WO2018036281A1 - 基于振动检测的智能家居监控方法及系统 - Google Patents

基于振动检测的智能家居监控方法及系统 Download PDF

Info

Publication number
WO2018036281A1
WO2018036281A1 PCT/CN2017/091671 CN2017091671W WO2018036281A1 WO 2018036281 A1 WO2018036281 A1 WO 2018036281A1 CN 2017091671 W CN2017091671 W CN 2017091671W WO 2018036281 A1 WO2018036281 A1 WO 2018036281A1
Authority
WO
WIPO (PCT)
Prior art keywords
person
characteristic parameter
vibration
signal
advance
Prior art date
Application number
PCT/CN2017/091671
Other languages
English (en)
French (fr)
Inventor
伍楷舜
王璐
陈文强
Original Assignee
深圳大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳大学 filed Critical 深圳大学
Publication of WO2018036281A1 publication Critical patent/WO2018036281A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B13/00—Burglar, theft or intruder alarms
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B15/00—Systems controlled by a computer
    • G05B15/02—Systems controlled by a computer electric
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B19/00—Alarms responsive to two or more different undesired or abnormal conditions, e.g. burglary and fire, abnormal temperature and abnormal rate of flow
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS], computer integrated manufacturing [CIM]
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/20—Pc systems
    • G05B2219/26—Pc applications
    • G05B2219/2642—Domotique, domestic, home control, automation, smart house

