out described step 4) time, common N step sample when gathering several patients and walking, wherein N is positive integer, each step of patient is a gait cycle, carry out spectral clustering for each eigenwert in gait cycle, it comprises the following steps:
1. adopt Euclidean distance formula to calculate any two step sample x in N step sample
ik, x
jkbetween distance:
Wherein, x
ik, x
jkbe proper vector, the dimension of p representation feature vector, k=1 represents to get one dimension, i=1,2 ..., N, j=1,2 ..., N;
2. N is walked to each step sample in sample and be set as a classification;
3. two nearest classifications are aggregated into a new classification, remaining classification is other classifications;
4. the new classification that adopts mean distance formula between class to calculate to obtain 3. from step mean distance respectively and between the class of other classifications:
Wherein, C
prepresent the new classification that step is obtained in 3., n
prepresent C
pin the number of samples that comprises, C
qrepresent any one classification in other classifications, n
qrepresent C
qin the number of samples that comprises;
If 5. step only has a new classification in 4., and without other classifications, enters next step; Otherwise, get back to step 3.;
6. according to step 1.~implementation 5., draw cluster pedigree chart;
7. according to drawn cluster pedigree chart, determine cluster number, it is different classes of that the patient that this cluster number is different gait features is divided into.
7. a kind of Approach for Gait Classification based on multi-sensor information fusion as claimed in claim 5, it is characterized in that: in the time carrying out described step 4, common N step sample when gathering several patients and walking, wherein N is positive integer, each step sample of patient is a gait cycle, carry out spectral clustering for each eigenwert in each gait cycle, it comprises the following steps:
1. adopt Euclidean distance formula to calculate any two step sample x in N step sample
ik, x
jkbetween distance:
Wherein, x
ik, x
jkbe proper vector, the dimension of p representation feature vector, k=1 represents to get one dimension, i=1,2 ..., N, j=1,2 ..., N;
2. N is walked to each step sample in sample and be set as a classification;
3. two nearest classifications are aggregated into a new classification, remaining classification is other classifications;
4. the new classification that adopts mean distance formula between class to calculate to obtain 3. from step mean distance respectively and between the class of other classifications:
Wherein, C
prepresent the new classification that step is obtained in 3., n
prepresent C
pin the number of samples that comprises, C
qrepresent any one classification in other classifications, n
qrepresent C
qin the number of samples that comprises;
If 5. step only has a new classification in 4., and without other classifications, enters next step; Otherwise, get back to step 3.;
6. according to step 1.~implementation 5., draw cluster pedigree chart;
7. according to drawn cluster pedigree chart, determine cluster number, it is different classes of that the patient that this cluster number is different gait features is divided into.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410257800.1A CN104008398B (en) | 2014-06-11 | 2014-06-11 | A kind of Approach for Gait Classification based on multi-sensor information fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410257800.1A CN104008398B (en) | 2014-06-11 | 2014-06-11 | A kind of Approach for Gait Classification based on multi-sensor information fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104008398A true CN104008398A (en) | 2014-08-27 |
CN104008398B CN104008398B (en) | 2017-07-07 |
Family
ID=51369047
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410257800.1A Active CN104008398B (en) | 2014-06-11 | 2014-06-11 | A kind of Approach for Gait Classification based on multi-sensor information fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104008398B (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104434123A (en) * | 2014-12-04 | 2015-03-25 | 黑龙江大学 | Clustering-algorithm-based human motion state judgment method |
CN104644191A (en) * | 2015-02-04 | 2015-05-27 | 宁德师范学院 | Foot pressure detecting system |
CN105125216A (en) * | 2015-08-29 | 2015-12-09 | 深圳市老年医学研究所 | Gait detection system based on sole pressure |
CN105411593A (en) * | 2014-09-12 | 2016-03-23 | 三星电子株式会社 | Method and apparatus for recognizing gait task |
CN106217352A (en) * | 2016-08-17 | 2016-12-14 | 尖叫智能科技(上海)有限公司 | Exoskeleton robot gait control method based on pressure transducer |
CN107092861A (en) * | 2017-03-15 | 2017-08-25 | 华南理工大学 | Lower limb movement recognition methods based on pressure and acceleration transducer |
