CN105590408A - Human body falling detection method and protection device - Google Patents
Human body falling detection method and protection device Download PDFInfo
- Publication number
- CN105590408A CN105590408A CN201610083726.5A CN201610083726A CN105590408A CN 105590408 A CN105590408 A CN 105590408A CN 201610083726 A CN201610083726 A CN 201610083726A CN 105590408 A CN105590408 A CN 105590408A
- Authority
- CN
- China
- Prior art keywords
- human body
- falling
- angle
- training
- obtaining
- 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
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0407—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis
- G08B21/043—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis detecting an emergency event, e.g. a fall
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0438—Sensor means for detecting
- G08B21/0446—Sensor means for detecting worn on the body to detect changes of posture, e.g. a fall, inclination, acceleration, gait
Abstract
The invention discloses a human body falling detection method and a protection device. The method comprises the steps of: obtaining a three-axis total acceleration, a roll angle and a pitch angle of each sampling time point; setting sliding time windows; obtaining a mean value and a standard difference of the three-axis total accelerations of the sampling points in each time window, obtaining the sum of absolute values of change amounts of the roll angle and the pitch angle at the last sampling point of the sliding time windows when a human body relatively stands, and using the threes parameters as characteristic values; utilizing a support vector machine algorithm to construct a classifier; obtaining falling samples and daily movement behavior samples to form a training set; training the classifier, and obtaining a trained classifier; and utilizing the trained classifier to carry out falling detection according to obtained human body practical sensing data. The detection rate of falling behaviors is 99.2%, the detection rate of daily movement behaviors is 96%, and the average lead time reaches 273 ms, so that reaction time is provided for real-time early warning of falling and starting of the protection device.
Description
Technical field
The present invention relates to a kind of detection method and protector of physical protection, be specifically related to the detection that fall on a kind of human body roadMethod and the device of in the time detecting that human body is fallen, human body being protected.
Background technology
Falling is the common a kind of accident of the elderly, and old man usually causes various damages, for example hip bone because fallingFolding, neck injury, brain injury and various soft tissue injury and other position fracture etc.
To human body, the fall detection of event is the prerequisite of protecting. In prior art, human body fall detection method is mainThere are 3 classes: the detection based on video image, the detection based on the surrounding signals such as audio frequency or radio and based on Wearable equipmentDetection. Advantage based on video images detection is that human body does not need to carry any equipment, and shortcoming is that video image is subject to light, ringThe impact in border etc. is larger, and detection range is limited and relate to individual privacy. Around the method detecting based on surrounding signals is subject toThe impact of environment is larger, cannot obtain very high precision, generally can only serve as aided detection method. Based on Wearable Equipment InspectionMethod because of cost low, detection range is large, is not subject to the impact of surrounding environment, is at present research and the maximum inspection of falling of practical applicationSurvey method.
For example, Chinese invention patent application CN102117533A discloses a kind of human body fall detection protection and the dress of reporting to the policePut, comprise fall detection part, fall detection partial message transmission, fall position detection part and protection air bag. It adopts wearsWear formula device, attempt to carry out starting protection air bag by obtaining fall detection signal, still, with current most fall detectionMethod is the same, adopts the method for setting threshold to detect. It has adopted acceleration transducer and gyroscope, definition signal vectorMould SVM and motion angular speed,(being three axle resultant accelerations), setting SVM threshold value is 1.8g-2.2g, motion angular speed threshold value is 0.5rad/s-0.55rad/s, in the time that SVM and human motion angular speed exceed setting threshold,Judge that human body falls. Adopt threshold detection method to belong to afterwards and detect, the protective device that makes to fall is difficult to play a role.
Method by training classifier is carried out body state judgement, likely earlier detects than common threshold methodThe trend that goes out to fall, detects thereby realize in advance. But, the characteristic value that How to choose grader uses, and adopt what pointCan class device be really realize the key point in advance detecting.
Therefore, how realizing the detection in advance that human body is fallen, is to realize at present fall protection institute of human body to be badly in need of asking of solutionTopic.
Summary of the invention
Goal of the invention of the present invention is to provide a kind of human body fall detection method, realizes the detection in advance that human body is fallen, withProvide the enough reaction time to protective device; Another goal of the invention of the present invention is to provide a kind of this detection method that adoptsThe human body protective device of falling.
To achieve the above object of the invention, the technical solution used in the present invention is: a kind of human body fall detection method, comprising:
1) obtaining of characteristic value:
1. the 3-axis acceleration instrument and the three-axis gyroscope that use human body to dress are sampled, and obtain the sampling of each sampling time pointData;
2. the sampled data 1. obtaining according to step, calculates respectively the three axle resultant accelerations that obtain each sampling time point, rollCorner, the angle of pitch，
Wherein,，ax、ay、azRespectively 3-axis acceleration, with both direction orthogonal in horizontal planeRespectively as x axle and y axle, roll angleThe attitude angle around x axle, the angle of pitchIt is the attitude angle around y axle;
3. the length of sliding time window and the stack rate of adjacent sliding time window are set, obtain in each sliding time windowThe average of three axle resultant accelerations of each sampled pointAnd standard deviation, last acquisition at sliding time window adoptedThe absolute value sum of the variable quantity of the roll angle of sampling point place human body during with respect to erectility and the angle of pitch, does with these three parametersFor characteristic value;
2) foundation of grader and training:
Adopt algorithm of support vector machine to build grader, three parameters that obtain using step 1) are as the input feature vector of graderValue;
Carry out respectively daily routines behavior and the different behaviors of falling by the personnel that train, obtain fall sample and daily workMoving behavior sample composing training collection, trains grader, obtains the grader after training;
3) adopt the grader after training to carry out fall detection according to the actual sensing data of human body obtaining.
In technique scheme, step 1) 1. in, sample frequency is conventionally relevant with the performance of the sensor chip of employing,Sample frequency is too low by the sampled point number deficiency causing in sliding time window, thereby affects determine effect. Therefore, sampling frequentlyRate is not less than 50Hz.
Preferred sample frequency is 100Hz.
In technique scheme, the length of sliding time window is 100～300ms, and the stack rate of window is 40%～60%.
The length of sliding time window is 100ms, and the stack rate of window is 50%.
The present invention provides a kind of human body protective device of falling simultaneously, comprises falling detection device, protection air bag and drives dressPut, wherein, described falling detection device is mainly made up of 3-axis acceleration instrument, three-axis gyroscope and controller, described controllerIn be provided with the grader after above-mentioned training.
Preferred technical scheme, described falling detection device is arranged on the waist of human body.
In technique scheme, described drive unit comprises the compressed gas cylinder being communicated with protection air bag through a control valve,Described control valve is opened and closed by described controller control.
Described protection air bag is made up of the multiple air bags that are worn on respectively the fragile position of people.
Because technique scheme is used, the present invention compared with prior art has following advantages:
1, the present invention, by selecting three relevant to acceleration and angular speed respectively parameters, adopts algorithm of support vector machine trainingGrader, has realized the detection in advance that human body is fallen, and experimental result shows, the verification and measurement ratio of the behavior of falling is 99.2%, daily routinesThe verification and measurement ratio of behavior is 96%, to comparatively violent daily routines behavior, also has higher verification and measurement ratio, and the average lead time reaches273ms, the reaction time providing for the startup of the real-time early warning of falling and protective device.
2, the present invention adopts the airing form of compressed gas cylinder to airbag aeration, has avoided available technology adopting explosive to startThe unexpected injury that the instant impact causing causes, air bag can be used as wearing equipment and is configured in the significant bits of the human body in clothesPut, such as hip, neck and other key positions, play a good protection to human body.
Brief description of the drawings
Fig. 1 is the method flow diagram of the embodiment of the present invention;
Fig. 2 is the resultant acceleration curve map of once falling forward in embodiment;
Fig. 3 is the roll angle change curve of once falling forward in embodiment;
Fig. 4 is acceleration and the change curve of a walking process in embodiment bis-.
Detailed description of the invention
Below in conjunction with drawings and Examples, the invention will be further described:
Embodiment: shown in Figure 1, the present embodiment provides a kind of human body fall detection method, comprising: obtain training data,Carry out feature extraction, obtain training sample, to disaggregated model training, obtain the grader after training. For detecting its effect,Obtain test data, carry out feature extraction, adopt the grader after training to carry out categorised decision.
Wherein, while carrying out feature extraction, select the factor of the concrete consideration of characteristic value to be expressed as follows:
Sample frequency in the present embodiment is 100Hz.
(1) selection of acceleration signature:
Falling in process at human body, can there is significant change in acceleration, considers the different directions of falling all around, gets three axlesResultant acceleration:
；
Shown in Fig. 2, be the resultant acceleration curve of once falling forward, can find out, in the process of falling, resultant acceleration first reducesRear increase, the moment of acceleration maximum is namely collided the moment on ground, falls and detects and will before this moment, examine exactly in advanceSurvey is fallen. From detecting that the time of falling collision ground is called the lead time. Correspondingly, carrying out SVM model trainingWith test time, the collection of the sample of falling also should be before this moment, and the reserved regular hour.
Consider size and the variation tendency of resultant acceleration, getting length is n(n sample point) sliding time window, extractionFollowing acceleration signature:
(1) average of resultant acceleration:
, i is the sequence number of sample point in a certain sliding time window.
(2) standard deviation of resultant acceleration:
In the present embodiment, n gets 10, and sliding time window length is 100ms, the stack rate of window is 50%, i.e. current windowRear 50% data are as front 50% data of a rear window.
(2) selection of angle character:
Falling in process at human body, also can there is significant change in the attitude angle of human body, taking eastwards as x axle, northwards as y axle, upwardsFor z axle is set up human body three-dimensional coordinate system, also referred to as sky, northeast coordinate system, be called rolling around the attitude angle of x-axis, y-axis and z-axisAngle γ, pitching angle theta and yaw angle ψ. While falling before and after human body, γ changes, and while falling in left and right, θ changes. Work as human bodyMove comparatively violent time, because sensor cannot be distinguished acceleration of gravity and self-acceleration, attitude angle can not be by addingSpeed is directly tried to achieve, and need to carry out attitude algorithm, and conventional attitude algorithm method is Quaternion Method.
Hypercomplex number is the number being made up of four units:
This instrument of hypercomplex number that Quaternion Method is introduced in algebra makes up the deficiency while describing angular velocity motions of rigid body by Eulerian angles.The basis of attitude algorithm is coordinate transform, and any attitude changes can be thought in moment by the angle of swaying for three timesSynthetic, this rotation order can not change, and the spatial attitude of human body can be regarded as successively around z axle, y axle, the basic rotation of x axle doAfter compound result. Through a series of mathematics conversions, the attitude angle representing by hypercomplex number is:
。
So, if hypercomplex number Q is definite, can calculate roll angle γ and pitching angle theta. Through mathematical derivation, hypercomplex numberThe differential equation can be expressed as:
Wherein:，、、Represent three axis angular rates, adopt and finish card this differential side of Algorithm for SolvingJourney, solves:
Wherein: I is unit matrix,，，、、For x, y, z axle existsAngle increment in sampling time interval.
Shown in Fig. 3, be the roll angle change curve of once falling forward, last angle is the angle in collision ground moment,Can find out, the variation of angle approaches 90 °, same, backward, left, while falling to the right, the variation of angle also approaches 90 °. ConsiderTo the different directions of falling, the absolute value sum of the variable quantity of roll angle and the angle of pitch is as angle character when relatively standing.
Thus, extract altogether 3 features, average, the standard deviation of acceleration in sliding time window, and relatively standTime roll angle and the angle of pitch the absolute value sum of variable quantity.
For the method for checking the present embodiment, prepare Wearable equipment. Wearable equipment is by MEMS sensor assembly, bluetoothModule and lithium battery composition. MEMS sensor assembly comprises inertial sensor MPU6050, and MPU6050 itself has integrated three axles and addedSpeed instrument and three-axis gyroscope. Sensor assembly gathers respectively 3-axis acceleration and three axis angular rates with the frequency of 100Hz, passes throughBluetooth module is passed to PC host computer, and PC host computer adopts visual c++ 6.0 to write data acquisition and data processor.
Because the daily routines at the positions such as the head of human body, wrist, large arm are frequent, sensor is worn on to these local appearancesEasily cause mistake alarm, often affect the subjective feeling of wearer and be positioned over chest, ergonomics data show, erectilityThe position of centre of gravity of lower human body is greatly about 56% place of Human Height, and it is proper therefore sensor being worn on to human body waist,The interference being brought by human body daily behavior is also less.
Adopt this device to carry out the collection of above-mentioned feature.
Gather respectively fall sample and daily routines behavior sample and carry out SVM model training and test. The present embodiment definition 4Plant the mode of falling: fall forward, fall back, fall and fall left to the right. Define 10 kinds of daily routines behaviors: stand, sitUnder, stand up from sitting down, squat down, stand up from squatting down, bend over, lie down, sit up, walk and jog from lying down. Experiment number is 5 people,Age 23 ± 1.3, height 171.4 ± 2.2cm, body weight 61.2 ± 4.6kg, the Sponge cushion of experiment at 4m × 1m × 0.2m of fallingOn complete. Fall sample and the daily routines behavior sample that obtain are divided into training sample set and test sample book collection, sample setComposition is in table 1.
The composition of table 1 sample set
| Sample set | The sample number of falling | Daily routines behavior sample number |
| Training set | 50 | 50 |
| Test set | 250 | 250 |
Training sample set is used for training SVM model, and test sample book collection is for assessment of the accuracy of classification. Adopt LIBSVM kitRealize SVM algorithm. The accuracy of fall detection can be weighed by following 2 indexs:
Wherein: the TP sample number that occurs and detect that represents to fall, FN represents the sample number of falling and occurring but fail to report, TN represents dayThe sample number that Chang Hangwei occurs and detects, FP represents to think daily behavior by mistake to be the sample number of falling.
The reserved different time gathers the sample of falling, and contrast simultaneously only adopts acceleration signature, only adopts angle characterAnd in conjunction with the situation of acceleration signature and angle character, carry out model training and test, experimental result is as follows:
Table 2 only adopts acceleration signature
| Index | 100ms | 150ms | 200ms | 250ms | 300ms |
| Sen/% | 98.8 | 98 | 98 | 99.2 | 99.2 |
| Spe/% | 65.2 | 69.2 | 71.6 | 64.8 | 42 |
Note: Sen=Sensitivity; Spe=Specificity
As can be seen from Table 2, while only adopting acceleration signature, the model standard that the sample training of falling that reserved 200ms gathers obtainsReally property is best.
Table 3 only adopts angle character
| Index | 100ms | 150ms | 200ms | 250ms | 300ms |
| Sen/% | 99.2 | 99.2 | 99.2 | 99.2 | 99.2 |
| Spe/% | 92 | 91.6 | 88.8 | 74 | 75.6 |
As can be seen from Table 3, while only adopting angle character, the model that the sample training of falling that reserved 100ms gathers obtains is accurateProperty is best.
Table 4 is in conjunction with acceleration signature and angle character
| Index | 100ms | 150ms | 200ms | 250ms | 300ms |
| Sen/% | 99.2 | 99.2 | 99.2 | 99.2 | 99.2 |
| Spe/% | 95.6 | 92.8 | 96 | 86 | 76 |
As can be seen from Table 4, during in conjunction with acceleration signature and angle character, the sample training of falling that reserved 200ms gathers obtainsModel accuracy best. Contrast the optimal detection result of these three kinds of situations, as shown in the table:
Table 5 comparing result
| Index | Only acceleration | Only angle | Acceleration and angle 5 --> |
| Sen/% | 98 | 99.2 | 99.2 |
| Spe/% | 71.6 | 92 | 96 |
As can be seen from Table 5, be better than adopting single acceleration signature or angle in conjunction with the result of acceleration signature and angle characterDegree feature. Analysis data find, while only adopting acceleration signature, squat down fast, the uniform acceleration of jogging changes more violent dailyIt is the behavior of falling that behavior is easily mistaken as, and while only adopting angle character, bends over, the equal angles of lying down changes daily behavior greatlyEasily be mistaken as is the behavior of falling.
This optimum SVM model is imported to PC host computer procedure, the sliding time window of 100ms is set, 50% window stackRate, the lead time of 250 groups of test sample books of falling of statistics, the results are shown in Table 6.
Table 6 lead time statistics
| Minimum | The highest | On average |
| 130ms | 570ms | 273ms |
Visible, in the present embodiment, adopt method of the present invention, the verification and measurement ratio of the behavior of falling is 99.2%, the inspection of daily routines behaviorSurvey rate is 96%, to comparatively violent daily routines behavior, also has higher verification and measurement ratio, and the average lead time reaches 273ms, forThe real-time early warning of falling and the startup of protective device provide the time.
Embodiment bis-: the method described in employing embodiment mono-, the length that changes sliding window is trained test.
Shown in accompanying drawing 4, be acceleration and the change curve of a walking process, first define length of window wWith the stack length o of window, for one section of time series data, first window is expressed as，Data in this window are extracted to acceleration and angular speed feature. Because the stack length of window is o, next windowData are expressed as, continue the data of this window to process, by that analogy, therefore onceDaily behavior can gather many group daily behavior samples. In experiment, the deadline of every kind of daily behavior is 5 seconds, daily behavior sampleThe sliding window length w of this collection gets 100ms equally, and window stack rate gets 50%, rear 50% data of current window as underFront 50% data of one window, therefrom select the sample that data variation is more violent, are easily mistaken for the sample of falling, and obtain altogether300 groups of daily routines behavior samples.
Adopt different sliding window length w to sample, in the full pattern of the behavior of falling, the sliding window of samplingTo actual setting aside some time as t of falling, result is as shown in the table:
| w/ms | t/ms | Sensitivity/% | Specificity/% | Lead time/ms |
| 100 | 200 | 99.6 | 96.8 | 283 |
| 150 | 150 | 100 | 96 | 269 |
| 200 | 200 | 100 | 78 | 296 |
| 250 | 150 | 99.6 | 100 | 271 |
| 300 | 150 | 99.6 | 99.6 | 266 |
As can be seen from the above table, when w gets 250ms, when the t that sets aside some time is 150ms, the fall detection model standard that sample training obtainsReally property is best, and using this fall detection model as optimum fall detection model, the behavior verification and measurement ratio of now falling is 99.6%, only has 1Group is fallen behavior less than detecting in face of the collision ground of falling, and daily routines behavior verification and measurement ratio is 100%, does not occur erroneous judgement, rightIn daily routines behavior more violent as jogging, also there is very high verification and measurement ratio. The average lead time of fall detection is271ms, the lead time distribution situation of 249 groups of behaviors of falling that detect sees the following form.
Lead time distributes
| Lead time distributed area | Sample group number 6 --> |
| <100ms | 0 |
| 100~200ms | 15 |
| 200~300ms | 185 |
| 300~400ms | 46 |
| >400ms | 3 |
Visible, the present embodiment can effectively carry out fall detection, and has enough lead times, can be for the human body guarantor that fallsThe control of protection unit.
Claims (9)
1. a human body fall detection method, is characterized in that, comprising:
1) obtaining of characteristic value:
1. the 3-axis acceleration instrument and the three-axis gyroscope that use human body to dress are sampled, and obtain the sampling of each sampling time pointData;
2. the sampled data 1. obtaining according to step, calculates respectively the three axle resultant accelerations that obtain each sampling time point, rollCorner, the angle of pitch，
Wherein,，ax、ay、azRespectively 3-axis acceleration, with both direction orthogonal in horizontal planeRespectively as x axle and y axle, roll angleThe attitude angle around x axle, the angle of pitchIt is the attitude angle around y axle;
3. the length of sliding time window and the stack rate of adjacent sliding time window are set, obtain in each sliding time windowThe average of three axle resultant accelerations of each sampled pointAnd standard deviation, last acquisition at sliding time window adoptedThe absolute value sum of the variable quantity of the roll angle of sampling point place human body during with respect to erectility and the angle of pitch, does with these three parametersFor characteristic value;
2) foundation of grader and training:
Adopt algorithm of support vector machine to build grader, three parameters that obtain using step 1) are as the input feature vector of graderValue;
Carry out respectively daily routines behavior and the different behaviors of falling by the personnel that train, obtain fall sample and daily workMoving behavior sample composing training collection, trains grader, obtains the grader after training;
3) adopt the grader after training to carry out fall detection according to the actual sensing data of human body obtaining.
2. human body fall detection method according to claim 1, is characterized in that: step 1) 1. in, sample frequency is notBe less than 50Hz.
3. human body fall detection method according to claim 2, is characterized in that: sample frequency is 100Hz.
4. human body fall detection method according to claim 1, is characterized in that: the length of sliding time window is 100～300ms, the stack rate of window is 40%～60%.
5. human body fall detection method according to claim 1, is characterized in that: the length of sliding time window is100ms, the stack rate of window is 50%.
6. the human body protective device of falling, comprises falling detection device, protection air bag and drive unit, it is characterized in that: instituteState falling detection device and mainly formed by 3-axis acceleration instrument, three-axis gyroscope and controller, in described controller, be provided with rightRequire the grader after the training described in 1.
7. the human body according to claim 6 protective device of falling, is characterized in that: described falling detection device is arranged on peopleThe waist of body.
8. the human body according to claim 6 protective device of falling, is characterized in that: described drive unit comprises through a controlThe compressed gas cylinder that valve processed is communicated with protection air bag, described control valve is opened and closed by described controller control.
9. the human body according to claim 6 protective device of falling, is characterized in that: described protection air bag is by being worn on respectivelyMultiple air bags at the fragile position of human body form.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610083726.5A CN105590408B (en) | 2016-02-06 | 2016-02-06 | A kind of tumble detection method for human body and protective device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610083726.5A CN105590408B (en) | 2016-02-06 | 2016-02-06 | A kind of tumble detection method for human body and protective device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105590408A true CN105590408A (en) | 2016-05-18 |
| CN105590408B CN105590408B (en) | 2018-05-04 |
Family
ID=55929958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610083726.5A Active CN105590408B (en) | 2016-02-06 | 2016-02-06 | A kind of tumble detection method for human body and protective device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105590408B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106227999A (en) * | 2016-07-16 | 2016-12-14 | 北京工业大学 | A kind of high-adaptability is fallen behavioral value method |
| CN106539587A (en) * | 2016-12-08 | 2017-03-29 | 浙江大学 | A kind of fall risk assessment and monitoring system and appraisal procedure based on sensor of doing more physical exercises |
| CN106910314A (en) * | 2017-02-03 | 2017-06-30 | 同济大学 | A kind of personalized fall detection method based on the bodily form |
| CN107495972A (en) * | 2017-08-14 | 2017-12-22 | 哈尔滨工业大学（威海） | One kind falls down detection algorithm |
| CN107933476A (en) * | 2017-11-14 | 2018-04-20 | 苏州数言信息技术有限公司 | The method and system of the sensing device of the anti-forgetting of general passenger inside the vehicle |
| CN108158591A (en) * | 2017-12-26 | 2018-06-15 | 深圳市康莱米电子股份有限公司 | Detect the method and device fallen down |
| WO2018107755A1 (en) * | 2016-12-15 | 2018-06-21 | 歌尔股份有限公司 | User behavior monitoring method and wearable device |
| CN108549900A (en) * | 2018-03-07 | 2018-09-18 | 浙江大学 | Tumble detection method for human body based on mobile device wearing position |
| CN109064710A (en) * | 2018-08-14 | 2018-12-21 | 东南大学 | With the security protection terminal and system for falling down identification function |
| CN109833048A (en) * | 2019-01-23 | 2019-06-04 | 中国民航大学 | A kind of alertness measurement method based on psychomotor abilities |
| CN109979162A (en) * | 2019-03-27 | 2019-07-05 | 苏州威斯德医疗科技有限公司 | A kind of anti-tumble alarm method and alarm system |
| CN109993312A (en) * | 2018-01-02 | 2019-07-09 | 中国移动通信有限公司研究院 | A kind of equipment and its information processing method, computer storage medium |
| CN110946585A (en) * | 2019-11-21 | 2020-04-03 | 上海理工大学 | Fall detection system and method based on data fusion and BP neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202650197U (en) * | 2012-07-09 | 2013-01-02 | 台州学院 | Fall prevention apparatus |
| US20130054180A1 (en) * | 2011-08-29 | 2013-02-28 | James R. Barfield | Method and system for detecting a fall based on comparing data to criteria derived from multiple fall data sets |
| CN104224182A (en) * | 2014-03-31 | 2014-12-24 | 桂林电子科技大学 | Method and device for monitoring human tumbling |
| CN104586398A (en) * | 2013-10-30 | 2015-05-06 | 广州华久信息科技有限公司 | Old man falling detecting method and system based on multi-sensor fusion |
| CN104622454A (en) * | 2015-01-23 | 2015-05-20 | 深圳市卡麦睿科技有限公司 | Multi-functional bracelet type human body intelligent monitor system |
-
2016
- 2016-02-06 CN CN201610083726.5A patent/CN105590408B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130054180A1 (en) * | 2011-08-29 | 2013-02-28 | James R. Barfield | Method and system for detecting a fall based on comparing data to criteria derived from multiple fall data sets |
| CN202650197U (en) * | 2012-07-09 | 2013-01-02 | 台州学院 | Fall prevention apparatus |
| CN104586398A (en) * | 2013-10-30 | 2015-05-06 | 广州华久信息科技有限公司 | Old man falling detecting method and system based on multi-sensor fusion |
| CN104224182A (en) * | 2014-03-31 | 2014-12-24 | 桂林电子科技大学 | Method and device for monitoring human tumbling |
| CN104622454A (en) * | 2015-01-23 | 2015-05-20 | 深圳市卡麦睿科技有限公司 | Multi-functional bracelet type human body intelligent monitor system |
Non-Patent Citations (2)
| Title |
|---|
| 刘磊: "基于Android平台的跌倒检测软件设计", 《万方数据》 * |
| 李路: "基于多传感器的人体运动模式识别研究", 《万方数据》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106227999A (en) * | 2016-07-16 | 2016-12-14 | 北京工业大学 | A kind of high-adaptability is fallen behavioral value method |
| CN106227999B (en) * | 2016-07-16 | 2018-10-09 | 北京工业大学 | A kind of high-adaptability tumble behavioral value method |
| CN106539587A (en) * | 2016-12-08 | 2017-03-29 | 浙江大学 | A kind of fall risk assessment and monitoring system and appraisal procedure based on sensor of doing more physical exercises |
| WO2018107755A1 (en) * | 2016-12-15 | 2018-06-21 | 歌尔股份有限公司 | User behavior monitoring method and wearable device |
| CN106910314A (en) * | 2017-02-03 | 2017-06-30 | 同济大学 | A kind of personalized fall detection method based on the bodily form |
| CN107495972A (en) * | 2017-08-14 | 2017-12-22 | 哈尔滨工业大学（威海） | One kind falls down detection algorithm |
| CN107933476A (en) * | 2017-11-14 | 2018-04-20 | 苏州数言信息技术有限公司 | The method and system of the sensing device of the anti-forgetting of general passenger inside the vehicle |
| CN107933476B (en) * | 2017-11-14 | 2019-08-13 | 苏州数言信息技术有限公司 | The method and system of the sensing device of the anti-forgetting of general passenger inside the vehicle |
| CN108158591A (en) * | 2017-12-26 | 2018-06-15 | 深圳市康莱米电子股份有限公司 | Detect the method and device fallen down |
| CN109993312A (en) * | 2018-01-02 | 2019-07-09 | 中国移动通信有限公司研究院 | A kind of equipment and its information processing method, computer storage medium |
| CN108549900A (en) * | 2018-03-07 | 2018-09-18 | 浙江大学 | Tumble detection method for human body based on mobile device wearing position |
| CN109064710A (en) * | 2018-08-14 | 2018-12-21 | 东南大学 | With the security protection terminal and system for falling down identification function |
| CN109833048A (en) * | 2019-01-23 | 2019-06-04 | 中国民航大学 | A kind of alertness measurement method based on psychomotor abilities |
| CN109979162A (en) * | 2019-03-27 | 2019-07-05 | 苏州威斯德医疗科技有限公司 | A kind of anti-tumble alarm method and alarm system |
| CN110946585A (en) * | 2019-11-21 | 2020-04-03 | 上海理工大学 | Fall detection system and method based on data fusion and BP neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105590408B (en) | 2018-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105590408A (en) | Human body falling detection method and protection device | |
| CN104598896B (en) | The falling over of human body automatic testing method followed the trail of based on Kinect skeletons | |
| CN109643499B (en) | System and method for swimming analysis | |
| Lee et al. | Inertial sensing-based pre-impact detection of falls involving near-fall scenarios | |
| Takhounts et al. | Development of brain injury criteria (BrIC) | |
| Wu et al. | Development of a wearable-sensor-based fall detection system | |
| Camarillo et al. | An instrumented mouthguard for measuring linear and angular head impact kinematics in American football | |
| Shi et al. | Mobile human airbag system for fall protection using MEMS sensors and embedded SVM classifier | |
| Pansiot et al. | Swimming stroke kinematic analysis with BSN | |
| CN106981174A (en) | A kind of Falls Among Old People detection method based on smart mobile phone | |
| Otanasap | Pre-impact fall detection based on wearable device using dynamic threshold model | |
| KR101872907B1 (en) | Motion analysis appratus and method using dual smart band | |
| Lai et al. | Adaptive body posture analysis using collaborative multi-sensors for elderly falling detection | |
| CN106618498A (en) | System and method for evaluating body balance | |
| US20160175681A1 (en) | Exercise analysis device, exercise analysis method, program, recording medium, and exercise analysis system | |
| Nakano et al. | Gaits classification of normal vs. patients by wireless gait sensor and Support Vector Machine (SVM) classifier | |
| CN107169512A (en) | The construction method of HMM SVM tumble models and the fall detection method based on the model | |
| Wang et al. | Head impact sensor triggering bias introduced by linear acceleration thresholding | |
| CN107662613B (en) | A kind of extreme driving behavior recognition methods and system based on mobile intelligent perception | |
| US20170028252A1 (en) | Swing diagnosis apparatus, swing diagnosis system, swing diagnosis method, and recording medium | |
| Aubin et al. | Support vector machine classification of Parkinson's disease, essential tremor and healthy control subjects based on upper extremity motion | |
| CN109758737A (en) | Posture correction device and method for underwater motion | |
| US11931636B2 (en) | Evaluation method, evaluation system and non-transitory computer-readable medium storing evaluation program | |
| Shi et al. | Development of a human airbag system for fall protection using mems motion sensing technology | |
| Woo et al. | A study on wearable airbag system applied with convolutional neural networks for safety of motorcycle |
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 |