CN103606248B - A kind of falling over of human body automatic testing method and system - Google Patents
A kind of falling over of human body automatic testing method and system Download PDFInfo
- Publication number
- CN103606248B CN103606248B CN201310456860.1A CN201310456860A CN103606248B CN 103606248 B CN103606248 B CN 103606248B CN 201310456860 A CN201310456860 A CN 201310456860A CN 103606248 B CN103606248 B CN 103606248B
- Authority
- CN
- China
- Prior art keywords
- condition information
- human body
- channel condition
- action
- falling over
- 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.)
- Active
Links
Landscapes
- Alarm Systems (AREA)
Abstract
A kind of falling over of human body automatic testing method and system, based on wireless network signal transmission technology, the step of described method includes: wireless receiving end receives the wireless signal from wireless transmitting terminals, and assesses channel condition information；Utilize the exception that Outlier Detection Algorithm identification channel condition information changes；Based on the one-class support vector machine that subject performance class and other action classes are made a distinction, cause abnormal patterns that channel condition information change for subject performance class with falling over of human body action, it may be judged whether generation falling over of human body action, the most then send and fall down alarm signal.The detection method rate of false alarm of the present invention is relatively low；Utilize existing wireless network and equipment, it is not necessary to other specific detection equipment is installed, there is high popularization；Detected person need not carry any additional sensor equipment simultaneously, it is to avoid detected person carries the inconvenience that detection equipment is caused.
Description
Technical field
The present invention relates to the information processing technology, particularly relate to a kind of falling over of human body automatic testing method and system.
Background technology
Nowadays, world population is in an aging trend that is unique and that do not allow change.As the topmost health threat of old people and the obstacle lived by oneself, have no emissary falling down and exacerbate senior health and fitness and the pressure of first aid cause.Therefore, we need to find a kind of can facilitating and the effective method that detection old people falls down and gives a response automatically in a hurry.In order to realize the accurate detection to action, there has been proposed the method utilizing environmental unit, vision or sensor to carry out detecting.But the detecting system itself using these methods to build all also exists all deficiencies.Falling down utilizing surrounding equipment to carry out in the system of detection, specific equipment and instrument needs to be installed in advance in the environment residing for detected person, and the vibration on video, audio and the floor experienced by equipment judges the action of detected person.Other sound around detected person or other environmental factorss are easy to the work of interference environment equipment, thus cause mistake fall down alarming result.And utilize the detecting system of falling down of vision, need high-resolution camera to be installed in the detection environment residing for detected person to shoot a large amount of image, and then determined whether the generation fallen down by captured image.But, camera is installed in the individual privacy that somewhat can invade detected person.Meanwhile, due to the impact of light factor, utilize falling down detecting system and effectively can not working under dark condition of vision.These systems all also exist by the big shortcoming of such environmental effects, and the limitation existing for these systems is all accurately and the detection that carries out easily falling down old people causes obstacle.
Along with the development of wireless communication technology, increasing wireless device is applied in the middle of life by people.Therefore, wireless communication technology is utilized to be remembered as a kind of effective and feasible method to the action of falling down detecting old people.Existing wireless communication technology proposes the method for the multiple various action of detection human body.Such as gather the three-dimensional acceleration information of metastomium on user by the 3-axis acceleration sensor of mobile device, and information is carried out fusion treatment, comprehensive human body is hit and is impacted and front and back gone up angle of inclination of body to judge whether human body is fallen down, the detection that many impact types are fallen down by the method and roll-down type is fallen down is still effective, has high detection rate；But use the method, detected person needs to carry mobile device, still can bring inconvenience.Separately there is one to fall down detection and location device and be equipped with inertial navigation module and GPS module, described inertial navigation module is by three-axis gyroscope, three axle magnetometers and three axis accelerometer composition, this falls down detection and location device relatives' hands machine by GSM/GPRS with binding tested personnel, monitoring station connects, and devise and based on neutral net and machine learning fall down detection algorithm, detection falling over of human body state and positional information exactly, can be accurately detected tested personnel falls down situation, and can be to its real-time track and localization, system is convenient and practical, accuracy rate is high, stability is strong；But it needs to be provided with specific professional equipment, cost is high and popularization is poor.Further, for utilize the sensor carried with or equipped with application program smart mobile phone fall down detecting system, although can carry out falling down detection in the case of not using other particular devices, but detected person cannot ensure carry sensors or mobile phone at any time, and carry sensors causes a certain degree of inconvenience can to the life of detected person.