CN107115102A (en) * | 2017-06-07 | 2017-09-01 | 西南科技大学 | A kind of osteoarticular function appraisal procedure and device |
CN107374643A (en) * | 2017-08-16 | 2017-11-24 | 南京信息职业技术学院 | A kind of juvenile walks appearance detection footwear and step posture detection method |
CN107536613A (en) * | 2016-06-29 | 2018-01-05 | 深圳光启合众科技有限公司 | Robot and its human body lower limbs Gait Recognition apparatus and method |
CN108601699A (en) * | 2016-06-30 | 2018-09-28 | 松下知识产权经营株式会社 | Device of walking aid and control method |
CN108731732A (en) * | 2018-05-22 | 2018-11-02 | 洛阳中科汇成科技有限公司 | Based on periodic mode information to the processing method and system of plantar pressure |
CN109239585A (en) * | 2018-09-06 | 2019-01-18 | 南京理工大学 | A kind of method for diagnosing faults based on the preferred wavelet packet of improvement |
CN109316190A (en) * | 2018-09-25 | 2019-02-12 | 快快乐动(北京)网络科技有限公司 | A kind of system and method for analyzing user's plantar pressure type |
CN110420029A (en) * | 2019-08-03 | 2019-11-08 | 苏州自如医疗器械有限公司 | A kind of walking step state wireless detecting system based on Multi-sensor Fusion |
CN110693501A (en) * | 2019-10-12 | 2020-01-17 | 上海应用技术大学 | Wireless walking gait detection system based on multi-sensor fusion |
CN111079651A (en) * | 2019-12-17 | 2020-04-28 | 河南水滴智能技术有限公司 | Re-optimization of gait classification method |
CN112998695A (en) * | 2021-02-03 | 2021-06-22 | 上海橙捷健康科技有限公司 | Data integration method and system for calcaneus angle quantification and pressure treadmill |
CN113729693A (en) * | 2021-09-17 | 2021-12-03 | 平安国际智慧城市科技股份有限公司 | Artificial intelligence-based exercise intensity detection method, device, equipment and medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050070834A1 (en) * | 2003-09-25 | 2005-03-31 | Massachusetts Institute Of Technology | Active Ankle Foot Orthosis |
CN101947151A (en) * | 2010-08-30 | 2011-01-19 | 北京大学 | Power below-knee prosthesis-based gait recognition method |
CN102670207A (en) * | 2012-05-15 | 2012-09-19 | 北京大学 | Gait analysis method based on plantar pressure |
CN103049741A (en) * | 2012-12-21 | 2013-04-17 | 中国科学院合肥物质科学研究院 | Foot-to-ground acting force-based gait feature extraction method and gait identification system |
CN103514302A (en) * | 2013-10-28 | 2014-01-15 | 深圳先进技术研究院 | Human body gait database and establishment method thereof |
-
2014
- 2014-06-11 CN CN201410257800.1A patent/CN104008398B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050070834A1 (en) * | 2003-09-25 | 2005-03-31 | Massachusetts Institute Of Technology | Active Ankle Foot Orthosis |
CN101947151A (en) * | 2010-08-30 | 2011-01-19 | 北京大学 | Power below-knee prosthesis-based gait recognition method |
CN102670207A (en) * | 2012-05-15 | 2012-09-19 | 北京大学 | Gait analysis method based on plantar pressure |
CN103049741A (en) * | 2012-12-21 | 2013-04-17 | 中国科学院合肥物质科学研究院 | Foot-to-ground acting force-based gait feature extraction method and gait identification system |
CN103514302A (en) * | 2013-10-28 | 2014-01-15 | 深圳先进技术研究院 | Human body gait database and establishment method thereof |
Non-Patent Citations (1)
Title |
---|
刘鑫朝: "聚类算法的研究及其应用", 《中国优秀硕士学位论文全文库》 * |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105411593A (en) * | 2014-09-12 | 2016-03-23 | 三星电子株式会社 | Method and apparatus for recognizing gait task |
CN104434123B (en) * | 2014-12-04 | 2017-01-04 | 黑龙江大学 | Human motion state determination methods based on clustering algorithm |
CN104434123A (en) * | 2014-12-04 | 2015-03-25 | 黑龙江大学 | Clustering-algorithm-based human motion state judgment method |
CN104644191A (en) * | 2015-02-04 | 2015-05-27 | 宁德师范学院 | Foot pressure detecting system |
CN104644191B (en) * | 2015-02-04 | 2015-12-02 | 宁德师范学院 | A kind of foot pressure detection system |
CN105125216A (en) * | 2015-08-29 | 2015-12-09 | 深圳市老年医学研究所 | Gait detection system based on sole pressure |
CN107536613B (en) * | 2016-06-29 | 2021-10-08 | 沭阳县成基实业有限公司 | Robot and human body lower limb gait recognition device and method thereof |
CN107536613A (en) * | 2016-06-29 | 2018-01-05 | 深圳光启合众科技有限公司 | Robot and its human body lower limbs Gait Recognition apparatus and method |
CN108601699A (en) * | 2016-06-30 | 2018-09-28 | 松下知识产权经营株式会社 | Device of walking aid and control method |
US10980699B2 (en) | 2016-06-30 | 2021-04-20 | Panasonic Intellectual Property Management Co., Ltd. | Walking assistance apparatus and control method |
CN106217352A (en) * | 2016-08-17 | 2016-12-14 | 尖叫智能科技(上海)有限公司 | Exoskeleton robot gait control method based on pressure transducer |
CN107092861B (en) * | 2017-03-15 | 2020-11-27 | 华南理工大学 | Lower limb action recognition method based on pressure and acceleration sensor |
CN107092861A (en) * | 2017-03-15 | 2017-08-25 | 华南理工大学 | Lower limb movement recognition methods based on pressure and acceleration transducer |
CN107115102A (en) * | 2017-06-07 | 2017-09-01 | 西南科技大学 | A kind of osteoarticular function appraisal procedure and device |
CN107374643A (en) * | 2017-08-16 | 2017-11-24 | 南京信息职业技术学院 | A kind of juvenile walks appearance detection footwear and step posture detection method |
CN108731732B (en) * | 2018-05-22 | 2021-08-13 | 洛阳中科汇成科技有限公司 | Method and system for processing plantar pressure based on periodic modal information |
CN108731732A (en) * | 2018-05-22 | 2018-11-02 | 洛阳中科汇成科技有限公司 | Based on periodic mode information to the processing method and system of plantar pressure |
CN109239585A (en) * | 2018-09-06 | 2019-01-18 | 南京理工大学 | A kind of method for diagnosing faults based on the preferred wavelet packet of improvement |
CN109316190A (en) * | 2018-09-25 | 2019-02-12 | 快快乐动(北京)网络科技有限公司 | A kind of system and method for analyzing user's plantar pressure type |
CN109316190B (en) * | 2018-09-25 | 2022-03-01 | 快快利华(北京)网络科技有限公司 | System and method for analyzing plantar pressure type of user |
CN110420029A (en) * | 2019-08-03 | 2019-11-08 | 苏州自如医疗器械有限公司 | A kind of walking step state wireless detecting system based on Multi-sensor Fusion |
CN110693501A (en) * | 2019-10-12 | 2020-01-17 | 上海应用技术大学 | Wireless walking gait detection system based on multi-sensor fusion |
CN111079651A (en) * | 2019-12-17 | 2020-04-28 | 河南水滴智能技术有限公司 | Re-optimization of gait classification method |
CN112998695A (en) * | 2021-02-03 | 2021-06-22 | 上海橙捷健康科技有限公司 | Data integration method and system for calcaneus angle quantification and pressure treadmill |
CN112998695B (en) * | 2021-02-03 | 2023-05-02 | 上海橙捷健康科技有限公司 | Calcaneus angle quantification and pressure treadmill data integration method and system |
CN113729693A (en) * | 2021-09-17 | 2021-12-03 | 平安国际智慧城市科技股份有限公司 | Artificial intelligence-based exercise intensity detection method, device, equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN104008398B (en) | 2017-07-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104008398A (en) | Gait classification method based on multi-sensor information fusion | |
Quante et al. | Practical considerations in using accelerometers to assess physical activity, sedentary behavior, and sleep | |
Moore et al. | Autonomous identification of freezing of gait in Parkinson's disease from lower-body segmental accelerometry | |
US10299736B2 (en) | Method, device, and system for diagnosing and monitoring frailty | |
Agostini et al. | Segmentation and classification of gait cycles | |
Rosati et al. | Muscle activation patterns during gait: A hierarchical clustering analysis | |
Nouredanesh et al. | Fall risk assessment in the wild: A critical examination of wearable sensor use in free-living conditions | |
CN109717833A (en) | A kind of neurological disease assistant diagnosis system based on human motion posture | |
Del Din et al. | Instrumented gait assessment with a single wearable: an introductory tutorial | |
Mei et al. | Sample entropy characteristics of movement for four foot types based on plantar centre of pressure during stance phase | |
Kour et al. | A survey of knee osteoarthritis assessment based on gait | |
Kim et al. | Saliency detection analysis of collective physiological responses of pedestrians to evaluate neighborhood built environments | |
Yu et al. | Assessing elderly’s functional balance and mobility via analyzing data from waist-mounted tri-axial wearable accelerometers in timed up and go tests | |
Baroudi et al. | Estimating walking speed in the wild | |
Selzler et al. | Neurodegenerative disease prediction based on gait analysis signals acquired with force-sensitive resistors | |
Aghanavesi et al. | A multiple motion sensors index for motor state quantification in Parkinson's disease | |
Lee et al. | Analysis of gait sub-movements to estimate ataxia severity using ankle inertial data | |
Huang et al. | Feature Selection, Construction, and Validation of a Lightweight Model for Foot Function Assessment During Gait With In-Shoe Motion Sensors | |
Wang et al. | Research on abnormal gait recognition algorithms for stroke patients based on array pressure sensing system | |
Xie et al. | Wearable multisource quantitative gait analysis of Parkinson's diseases | |
Liu et al. | A novel interpretable feature set optimization method in blood pressure estimation using photoplethysmography signals | |
Xu et al. | A model for medical diagnosis based on plantar pressure | |
Lo et al. | Bayesian analysis of sub-plantar ground reaction force with BSN | |
Donahue et al. | Estimation of ground reaction force waveforms during fixed pace running outside the laboratory | |
Rodrigo et al. | Application of Kohonen maps to kinetic analysis of human gait |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |