CN206697008U - A kind of falling detection device based on Fusion - Google Patents
A kind of falling detection device based on Fusion Download PDFInfo
- Publication number
- CN206697008U CN206697008U CN201720107403.5U CN201720107403U CN206697008U CN 206697008 U CN206697008 U CN 206697008U CN 201720107403 U CN201720107403 U CN 201720107403U CN 206697008 U CN206697008 U CN 206697008U
- Authority
- CN
- China
- Prior art keywords
- human body
- module
- tumble
- detection device
- emergency
- 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.)
- Expired - Fee Related
Links
Abstract
The utility model discloses a kind of falling detection device based on Fusion, the falling detection device is made up of pressure sensor, acceleration transducer, wireless sending module, wireless receiving module, microprocessor, emergency module etc..Wherein, the method that the fall detection method uses pattern-recognition, human body behavior is divided into tumble pattern and ADL pattern classes, the characteristic vector for being used as human body tumble criterion is filtered out by the machine learning method based on SVMs, human body tumble behavior is detected according to characteristic vector and human motion state data aggregate.By the fall detection method, monitoring in real time and processing are carried out to the pressure data during tumble, acceleration information, correctly identify human body tumble behavior, and positional information is sent to family members or caregiver, so as to give treatment in time.The utility model is based on multi-sensor data extraction characteristic vector, effectively increases the recognition capability of tumble.
Description
Technical field
It the utility model is related to Automatic Measurement Technique field, and in particular to a kind of tumble based on Fusion
Detection means.
Background technology
With the development of the society, aging population is the trend that human social development can not prevent, also turn into China already
Very important development problem.It data show, at the year end of cut-off 2014, more than the 60 years old the elderly's population in China accounts for up to 2.1 hundred million
The 15.5% of total population, over-65s population 1.3 hundred million, account for the 10.1% of total population.
Because each organ dysfunction of old man's body starts aging, physical function is caused to decline, reaction is insensitive, often anticipates
Outer tumble situation, this has become one of major reason for threatening the elderly's life and health.According to statistics, in the old age of over-65s
In crowd, have more than 1/3 people has tumble to undergo every year, such as cannot in time give treatment to, easily trigger complication, gently then cause
Limb injury, it is heavy then cause paralysis in addition lose life.Therefore, it is monitored to lacking the elderly's daily behavior looked after, it is real
When detect body posture, find to fall in time and alarm and seeks help so that the carry out outdoor activities that the elderly can trust, enhancing are old
The sense of security and self-confidence of year people, it is significant with physical and mental health so as to improve life of elderly person quality.
Domestic and foreign scholars have had the research of long period with tumble detection method for human body, and the method for fall detection also has
Many types.Research both domestic and external is summarized, following 3 class is broadly divided into for the method for fall detection：(1) based on the vibration such as audio frequency
The fall detection method of signal.By installing the audio signal in multiple acoustic sensor arrays acquisition surrounding environment indoors,
Tumble event can be judged by vibration frequency caused by analytic activity.Such detecting system is typically only applicable to interior, detection
Scope be limited.(2) fall detection method based on video image.Mainly obtained by the camera installed indoors monitored
The action message of people, such detection method have been similarly subjected to the limitation of monitoring range, and larger with fund input.(3) it is based on
The fall detection method of wearable device.The fall detection method of wearable device is mainly by being worn on the person
Lightweight sensor carries out the data acquisition of human body attitude, can detect and gather the acceleration of human body, inclination angle, vibrations, impact
With the data such as multifreedom motion, judge whether to fall by tumble algorithm process.Algorithm based on threshold value identification is in algorithm
It is relatively more directly perceived in design, it is easy to accomplish, it is method more conventional in current fall detection, its deficiency is the setting pair of threshold value
Discrimination has a great influence, and the setting of threshold value needs to establish on the basis of intuitive judgment or mass data, and due to of people
For body difference, it is necessary to set different threshold values, the complexity of identification model is higher.
Utility model content
The purpose of this utility model is to solve drawbacks described above of the prior art, there is provided one kind is based on multisensor number
According to the falling detection device of fusion, the device is by based on SVMs and the dual judgement human body tumble behavior of threshold values, having
Higher reliability and distinguishing speed.
The purpose of this utility model can be reached by adopting the following technical scheme that：
A kind of falling detection device based on Fusion, the falling detection device include, pressure sensing
Device, 3-axis acceleration sensor, wireless sending module, wireless receiving module, microprocessor, emergency module,
Wherein, the pressure sensor and the 3-axis acceleration sensor are connected with the wireless sending module respectively,
The wireless sending module passes through wireless connection, the wireless receiving module and the microprocessor with the wireless receiving module
Connection, the micro treatment module carry out analysis judgement according to pressure sensor and 3-axis acceleration data, judge whether human body is sent out
It is raw to fall；The microprocessor is connected with the emergency module, and the emergency module sends an SOS according to result of determination.
Further, the pressure sensor and the 3-axis acceleration sensor are arranged at Human Sole.
Further, the emergency module includes the one or two of short message emergency function or alarm emergency function.
Further, at the pressure and acceleration information that the microprocessor is received to the wireless receiving module
Reason, according to human body tumble model eigenvectors and acceleration threshold values based on SVMs, comprehensive descision human body real time kinematics
State.
Further, the emergency module sends an SOS after judging to fall, if generation is failed to report or reports situation by mistake,
Can be by manually opened, or can manual abort after sending an SOS.
The utility model is had the following advantages relative to prior art and effect：
1. the utility model is carried out real using pressure sensor and 3-axis acceleration sensor to human motion state data
When monitor, collection plantar pressure data can be crossed and establish the identification model based on SVMs, as judge to fall first
Part；By gathering 3-axis acceleration data setting acceleration threshold values, as the second condition for judging to fall；Comprehensive two conditions are equal
When being true, just carry out falling down alarm, further increase the reliability for judging to fall, reduce wrong report.
2. by pattern-recognition and the dual judgement of threshold values, the recognition capability to different tumble states can be effectively improved.
Brief description of the drawings
Fig. 1 is the process step of the fall detection tumble method based on Fusion disclosed in the utility model
Figure；
Fig. 2 is to judge the overhaul flow chart whether human body falls；
Fig. 3 is the structural representation of the fall detection tumble device based on Fusion disclosed in the utility model
Figure.
Embodiment
It is new below in conjunction with this practicality to make the purpose, technical scheme and advantage of the utility model embodiment clearer
Accompanying drawing in type embodiment, the technical scheme in the embodiment of the utility model is clearly and completely described, it is clear that is retouched
The embodiment stated is the utility model part of the embodiment, rather than whole embodiments.Based on the implementation in the utility model
Example, the every other embodiment that those of ordinary skill in the art are obtained under the premise of creative work is not made, is belonged to
The scope of the utility model protection.
Embodiment one
As shown in figure 3, a kind of falling detection device based on Fusion disclosed in the present embodiment, including,
Pressure sensor, 3-axis acceleration sensor, wireless sending module, wireless receiving module, microprocessor, emergency module, it is described
Pressure sensor, the 3-axis acceleration sensor are connected with the wireless sending module respectively, the wireless sending module with
The wireless receiving module is connected by wireless connection, the wireless receiving module with the microprocessor, the microprocessor mould
Root tuber carries out analysis judgement according to pressure sensor and 3-axis acceleration data, judges whether human body falls；The microprocessor
Device is connected with the emergency module, and the emergency module sends an SOS according to result of determination.
In embodiment, pressure sensor and 3-axis acceleration sensor are arranged at Human Sole.
In embodiment, emergency module includes the one or two of short message emergency function or alarm emergency function.
In embodiment, at the pressure and acceleration information that microprocessor is received to wireless receiving module
Reason, according to human body tumble model eigenvectors and acceleration threshold values based on SVMs, comprehensive descision human body real time kinematics
State.
In embodiment, sent an SOS after judging to fall, situation is failed to report or reported by mistake to emergency module in generation
When, can be by manually opened, or can manual abort after sending an SOS.
In embodiment, the microprocessor multi-feature vector and acceleration information, based on multi-sensor data
The fall detection method of fusion, correctly identify human body tumble behavior.
As depicted in figs. 1 and 2, the above-mentioned fall detection method based on Fusion, comprises the following steps：
S1, the sufficient amount of human motion state data sample of collection.
In order to build the human body tumble identification model based on SVMs, supporting vector machine model is instructed
Practice, gather the sample data of multigroup human motion state respectively, include the tumble action data and ADL of all ages and classes stage crowd
(activity of human body daily behavior, Activity of daily livings) data.
Wherein tumble action include fall forward, fall back, fall, fall to the right to the left, face upward lie fall and prostrate fall
Wait tumble type.ADL data such as include normal walking, stair activity, running and standing, sit down, squat down at the conversion of posture
Process.
S2, human body tumble plantar pressure characteristic vector is extracted from sample data, establish the human body based on SVMs
Tumble identification model.As shown in S2 in Fig. 1, including step：
S21, construction feature vector storehouse：The human body movement data sample collected is pre-processed, forms human motion
State sample characteristic vector, construction feature vector storehouse；
S22, training grader：The pressure data in human body movement data sample is extracted, according to former based on SVMs
Reason, it is trained with human body in the characteristic vector that the data of different motion state are formed, distinguishes human body tumble state and ADL shapes
State, obtain detecting the supporting vector machine model of body state.
The kernel function of the present embodiment selection is RBF functions.
Radial basis kernel function (RBF):
In the present embodiment, σ=0.5 of Radial basis kernel function is set, but above-mentioned σ value is not formed to the technical program
Limitation.
After obtaining normalized characteristic vector, characteristic vector model is carried out using support vector machine classifier design method
Training, and the Detection results of characteristic vector model are constantly verified by new samples, so as to which sample is demarcated in amendment repeatedly.Final instruction
Optimal classification function is obtained after white silk is：
S23, structure identification model：Extraction judges to distinguish the feature that human body is fallen with normal condition from training characteristics sample
Vector, build tumble model.
Standard feature vector in tumble characteristic vector storehouse is exactly the SVMs determined in training process, by feature to
Amount sample is contrasted with standard feature vector, you can identifies the Status Type belonging to characteristic vector sample.
S3, monitoring human motion state data, according to the human body tumble model eigenvectors based on SVMs, identification
Whether human body falls.
The characteristic vector sample that the Human Sole pressure data monitored is obtained, compared with standard feature vector, by
Radial basis kernel function (RBF) similarity measurement, judge whether the characteristic vector sample for belonging to tumble Status Type, judge that human body is
It is no to fall.
Wherein, whether the human body tumble model eigenvectors based on SVMs, identification human body fall, and only judge to fall
First condition.
S4, extraction human motion state acceleration information, judge whether the acceleration is more than the acceleration threshold values of setting.
Acceleration transducer data are sent to microprocessor by wireless module, are compared with the acceleration threshold values of setting,
If being more than threshold values, tumble type just it has been judged as.
Wherein, judge whether human body falls by acceleration threshold values, to judge the second condition fallen.
S5, multi-feature vector and acceleration information, judge to send an SOS after falling.
Wherein, distress signal is sent by emergency module, and can be by manually opened, or can be manual after sending an SOS
Stop.
Above-described embodiment is the preferable embodiment of the utility model, but embodiment of the present utility model is not by above-mentioned
The limitation of embodiment, it is other it is any without departing from Spirit Essence of the present utility model with made under principle change, modify, replace
Generation, combination, simplify, should be equivalent substitute mode, be included within the scope of protection of the utility model.
Claims (4)
- A kind of 1. falling detection device based on Fusion, it is characterised in that the falling detection device includes, Pressure sensor, 3-axis acceleration sensor, wireless sending module, wireless receiving module, microprocessor, emergency module,Wherein, the pressure sensor and the 3-axis acceleration sensor are connected with the wireless sending module respectively, described Wireless sending module is connected with the wireless receiving module by wireless connection, the wireless receiving module with the microprocessor Connect, the micro treatment module carries out analysis judgement according to pressure sensor and 3-axis acceleration data, judges whether human body occurs Fall；The microprocessor is connected with the emergency module, and the emergency module sends an SOS according to result of determination；It is described Pressure sensor and the 3-axis acceleration sensor are arranged at Human Sole.
- A kind of 2. falling detection device based on Fusion according to claim 1, it is characterised in that institute Stating emergency module includes the one or two of short message emergency function or alarm emergency function.
- A kind of 3. falling detection device based on Fusion according to claim 1, it is characterised in that institute State pressure that microprocessor received to the wireless receiving module and acceleration information is handled, according to being based on supporting vector The human body tumble model eigenvectors and acceleration threshold values of machine, comprehensive descision human body real time kinematics state.
- A kind of 4. falling detection device based on Fusion according to claim 1, it is characterised in that institute Emergency module is stated to send an SOS after judging to fall, can be by manually opened if producing when failing to report or reporting situation by mistake, or sending out Going out distress signal afterwards can manual abort.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720107403.5U CN206697008U (en) | 2017-02-04 | 2017-02-04 | A kind of falling detection device based on Fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720107403.5U CN206697008U (en) | 2017-02-04 | 2017-02-04 | A kind of falling detection device based on Fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN206697008U true CN206697008U (en) | 2017-12-01 |
Family
ID=60439316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201720107403.5U Expired - Fee Related CN206697008U (en) | 2017-02-04 | 2017-02-04 | A kind of falling detection device based on Fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN206697008U (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106781278A (en) * | 2017-02-04 | 2017-05-31 | 华南理工大学 | A kind of fall detection method and device based on Fusion |
| CN112400191A (en) * | 2018-06-29 | 2021-02-23 | 皇家飞利浦有限公司 | Fall detection apparatus, method of detecting a fall of an object and computer program product for implementing the method |
| CN112842277A (en) * | 2021-02-08 | 2021-05-28 | 上海理工大学 | Fall detection method and device based on multiple sequential probability ratio detection |
-
2017
- 2017-02-04 CN CN201720107403.5U patent/CN206697008U/en not_active Expired - Fee Related
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106781278A (en) * | 2017-02-04 | 2017-05-31 | 华南理工大学 | A kind of fall detection method and device based on Fusion |
| CN112400191A (en) * | 2018-06-29 | 2021-02-23 | 皇家飞利浦有限公司 | Fall detection apparatus, method of detecting a fall of an object and computer program product for implementing the method |
| CN112842277A (en) * | 2021-02-08 | 2021-05-28 | 上海理工大学 | Fall detection method and device based on multiple sequential probability ratio detection |
| CN112842277B (en) * | 2021-02-08 | 2022-08-09 | 上海理工大学 | Fall detection method and device based on multiple sequential probability ratio detection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106781278A (en) | A kind of fall detection method and device based on Fusion | |
| CN102302370B (en) | Method and device for detecting tumbling | |
| WO2017125094A1 (en) | Body fall smart control system and method therefor | |
| CN206697008U (en) | A kind of falling detection device based on Fusion | |
| CN105310696B (en) | A kind of fall detection model building method and corresponding fall detection method and device | |
| CN103886715B (en) | A kind of tumble detection method for human body | |
| US11047706B2 (en) | Pedometer with accelerometer and foot motion distinguishing method | |
| CN103810817B (en) | A kind of detection alarm method of the wearable human paralysis device of falling detection alarm | |
| US20130054180A1 (en) | Method and system for detecting a fall based on comparing data to criteria derived from multiple fall data sets | |
| CN205050303U (en) | Human paralysis of intelligence is monitoring devices | |
| CN103211599A (en) | Method and device for monitoring tumble | |
| CN106887115A (en) | A kind of Falls Among Old People monitoring device and fall risk appraisal procedure | |
| Zhao et al. | Recognition of human fall events based on single tri-axial gyroscope | |
| CN103177530B (en) | Human fall detection method based on triaxial acceleration | |
| CN110047248A (en) | It is a kind of based on wearable Falls Among Old People detection method and equipment | |
| CN105286240B (en) | A kind of intelligent crutch device and its monitor system | |
| CN106407996A (en) | Machine learning based detection method and detection system for the fall of the old | |
| CN112489368A (en) | Intelligent falling identification and detection alarm method and system | |
| KR102337861B1 (en) | Apparatus for fall prediction using lead time and method thereof | |
| CN105469546A (en) | Tumbling alarm system and method | |
| CN114469076A (en) | Identity feature fused old solitary people falling identification method and system | |
| Shi et al. | Fall detection system based on inertial mems sensors: Analysis design and realization | |
| US11749086B2 (en) | Fall detector incorporating physiological sensing | |
| CN107019501B (en) | Remote tumble detection method and system based on genetic algorithm and probabilistic neural network | |
| CN111626273B (en) | Fall behavior recognition system and method based on atomic action time sequence characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171201 Termination date: 20190204 |