Definitions

  • the invention belongs to the field of smart home technology, and in particular relates to an indoor positioning and monitoring method of a smart home.
  • WIFI-based human behavior detection many researchers in the field of Internet of Things technology are now engaged in WIFI-based human behavior detection.
  • the signal modeling of WIFI or the method of machine learning can identify human beings walking, running, falling, sitting down, Brushing, cooking, etc., using the energy diminishing characteristics or propagation speed of WIFI can achieve indoor positioning of people and obtain the walking trajectory of people.
  • the advantage of WIFI lies in its popularity, which is basically every household.
  • the WIFI signal has multipath and non-line-of-sight problems. When the wall is separated, the signal strength will decrease sharply, which greatly affects the accuracy of indoor positioning and the recognition accuracy of human behavior.
  • WIFI-based technology requires a fixed environment.
  • the present invention provides a smart home monitoring method based on vibration detection, which collects structural vibration signals caused by human behavior and analyzes them, especially the ground caused by footsteps.
  • the vibration signal solves the above-mentioned shortcomings of the existing monitoring technology, and can locate, track and conduct behavior monitoring of indoor characters.
  • a smart home monitoring method based on vibration detection includes the following steps: a data acquisition module collects structural vibration information generated by a person's behavioral activities, and intelligently monitors the behavior of the family by analyzing and processing the signal and alerts the abnormality.
  • the analysis process is specifically:
  • the characteristic parameter is a set including a plurality of elements
  • the characteristic parameter of the legal person stored in advance is obtained by collecting the action of the legal person in the house in advance, and has the same structure as the currently obtained characteristic parameter.
  • step B) further comprises:
  • step C) further includes the following steps:
  • step C3) Returning to step C1) Obtaining the characteristic parameters of another previously stored legal person who has not been compared. The above steps are repeated; if all the characteristic parameters of the legal persons stored in advance have been compared, it is determined that the current person is an illegal person.
  • the invention also provides a smart home monitoring system based on vibration detection, comprising:
  • a signal acquisition unit configured to collect vibration information of a house structure caused by an action of a person in the house by using a vibration sensor, to obtain a vibration signal;
  • a feature parameter forming unit configured to process the obtained vibration signal to obtain a characteristic parameter of the vibration signal indicating an active person identity; the feature parameter is a set including a plurality of elements;
  • the comparing unit is configured to compare the obtained characteristic parameter with the characteristic parameter of the at least one legal person stored in advance, and determine whether the currently obtained characteristic parameter matches one of the characteristic parameters of the legal person stored in advance, and if yes, determine the current activity.
  • the person is a legal person; otherwise, it is judged that an illegal person enters the house and sends an alarm signal;
  • the characteristic parameter of the legal person stored in advance is obtained by collecting the action of the legal person in the house in advance, and has the same structure as the currently obtained characteristic parameter.
  • the characteristic parameter forming unit further includes:
  • a signal acquisition module configured to obtain, in the oscillating signal, a signal including a first set number of vibrations caused by actions of a person in the house as a feature parameter extraction signal;
  • a parameter obtaining module configured to respectively obtain an amplitude of each vibration in the characteristic parameter extraction signal; and obtain an interval time of two adjacent vibrations in the characteristic parameter extraction signal;
  • the feature parameter forming module is configured to arrange the amplitude and the interval time obtained in the above steps as a set of elements, and arrange them in a set order to obtain a characteristic parameter of the current vibration signal.
  • the comparing unit further comprises:
  • Comparing module for selecting a characteristic parameter of a legal person stored in advance, comparing its elements one by one with the corresponding element in the currently obtained characteristic parameter, determining whether the two are the same, and recording the same element;
  • a judging module determining whether the number of the same elements exceeds a set threshold, and if the comparison is stopped, determining that the current person is a person corresponding to the characteristic parameter of the legal person stored in advance; if not, performing the next step;
  • Selection module used to compare the characteristic parameters of another legal person who has not been compared beforehand; if all the characteristic parameters of the legal person stored in advance have been compared, it is determined that the current person is an illegal person.
  • a feedback unit configured to add a feature parameter that has been determined to be a legal person's action to a characteristic parameter of the legal person stored in advance, and corresponding to the legal person;
  • a position determining unit configured to determine a current person's position within the house by a current characteristic parameter received by a vibration sensor installed at a plurality of different locations within the house.
  • the invention has the beneficial effects that the privacy problem of the camera indoor monitoring is solved, the user's life is not affected, and the user is more convenient to use; compared to the video camera, the collected data is converted into digital text information by the ADC, and the required storage is required. The space is small and the storage period is long; compared to the video camera, the collected data is processed and analyzed, and the invention can automatically detect the accident and issue an alarm without manually spending a lot of time watching the video.
  • the vibration signal is transmitted by solids of the same material. Compared with electromagnetic signals such as WIFI, solids of the same material (such as the floor) will not be reflected and will not be blocked by the wall, and there is no non-line-of-sight and multipath effect, and the propagation range is wide. The signal stability is strong, so the more accurate original signal can be obtained, which is helpful for improving the accuracy of indoor positioning and behavior recognition in the subsequent data analysis.
  • FIG. 1 is a schematic structural view of a smart home monitoring system based on vibration detection according to the present invention.
  • a smart home monitoring method based on vibration detection includes the following steps: a data acquisition module collects structural vibration information generated by a person's behavioral activities, and intelligently monitors the behavior of the family by analyzing and processing the signal and alerts the abnormality.
  • the analysis process is specifically as follows:
  • S1 analyzes the ground footstep vibration signal, identifies the person, and alerts the illegal entry event
  • S2 analyzes and processes the ground footstep vibration signal, and uses TDOA three-point positioning to obtain the precise position information of the person;
  • S3 analyzes the vibration signal, pre-models or extracts feature classifications for different vibration types, and identifies human behavioral activities.
  • the human behavioral activities are monitored by recognizing the vibration signals generated by human activities, and the vibration signals are analyzed and processed to recognize human behavioral activities.
  • Structural vibration is a ground vibration signal generated by human walking.
  • the data acquisition module includes a geophone, a preamplifier, a filter, a post amplifier, an analog to digital conversion module, and a microcontroller, and the data acquisition module includes a geophone, a preamplifier, a filter, and a post amplifier.
  • the analog-to-digital conversion module and the microcontroller are connected in sequence.
  • the geophone acquires the direction angle of the footstep vibration signal, and uses different angles to determine different footstep vibration signals, thereby determining the number of people and achieving simultaneous positioning by multiple people.
  • the main signal collected by the present invention is a footstep sound vibration signal generated by a resident walking indoors.
  • geophones can also be installed elsewhere to obtain structural vibrations due to human behavior.
  • the time-frequency analysis of the number uses techniques such as fast Fourier transform and wavelet transform to extract features and use machine learning techniques such as SVM classifiers to achieve human identification.
  • the invention also provides a smart home monitoring system based on vibration detection, comprising a data acquisition module, a data analysis module and a danger alarm module, and the data acquisition module sends the collected structural vibration information to the data analysis module, and the data analysis module analyzes The result is sent to the hazard alert module.
  • the data acquisition module includes a geophone, a preamplifier, a filter, a post amplifier, an analog to digital conversion module, and a microcontroller
  • the data acquisition module includes a geophone, a preamplifier, a filter, and a post amplifier.
  • the analog-to-digital conversion module and the micro-controller are sequentially connected; the geophone acquires the direction angle of the footstep vibration signal, and uses different angles to determine different footstep vibration signals, thereby determining the number of people and achieving simultaneous positioning by multiple people.
  • the data analysis module performs indoor positioning, behavior detection, and identity authentication processing.
  • the position information of the occupant can be calculated.
  • vibrations By pre-modeling different types of vibrations, or by machine learning, different types of vibrations can be categorized to identify human behavior, such as an old man falling.
  • an alarm can be sent to the user's mobile phone to allow the family member who is out of the door to promptly help the person who falls at home.
  • Combining people's location information into classification technology can identify more behavior categories, such as sleeping, cooking, going to the toilet, and so on.
  • the function can be continuously upgraded to identify more and more human behaviors. .
  • the invention also provides a smart home monitoring system based on vibration detection, which is composed of three modules.
  • Data acquisition module can select the geophone geophone (but not limited to geophone).
  • Geophones are highly sensitive instruments that detect vibration and are more sensitive than vibration sensors.
  • the data acquisition module consists of a geophone, a preamplifier, a filter, a post amplifier, an analog to digital converter module, and a microcontroller.
  • the data processing module is mainly divided into three functions: indoor positioning, behavior detection and identification.
  • the current noise is first removed by notch filtering, and the signal-to-noise ratio is improved by Wiener filtering. It is processed by techniques such as fast Fourier transform and wavelet transform, and then the segmentation algorithm is used to cut the desired signal.
  • three-point positioning can be used to obtain a person's location information and know that the occupant is at a specific location.
  • Continuous detection can obtain the walking trajectory of the family.
  • the microcontroller sets a higher sampling frequency. Due to the signal propagation advantage of solid vibration, there is no multipath and non-line of sight problem, and a positioning accuracy of several centimeters can be achieved.
  • the method of classification can know the number of people at home.
  • the three-part geophone group collects signals to obtain the direction angle of the footstep vibration signal. Use no At the same angle, different footstep vibration signals can be determined, thereby determining the number of people and even multi-person positioning at the same time.
  • the proximity detector will obtain a higher energy value signal.
  • the algorithm By selecting the algorithm, only three high energy value detector data are selected, and the position information of the person can be calculated independently. Multi-person positioning can achieve the effect of single-person positioning.
  • some basic vibration types are trained, such as the vibration signal of the fall, and the training model is obtained.
  • the collected vibration signal matches the falling sample, it is considered to be a fall event.
  • a classification method is provided for reference, but there are many classification methods: using the Mel Cepstral Coefficient (MFCC) as the feature, and establishing a corresponding probability model for each individual in the feature set of each vibration signal by the Gaussian mixture model.
  • the abstraction of the individual features of each vibration signal in the feature space is abstracted as the result of the random generation of the probability model.
  • the estimation of the GMM parameters using the EM algorithm makes the log likelihood function have a maximum value, and the logarithm can be obtained through testing. The value, and the degree of similarity of its simulation can be measured by the range of log likelihood values.
  • Various vibration signals can be classified by the range of log likelihood values.
  • each person has a unique walking pattern, such as the physical characteristics of the individual, the position of the center of gravity when walking, and the way the foot touches the ground.
  • a unique walking pattern such as the physical characteristics of the individual, the position of the center of gravity when walking, and the way the foot touches the ground.
  • the danger alert module has an occupant accident alarm and a thief intrusion alert.
  • Alerts can be sent to users via SMS or mobile app.
  • the steps adopted include: collecting vibration information of the structure of the house caused by the action of a person in the house by the vibration sensor, obtaining a vibration signal; and processing the obtained vibration signal to obtain the a characteristic parameter representing an active person identity in the vibration signal; the feature parameter is a set including a plurality of elements; and comparing the obtained feature parameter with at least one previously stored legal person's characteristic parameter one by one to determine the currently obtained feature Whether the parameter matches one of the characteristic parameters of the legal person stored in advance, and if so, determining that the current active person is a legal person; otherwise, determining that an illegal person enters the house and issuing an alarm signal; wherein, the characteristics of the legal person stored in advance
  • the parameters are obtained by collecting the actions of legal persons in the house in advance, and have the same structure as the currently obtained characteristic parameters.
  • the step of obtaining the characteristic parameter further includes: acquiring, in the oscillating signal, a signal including a first set number of vibrations caused by actions of a person in the house as a feature parameter extraction letter; The characteristic parameter extracts the amplitude of each vibration in the signal; obtains the interval time of the adjacent two vibrations in the characteristic parameter extraction signal; and takes the amplitude and interval time obtained in the above step as a set of elements according to the set order Arranged together to obtain the characteristic parameters of the current vibration signal.
  • the method may further include the following steps: adding a feature parameter that has been determined to be a legal person's action to the feature parameter of the legal person stored in advance, and corresponding to the legal person. And through A current characteristic parameter received by a vibration sensor installed at a plurality of different locations within the house determines a location of a current person within the house.
  • system further includes:
  • a signal acquisition unit configured to collect vibration information of a house structure caused by an action of a person in the house by using a vibration sensor, to obtain a vibration signal;
  • a feature parameter forming unit configured to process the obtained vibration signal to obtain a characteristic parameter of the vibration signal indicating an active person identity; the feature parameter is a set including a plurality of elements;
  • the comparing unit is configured to compare the obtained characteristic parameter with the characteristic parameter of the at least one legal person stored in advance, and determine whether the currently obtained characteristic parameter matches one of the characteristic parameters of the legal person stored in advance, and if yes, determine the current activity.
  • the person is a legal person; otherwise, it is judged that an illegal person enters the house and sends an alarm signal;
  • the characteristic parameter of the legal person stored in advance is obtained by collecting the action of the legal person in the house in advance, and has the same structure as the currently obtained characteristic parameter.
  • the feature parameter forming unit further includes:
  • a signal acquisition module configured to obtain, in the oscillating signal, a signal including a first set number of vibrations caused by actions of a person in the house as a feature parameter extraction signal;
  • a parameter obtaining module configured to respectively obtain an amplitude of each vibration in the characteristic parameter extraction signal; and obtain an interval time of two adjacent vibrations in the characteristic parameter extraction signal;
  • the feature parameter forming module is configured to arrange the amplitude and the interval time obtained in the above steps as a set of elements, and arrange them in a set order to obtain a characteristic parameter of the current vibration signal.
  • the comparing unit further includes:
  • Comparing module for selecting a characteristic parameter of a legal person stored in advance, comparing its elements one by one with the corresponding element in the currently obtained characteristic parameter, determining whether the two are the same, and recording the same element;
  • a judging module determining whether the number of the same elements exceeds a set threshold, and if the comparison is stopped, determining that the current person is a person corresponding to the characteristic parameter of the legal person stored in advance; if not, performing the next step;
  • Selection module used to compare the characteristic parameters of another legal person who has not been compared beforehand; if all the characteristic parameters of the legal person stored in advance have been compared, it is determined that the current person is an illegal person.
  • the smart home monitoring system based on the vibration detection further includes a feedback unit and a location determining unit, wherein the feedback unit is configured to add the feature parameter obtained by the determined legal person action to the Pre-stored in the characteristic parameter of the legal person and corresponding to the legal person; the position determining unit is configured to determine the current person in the current characteristic parameter received by the vibration sensor installed at a plurality of different positions in the house The location within the house.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Manufacturing & Machinery (AREA)
  • Quality & Reliability (AREA)
  • Alarm Systems (AREA)
  • Emergency Alarm Devices (AREA)
  • Burglar Alarm Systems (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

一种基于振动检测的智能家居监控方法，包括如下步骤：数据采集模块采集人的行为活动产生的结构振动信息，通过分析处理该信号来智能监控家里人的行为活动并对异常做出警报。本发明解决了摄像头室内监控的隐私问题，不对用户的生活带来影响，让用户更方便地使用。

Description

基于振动检测的智能家居监控方法及系统 技术领域
本发明属于智能家居技术领域，尤其涉及智能家居的室内定位及监控方法。
背景技术
随着物联网的发展和智慧城市的建设，智能家居越来越受到人们的欢迎。人们希望不在家时同样可以监控到家里的情况，比如老人小孩是否安全、是否有盗贼入侵、保姆是否有对小孩的不当行为等。目前的监控方法普遍是摄像头技术。但是，在家里安装摄像头涉及到隐私甚至法律问题，更不可能在卧室和洗手间安装摄像头，而洗手间和卧室却是独自在家的老人小孩发生意外的高危场所。除了隐私问题，摄像头连续工作拍摄的视频需要极大的储存空间，这就导致视频存档的覆盖，无法获取保存周期前的视频源。另外，由于视频的自我识别分类及报警技术实现难度较大，需要当家里意外发生后，人工花费大量时间观看视频来找寻人们需要的信息。
鉴于以上原因，现在的物联网技术领域很多科研工作者投身于基于WIFI的人类行为检测，对WIFI的信号建模或通过机器学习的方法，可以识别人类在走路、跑步、摔倒、坐下、刷牙、做饭等行为，利用WIFI的能量递减特性或传播速度可以实现对人的室内定位，获取人的行走轨迹。WIFI的优势在于普及性，基本每家每户都有。但是，WIFI信号存在多径和非视距问题，当隔了墙时，信号强度会骤减，这大大影响了室内定位的精度和人类行为的识别准确度。另外，基于WIFI的技术要求固定不变的环境，如果移动WIFI设备或改变室内环境(如家具摆放位置)，这会导致训练模型不再有效，也就无法再进行行为识别。其次，利用WIFI对人进行行为识别仅能在一个人时有较高的准确性，当有其他人或者宠物在时，会影响WIFI信号，导致无法进行行为识别。所以，利用WIFI对室内进行监控的技 术并未商业化。
发明内容
为了克服上述所指的现有技术中的不足之处，本发明提供一种基于振动检测的智能家居监控方法，通过采集人的行为引起的结构振动信号并做分析，尤其是脚步声引起的地面振动信号，解决了上述的现有监控技术缺点，可以对室内人物进行定位、追踪和行为监控。
本发明是通过以下技术方案实现的:
一种基于振动检测的智能家居监控方法，包括如下步骤：数据采集模块采集人的行为活动产生的结构振动信息，通过分析处理该信号来智能监控家里人的行为活动并对异常做出警报。
作为本发明的进一步改进：分析处理具体为：
A)通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；
B)对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；
C)将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；
其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
作为本发明的进一步改进：所述步骤B)中进一步包括：
B1)在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信号；
B2)分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；
B3)将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列 在一起，得到当前振动信号的特征参量。
作为本发明的进一步改进：所述步骤C)中进一步包括如下步骤：
C1)选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；
C2)判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，执行下一步骤；
C3)返回步骤C1)取得另一个未进行比较的事先存储的合法人员的特征参量重复上述步骤；如所有事先存储的合法人员的特征参量均已比较，则判断当前人员为非法人员。
作为本发明的进一步改进：还包括如下步骤：
D)将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应。
作为本发明的进一步改进：还包括如下步骤：
E)通过安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
本发明同时提供了一种基于振动检测的智能家居监控系统，包括：
信号取得单元：用于通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；
特征参量形成单元：用于对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；
比较单元：用于将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；
其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
作为本发明的进一步改进：所述特征参量形成单元进一步包括：
信号取得模块：用于在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信号；
参数取得模块：用于分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；
特征参量形成模块：用于将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列在一起，得到当前振动信号的特征参量。
作为本发明的进一步改进：所述比较单元进一步包括：
比较模块：用于选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；
判断模块：用于判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，执行下一步骤；
选择模块：用于取得另一个未进行比较的事先存储的合法人员的特征参量进行比较；如所有事先存储的合法人员的特征参量均已比较，则判断当前人员为非法人员。
作为本发明的进一步改进：还包括：
反馈单元：用于将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应；
位置确定单元：用于通过安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
本发明的有益效果：解决了摄像头室内监控的隐私问题，不对用户的生活带来影响，让用户更方便地使用；相比视频摄像，采集的数据通过ADC转化成数字文本信息保存，所需储存空间小，保存周期长；相比视频摄像，采集的数据经过了处理分析，无需人工花费大量时间观看视频，该发明能自动检测意外发生并发出警报。振动信号是通过相同材质的固体来传播，相比WIFI等电磁波信号，相同材质的固体(例如地板)不会反射，也不会被墙壁阻挡，不存在非视距和多径效应，传播范围广、信号稳定性强，所以可以获取到更精准的原始信号，对后面数据分析时提高室内定位和行为识别的精度有很大帮助。
附图说明
图1为本发明中基于振动检测的智能家居监控系统的结构示意图。
具体实施方式
下面结合附图和实施例对本发明作进一步的描述。
一种基于振动检测的智能家居监控方法，包括如下步骤：数据采集模块采集人的行为活动产生的结构振动信息，通过分析处理该信号来智能监控家里人的行为活动并对异常做出警报。
分析处理具体为：
S1分析处理地面脚步声振动信号，对人进行身份鉴定，对非法入室事件做出警报；
S2分析处理地面脚步声振动信号，利用TDOA三点定位得出人的精确位置信息；
S3分析处理振动信号，对不同的振动类型预先建模或提取特征分类，识别人的行为活动。
通过识别人类活动产生的振动信号来监控人类的行为活动，分析处理振动信号识别人类行为活动。
结构振动为人走路产生的地面振动信号。
所述数据采集模块包括地震检波器、前级放大器、滤波器、后级放大器、模数转换模块和微控制器，所述数据采集模块包括地震检波器、前级放大器、滤波器、后级放大器、模数转换模块和微控制器依次连接。
地震检波器获取脚步声振动信号的方向角度，利用不同的角度，确定不同的脚步声振动信号，进而确定人数并做到多人同时定位
本发明主要采集的信号是居住人在室内行走产生的脚步声振动信号。为了升级对居住者的行为识别能力，也可以在其他地方安装地震检波器来获取由于人类行为产生的结构振动。
由于人与人的身高体重不同，体态和行走习惯速度等都不一样，通过对脚步声振动信 号的时频分析，采用快速傅里叶转换和小波变换等技术，提取特征，利用SVM分类器等机器学习技术，可以做到对人的身份识别功能。
本发明同时提供了一种基于振动检测的智能家居监控系统，包括数据采集模块、数据分析模块以及险情警报模块，数据采集模块将采集到的结构振动信息发送至数据分析模块，数据分析模块将分析结果发送至险情警报模块。
所述数据采集模块包括地震检波器、前级放大器、滤波器、后级放大器、模数转换模块和微控制器，所述数据采集模块包括地震检波器、前级放大器、滤波器、后级放大器、模数转换模块和微控制器依次连接；所述地震检波器获取脚步声振动信号的方向角度，利用不同的角度，确定不同的脚步声振动信号，进而确定人数并做到多人同时定位。
所述数据分析模块进行室内定位、行为检测和身份鉴定处理。
当用户离开家里需要开启防盗模式时，可以选择有人进去家里和在家里活动时报警。
同样可以设置更高级的防盗功能，设定合法的脚步声振动信号，当陌生人进去时才报警。
利用TDOA定位算法，在三个以上的地震检波器获取到脚步声振动信号以后，可以计算出居住者的位置信息。
通过多种方法，可以做到检测家里的人数和多人同时室内定位。
通过对不同的振动类型预先建模，或通过机器学习的办法，可以对不同的振动类型做分类，识别人的行为，比如老人摔倒。
当检测到有摔倒类型的振动信号时，可以通过给用户的手机发送警报，让出门在外的家人及时帮助到家里摔倒的人。
结合人的位置信息做分类技术，可以识别更多的行为类别，如睡觉、做饭、上厕所等。
当家里人睡觉的状态时间连续超过一个时间段，也可以设置成一个意外警报事件发送给远程用户。
由于人的大部分室内活动行为都会与家里的物体发生接触振动，如地面、灶台等，通过对更多的振动信号做训练或建模，可以不断升级功能，识别越来越多的人类行为。
本发明同时提供了一种基于振动检测的智能家居监控系统，由三个模块组成。数据采集模块、数据处理模块和险情警报模块。数据采集模块可以选择地震检波器geophone(但不限于geophone)。地震检波器是一种高灵敏度的检测振动的仪器，要比振动传感器更灵敏。
地震检波器经过几十年的技术发展，已经相对成熟且成本很低，可以大面积布置。
数据采集模块由地震检波器、前级放大器、滤波器、后级放大器、模数转换模块以及微控制器组成。
由于地震检波器直接采集到的未加工的信号太微弱，需要放大器提高信噪比。对于噪声，我们需要滤波器来去噪。利用模数转换器ADC将模拟信号转换成数字信号，最后传送到微控制器准备数据处理分析工作。
数据处理模块主要分三个功能：室内定位、行为检测和身份鉴定。
对原始信号，首先用陷波滤波去除电流噪声，用维纳滤波等技术提高信噪比。采用快速傅里叶变化和小波变换等技术处理，再用分段算法切割所需信号。
室内定位：
通过TDOA算法，利用三点定位可以获取一个人的位置信息，知道居住者在具体的位置。连续检测可以获取家里人的行走轨迹。
微控制器设置较高的采样频率，由于固体振动的信号传播优势，没有多径和非视距问题，可以达到几厘米的定位精度。
当家里有不止一个人时，如果不同人的脚步声振动信号存在交叉重叠的部分，由于多人的脚步声振动和单人的脚步声振动信号持续时间等特征不一样，通过训练，利用机器学习的方法分类，可以知道家里的人数。
采用三部件的地震检波器组采集信号，可以获取脚步声振动信号的方向角度。利用不 同的角度，可以确定不同的脚步声振动信号，从而确定人数甚至做到多人同时定位。
除了上述两种方法，还有多种信号分离的方法，可以达到多人同时定位的效果。
另外，当家里人较为分散活动时，距离近的检波器将获得更高能量值的信号，通过选择算法，仅分别选择3个高能量值的检波器数据，可以独立地计算人的位置信息，多人定位可以获得单人定位的效果。
行为检测：
首先对一些基本的振动类型做训练，比如摔倒的振动信号，获取训练模型，当采集的振动信号和摔倒样本匹配时，则认为是摔倒事件。
这里提供一种分类方法供参考，但还有很多分类方法：利用梅尔倒谱系数(MFCC)作特征,通过高斯混合模型对每个振动信号的特征集合中的每个个体建立对应的概率模型,把每个振动信号的个体特征在特征空间的分布抽象为该概率模型随机产生的结果,对GMM参数利用EM算法的估计使得对数似然函数有最大值,通过测试可以得出对数似然值,而其模拟的相似程度则可以用对数似然值的范围来衡量。通过对数似然值的范围可以分类各种振动信号。
因为家里的行为很多会限制在特定地点，比如在厨房在能做饭，在洗手间上厕所。所以，结合位置信息和行走轨迹，可以识别更多的行为活动。
为了升级行为检测功能，可以不止在地面安装地震检波器，还可以在更多人类可能接触到的地方安装，可以通过训练振动信号识别更多人类行为活动。
身份鉴定：
由于许多因素，每个人都有一个独特的行走模式，比如个人的身体特征,走路时的重心位置,脚接触地的方式等。在时域上、频域上、或者时频结合，都可以找出每个人产生的特有的脚步声振动特征。利用这些特征，可以做到身份识别。
因为每个人脚步声的持续时间与脚步声的间隔时间都不一样，利用这个点作特征，再 利用KNN分类器来做身份鉴定也是一种可行方案。
险情警报模块有居住者意外发生警报和盗贼入侵警报。
当居住者发生一些意外危险行为时，例如摔倒，长时间睡觉等行为，发出警报。
当有非法身份在设定时间里出现在家里时，发出警报。
警报可以通过短信或者手机app等方式向用户推送警报消息。
在本实施例，从总体上来看，采用的步骤包括：通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；以及将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
而得到上述特征参量的步骤又包括：在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信；分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列在一起，得到当前振动信号的特征参量。
在进行特征参量对比时，通过选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；然后，再判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，则取得另一个未进行比较的事先存储的合法人员的特征参量重复上述步骤；如所有事先存储的合法人员的特征参量均已比较，但认为找到相对应的特征参数，则判断当前人员为非法人员。
在本实施例，除了上述步骤外，还可以包括如下步骤：将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应。以及通过 安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
此外，在本实施例，除了上述必要的基础硬件外，该系统还包括：
信号取得单元：用于通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；
特征参量形成单元：用于对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；
比较单元：用于将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；
其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
所述特征参量形成单元进一步包括：
信号取得模块：用于在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信号；
参数取得模块：用于分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；
特征参量形成模块：用于将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列在一起，得到当前振动信号的特征参量。
所述比较单元进一步包括：
比较模块：用于选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；
判断模块：用于判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，执行下一步骤；
选择模块：用于取得另一个未进行比较的事先存储的合法人员的特征参量进行比较；如所有事先存储的合法人员的特征参量均已比较，则判断当前人员为非法人员。
同时，在本实施例中，上述基于振动检测的智能家居监控系统，还包括反馈单元和位置确定单元；其中，所述反馈单元：用于将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应；所述位置确定单元：用于通过安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
以上内容是结合具体实现方式对本发明做的进一步阐述，不应认定本发明的具体实现只局限于以上说明。对于本技术领域的技术人员而言，在不脱离本发明构思的前提下，还可以做出若干简单推演或替换，均应视为有本发明所提交的权利要求确定的保护范围之内。

Claims (10)

  1. 一种基于振动检测的智能家居监控方法，其特征在于：包括如下步骤：采集人的行为活动产生的结构振动信息，通过分析处理该信号来智能监控家里人的行为活动并对异常做出警报。
  2. 根据权利要求1所述的基于振动检测的智能家居监控方法，其特征在于，分析处理具体为：
    A)通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；
    B)对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；
    C)将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；
    其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
  3. 根据权利要求2所述的基于振动检测的智能家居监控方法，其特征在于，所述步骤B)中进一步包括：
    B1)在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信号；
    B2)分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；
    B3)将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列在一起，得到当前振动信号的特征参量。
  4. 根据权利要求3所述的基于振动检测的智能家居监控方法，其特征在于:所述步骤C)中进一步包括如下步骤：
    C1)选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特 征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；
    C2)判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，执行下一步骤；
    C3)返回步骤C1)取得另一个未进行比较的事先存储的合法人员的特征参量重复上述步骤；如所有事先存储的合法人员的特征参量均已比较，则判断当前人员为非法人员。
  5. 根据权利要求1所述的基于振动检测的智能家居监控方法，其特征在于:还包括如下步骤：
    D)将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应。
  6. 根据权利要求5所述的基于振动检测的智能家居监控方法，其特征在于:还包括如下步骤：
    E)通过安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
  7. 一种基于振动检测的智能家居监控系统，其特征在于：包括：
    信号取得单元：用于通过振动传感器采集房屋内的人的行动导致的房屋结构的振动信息，得到振动信号；
    特征参量形成单元：用于对得到的所述振动信号进行处理，得到所述振动信号中表示活动的人身份的特征参量；所述特征参量是一个包括多个元素的集合；
    比较单元：用于将得到的特征参量与至少一个事先存储的合法人员的特征参量逐个进行比较，判断当前得到特征参量是否与事先存储的合法人员的特征参量中的一个匹配，如是，判断当前活动人员是合法人员；否则，判断有非法人员进入房屋，发出告警信号；
    其中，所述事先存储的合法人员的特征参量是事先对合法人员在房屋中的行动进行采集而得到的，其与当前得到的特征参量具有相同的结构。
  8. 根据权利要求7所述的基于振动检测的智能家居监控系统，其特征在于：所述特征参量形成单元进一步包括：
    信号取得模块：用于在所述振荡信号中取得一段包括第一设定数量的、由所述房屋内的人的行动导致的振动的信号，作为特征参量提取信号；
    参数取得模块：用于分别取得所述特征参量提取信号中每次振动的振幅；取得所述特征参量提取信号中相邻两次振动的间隔时间；
    特征参量形成模块：用于将上述步骤中得到的振幅和间隔时间作为一个集合的元素，按照设定的顺序排列在一起，得到当前振动信号的特征参量。
  9. 根据权利要求8所述的基于振动检测的智能家居监控系统，其特征在于：所述比较单元进一步包括：
    比较模块：用于选择一个事先存储的合法人员的特征参量，将其元素逐个与所述当前得到的特征参量中的相应的元素对比，判断二者是否相同，并记录相同的元素；
    判断模块：用于判断相同的元素的数量是否超过设定阈值，如是停止比较，判断当前人员为所述事先存储的合法人员的特征参量所对应的人员；如否，执行下一步骤；
    选择模块：用于取得另一个未进行比较的事先存储的合法人员的特征参量进行比较；如所有事先存储的合法人员的特征参量均已比较，则判断当前人员为非法人员。
  10. 根据权利要求/9所述的基于振动检测的智能家居监控系统，其特征在于：还包括：
    反馈单元：用于将已判断为合法人员行动得到的特征参量加入所述事先存储的合法人员的特征参量中，并与该合法人员对应；
    位置确定单元：用于通过安装在所述房屋内多个不同位置的振动传感器接收到的当前特征参量，确定当前人员在所述房屋内的位置。
PCT/CN2017/091671 2016-08-26 2017-07-04 基于振动检测的智能家居监控方法及系统 WO2018036281A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610743598.2A CN106251572A (zh) 2016-08-26 2016-08-26 基于振动检测的智能家居监控方法及系统
CN201610743598.2 2016-08-26

Publications (1)

Publication Number Publication Date
WO2018036281A1 true WO2018036281A1 (zh) 2018-03-01

Family

ID=57596590

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/091671 WO2018036281A1 (zh) 2016-08-26 2017-07-04 基于振动检测的智能家居监控方法及系统

Country Status (2)

Country Link
CN (2) CN106251572A (zh)
WO (1) WO2018036281A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3094544A1 (fr) * 2019-03-29 2020-10-02 Orange Procédé de localisation d’un individu ou d’un animal dans un bâtiment, procédé de contrôle d’au moins un objet communicant dans un bâtiment, dispositif de localisation, système, équipement d’accès et programmes d’ordinateur correspondants
US11194303B2 (en) 2018-12-03 2021-12-07 International Business Machines Corporation Method and system for anomaly detection and notification through profiled context
CN113973263A (zh) * 2021-10-25 2022-01-25 合肥工业大学 一种基于wifi定位的智慧校园系统
EP3920794A4 (en) * 2019-02-06 2023-01-25 University of Georgia Research Foundation, Inc. SYSTEM FOR INTERIOR OCCUPANCY ESTIMATION, PATH TRACKING, AND EVENT MONITORING AND TRACKING

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106251572A (zh) * 2016-08-26 2016-12-21 深圳大学 基于振动检测的智能家居监控方法及系统
CN107290721B (zh) * 2017-06-01 2019-01-11 深圳大学 一种室内的定位方法及系统
CN107742385B (zh) * 2017-09-13 2020-05-01 武汉安保通科技有限公司 一种微振动报警系统和方法
CN107564215B (zh) * 2017-09-13 2019-08-20 武汉安保通科技有限公司 一种周界安防系统和方法
CN108899083A (zh) * 2018-07-27 2018-11-27 滁州学院 一种具有分析功能的健康Toilet系统
CN110908289A (zh) * 2018-09-17 2020-03-24 珠海格力电器股份有限公司 智能家居的控制方法及装置
CN110703699A (zh) * 2018-12-07 2020-01-17 上海产业技术研究院 基于nb-iot通信技术的行为监测系统、监测器和存储介质
CN111311860B (zh) * 2018-12-12 2022-05-03 杭州海康威视数字技术股份有限公司 一种区域入侵检测方法及装置
WO2020147098A1 (zh) * 2019-01-18 2020-07-23 深圳大学 基于地面振动信号的人体摔倒检测系统
CN109830076A (zh) * 2019-03-26 2019-05-31 中山绿威科技有限公司 室内防盗的检测方法、装置及计算机可读存储介质
CN112435423B (zh) * 2019-08-22 2022-06-03 杭州海康威视数字技术股份有限公司 一种监控方法及装置
CN111508203B (zh) * 2020-04-15 2022-03-01 广东小天才科技有限公司 一种地震预警提示方法及装置、可穿戴设备
CN113781728A (zh) * 2021-08-02 2021-12-10 盐城市湛安智感科技有限公司 一种基于群体智能优化的震动感知系统及方法
CN114170754A (zh) * 2021-12-09 2022-03-11 中科计算技术西部研究院 基于大数据的林业维护管理系统
CN114915725A (zh) * 2022-04-25 2022-08-16 歌尔科技有限公司 摄像头控制方法、装置、系统及计算机可读存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101261759A (zh) * 2007-02-13 2008-09-10 霍尼韦尔国际公司 检测人类入侵者的传感器和安全系统
CN101393660A (zh) * 2008-10-15 2009-03-25 中山大学 一种基于脚步声识别的智能门禁系统
CN103558851A (zh) * 2013-10-10 2014-02-05 杨松 精确感知室内活动的方法及装置
WO2016130571A1 (en) * 2015-02-09 2016-08-18 Carnegie Mellon University Indoor identification of individuals through footstep induced structural vibration
CN205581626U (zh) * 2016-02-17 2016-09-14 南京物联传感技术有限公司 可控制智能家居设备的智能地板
CN106251572A (zh) * 2016-08-26 2016-12-21 深圳大学 基于振动检测的智能家居监控方法及系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9966085B2 (en) * 2006-12-30 2018-05-08 Google Technology Holdings LLC Method and noise suppression circuit incorporating a plurality of noise suppression techniques
CN101046913B (zh) * 2007-05-08 2010-09-29 陈华鹏 多媒体无线安防装置
CN102324232A (zh) * 2011-09-12 2012-01-18 辽宁工业大学 基于高斯混合模型的声纹识别方法及系统
CN105629936B (zh) * 2016-02-17 2018-12-11 南京物联传感技术有限公司 一种可控制智能家居设备的智能地板
CN107240397A (zh) * 2017-08-14 2017-10-10 广东工业大学 一种基于声纹识别的智能锁及其语音识别方法和系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101261759A (zh) * 2007-02-13 2008-09-10 霍尼韦尔国际公司 检测人类入侵者的传感器和安全系统
CN101393660A (zh) * 2008-10-15 2009-03-25 中山大学 一种基于脚步声识别的智能门禁系统
CN103558851A (zh) * 2013-10-10 2014-02-05 杨松 精确感知室内活动的方法及装置
WO2016130571A1 (en) * 2015-02-09 2016-08-18 Carnegie Mellon University Indoor identification of individuals through footstep induced structural vibration
CN205581626U (zh) * 2016-02-17 2016-09-14 南京物联传感技术有限公司 可控制智能家居设备的智能地板
CN106251572A (zh) * 2016-08-26 2016-12-21 深圳大学 基于振动检测的智能家居监控方法及系统

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11194303B2 (en) 2018-12-03 2021-12-07 International Business Machines Corporation Method and system for anomaly detection and notification through profiled context
EP3920794A4 (en) * 2019-02-06 2023-01-25 University of Georgia Research Foundation, Inc. SYSTEM FOR INTERIOR OCCUPANCY ESTIMATION, PATH TRACKING, AND EVENT MONITORING AND TRACKING
FR3094544A1 (fr) * 2019-03-29 2020-10-02 Orange Procédé de localisation d’un individu ou d’un animal dans un bâtiment, procédé de contrôle d’au moins un objet communicant dans un bâtiment, dispositif de localisation, système, équipement d’accès et programmes d’ordinateur correspondants
WO2020201006A1 (fr) * 2019-03-29 2020-10-08 Orange Procede de localisation d'un individu ou d'un animal dans un batiment, procede de controle d'au moins un objet communicant dans un batiment, dispositif de localisation, systeme, equipement d'acces et programmes d'ordinateur correspondants domaine de l'invention
CN113973263A (zh) * 2021-10-25 2022-01-25 合肥工业大学 一种基于wifi定位的智慧校园系统

Also Published As

Publication number Publication date
CN107422643B (zh) 2021-04-23
CN107422643A (zh) 2017-12-01
CN106251572A (zh) 2016-12-21

Similar Documents

Publication Publication Date Title
WO2018036281A1 (zh) 基于振动检测的智能家居监控方法及系统
US10614299B2 (en) Indoor identification of individuals through footstep induced structural vibration
Clemente et al. Smart seismic sensing for indoor fall detection, location, and notification
US10785643B2 (en) Deep learning neural network based security system and control method therefor
Su et al. Doppler radar fall activity detection using the wavelet transform
Pan et al. Indoor person identification through footstep induced structural vibration
US7825813B2 (en) Identifying activity in an area utilizing sound detection and comparison
US8258932B2 (en) Occupant detection system for vehicle
Clemente et al. Indoor person identification and fall detection through non-intrusive floor seismic sensing
Huang et al. G-fall: device-free and training-free fall detection with geophones
Toreyin et al. Falling person detection using multi-sensor signal processing
WO2016155789A1 (en) Fall detection system and method
JP2011237865A (ja) 生活空間の見守りシステム
US8271204B2 (en) Human recognition apparatus and human recognition method
Han et al. Smart home occupant identification via sensor fusion across on-object devices
CN105139869B (zh) 一种基于区间差分特征的婴儿啼哭检测方法
Zieger et al. Acoustic based surveillance system for intrusion detection
CN109345771A (zh) 一种检测人体跌倒的方法及系统
JP6297611B2 (ja) 転倒検知装置及び転倒判定方法
Ciabattoni et al. Complex activity recognition system based on cascade classifiers and wearable device data
Valero et al. Reprint of: Vibration sensing-based human and infrastructure safety/health monitoring: A survey
Ntalampiras Audio surveillance
JP6766346B2 (ja) 情報処理装置、活動状態検出プログラムおよび活動状態検出方法
Diraco et al. Radar sensing technology for fall detection under near real-life conditions
Pipanmaekaporn et al. Mining Acceleration Data for Smartphone-based Fall Detection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17842704

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/06/2019)

122 Ep: pct application non-entry in european phase

Ref document number: 17842704

Country of ref document: EP

Kind code of ref document: A1