Summary of the invention
In order to overcome the weak point of the prior art of above-mentioned indication, the present invention provides a kind of falling over of human body automatic testing method and system, with in specific indoor environment, by utilizing existing wireless network and equipment, realize the effective detection to falling over of human body action, and reach and alarm and the function of feedback.
The present invention is achieved by the following technical solutions:
A kind of falling over of human body automatic testing method, based on wireless network signal transmission technology, its step includes:
S1, wireless receiving end receive the wireless signal from wireless transmitting terminals, and assess channel condition information；
S2, utilize the exception that Outlier Detection Algorithm identification channel condition information changes；
S3, based on the one-class support vector machine that subject performance class and other action classes are made a distinction, cause abnormal patterns that channel condition information changes for subject performance class with falling over of human body action, judge whether falling over of human body action, the most then send and fall down alarm signal.
Further, the number of described wireless transmitting terminals is one or more, and the number of described wireless receiving end is one or more.
Described step S1 assessment channel condition information includes:
S11, collection initial channel status data, based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
S12, to each spatial flow, ask for the meansigma methods of the P on same time point the CSI value of subcarrier continuously, using this meansigma methods as channel condition information, P is the natural number being less than M more than 1；
S13, utilize data filtering techniques and moving average method that channel condition information is smoothed.
The exception that described step S2 utilizes Outlier Detection Algorithm identification channel condition information to change is time series Outlier Detection Algorithm based on local outlier factor.
Further, described step S2 includes:
S21, the time series of channel condition information is implemented data segmentation obtain subsequence, calculate the local outlier factor of subsequence；
S22, when described local outlier factor more than or equal to predetermined threshold value time, using subsequence as abnormal patterns export.
Further, described step S3 includes:
S31, based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
S32, map to the abnormal patterns that step S23 exports, in the high dimensional feature model of one-class support vector machine, isolate subject performance class；
There is falling over of human body action in S33, judgement, sends and fall down alarm signal.
The falling over of human body automatic testing method of the present invention also includes: step S4, feeds back for the response message falling down alarm signal, adjusts the high dimensional feature model of one-class support vector machine.
Present invention also offers a kind of falling over of human body automatic checkout system, based on wireless network signal transmission technology, including:
CSI acquisition module, receives the wireless signal from wireless transmitting terminals for wireless receiving end, and assesses channel condition information；
Abnormality detection module, for the exception utilizing Outlier Detection Algorithm identification channel condition information to change；
Action judge module, for based on one-class support vector machine subject performance class and other action classes made a distinction, the abnormal patterns caused with falling over of human body action is for subject performance class, it may be judged whether falling over of human body action occurs；
Alarm modules, for when judging to occur falling over of human body action, sending and fall down alarm signal.
Further, described CSI acquisition module includes:
Sensing unit, is used for gathering initial channel status data, and based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
Data processing unit, for each spatial flow, asking for the meansigma methods of the CSI value of the P on same time point continuous subcarrier, using this meansigma methods as channel condition information, P is the natural number more than 1 less than M；
Smooth unit, is used for utilizing data filtering techniques and moving average method to be smoothed channel condition information.
Further, described abnormality detection module time series based on local outlier factor Outlier Detection Algorithm, including:
LOF computing unit, obtains subsequence for the time series of channel condition information is implemented data segmentation, calculates the local outlier factor of subsequence；
Abnormal output unit, for when described local outlier factor is more than or equal to predetermined threshold value, exporting subsequence as abnormal patterns.
Further, described action judge module includes:
Set up model unit, for based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
Action recognition unit, in the high dimensional feature model that the abnormal patterns that abnormal output unit is exported maps to one-class support vector machine, isolates subject performance class.
The falling over of human body automatic checkout system of the present invention also includes a feedback module, for feeding back for the response message falling down alarm signal, adjusts the high dimensional feature model of one-class support vector machine.
Compared with prior art, the present invention is based on the radio propagation mechanism under indoor environment, channel condition information CSI is set up with human action and contacts, the action that human body occurs is judged by the change of CSI, determine the action whether detected person falls down, thus realize the effective function reported to the police, its beneficial effect includes: in the less indoor environment of ornament (such as laboratory), the Detection accuracy of detected action is 84%～94%, and in the indoor environment (such as dormitory) that ornament is more, Detection accuracy also reaches 78%, rate of false alarm is relatively low, only have 15%～22%；It is capable of falling down signal an alert after action judges, and utilizes the self-learning function of system to process wrong report situation, reduce rate of false alarm further；This detection method is on the basis of existing wireless network and equipment, carry out falling down the detection work of action, without installing other specific detection equipment in detected environment, can use in any environment of family, there is high popularization, detected person need not carry any additional sensor equipment simultaneously, it is to avoid detected person carries the inconvenience that detection equipment is caused, and provides convenience for its life.
Accompanying drawing explanation
Accompanying drawing 1 is that the hardware of the falling over of human body automatic checkout system of a kind of embodiment of the present invention constitutes schematic diagram；
Accompanying drawing 2 be the falling over of human body automatic testing method of the present invention realize general flow chart；
Accompanying drawing 3 be the falling over of human body automatic testing method of a kind of embodiment of the present invention realize schematic flow sheet；
Accompanying drawing 4 is the frame diagram of the falling over of human body automatic checkout system of a kind of embodiment of the present invention.
Detailed description of the invention
For the ease of the understanding of those skilled in the art, the invention will be further described with embodiment below in conjunction with the accompanying drawings.
A kind of falling over of human body automatic testing method, based on wireless network signal transmission technology, its step includes:
S1, wireless receiving end receive the wireless signal from wireless transmitting terminals, and assess channel condition information；
S2, utilize the exception that Outlier Detection Algorithm identification channel condition information changes；
S3, based on the one-class support vector machine that subject performance class and other action classes are made a distinction, cause abnormal patterns that channel condition information changes for subject performance class with falling over of human body action, judge whether falling over of human body action, the most then send and fall down alarm signal.
In actual applications, the falling over of human body automatic testing method of the present invention can realize on the application server.Preferably, described wireless receiving end is wireless network card, described wireless transmitting terminals is wireless router, the method is based on the radio propagation mechanism under indoor environment, set up the relation of wireless signal and human action, have only to use the existing Wireless Communication Equipment of family, the change of the wireless signal that i.e. can cause by changing detected person's action is analyzed, judge the action fallen down that whether detected person occur and report to the police, reduce the ambient environmental factors impact on falling down activity detection, by be greatly improved fall down detection accuracy.In specific indoor environment, the abundant channel condition information of wireless network can be collected by wireless network card.In the present invention, the number of described wireless transmitting terminals is one or more, and the number of described wireless receiving end is one or more.System is sent by many antennas and received wireless signal respectively；The wireless network card that system is used can be with receiving channel state information.Preferably, the number of wireless transmitting terminals and wireless receiving end is 2 or 3.As described in Figure 1, there are two wireless transmitting terminals (the first emitter and the second emitter) and two wireless receiving ends (the first receiving terminal and the second receiving terminal) in detected environment.Wherein the first receiving terminal receives the abbreviation of the CSI(Channel State Information from the first emitter, i.e. channel condition information, at wireless communication field, CSI is exactly the channel attribute of communication link, describe the signal weak factor on every transmission paths), the second receiving terminal receives the CSI from the second emitter.In detected environment, detected person is without carrying other extras.The action of detected person is detected by system by the CSI utilizing two wireless receiving ends to be received, and thus judges whether detected person falls down.
In order to set up the contact between wireless signal and human action, the present invention uses the channel condition information CSI of wireless network as indicant.CSI can be depicted under the joint effect of time delay, amplitude decay and phase shift, the route of transmission of a signal.Based on the radio propagation model under indoor environment, the present invention establishes contacting between CSI and human action.(such as a room) in a specific indoor environment, there is a main propagation path and multiple reflection path produced because of surrounding (ceiling, floor and wall) impact.In detected person is in this room, his health can produce a plurality of scattering path.When detected person keeps resting state in this context, it is in the receiving terminal in this environment and can receive stable Propagation of Energy.And when the generation action change of detected person, such as when detected person falls down, human influence the scattering point producing scattering path can change position rapidly, and the energy that this sudden change can cause receiving terminal to be received changes.And by this change, the present invention will judge which kind of action change human body specifically there occurs.The present invention utilizes orthogonal frequency carrier multiplexing (Orthogonal
Frequency Division Multiplex, OFDM) obtain with the CSI of subcarrier (subcarrier) form existence.And make to set up between the CSI obtained in this way and human action to contact to improve the accuracy rate that action is judged.
Specifically, in described step S1, assessment channel condition information includes:
S11, collection initial channel status data, based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
S12, to each spatial flow, ask for the meansigma methods of the P on same time point the CSI value of subcarrier continuously, using this meansigma methods as channel condition information, P is the natural number being less than M more than 1；
S13, utilize data filtering techniques and moving average method that channel condition information is smoothed.
When the system starts of the present invention, wireless transmitting terminals propagates wireless network signal, and the wireless receiving end being simultaneously in specific region (as equipped with the computer of network interface card) can collect CSI as initial channel status data, then carries out data process.As a example by Multiple Input Multiple Output (Multiple-Input Multiple-Out-put, MIMO) with use 3 × 3, the initial channel status data obtained at phase of sensitization can be divided into 9 spatial flows, and has 30 subcarriers in each flows.Finding through experiment, the data that different spatial flows can be comprised by the change of human action produce impact, and all subcarriers in each spatial flow then can be produced similar impact.Meanwhile, experiment shows, environmental factors (such as temperature, the setting in room) will also result in the CSI of collection and risen and fallen.Therefore, in the present invention, the CSI value of 30 subcarriers in each independent spatial flow is aggregated into a single channel condition information.Preferably, to each spatial flow, first try to achieve the CSI meansigma methods of continuous 5 subcarriers, and 9 spatial flows are taken respectively the CSI meansigma methods of same time point as channel condition information.In order to reduce the interference of environmental factors, the present invention utilizes data filtering techniques and moving average method, specifically, uses the method for weighted moving average (weighted
Moving average) it is smoothed by the channel condition information of above-mentioned process, to reduce the noise in data.
Described step S2 is intended to detect the abnormal change of wireless signal.Although different fixing human postures can cause different signal intensitys, but it has identical signal intensity tendency in the time domain, and the most gradually tends towards stability.And the action of human body, as walking, upright, fall down and can cause the unexpected fluctuating of obvious channel condition information.Preferably, the present invention utilizes based on local outlier factor (Local
Outlier Factor) Outlier Detection Algorithm, first normal profile is estimated, then calculates local outlier factor, thus isolate the abnormal patterns caused by various human actions.
The exception that described step S2 utilizes Outlier Detection Algorithm identification channel condition information to change is time series Outlier Detection Algorithm based on local outlier factor, further, and including:
S21, the time series of channel condition information is implemented data segmentation obtain some subsequences, calculate the local outlier factor of subsequence；
S22, when described local outlier factor more than or equal to predetermined threshold value time, using subsequence as abnormal patterns export.
Preferably, utilize time series Outlier Detection Algorithm, can limit by more accurately detecting standard, by separating with the time series corresponding to two kinds of postures of standing on foot, get rid of the abnormal patterns that this two kinds of common human actions cause.
After completing abnormality detection module, groups of people's body action will export the abnormal patterns of correspondence because of causing the significant change of channel condition information to be detected.Then, these abnormal patterns will be entered motion analysis, thus judge which kind of action abnormal patterns is belonging to.Fall down action to distinguish from these patterns, present invention uses one-class support vector machine (one-class based on the feature extracted from abnormal patterns
Support Vector Machine, one-class SVM).One class SVM is the algorithm of support vector machine of a kind of extension.In class SVM, all of sample is divided into target class and other classes.In order to solve the problem of Nonlinear Classification, input sample is mapped to a dimensional images.In the present invention, the abnormal patterns falling down action is considered subject performance class, and the abnormal patterns of other actions is regarded as other action classes.The abnormal patterns falling down action has been mapped to a dimensional images the most in advance.By utilizing class SVM to judge, can isolate from the abnormal patterns of previous step output and fall down action, depend on by the abnormal patterns exported and selected model, thus determine to there occurs which kind of action.
Further, described step S3 includes:
S31, based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
S32, map to the abnormal patterns that step S23 exports, in the high dimensional feature model of one-class support vector machine, isolate subject performance class；
There is falling over of human body action in S33, judgement, sends and fall down alarm signal.
The falling over of human body automatic testing method of the present invention also includes: step S4, feeds back for the response message falling down alarm signal, adjusts the high dimensional feature model of one-class support vector machine.After falling down action and being detected, the alarm processing stage will be entered.When certain abnormal patterns is identified as falling down action, and system can occur alarm.If alarm is shut off promptly, then explanation detected person need not or have been obtained for help, now system utilizes the response message of feedback, update the historical record that falling over of human body action judges, and for adjusting and improving the high dimensional feature model of one-class support vector machine, it is provided that the system feedback of detection and decision making algorithm can be optimized.If alarm is closed the most in time, system then can send help information by other equipment that signal associates to other people, such as sends instant messages or note etc. by third-party application and seeks help.
Flow chart as shown in Figure 2, discloses three important steps of the detection method of the present invention, including: the process of CSI data, abnormality detection and the classification of motion.
Specifically, as shown in Figure 3, present invention also offers a kind of embodiment falling over of human body automatic testing method realize flow process, its step includes:
S301, wireless receiving end receive the wireless signal from wireless transmitting terminals, gather initial channel status data simultaneously；
S302, ask for merge subcarrier CSI meansigma methods as channel condition information；
S303, channel condition information is smoothed；
S304, normal profile to channel condition information are estimated, utilize Outlier Detection Algorithm to calculate local outlier factor；
S305, output abnormality pattern；
S306, the abnormal patterns of output is mapped in the high dimensional feature model of one-class support vector machine；
S307, utilize one-class support vector machine to classify, wherein cause abnormal patterns that channel condition information changes for subject performance class with falling over of human body action；
S308, judge whether to isolate subject performance class, if so, perform step S209, otherwise return step S301；
There is falling over of human body action in S309, judgement, sends and fall down alarm signal；
S310, judge that whether alarm is shut off promptly in one sets time interval, if so, perform step S311, otherwise, perform step S312；
S311, to system feedback shutoff operation and abnormal patterns parameter, optimize detection and action evaluation algorithm；
S312, it is sent out help information.
Present invention also offers a kind of falling over of human body automatic checkout system, as shown in Figure 4, based on wireless network signal transmission technology, including:
CSI acquisition module 41, receives the wireless signal from wireless transmitting terminals for wireless receiving end, and assesses channel condition information；
Abnormality detection module 42, for the exception utilizing Outlier Detection Algorithm identification channel condition information to change；
Action judge module 43, for based on one-class support vector machine subject performance class and other action classes made a distinction, the abnormal patterns caused with falling over of human body action is for subject performance class, it may be judged whether falling over of human body action occurs；
Alarm modules 44, for when judging to occur falling over of human body action, sending and fall down alarm signal.
In the falling over of human body automatic checkout system of the present invention, the number of described wireless transmitting terminals is one or more, and the number of described wireless receiving end is one or more.Preferably, the number of wireless transmitting terminals and wireless receiving end is 2 or 3.
Further, described CSI acquisition module includes:
Sensing unit 411, is used for gathering initial channel status data, and based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
Data processing unit 412, for each spatial flow, asking for the meansigma methods of the CSI value of the P on same time point continuous subcarrier, using this meansigma methods as channel condition information, P is the natural number more than 1 less than M；
Smooth unit 413, is used for utilizing data filtering techniques and moving average method to be smoothed channel condition information.
Further, described abnormality detection module time series based on local outlier factor Outlier Detection Algorithm, including:
LOF computing unit 421, obtains some subsequences for the time series of channel condition information is implemented data segmentation, calculates the local outlier factor of subsequence；
Abnormal output unit 422, for when described local outlier factor is more than or equal to predetermined threshold value, exporting subsequence as abnormal patterns.
Further, described action judge module includes:
Set up model unit 431, for based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
Action recognition unit 432, in the high dimensional feature model that the abnormal patterns that abnormal output unit is exported maps to one-class support vector machine, isolates subject performance class.
The falling over of human body automatic checkout system of the present invention also includes a feedback module 45, for feeding back for the response message falling down alarm signal, adjusts the high dimensional feature model of one-class support vector machine.
Above content is to combine concrete optimal way further description made for the present invention, should not assert the present invention be embodied as be confined to described above.For those skilled in the art, without departing from the inventive concept of the premise, it is also possible to make some simple deduction or replace, within the protection domain that the claim being regarded as being submitted to by the present invention determines.
Claims (9)
1. a falling over of human body automatic testing method, based on wireless network signal transmission technology, its step includes:
S1, wireless receiving end receive the wireless signal from wireless transmitting terminals, and assess channel condition information；
S2, utilize the exception that Outlier Detection Algorithm identification channel condition information changes；
S3, based on the one-class support vector machine that subject performance class and other action classes are made a distinction, cause abnormal patterns that channel condition information changes for subject performance class with falling over of human body action, judge whether falling over of human body action, the most then send and fall down alarm signal；
The number of described wireless transmitting terminals is one or more, and the number of described wireless receiving end is one or more；
Described step S1 assessment channel condition information includes:
S11, collection initial channel status data, based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
S12, to each spatial flow, ask for the meansigma methods of the P on same time point the CSI value of subcarrier continuously, using this meansigma methods as channel condition information, P is the natural number being less than M more than 1；
S13, utilize data filtering techniques and moving average method that channel condition information is smoothed.
Falling over of human body automatic testing method the most according to claim 1, it is characterised in that: the exception that described step S2 utilizes Outlier Detection Algorithm identification channel condition information to change is time series Outlier Detection Algorithm based on local outlier factor.
Falling over of human body automatic testing method the most according to claim 2, it is characterised in that described step S2 includes:
S21, the time series of channel condition information is implemented data segmentation obtain subsequence, calculate the local outlier factor of subsequence；
S22, when described local outlier factor more than or equal to predetermined threshold value time, using subsequence as abnormal patterns export.
Falling over of human body automatic testing method the most according to claim 3, it is characterised in that: described step S3 includes:
S31, based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
S32, map to the abnormal patterns that step S22 exports, in the high dimensional feature model of one-class support vector machine, isolate subject performance class；
There is falling over of human body action in S33, judgement, sends and fall down alarm signal.
Falling over of human body automatic testing method the most according to claim 4, it is characterized in that, also include: step S4, feedback is for the response message falling down alarm signal, adjust the high dimensional feature model of one-class support vector machine, the definition of described high dimensional feature model is based on Statistical Learning Theory, pre-builds to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample.
6. a falling over of human body automatic checkout system, based on wireless network signal transmission technology, it is characterised in that including:
CSI acquisition module, receives the wireless signal from wireless transmitting terminals for wireless receiving end, and assesses channel condition information；
Abnormality detection module, for the exception utilizing Outlier Detection Algorithm identification channel condition information to change；
Action judge module, for based on one-class support vector machine subject performance class and other action classes made a distinction, the abnormal patterns caused with falling over of human body action is for subject performance class, it may be judged whether falling over of human body action occurs；
Alarm modules, for when judging to occur falling over of human body action, sending and fall down alarm signal；
Described CSI acquisition module includes:
Sensing unit, is used for gathering initial channel status data, and based on MIMO technique, described initial channel status data includes the CSI value of M subcarrier in N number of spatial flow, N and M is the natural number more than 1；
Data processing unit, for each spatial flow, asking for the meansigma methods of the CSI value of the P on same time point continuous subcarrier, using this meansigma methods as channel condition information, P is the natural number more than 1 less than M；
Smooth unit, is used for utilizing data filtering techniques and moving average method to be smoothed channel condition information.
Falling over of human body automatic checkout system the most according to claim 6, it is characterised in that: described abnormality detection module time series based on local outlier factor Outlier Detection Algorithm, including:
LOF computing unit, obtains subsequence for the time series of channel condition information is implemented data segmentation, calculates the local outlier factor of subsequence；
Abnormal output unit, for when described local outlier factor is more than or equal to predetermined threshold value, exporting subsequence as abnormal patterns.
Falling over of human body automatic checkout system the most according to claim 7, it is characterised in that: described action judge module includes:
Set up model unit, for based on Statistical Learning Theory, pre-build to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample；
Action recognition unit, in the high dimensional feature model that the abnormal patterns that abnormal output unit is exported maps to one-class support vector machine, isolates subject performance class.
Falling over of human body automatic checkout system the most according to claim 8, it is characterized in that, also include a feedback module, for feedback for the response message falling down alarm signal, adjust the high dimensional feature model of one-class support vector machine, the definition of described high dimensional feature model is based on Statistical Learning Theory, pre-builds to set the abnormal patterns that causes channel condition information to change due to every human action in the space high dimensional feature model as training sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310456860.1A CN103606248B (en) | 2013-09-30 | 2013-09-30 | A kind of falling over of human body automatic testing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310456860.1A CN103606248B (en) | 2013-09-30 | 2013-09-30 | A kind of falling over of human body automatic testing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103606248A CN103606248A (en) | 2014-02-26 |
| CN103606248B true CN103606248B (en) | 2016-08-10 |
Family
ID=50124467
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310456860.1A Active CN103606248B (en) | 2013-09-30 | 2013-09-30 | A kind of falling over of human body automatic testing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103606248B (en) |
Families Citing this family (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927851B (en) * | 2014-04-04 | 2016-06-29 | 余志峰 | A kind of individualized multi thresholds fall detection method and system |
| CN103971108A (en) * | 2014-05-28 | 2014-08-06 | 北京邮电大学 | Wireless communication-based human body posture recognition method and device |
| CN104065928B (en) * | 2014-06-26 | 2018-08-21 | 北京小鱼在家科技有限公司 | A kind of behavior pattern statistic device and method |
| CN104361361B (en) * | 2014-11-14 | 2018-04-03 | 北京天地弘毅科技有限公司 | The method and system for judging to fall down by cloud computing and machine learning algorithm |
| CN104615244A (en) * | 2015-01-23 | 2015-05-13 | 深圳大学 | Automatic gesture recognizing method and system |
| CN104766427B (en) * | 2015-04-27 | 2017-04-05 | 太原理工大学 | A kind of house illegal invasion detection method based on Wi Fi |
| CN105761407B (en) * | 2016-01-06 | 2019-05-10 | 深圳大学 | Indoor detection fire behavior and alarm method and system based on wireless network signal transmission |
| CN105933080B (en) * | 2016-01-20 | 2020-11-03 | 北京大学 | Fall detection method and system |
| CN105809110A (en) * | 2016-02-24 | 2016-07-27 | 南京大学 | Behavior identification system and method based on wireless signal identity |
| CN106198868A (en) * | 2016-07-05 | 2016-12-07 | 深圳大学 | The method and system of Humidity Detection based on wireless aware |
| CN106802404B (en) * | 2017-01-18 | 2018-10-16 | 深圳大学 | A kind of human body indoor positioning automatic testing method and system |
| CN107146377A (en) * | 2017-04-14 | 2017-09-08 | 深圳大学 | Fall detection method and device before collision |
| CN107749143B (en) * | 2017-10-30 | 2023-09-19 | 安徽工业大学 | WiFi signal-based system and method for detecting falling of personnel in through-wall room |
| CN109745026A (en) * | 2017-11-07 | 2019-05-14 | 深圳欧德蒙科技有限公司 | A kind of method for measuring heart rate and system |
| CN108833036B (en) * | 2018-05-31 | 2020-11-03 | 湖南大学 | Human body falling detection method |
| CN109190605B (en) * | 2018-10-30 | 2020-05-15 | 宁波大学 | Human body continuous action counting method based on CSI |
| CN109394229A (en) * | 2018-11-22 | 2019-03-01 | 九牧厨卫股份有限公司 | A kind of fall detection method, apparatus and system |
| TWI761671B (en) | 2019-04-02 | 2022-04-21 | 緯創資通股份有限公司 | Living body detection method and living body detection system |
| CN110706463A (en) * | 2019-06-11 | 2020-01-17 | 南京信息工程大学 | WIFI passive sensing method and system suitable for tumble monitoring |
| CN112545496A (en) * | 2020-11-09 | 2021-03-26 | 金茂智慧科技(广州)有限公司 | Wireless technology identification method for detecting human body tumble |
| CN112446426A (en) * | 2020-11-23 | 2021-03-05 | 中国科学技术大学 | Tumble detection method and device, electronic device and storage medium |
| EP4297645A1 (en) * | 2021-02-24 | 2024-01-03 | Signify Holding B.V. | Fall-classification device and arrangement |
| CN113499064A (en) * | 2021-07-07 | 2021-10-15 | 郑州大学 | Wi-Fi perception human body tumbling detection method and system in bathroom scene |
| CN113870528A (en) * | 2021-09-17 | 2021-12-31 | 四川华能宝兴河水电有限责任公司 | Method for rapidly sensing dangerous events through intelligent safety helmet |
| CN114758476B (en) * | 2022-04-18 | 2024-04-26 | 上海物骐微电子有限公司 | Activity information monitoring method, device, system, equipment and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103211599A (en) * | 2013-05-13 | 2013-07-24 | 桂林电子科技大学 | Method and device for monitoring tumble |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7076102B2 (en) * | 2001-09-27 | 2006-07-11 | Koninklijke Philips Electronics N.V. | Video monitoring system employing hierarchical hidden markov model (HMM) event learning and classification |
| TWI356357B (en) * | 2007-12-24 | 2012-01-11 | Univ Nat Chiao Tung | A method for estimating a body pose |
| CN102136180B (en) * | 2011-03-11 | 2013-09-18 | 北京航空航天大学 | Device for detecting and alarming human body tumble |
| CN202033874U (en) * | 2011-03-29 | 2011-11-09 | 江洪 | Tumbling alarm system |
| CN102289911A (en) * | 2011-07-19 | 2011-12-21 | 中山大学深圳研究院 | Falling detection system based on pyroelectric infrared rays |
| CN102421189B (en) * | 2011-11-22 | 2014-04-30 | 中山大学 | Physical-layer-based method and device for realizing indoor positioning of wireless network |
| CN102595324A (en) * | 2012-03-31 | 2012-07-18 | 上海交通大学 | Wireless sensor network-based regional falling detection positioning system and method |
-
2013
- 2013-09-30 CN CN201310456860.1A patent/CN103606248B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103211599A (en) * | 2013-05-13 | 2013-07-24 | 桂林电子科技大学 | Method and device for monitoring tumble |
Non-Patent Citations (1)
| Title |
|---|
| 基于室内无线传感器网络射频信号的老年人跌倒检测研究;霍宏伟等;《电子学报》;20110131;第39卷(第1期);第195-200页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103606248A (en) | 2014-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103606248B (en) | A kind of falling over of human body automatic testing method and system | |
| CN105310696B (en) | A kind of fall detection model building method and corresponding fall detection method and device | |
| Gu et al. | Paws: Passive human activity recognition based on wifi ambient signals | |
| Zhang et al. | Anti-fall: A non-intrusive and real-time fall detector leveraging CSI from commodity WiFi devices | |
| CN106802404B (en) | A kind of human body indoor positioning automatic testing method and system | |
| Lim et al. | Fall‐detection algorithm using 3‐axis acceleration: combination with simple threshold and hidden Markov model | |
| CN104615244A (en) | Automatic gesture recognizing method and system | |
| CN109816933B (en) | Intelligent monitoring system and monitoring method for preventing old people from falling down based on composite sensor | |
| CN107331136A (en) | Indoor human body activity detection approach and system based on WiFi | |
| CN105913614A (en) | Method and device of monitoring tumbling, intelligent wearable device and tumbling alarm system | |
| CN105336102A (en) | Fall detecting and locating method and device | |
| Wittenburg et al. | Cooperative event detection in wireless sensor networks | |
| Tung et al. | Use of phone sensors to enhance distracted pedestrians’ safety | |
| US10061287B2 (en) | Apparatus, a system and a method for monitoring a usage of an item | |
| CN105225419A (en) | Fall detection method, system and the autoalarm of falling based on this system | |
| CN103281779B (en) | Based on the radio frequency tomography method base of Background learning | |
| CN105741491B (en) | Fall detection warning system and method based on Kalman filtering Yu KNN algorithms | |
| CN109920192A (en) | Fire alarm method, system and computer readable storage medium | |
| CN106650300B (en) | Old man monitoring system and method based on extreme learning machine | |
| CN110737201B (en) | Monitoring method and device, storage medium and air conditioner | |
| CN110706463A (en) | WIFI passive sensing method and system suitable for tumble monitoring | |
| CN106095293A (en) | A kind of application operating method and apparatus of mobile device | |
| CN106488014A (en) | The safety prompt function method and system of mobile terminal | |
| US20130182115A1 (en) | Apparatus, System and Method for Electronic Motion Detection | |
| CN107003900A (en) | A kind of switching method and portable electric appts |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |