WO2015043546A1 - 人体摔倒检测的方法、装置及移动终端系统 - Google Patents

人体摔倒检测的方法、装置及移动终端系统 Download PDF

Info

Publication number
WO2015043546A1
WO2015043546A1 PCT/CN2014/087996 CN2014087996W WO2015043546A1 WO 2015043546 A1 WO2015043546 A1 WO 2015043546A1 CN 2014087996 W CN2014087996 W CN 2014087996W WO 2015043546 A1 WO2015043546 A1 WO 2015043546A1
Authority
WO
WIPO (PCT)
Prior art keywords
acceleration
human body
threshold
fall
time
Prior art date
Application number
PCT/CN2014/087996
Other languages
English (en)
French (fr)
Inventor
胡琨
Original Assignee
胡琨
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 胡琨 filed Critical 胡琨
Publication of WO2015043546A1 publication Critical patent/WO2015043546A1/zh
Priority to US15/084,508 priority Critical patent/US9928717B2/en

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/02—Alarms for ensuring the safety of persons
    • G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
    • G08B21/0407—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis
    • G08B21/043—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis detecting an emergency event, e.g. a fall
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1116—Determining posture transitions
    • A61B5/1117—Fall detection

Definitions

  • the invention relates to a human body fall detection technology, in particular to a method, a device and a mobile terminal system for human body fall detection.
  • the main object of the present invention is to provide a method for detecting a fall of a human body, which can improve the accuracy of the fall judgment.
  • the invention provides a method for human body fall detection, comprising:
  • a 0 is the baseline two-dimensional coordinates, the coordinates calculating I 0 with the reference line curve surrounded reference line, the area under the curve;
  • the method before the step of acquiring the resultant acceleration sequence I 0 in the fixed time when the resultant acceleration of the human body is less than the low acceleration threshold a 0 , the method further includes:
  • the speed information includes a resultant acceleration, speed, and time;
  • the analyzing the characteristics of the speed information, calculating and extracting the low acceleration threshold a 0 , the combined acceleration sequence I 0 , the curve area threshold ⁇ S, and the low acceleration time threshold ⁇ T in the fixed time further includes:
  • Human body status information includes height, weight, and/or exercise status.
  • the method further includes:
  • the method before the step of acquiring the resultant acceleration sequence I 0 in the fixed time when the resultant acceleration of the human body is less than the low acceleration threshold a 0 , the method further includes:
  • the method further includes:
  • the current speed sequence of the human body is collected; when the current speed is lower than the speed threshold exceeds the set time, an alarm prompt message is generated; when the alarm prompt information is confirmed or not confirmed for more than a certain time, the alarm may be performed.
  • the invention also provides a device for detecting human body fall, comprising:
  • Acquisition determination means for detecting a human body in a force acceleration is less than a low acceleration threshold a 0, collected together in a fixed time sequence of acceleration I 0;
  • the area calculation module is configured to calculate a curve area in the I 0 coordinate and a reference line surrounded by the reference line in a two-dimensional coordinate with the resultant acceleration and time as the coordinate axis and a 0 as the reference line;
  • the fall determination module is configured to use the difference between the curve area on the reference line and the curve area under the reference line to be smaller than the curve area threshold ⁇ S, and when the time of the resultant acceleration under the reference line is greater than the low acceleration time threshold ⁇ T, it is determined that the human body falls.
  • the device further comprises:
  • a sample collection module configured to collect speed information during a body fall in one or more individual body fall samples; the speed information includes a resultant acceleration, speed, and time;
  • An analysis extraction module is configured to analyze characteristics of the velocity information, calculate and extract a low acceleration threshold a 0 , a resultant acceleration sequence I 0 in a fixed time, a high acceleration threshold a 1 , a curve area threshold ⁇ S, and a low acceleration time threshold ⁇ T, establish a fall detection mechanism.
  • the analysis extraction module is further configured to:
  • Human body status information includes height, weight, and/or exercise status.
  • the device further comprises:
  • a high acceleration determination module configured to determine whether a resultant acceleration greater than a high acceleration threshold a 1 is collected in the sequence I 0 when the acquisition of the resultant acceleration sequence I 0 is performed; if yes, calculate the curve area by the area calculation module .
  • the device further comprises:
  • the receiving module is configured to receive the setting of the human body state information, and adjust a 0 , ⁇ S and/or ⁇ T according to the set human body state information.
  • the device further comprises:
  • the alarm prompting module is configured to collect the current speed sequence of the human body after determining that the fall occurs; when the current speed is lower than the speed threshold exceeds the set time, generate an alarm prompt message; when the alarm prompt information is confirmed or not confirmed for more than a certain time, The alarm can be performed.
  • the invention also provides a mobile terminal system, comprising:
  • An information acquisition module configured to acquire speed information by using a three-axis acceleration sensor
  • Acquisition determination means for detecting a human body in a force acceleration is less than a low acceleration threshold a 0, collected together in a fixed time sequence of acceleration I 0;
  • the area calculation module is configured to calculate a curve area in the I 0 coordinate and a reference line surrounded by the reference line in a two-dimensional coordinate with the resultant acceleration and time as the coordinate axis and a 0 as the reference line;
  • the fall determination module is configured to use the difference between the curve area on the reference line and the curve area under the reference line to be smaller than the curve area threshold ⁇ S, and the time of the resultant acceleration under the reference line is greater than the low acceleration time threshold ⁇ T, and it is determined that the human body falls;
  • the alarm prompting module is configured to collect the current speed sequence of the human body after determining that the fall occurs; when the current speed is lower than the speed threshold exceeds the set time, generate an alarm prompt message; when the alarm prompt information is confirmed or not confirmed for more than a certain time, The alarm can be made through the mobile communication component.
  • the mobile terminal system further includes:
  • a high acceleration determination module configured to determine whether a resultant acceleration greater than a high acceleration threshold a 1 is collected in the sequence I 0 when the acquisition of the resultant acceleration sequence I 0 is performed; if yes, calculate the curve area by the area calculation module .
  • the invention can be based on a device with acceleration detection and communication function, such as a smart phone, which is based on the difference between the physical safety state of the human body and the kinematics and dynamics characteristics of the fall, as long as the user carries the smart phone with the applied body fall detection.
  • the application will automatically collect and analyze the human body dynamics information to determine whether the human body has fallen, and use the mobile phone communication advantages to make alarm notifications such as SMS and telephone.
  • a device with acceleration detection and communication function such as a smart phone
  • the application will automatically collect and analyze the human body dynamics information to determine whether the human body has fallen, and use the mobile phone communication advantages to make alarm notifications such as SMS and telephone.
  • the key is to fully consider the characteristics of the human body's sports behavior, improve the detection rate, and reduce False positive rate.
  • FIG. 1 is a schematic flow chart showing the steps of an embodiment of a method for detecting human fall in the present invention
  • FIG. 2 is a schematic flow chart of another step of an embodiment of a method for detecting body fall detection according to the present invention
  • FIG. 3 is a schematic flow chart showing the steps of another embodiment of the method for detecting human fall in the present invention.
  • FIG. 4 is a schematic flow chart showing the steps of still another embodiment of the method for detecting human fall in the present invention.
  • FIG. 5 is a schematic flow chart showing the steps of still another embodiment of the method for detecting human fall fall in the present invention.
  • FIG. 6 is a schematic diagram of two-dimensional coordinates in which the time is the abscissa and the combined acceleration is the ordinate in an embodiment of the method for detecting human fall in the present invention
  • FIG. 7 is a schematic structural view of an embodiment of a device for detecting human fall in the present invention.
  • FIG. 8 is another schematic structural view of an embodiment of a device for detecting human fall in the present invention.
  • FIG. 9 is a schematic structural view of another embodiment of the device for detecting human fall in the present invention.
  • FIG. 10 is a schematic structural view of still another embodiment of the device for detecting human fall in the present invention.
  • FIG. 11 is a schematic structural view of still another embodiment of the device for detecting human fall in the present invention.
  • FIG. 12 is a schematic structural diagram of an embodiment of a mobile terminal system according to the present invention.
  • the method for detecting body fall detection may include:
  • Step S11 determining whether the resultant acceleration of the human body is less than the low acceleration threshold a 0 ; if yes, proceeding to step S12; otherwise, continuing to perform the detection;
  • Step S12 collecting the resultant acceleration sequence I 0 in a fixed time
  • Step S13 calculating the curve area in the I 0 coordinate and the reference line surrounded by the reference line in the two-dimensional coordinates with the resultant acceleration and time as the coordinate axes and a 0 as the reference line;
  • Step S14 determining whether the difference between the curve area on the reference line and the curve area under the reference line is smaller than the curve area threshold ⁇ S, and the time occupied by the resultant force under the reference line is greater than the low acceleration time threshold ⁇ T; if yes, Go to step S15; otherwise, continue to perform the detection;
  • step S15 it is determined that the human body falls.
  • the method for detecting the fall of the human body proposed by the invention can automatically detect the acceleration information and the speed information of the human body through the sensor, and integrate the acceleration, the speed, the movement time and the motion state of the human body. Correctly judge whether the human body has fallen, and remind whether it is necessary to call for help, and at the same time, the operation of confirming and canceling the rescue can be performed for the reminder.
  • the device for realizing the above method for detecting body fall detection needs at least an acceleration sensor (such as a three-axis acceleration sensor), and a communication module (such as a GSM module in a mobile communication module, CDMA) is required when a rescue function is required. Module, etc.).
  • an acceleration sensor such as a three-axis acceleration sensor
  • a communication module such as a GSM module in a mobile communication module, CDMA
  • the above device can be a mobile terminal, such as a commonly used mobile terminal device such as a smart phone or a tablet computer. Since the commonly used mobile terminal devices are usually provided with a communication module and an acceleration sensor, the direct use can be directly utilized without additional configuration.
  • the method may further include:
  • Step S100 Collect speed information during a fall of a human body in one or more personal fall samples; the speed information includes a resultant acceleration, speed, and time;
  • Step S101 Analyze the characteristics of the speed information, calculate and extract a low acceleration threshold a 0 , a resultant acceleration sequence I 0 , a high acceleration threshold a 1 , a curve area threshold ⁇ S, and a low acceleration time threshold ⁇ T. Fall detection mechanism.
  • the speed information during the fall of the human body is collected; the speed information includes the resultant acceleration, speed and time; and the acceleration time sequence before the crash is also recorded, and the extraction is performed.
  • Some acceleration features establish a fall detection model that can be continuously trained and optimized by falling data.
  • the establishment of the fall detection model is the most important part to achieve accurate fall detection.
  • the acceleration data related to the fall process is obtained for analysis, filtering, fusion, and extraction of human fall kinematics and power. Learning characteristics, such as: long low acceleration before falling, may appear squat before falling, etc., thus establishing a fall detection model, the model can be continuously trained and self-adjusted.
  • the movement state of the human body and the fall can be detected according to the change of the three-axis acceleration.
  • the extractable features are as follows: falling low acceleration state, speed, high speed state before collision, high speed state of collision And the time of maintenance of various states, etc., and then based on the above characteristics to establish a model for dynamically detecting the fall of the human body to match the triaxial acceleration value of the human body motion, and determine whether to fall according to the output probability, and some parameters in the model can be based on the height of the human body. , weight, amount of exercise and real-time movement adjustment of the human body.
  • the acceleration sequence I 0 , the high acceleration threshold a 1 , the curve area threshold ⁇ S, and the low acceleration time threshold ⁇ T are associated with the human body state information corresponding to the sample; the human body state information may include height, weight, and/or motion state.
  • the height and weight can be an interval value.
  • the method may further include:
  • Step S102 Receive setting of human body state information, and adjust a 0 , ⁇ S and/or ⁇ T according to the set human body state information.
  • model parameters can be different according to the height and weight of the human body, and the real-time motion state of the human body, each user can set according to the different human body state information, and the detecting device will be based on the set human body state information. And matching the corresponding parameters such as a 0 , ⁇ S and / or ⁇ T.
  • the method may further include: incorporating the human body falling sample detected by itself into the fall detection mechanism, and associating the speed information of the sample with the human body state information.
  • the fall event may be added to the human fall detection model according to the user's confirmation. For example, collecting speed information during the fall and analyzing, extracting a low acceleration threshold a 0 , a combined acceleration sequence I 0 , a high acceleration threshold a 1 , a curve area threshold ⁇ S, and a low acceleration time threshold ⁇ T And associating with the human body state information corresponding to the sample, the revision has established a fall detection mechanism, thereby realizing the self-learning mechanism of the above-mentioned human fall detection model.
  • the method may further include:
  • Step S110 the acceleration force when the acquisition sequence I 0 is performed, determining whether there is the sequence I is greater than 0 collected to force a high acceleration threshold acceleration 1; if so, proceeds step S12, otherwise continue detection.
  • the fall determination may proceed to the next step, or may be determined not occurred Fall down and return to the initial combined force detection and judgment.
  • the method may further include:
  • Step S16 After determining that the fall occurs, collecting the current speed sequence of the human body; when the current speed is lower than the speed threshold exceeds the set time, generating an alarm prompt message; when the alarm prompt information is confirmed or not confirmed for more than a certain time, the operation may be performed. Alarm.
  • the initial initial velocity of the human body is zero, and the gravity acceleration is considered according to the change of the three-axis acceleration, and the speed of each moment of the human body can be calculated.
  • the triaxial acceleration will continue to change, and the speed will continue to change.
  • an acceleration time series of a fixed time length can be obtained, which can completely record the acceleration value of a fall process (including a period before and after the fall). Since the human body has a low acceleration state before falling, a low acceleration threshold can also be obtained from the experimental data. When the acceleration of the human body is lower than the low acceleration threshold during the movement, the acceleration data can be collected for the fall.
  • the detection model is inverted to further determine whether to continue collecting data or clearing the data and recording the time (ie, entering the above time series). Since the human body may have a high acceleration state before the collision due to defects, etc., the above-mentioned low acceleration may be judged by further high acceleration; a high acceleration threshold is obtained through experimental data, during the movement. When the acceleration of the human body is higher than the high acceleration threshold, the detection model can be used to match the acceleration data that has been collected.
  • the operation is performed and matched with the parameters in the model; if the matching is successful, The description has fallen, and the alarm is judged according to the speed information of the human body. If the state of the human body is below the speed threshold exceeds the set time, an alarm prompt message is generated, and the user selects whether to alarm according to the actual situation, and does not perform any operation for a certain period of time. , the alarm is executed, and the short message notification and/or the telephone notification are performed according to the preset contact telephone number.
  • the set body state information such as body height, weight and/or motion state, and the speed and acceleration information of the pre-fall and fall that have been collected
  • another high acceleration threshold can be obtained.
  • the combined acceleration of the human body is greater than the threshold, it is very likely that the user is subjected to an instantaneous strong impact, such as a car accident. If the real-time monitoring acceleration of the human body is greater than the threshold, an alarm prompt message is directly generated to perform an alarm operation when the user confirms or prompts a timeout.
  • a two-dimensional coordinate diagram with time (10ms) as the abscissa and combined (force) acceleration (m/s2) as the ordinate is used, wherein the resultant acceleration a0 is taken as the reference line.
  • the acceleration of the human body force detected during the fall changes with time: when the time is from 100 to about 190, the resultant acceleration is always around 10, indicating that the user may be in a standing state at this time; From about 190 to about 430, the resultant acceleration changes uniformly and regularly at 10, and the resultant acceleration is above a0, indicating that the user may be in a normal walking state; at about 450 490, the first force occurs first.
  • the acceleration is lower than a0, and lasts between 450 and 480, and then there is a relatively short high acceleration that is several times higher than the normal (10), which lasts between 480 and 490.
  • the above low acceleration indicates the use.
  • the person may be in the process of falling before the fall.
  • the high acceleration indicates that the user may be in the process of falling after the fall; it is obvious in the figure that the resultant acceleration is formed in the coordinates during the fall.
  • the area of the curve under the reference line enclosed by the curve and the reference line is larger than the area of the curve on the reference line.
  • the calculation and matching method in the fall detection model is as follows: firstly, the low acceleration threshold a 0 is determined by experiments and research. When the resultant acceleration value generated during the movement of the human body is lower than a 0 , the trigger starts to collect the triaxial acceleration sensor data to one.
  • the fixed acceleration time series I 0 simultaneous calculation is performed: calculating the area of the acceleration curve based on the reference line a 0 from the time of collecting the data, and the area enclosed by the acceleration and the reference line a 0 is positive on the reference line. On the contrary, it is negative.
  • the total curve area is the area of the curve on the reference line plus the area of the curve under the baseline (that is, the difference between the area of the curve on the reference line and the area of the curve below the baseline when both are positive)
  • ⁇ S the sum of the acceleration times lower than the acceleration a 0 in I 0 is larger than ⁇ T
  • ⁇ T, ⁇ S, a 0 can be dynamically adjusted according to the user's weight, height, and exercise state (the amount of exercise).
  • the detection can also be performed according to the acceleration and the speed.
  • the detected acceleration is below the low acceleration threshold, there is first a sequence of accelerations and speeds for a fixed time interval, the length of the sequence being sufficient to include the normal movement of the user before the complete fall has fallen.
  • the default user initial speed is zero, and the approximate speed of the human body at a certain moment can be calculated based on the three-axis acceleration and time.
  • the acceleration and velocity sequence of the fixed time length can obtain the current exercise intensity and state of the user, and accordingly, some related parameters of the fall verification can be adjusted, so that the model can more accurately detect the state before the fall.
  • the above method for detecting body fall can be based on a device having an acceleration detecting and communication function, such as a smart phone, which is based on the difference between the kinematics and dynamic characteristics of the human body in a safe state of motion and the fall, as long as the user carries the applied body and falls.
  • a device having an acceleration detecting and communication function such as a smart phone
  • the application will automatically collect and analyze the human body dynamics information, determine whether the human body falls, and use the mobile phone communication advantages to send SMS, telephone and other alarm notifications.
  • it Compared with other fall detection equipment that needs to be purchased and equipped, it has a wide range of use, low price and convenient carrying, so it is relatively more practical; the key is to fully consider the characteristics of the human body's sports behavior, improve the detection rate, and reduce False positive rate.
  • the device 20 may include: a determination acquisition module 21, an area calculation module 22, and a fall determination module 23; the determination acquisition module 21 is configured to collect the resultant force in a fixed time when detecting that the combined acceleration of the human body is less than the low acceleration threshold a 0 The acceleration sequence I 0 ; the area calculation module 22 is configured to calculate the curve in the I 0 coordinate and the reference line surrounded by the reference line in the two-dimensional coordinates with the resultant acceleration and time as the coordinate axes and a 0 as the reference line.
  • the lower curve area; the fall determination module 23 is configured to: when the difference between the curve area on the reference line and the curve area under the reference line is smaller than the curve area threshold ⁇ S, and the time of the resultant force under the reference line is greater than the low acceleration When the time threshold is ⁇ T, it is determined that the human body falls.
  • the device 20 for human body fall detection provided by the present invention can automatically detect the acceleration information and speed information of the human body through the sensor, and integrate the human body acceleration, speed, exercise time and The state of motion correctly determines whether the human body has fallen, and reminds whether it is necessary to call for help, and at the same time, the operation of confirming and canceling the rescue can be performed for the reminder.
  • the device for realizing the above-mentioned human body fall detection device 20 is required to be provided with at least an acceleration sensor (such as a three-axis acceleration sensor, etc.), and a communication module (such as a GSM module in the mobile communication module) is required to be provided when the rescue function needs to be implemented. CDMA module, etc.).
  • the above device can be a mobile terminal, such as a commonly used mobile terminal device such as a smart phone or a tablet computer.
  • the device 20 for detecting the body fall can be disposed in the mobile terminal. Since the commonly used mobile terminal device is usually provided with a communication module and an acceleration sensor, the direct use can be directly utilized without additional configuration.
  • the device 20 may further include: a sample collection module 24 and an analysis extraction module 25; the sample collection module 24 is configured to collect speed information during a body fall in one or more individual body fall samples; The speed information includes a resultant acceleration, speed, and time; the analysis extraction module 25 is configured to analyze the characteristics of the speed information, calculate and extract a low acceleration threshold a 0 , a combined acceleration sequence I 0 in a fixed time, and a high acceleration threshold a 1
  • the curve area threshold ⁇ S and the low acceleration time threshold ⁇ T establish a fall detection mechanism.
  • the speed information during the fall of the human body is collected; the speed information includes the resultant acceleration, speed and time; and the acceleration time sequence before the crash is also recorded, and the extraction is performed.
  • Some acceleration features establish a fall detection model that can be continuously trained and optimized by falling data.
  • the establishment of the fall detection model is the most important part to achieve accurate fall detection.
  • the acceleration data related to the fall process is obtained for analysis, filtering, fusion, and extraction of human fall kinematics and power. Learning characteristics, such as: long low acceleration before falling, may appear squat before falling, etc., thus establishing a fall detection model, the model can be continuously trained and self-adjusted.
  • the movement state of the human body and the fall can be detected according to the change of the three-axis acceleration.
  • the extractable features are as follows: falling low acceleration state, speed, high speed state before collision, high speed state of collision And the time of maintenance of various states, etc., and then based on the above characteristics to establish a model for dynamically detecting the fall of the human body to match the triaxial acceleration value of the human body motion, and determine whether to fall according to the output probability, and some parameters in the model can be based on the height of the human body. , weight, amount of exercise and real-time movement adjustment of the human body.
  • the above analysis extraction module 25 can also be used to: calculate and extract a low acceleration threshold a 0 , a fixed time acceleration sequence I 0 , a high acceleration threshold a 1 , a curve area threshold ⁇ S, and a low acceleration time threshold ⁇ T and samples Corresponding human body state information is associated; the human body state information includes height, weight, and/or motion state. The height and weight can be an interval value.
  • Setting receiving module 26 for receiving the setting state information of the human body, the human body information adjustment according to a 0 state set, ⁇ S and: Referring to FIG. 9, for example, the apparatus 20 may also include the above-described another embodiment of the present invention, / or ⁇ T.
  • model parameters can be different according to the height and weight of the human body, and the real-time motion state of the human body, each user can set according to the different human body state information, and the detecting device will be based on the set human body state information. And matching the corresponding parameters such as a 0 , ⁇ S and / or ⁇ T.
  • the device 20 may further include: a self-learning module 30, configured to include the human body fall sample detected by itself into the fall detection mechanism, and associate the speed information of the sample with the human body state information.
  • a self-learning module 30 configured to include the human body fall sample detected by itself into the fall detection mechanism, and associate the speed information of the sample with the human body state information.
  • the fall event may be added to the human fall detection model according to the user's confirmation. For example, collecting speed information during the fall and analyzing, extracting a low acceleration threshold a 0 , a combined acceleration sequence I 0 , a high acceleration threshold a 1 , a curve area threshold ⁇ S, and a low acceleration time threshold ⁇ T And associating with the human body state information corresponding to the sample, the revision has established a fall detection mechanism, thereby realizing the self-learning mechanism of the above-mentioned human fall detection model.
  • the apparatus 20 further comprising: a high acceleration determination module 27, for the acceleration force during acquisition sequence I 0 is performed, determining whether there is a collection of sequence I 0 To a resultant acceleration greater than the high acceleration threshold a 1 ; if any, the area of the curve is calculated by the area calculation module 22.
  • the fall determination may proceed to the next step, or may be determined not occurred Fall down and return to the initial combined force detection and judgment.
  • the device 20 may further include: an alarm prompting module 28, configured to collect a current speed sequence of the human body after determining that the fall is exceeded; and the current speed is lower than the speed threshold exceeds the setting. At the time of the time, an alarm prompt message is generated; when the alarm prompt information is confirmed or not confirmed for a certain period of time, an alarm can be generated.
  • an alarm prompting module 28 configured to collect a current speed sequence of the human body after determining that the fall is exceeded; and the current speed is lower than the speed threshold exceeds the setting.
  • the initial initial velocity of the human body is zero, and the gravity acceleration is considered according to the change of the three-axis acceleration, and the speed of each moment of the human body can be calculated.
  • the triaxial acceleration will continue to change, and the speed will continue to change.
  • an acceleration time series of a fixed time length can be obtained, which can completely record the acceleration value of a fall process (including a period before and after the fall). Since the human body has a low acceleration state before falling, a low acceleration threshold can also be obtained from the experimental data. When the acceleration of the human body is lower than the low acceleration threshold during the movement, the acceleration data can be collected for the fall.
  • the detection model is inverted to further determine whether to continue collecting data or clearing the data and recording the time (ie, entering the above time series). Since the human body may have a high acceleration state before the collision due to defects, etc., the above-mentioned low acceleration may be judged by further high acceleration; a high acceleration threshold is obtained through experimental data, during the movement. When the acceleration of the human body is higher than the high acceleration threshold, the detection model can be used to match the acceleration data that has been collected.
  • the operation is performed and matched with the parameters in the model; if the matching is successful, The description has fallen, and the alarm is judged according to the speed information of the human body. If the state of the human body is below the speed threshold exceeds the set time, an alarm prompt message is generated, and the user selects whether to alarm according to the actual situation, and does not perform any operation for a certain period of time. , the alarm is executed, and the short message notification and/or the telephone notification are performed according to the preset contact telephone number.
  • the set body state information such as body height, weight and/or motion state, and the speed and acceleration information of the pre-fall and fall that have been collected
  • another high acceleration threshold can be obtained.
  • the combined acceleration of the human body is greater than the threshold, it is very likely that the user is subjected to an instantaneous strong impact, such as a car accident. If the real-time monitoring of the human body acceleration is greater than the threshold, the alarm prompt information is directly generated to perform an alarm operation through the communication component when the user confirms or prompts the timeout.
  • a two-dimensional coordinate diagram with time (10ms) as the abscissa and combined (force) acceleration (m/s2) as the ordinate is used, wherein the resultant acceleration a0 is taken as the reference line.
  • the acceleration of the human body force detected during the fall changes with time: when the time is from 100 to about 190, the resultant acceleration is always around 10, indicating that the user may be in a standing state at this time; From about 190 to about 430, the resultant acceleration changes uniformly and regularly at 10, and the resultant acceleration is above a0, indicating that the user may be in a normal walking state; at about 450 490, the first force occurs first.
  • the acceleration is lower than a0, and lasts between 450 and 480, and then there is a relatively short high acceleration that is several times higher than the normal (10), which lasts between 480 and 490.
  • the above low acceleration indicates the use.
  • the person may be in the process of falling before the fall.
  • the high acceleration indicates that the user may be in the process of falling after the fall; it is obvious in the figure that the resultant acceleration is formed in the coordinates during the fall.
  • the area of the curve under the reference line enclosed by the curve and the reference line is larger than the area of the curve on the reference line.
  • the calculation and matching method in the fall detection model is as follows: firstly, the low acceleration threshold a 0 is determined by experiments and research. When the resultant acceleration value generated during the movement of the human body is lower than a 0 , the trigger starts to collect the triaxial acceleration sensor data to one. Fixed acceleration time series I 0 ; simultaneous calculation is performed: calculating the area of the acceleration curve based on the reference line a 0 from the time of collecting the data, and the area enclosed by the acceleration and the reference line a0 is positive on the reference line. Otherwise, it is negative.
  • the total curve area is the curve area on the reference line plus the curve area under the baseline (that is, the difference between the curve area on the reference line and the curve area under the baseline when both are positive);
  • ⁇ S the sum of the acceleration times lower than the acceleration a 0 in I 0 is larger than ⁇ T, it can be considered that this is a process before the collision occurs.
  • ⁇ T, ⁇ S, a 0 can be dynamically adjusted according to the user's weight, height, and exercise state (the amount of exercise).
  • the detection can also be performed according to the acceleration and the speed.
  • the detected acceleration is below the low acceleration threshold, there is first a sequence of accelerations and speeds for a fixed time interval, the length of the sequence being sufficient to include the normal movement of the user before the complete fall has fallen.
  • the default user initial speed is zero, and the approximate speed of the human body at a certain moment can be calculated based on the three-axis acceleration and time.
  • the acceleration and velocity sequence of the fixed time length can obtain the current exercise intensity and state of the user, and accordingly, some related parameters of the fall verification can be adjusted, so that the model can more accurately detect the state before the fall.
  • the device 20 for detecting the fall of the human body may be based on a device having an acceleration detecting and communication function, such as a smart phone, and is based on a difference between the kinematics and dynamic characteristics of the human body in a safe state of motion and the fall, as long as the user carries the applied body with the body.
  • the application will automatically collect and analyze human body dynamics information to determine whether the human body has fallen, and use the mobile phone communication advantages to make alarm notifications such as SMS and telephone.
  • the mobile terminal system 40 can include an information acquisition module 29, a determination acquisition module 21, an area calculation module 22, a fall determination module 23, and an alarm prompt module 28.
  • the information acquisition module 29 is configured to acquire speed information by using a three-axis acceleration sensor.
  • the determination acquisition module 21 is configured to collect the resultant acceleration sequence I 0 in a fixed time when detecting that the combined acceleration of the human body is less than the low acceleration threshold a 0 .
  • the area calculation module 22 is configured to calculate the curve area in the I 0 coordinate and the reference line surrounded by the reference line in the two-dimensional coordinates with the resultant acceleration and time as the coordinate axes and a 0 as the reference line.
  • the fall determination module 23 is configured to: when the difference between the curve area on the reference line and the curve area under the reference line is smaller than the curve area threshold ⁇ S, and the time of the resultant force under the reference line is greater than the low acceleration time threshold ⁇ T
  • the alarm prompting module 28 is configured to collect the current speed sequence of the human body after determining that the fall is exceeded; when the current speed is lower than the speed threshold exceeds the set time, generate an alarm prompt message; when the alarm prompt information is After confirming or not being confirmed for more than a certain period of time, an alarm can be made through the mobile communication component.
  • the mobile terminal system 40 may further include: a sample collection module 24, an analysis extraction module 25, a high acceleration determination module 27, a setup reception module 26, a self-learning module 30, and the like.
  • the determination acquisition module 21, the area calculation module 22, the fall determination module 23, the sample collection module 24, the analysis extraction module 25, the high acceleration determination module 27, the setup receiving module 26, the alarm prompt module 28, and the self-learning module in this embodiment 30 can be specifically as described in the above embodiments.
  • the mobile terminal system 40 can be based on a device having an acceleration detecting and communication function, such as a smart phone, and is based on the difference between the kinematics and dynamic characteristics of the human body in the state of motion and the fall, as long as the user carries the applied body fall detection.
  • the smart phone, the application will automatically collect and analyze the human body dynamics information, determine whether the human body falls, and use the mobile phone communication advantages to send SMS, telephone and other alarm notifications.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Emergency Management (AREA)
  • Business, Economics & Management (AREA)
  • Gerontology & Geriatric Medicine (AREA)
  • Social Psychology (AREA)
  • Psychology (AREA)
  • Psychiatry (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Dentistry (AREA)
  • Physiology (AREA)
  • Emergency Alarm Devices (AREA)
  • Alarm Systems (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

一种人体摔倒检测的方法、装置及移动终端系统。该方法可包括：在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值ΔS，且基准线下的合力加速度所占时间大于低加速度时间阈值ΔT时，判定人体摔倒。所述装置使用范围广，价格便宜，携带方便，相对实用性较高；关键是充分考虑到了人体的运动行为特点，提高检出率，减少误判率。

Description

人体摔倒检测的方法、装置及移动终端系统 技术领域
本发明涉及到人体摔倒检测技术，特别是涉及到一种人体摔倒检测的方法、装置及移动终端系统。
背景技术
根据世界疾病控制和预防组织统计，世界上超过65岁的老人，每年有三分之一会摔倒，其中有一半为再摔倒。一次性摔倒中近10%会引起严重伤害和疾病，造成巨大的医疗负担和健康伤害。2014年中国老年人口将超过2亿，2025年达到3亿，2042年老年人口比例将超过30%。老人摔倒已经成为当前一个重大的医疗问题和社会问题，减少老年人摔倒带来的伤害已经成为国内外新的研究热点减轻医疗保障系统和老人子女的医疗负担；尤其对于子女在外独居，或者经常外出走动的老人，具有较为重要的应用价值。
目前国已经有一些摔倒检测装置，大多都是基于一个特殊设备，老人需要额外穿戴在身上，非常不方便。相关的一些检测方法由于信息处理方法、设备等限制，或者没有充分考虑到人体运动行为等原因，使得误判率较高。
发明内容
本发明的主要目的为提供一种人体摔倒检测的方法，可提升摔倒判断的准确性。
本发明提出一种人体摔倒检测的方法，包括：
在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒。
优选地，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之前，还包括：
采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；所述速度信息包括合力加速度、速度以及时间；
分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
优选地，所述分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、曲线面积阈值△S以及低加速度时间阈值△T还包括：
将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；所述人体状态信息包括身高、体重和/或运动状态。
优选地，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之后，还包括：
在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则进行下一步骤。
优选地，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之前，还包括：
接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
优选地，所述方法之后，还包括：
在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
本发明还提出一种人体摔倒检测的装置，包括：
判定采集模块，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
面积计算模块，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
摔倒判定模块，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒。
优选地，所述装置还包括：
样本收集模块，用于采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；所述速度信息包括合力加速度、速度以及时间；
分析提取模块，用于分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
优选地，所述分析提取模块还用于：
将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；所述人体状态信息包括身高、体重和/或运动状态。
优选地，所述装置还包括：
高加速判断模块，用于在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则通过面积计算模块计算曲线面积。
优选地，所述装置还包括：
设置接收模块，用于接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
优选地，所述装置还包括：
告警提示模块，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
本发明还提出一种移动终端系统，包括：
信息获取模块，用于通过三轴加速度感应器获取速度信息；
判定采集模块，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
面积计算模块，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
摔倒判定模块，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒；
告警提示模块，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可通过移动通信组件进行告警。
优选地，所述移动终端系统还包括：
高加速判断模块，用于在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则通过面积计算模块计算曲线面积。
本发明可以基于智能手机等具有加速度检测及通讯功能的设备，是基于人体安全运动状态与摔倒时运动学和动力学特性的不同，只要用户随身携带装有应用人体摔倒检测的智能手机，应用将自动采集和分析人体动力学信息，判断人体是否摔倒，并使用手机通讯优势进行短信、电话等报警通知。相对于其他需要额外购买和配备的摔倒检测装备来说，使用范围广，价格便宜，携带方便，所以相对实用性较高；关键是充分考虑到了人体的运动行为特点，提高检出率，减少误判率。
附图说明
图1 是本发明人体摔倒检测的方法一实施例的步骤流程示意图；
图2 是本发明人体摔倒检测的方法一实施例的另一步骤流程示意图；
图3 是本发明人体摔倒检测的方法另一实施例的步骤流程示意图；
图4 是本发明人体摔倒检测的方法又一实施例的步骤流程示意图；
图5 是本发明人体摔倒检测的方法还一实施例的步骤流程示意图；
图6 是本发明人体摔倒检测的方法一实施例中以时间为横坐标、合加速度为纵坐标的二维坐标示意图；
图7 是本发明人体摔倒检测的装置一实施例的结构示意图；
图8 是本发明人体摔倒检测的装置一实施例的另一结构示意图；
图9 是本发明人体摔倒检测的装置另一实施例的结构示意图；
图10 是本发明人体摔倒检测的装置又一实施例的结构示意图；
图11 是本发明人体摔倒检测的装置还一实施例的结构示意图；
图12 是本发明移动终端系统一实施例的结构示意图。
本发明目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
参照图1，提出本发明一种人体摔倒检测的方法一实施例。该人体摔倒检测的方法可包括：
步骤S11、判断检测到人体的合力加速度是否小于低加速度阈值a0；如是，则进行步骤S12；否则，继续执行检测；
步骤S12、采集固定时间内的合力加速度序列I0；
步骤S13、以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
步骤S14、判断基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值△S、且基准线下的合力加速度所占时间大于低加速度时间阈值△T是否成立；如成立，则进行步骤S15；否则，继续执行检测；
步骤S15、判定人体摔倒。
针对现有的摔倒检测技术和检测装置存在的问题，本发明提出的人体摔倒检测的方法，可以通过传感器自动检测人体的加速度信息及速度信息，综合人体加速度、速度、运动时间以及运动状态正确判断人体是否已经摔倒，以及提醒是否需要报警求救，同时针对该提醒可进行确认及取消救助的操作。
实现上述人体摔倒检测的方法的设备至少需配套设置有加速度传感器（比如三轴加速度传感器等），在需要实现求救功能时还需配套设置有通讯模块（比如移动通讯模块中的GSM模块、CDMA模块等）。为便于随身携带及方便使用，上述设备可为移动终端，比如智能手机、平板电脑等常用移动终端设备。由于该种常用移动终端设备通常都设置有通讯模块及加速度传感器，从而可直接利用无需另外再配置。
参照图2，上述步骤S11之前还可包括：
步骤S100、采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；该速度信息包括合力加速度、速度以及时间；
步骤S101、分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
在实现摔倒检测之前，首先通过若干摔倒实验，采集人体摔倒过程中的速度信息；该速度信息包括合力加速度、速度以及时间等；以及还可记录摔倒碰撞前的加速度时间序列，提取一些加速度特征，建立摔倒检测模型，该模型可以不断的通过摔倒数据训练，优化。
摔倒检测模型的建立是实现准确摔倒检测至关重要的部分，首先根据一些实验以及相关摔倒数据来获取摔倒过程相关加速度数据进行分析，过滤，融合，提取人体摔倒运动学及动力学特征，如：摔倒前的长时间的低加速度，摔倒前可能出现踉跄等，由此建立摔倒检测模型，模型可以不断的训练和自我调整。在实验中可根据三轴加速度的变化来检测人体运动状态以及摔倒，其中可提取的特征有如下几个：摔倒低加速度状态、速度、碰撞前踉跄的高速度状态、碰撞的高速度状态以及各种状态维持的时间等，然后依据上述特征来建立一个动态检测人体摔倒的模型去匹配人体运动的三轴加速度值，依据输出概率来判断是否摔倒，模型中一些参数可以根据人体身高，体重，运动量以及人体的实时运动状态调整。
针对计算取得的低加速度阈值a0、固定时间内的合力加速度序列I0、曲线面积阈值△S以及低加速度时间阈值△T：可将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；该人体状态信息可包括身高、体重和/或运动状态等。该身高、体重可以是一个区间值。
参照图3，在本发明另一实施例中，上述步骤S11之前及步骤S101之后，还可包括：
步骤S102、接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
由于模型参数可根据人体身高、体重、以及人体的实时运动状态的不同而有所区别，因此每个使用者可根据自身人体状态信息的不同而进行设置，检测设备将根据所设置的人体状态信息而匹配相应的a0、△S和/或△T等参数。
在上述方法在具体的使用过程中还可包括：将自身所检测的人体摔倒样本纳入摔倒检测机制中，并将样本的速度信息与人体状态信息关联。
当人体摔倒检测模型准确判定一次人体摔倒事件发生后，根据使用者的确认可将该次摔倒事件加入至该人体摔倒检测模型中。比如采集该次摔倒过程中的速度信息并分析，从中提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，以及与样本对应的人体状态信息关联，修订已建立摔倒检测机制，以此实现上述人体摔倒检测模型的自我学习机制。
参照图4，在本发明又一实施例中，上述步骤S11之后还可包括：
步骤S110、在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则进行步骤S12，否则继续执行检测。
由于人体摔倒之前总会有一个相对较长时间的低加速度过程，同时可能由于一些踉跄等其他外因导致摔倒碰撞前出现较高加速度的情况。因此，可在检测到出现低加速度过程后，再进行高加速度的检测；如出现高合力加速度大于高加速度阈值a1的情况，则可继续进行下一步骤的摔倒判断，否则可判定未发生摔倒，返回至初始的合力加速度检测及判断。
参照图5，在本发明还一实施例中，上述步骤S15之后，还可包括：
步骤S16、在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
本实施例的人体摔倒检测中，首先可默认人体初始速度为零，根据三轴加速度的变化已经考虑重力加速度，可以计算出人体每个时刻的速度是多少。随着人体的运动，三轴加速度会不断的变化，速度也跟着不断的变化。根据实验数据可获取一个固定时间长度的加速度时间序列，该时间序列可以完全记录一次摔倒过程（包括摔倒前、后的一段时间）的加速度数值。由于人体在摔倒前有一个低加速度状态，同样可从实验数据中得出一个低加速度阀值，当运动过程中人体的加速度低于该低加速度阈值时，即可开始收集加速度数据以供摔倒检测模型检测，以便进一步判断是继续收集数据还是清空数据，并记录时间（即进入上述时间序列）。由于人体在摔倒中可能由于踉跄等原因而在碰撞前出现高加速度状态，上述低加速度的判断之后即可进行进一步进行高加速度的判断；通过实验数据得到一个高加速度阀值，当运动过程中人体的加速度高于该高加速度阈值时，即可开始用检测模型匹配上述已经收集的加速度数据。
然后可根据设置的人体身高、体重和/或运动状态等人体状态信息，以及已经收集的摔倒前和摔倒的速度和加速度信息进行运算，并与模型中的参数进行匹配；如果匹配成功，说明已经摔倒，根据人体速度信息判断是否报警，如果人体处于低于速度阈值的状态超过设定时间则产生告警提示信息，由使用者根据实际情况选择是否报警，若一定时间内不做任何操作，则执行报警，根据预设的联系电话进行短信通知和/或电话通知等。
同时，根据实验数据还可以得出另外一个高加速度阀值，当人体的合力加速度大于该阀值时，即很有可能是使用者受到瞬间强烈撞击，如车祸等事件发生。如果实时监测的人体合力加速度大于该阀值时，直接产生告警提示信息，以在使用者确认或提示超时的情况下执行报警操作。
参照图6，为以时间（10ms）为横坐标、合（力）加速度（m/s2）为纵坐标的二维坐标示意图，其中以合力加速度a0为基准线。显示在摔倒过程中所检测到的人体合力加速度随时间的推移而变化：在时间为100至约190时，合力加速度一直处在10左右，说明此时使用者可能处于站立状态；在时间为约190至约430时，合力加速度在10上、下均匀且有规律地变化，此时合力加速度都在a0之上说明使用者可能处于正常行走状态；在时间约450约490时，首先出现合力加速度低于a0的情况，且持续于450至480之间，然后再出现较为短暂的高出常态（10）数倍的高加速度情况，持续于480至490之间，上述低加速度的情况说明使用者可能是处于摔倒前的跌落过程，该高加速度的情况说明使用者可能是处于摔倒后撞击过程；图示中可明显看出，在摔倒过程中，合力加速度在坐标中所形成的曲线与基准线围成的基准线下的曲线面积大于基准线上的曲线面积。
摔倒检测模型中计算与匹配的方式为：首先由实验和研究确定低加速度阀值a0，当人体运动过程中产生的合力加速度值低于a0时触发开始采集三轴加速度传感器数据到一个固定加速度时间序列I0；同时进行如下计算：计算从采集数据时刻起，基于基准线a0的加速度曲线面积，加速度与基准线a0所围成的面积在基准线上时则取值为正，反之则为负，总的曲线面积为基准线上曲线面积加上基准线下曲线面积（即等同于在两者都取正的情况下基准线上曲线面积与基准线下曲线面积之差）；只要所得曲线面积小于△S，且I0中低于加速度a0的加速度时间之和大于△T，则可以认为这是一次摔倒产生碰撞前的过程。其中△T，△S，a0可以根据用户体重，身高，运动状态（运动量）进行动态调整。
上述是针对加速度进行的人体摔倒检测，本实施例中还可根据加速度和速度进行检测。当检测到的加速度低于低加速度阈值时，首先有一个固定时间区间的加速度和速度序列，该序列的时间长度可以足够包含一次完整的摔倒过程已经摔倒前使用者正常运动的过程。默认使用者初始速度为零，可根据三轴加速度以及时间来计算出某个时刻人体的大概速度。同时固定时间长度的加速度和速度序列可以获得使用者当前的运动强度和状态，据此可以调整摔倒验证的一些相关参数，让模型能更准确的检测出摔倒前的状态。
上述人体摔倒检测的方法，可以基于智能手机等具有加速度检测及通讯功能的设备，是基于人体安全运动状态与摔倒时运动学和动力学特性的不同，只要用户随身携带装有应用人体摔倒检测的智能手机，应用将自动采集和分析人体动力学信息，判断人体是否摔倒，并使用手机通讯优势进行短信、电话等报警通知。相对于其他需要额外购买和配备的摔倒检测装备来说，使用范围广，价格便宜，携带方便，所以相对实用性较高；关键是充分考虑到了人体的运动行为特点，提高检出率，减少误判率。
参照图7，提出本发明一种人体摔倒检测的装置20一实施例。该装置20可包括：判定采集模块21、面积计算模块22以及摔倒判定模块23；该判定采集模块21，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；该面积计算模块22，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；该摔倒判定模块23，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒。
针对现有的摔倒检测技术和检测装置20存在的问题，本发明提出的人体摔倒检测的装置20，可以通过传感器自动检测人体的加速度信息及速度信息，综合人体加速度、速度、运动时间以及运动状态正确判断人体是否已经摔倒，以及提醒是否需要报警求救，同时针对该提醒可进行确认及取消救助的操作。
实现上述人体摔倒检测的装置20的设备至少需配套设置有加速度传感器（比如三轴加速度传感器等），在需要实现求救功能时还需配套设置有通讯模块（比如移动通讯模块中的GSM模块、CDMA模块等）。为便于随身携带及方便使用，上述设备可为移动终端，比如智能手机、平板电脑等常用移动终端设备。上述人体摔倒检测的装置20可设置于移动终端中，由于该种常用移动终端设备通常都设置有通讯模块及加速度传感器，从而可直接利用无需另外再配置。
参照图8，上述装置20还可包括：样本收集模块24以及分析提取模块25；该样本收集模块24，用于采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；所述速度信息包括合力加速度、速度以及时间；该分析提取模块25，用于分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
在实现摔倒检测之前，首先通过若干摔倒实验，采集人体摔倒过程中的速度信息；该速度信息包括合力加速度、速度以及时间等；以及还可记录摔倒碰撞前的加速度时间序列，提取一些加速度特征，建立摔倒检测模型，该模型可以不断的通过摔倒数据训练，优化。
摔倒检测模型的建立是实现准确摔倒检测至关重要的部分，首先根据一些实验以及相关摔倒数据来获取摔倒过程相关加速度数据进行分析，过滤，融合，提取人体摔倒运动学及动力学特征，如：摔倒前的长时间的低加速度，摔倒前可能出现踉跄等，由此建立摔倒检测模型，模型可以不断的训练和自我调整。在实验中可根据三轴加速度的变化来检测人体运动状态以及摔倒，其中可提取的特征有如下几个：摔倒低加速度状态、速度、碰撞前踉跄的高速度状态、碰撞的高速度状态以及各种状态维持的时间等，然后依据上述特征来建立一个动态检测人体摔倒的模型去匹配人体运动的三轴加速度值，依据输出概率来判断是否摔倒，模型中一些参数可以根据人体身高，体重，运动量以及人体的实时运动状态调整。
上述分析提取模块25还可用于：将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；该人体状态信息包括身高、体重和/或运动状态等。该身高、体重可以是一个区间值。
参照图9，在本发明的另一实施例中，上述装置20还可包括：设置接收模块26，用于接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
由于模型参数可根据人体身高、体重、以及人体的实时运动状态的不同而有所区别，因此每个使用者可根据自身人体状态信息的不同而进行设置，检测设备将根据所设置的人体状态信息而匹配相应的a0、△S和/或△T等参数。
上述装置20还可包括：自学习模块30，用于将自身所检测的人体摔倒样本纳入摔倒检测机制中，并将样本的速度信息与人体状态信息关联。
当人体摔倒检测模型准确判定一次人体摔倒事件发生后，根据使用者的确认可将该次摔倒事件加入至该人体摔倒检测模型中。比如采集该次摔倒过程中的速度信息并分析，从中提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，以及与样本对应的人体状态信息关联，修订已建立摔倒检测机制，以此实现上述人体摔倒检测模型的自我学习机制。
参照图10，在本发明的又一实施例中，上述装置20还可包括：高加速判断模块27，用于在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则通过面积计算模块22计算曲线面积。
由于人体摔倒之前总会有一个相对较长时间的低加速度过程，同时可能由于一些踉跄等其他外因导致摔倒碰撞前出现较高加速度的情况。因此，可在检测到出现低加速度过程后，再进行高加速度的检测；如出现高合力加速度大于高加速度阈值a1的情况，则可继续进行下一步骤的摔倒判断，否则可判定未发生摔倒，返回至初始的合力加速度检测及判断。
参照图11，在本发明的还一实施例中，上述装置20还可包括：告警提示模块28，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
本实施例的人体摔倒检测中，首先可默认人体初始速度为零，根据三轴加速度的变化已经考虑重力加速度，可以计算出人体每个时刻的速度是多少。随着人体的运动，三轴加速度会不断的变化，速度也跟着不断的变化。根据实验数据可获取一个固定时间长度的加速度时间序列，该时间序列可以完全记录一次摔倒过程（包括摔倒前、后的一段时间）的加速度数值。由于人体在摔倒前有一个低加速度状态，同样可从实验数据中得出一个低加速度阀值，当运动过程中人体的加速度低于该低加速度阈值时，即可开始收集加速度数据以供摔倒检测模型检测，以便进一步判断是继续收集数据还是清空数据，并记录时间（即进入上述时间序列）。由于人体在摔倒中可能由于踉跄等原因而在碰撞前出现高加速度状态，上述低加速度的判断之后即可进行进一步进行高加速度的判断；通过实验数据得到一个高加速度阀值，当运动过程中人体的加速度高于该高加速度阈值时，即可开始用检测模型匹配上述已经收集的加速度数据。
然后可根据设置的人体身高、体重和/或运动状态等人体状态信息，以及已经收集的摔倒前和摔倒的速度和加速度信息进行运算，并与模型中的参数进行匹配；如果匹配成功，说明已经摔倒，根据人体速度信息判断是否报警，如果人体处于低于速度阈值的状态超过设定时间则产生告警提示信息，由使用者根据实际情况选择是否报警，若一定时间内不做任何操作，则执行报警，根据预设的联系电话进行短信通知和/或电话通知等。
同时，根据实验数据还可以得出另外一个高加速度阀值，当人体的合力加速度大于该阀值时，即很有可能是使用者受到瞬间强烈撞击，如车祸等事件发生。如果实时监测的人体合力加速度大于该阀值时，直接产生告警提示信息，以在使用者确认或提示超时的情况下通过通信组件执行报警操作。
参照图6，为以时间（10ms）为横坐标、合（力）加速度（m/s2）为纵坐标的二维坐标示意图，其中以合力加速度a0为基准线。显示在摔倒过程中所检测到的人体合力加速度随时间的推移而变化：在时间为100至约190时，合力加速度一直处在10左右，说明此时使用者可能处于站立状态；在时间为约190至约430时，合力加速度在10上、下均匀且有规律地变化，此时合力加速度都在a0之上说明使用者可能处于正常行走状态；在时间约450约490时，首先出现合力加速度低于a0的情况，且持续于450至480之间，然后再出现较为短暂的高出常态（10）数倍的高加速度情况，持续于480至490之间，上述低加速度的情况说明使用者可能是处于摔倒前的跌落过程，该高加速度的情况说明使用者可能是处于摔倒后撞击过程；图示中可明显看出，在摔倒过程中，合力加速度在坐标中所形成的曲线与基准线围成的基准线下的曲线面积大于基准线上的曲线面积。
摔倒检测模型中计算与匹配的方式为：首先由实验和研究确定低加速度阀值a0，当人体运动过程中产生的合力加速度值低于a0时触发开始采集三轴加速度传感器数据到一个固定加速度时间序列I0；同时进行如下计算：计算从采集数据时刻起，基于基准线a0的加速度曲线面积，加速度与基准线a0所围成的面积在基准线上时则取值为正，反之则为负，总的曲线面积为基准线上曲线面积加上基准线下曲线面积（即等同于在两者都取正的情况下基准线上曲线面积与基准线下曲线面积之差）；只要所得曲线面积小于△S，且I0中低于加速度a0的加速度时间之和大于△T，则可以认为这是一次摔倒产生碰撞前的过程。其中△T，△S，a0可以根据用户体重，身高，运动状态（运动量）进行动态调整。
上述是针对加速度进行的人体摔倒检测，本实施例中还可根据加速度和速度进行检测。当检测到的加速度低于低加速度阈值时，首先有一个固定时间区间的加速度和速度序列，该序列的时间长度可以足够包含一次完整的摔倒过程已经摔倒前使用者正常运动的过程。默认使用者初始速度为零，可根据三轴加速度以及时间来计算出某个时刻人体的大概速度。同时固定时间长度的加速度和速度序列可以获得使用者当前的运动强度和状态，据此可以调整摔倒验证的一些相关参数，让模型能更准确的检测出摔倒前的状态。
上述人体摔倒检测的装置20，可以基于智能手机等具有加速度检测及通讯功能的设备，是基于人体安全运动状态与摔倒时运动学和动力学特性的不同，只要用户随身携带装有应用人体摔倒检测的智能手机，应用将自动采集和分析人体动力学信息，判断人体是否摔倒，并使用手机通讯优势进行短信、电话等报警通知。相对于其他需要额外购买和配备的摔倒检测装备来说，使用范围广，价格便宜，携带方便，所以相对实用性较高；关键是充分考虑到了人体的运动行为特点，提高检出率，减少误判率。
参照图12，提出本发明一种移动终端系统40一实施例。该移动终端系统40可包括：信息获取模块29、判定采集模块21、面积计算模块22、摔倒判定模块23以及告警提示模块28。该信息获取模块29，用于通过三轴加速度感应器获取速度信息；该判定采集模块21，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；该面积计算模块22，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；该摔倒判定模块23，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒；该告警提示模块28，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可通过移动通信组件进行告警。
上述移动终端系统40还可包括：样本收集模块24、分析提取模块25、高加速判断模块27、设置接收模块26以及自学习模块30等。本实施例中的判定采集模块21、面积计算模块22、摔倒判定模块23、样本收集模块24、分析提取模块25、高加速判断模块27、设置接收模块26、告警提示模块28以及自学习模块30具体可如上述实施例中所述。
上述移动终端系统40，可以基于智能手机等具有加速度检测及通讯功能的设备，是基于人体安全运动状态与摔倒时运动学和动力学特性的不同，只要用户随身携带装有应用人体摔倒检测的智能手机，应用将自动采集和分析人体动力学信息，判断人体是否摔倒，并使用手机通讯优势进行短信、电话等报警通知。相对于其他需要额外购买和配备的摔倒检测装备来说，使用范围广，价格便宜，携带方便，所以相对实用性较高；关键是充分考虑到了人体的运动行为特点，提高检出率，减少误判率。
以上所述仅为本发明的优选实施例，并非因此限制本发明的专利范围，凡是利用本发明说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本发明的专利保护范围内。

Claims (14)

  1. 一种人体摔倒检测的方法，其特征在于，包括：
    在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
    以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
    当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒。
  2. 跟据权利要求1所述的人体摔倒检测的方法，其特征在于，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之前，还包括：
    采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；所述速度信息包括合力加速度、速度以及时间；
    分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
  3. 跟据权利要求2所述的人体摔倒检测的方法，其特征在于，所述分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、曲线面积阈值△S以及低加速度时间阈值△T还包括：
    将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；所述人体状态信息包括身高、体重和/或运动状态。
  4. 跟据权利要求1所述的人体摔倒检测的方法，其特征在于，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之后，还包括：
    在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则进行下一步骤。
  5. 跟据权利要求1所述的人体摔倒检测的方法，其特征在于，所述在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0的步骤之前，还包括：
    接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
  6. 跟据权利要求1所述的人体摔倒检测的方法，其特征在于，所述方法之后，还包括：
    在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；
    当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
  7. 一种人体摔倒检测的装置，其特征在于，包括：
    判定采集模块，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
    面积计算模块，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
    摔倒判定模块，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒。
  8. 跟据权利要求7所述的人体摔倒检测的装置，其特征在于，所述装置还包括：
    样本收集模块，用于采集一个及以上个人体摔倒样本中人体摔倒过程中的速度信息；所述速度信息包括合力加速度、速度以及时间；
    分析提取模块，用于分析所述速度信息的特征，计算并提取低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T，建立摔倒检测机制。
  9. 跟据权利要求8所述的人体摔倒检测的装置，其特征在于，所述分析提取模块还用于：
    将计算并提取的低加速度阈值a0、固定时间内的合力加速度序列I0、高加速度阈值a1、曲线面积阈值△S以及低加速度时间阈值△T与样本对应的人体状态信息关联；所述人体状态信息包括身高、体重和/或运动状态。
  10. 跟据权利要求7所述的人体摔倒检测的装置，其特征在于，所述装置还包括：
    高加速判断模块，用于在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则通过面积计算模块计算曲线面积。
  11. 跟据权利要求7所述的人体摔倒检测的装置，其特征在于，所述装置还包括：
    设置接收模块，用于接收人体状态信息的设定，根据设定的人体状态信息调整a0、△S和/或△T。
  12. 跟据权利要求7所述的人体摔倒检测的装置，其特征在于，所述装置还包括：
    告警提示模块，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时， 产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可进行告警。
  13. 一种移动终端系统，其特征在于，包括：
    信息获取模块，用于通过三轴加速度感应器获取速度信息；
    判定采集模块，用于在检测人体的合力加速度小于低加速度阈值a0时，采集固定时间内的合力加速度序列I0；
    面积计算模块，用于以合力加速度与时间为坐标轴、a0为基准线的二维坐标中，计算I0坐标中的曲线与基准线围成的基准线上、下的曲线面积；
    摔倒判定模块，用于当基准线上的曲线面积与基准线下的曲线面积之差小于曲线面积阈值 △S，且基准线下的合力加速度所占时间大于低加速度时间阈值△T时，判定人体摔倒；
    告警提示模块，用于在判定摔倒后，采集人体当前速度序列；在当前速度低于速度阈值超过设定时间时，产生告警提示信息；当告警提示信息被确认或者未被确认超过一定时间，即可通过移动通信组件进行告警。
  14. 跟据权利要求13所述的移动终端系统，其特征在于，所述移动终端系统还包括：
    高加速判断模块，用于在进行合力加速度序列I0的采集时，判断所述序列I0中是否有采集到大于高加速度阈值a1的合力加速度；如有，则通过面积计算模块计算曲线面积。
PCT/CN2014/087996 2013-09-30 2014-09-30 人体摔倒检测的方法、装置及移动终端系统 WO2015043546A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/084,508 US9928717B2 (en) 2013-09-30 2016-03-30 Human body tumbling detection method and device and mobile terminal system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310469682.6A CN103581852B (zh) 2013-09-30 2013-09-30 人体摔倒检测的方法、装置及移动终端系统
CN201310469682.6 2013-09-30

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/084,508 Continuation US9928717B2 (en) 2013-09-30 2016-03-30 Human body tumbling detection method and device and mobile terminal system

Publications (1)

Publication Number Publication Date
WO2015043546A1 true WO2015043546A1 (zh) 2015-04-02

Family

ID=50052592

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/087996 WO2015043546A1 (zh) 2013-09-30 2014-09-30 人体摔倒检测的方法、装置及移动终端系统

Country Status (3)

Country Link
US (1) US9928717B2 (zh)
CN (1) CN103581852B (zh)
WO (1) WO2015043546A1 (zh)

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103581852B (zh) * 2013-09-30 2018-03-06 吴家宝 人体摔倒检测的方法、装置及移动终端系统
CN103577836B (zh) * 2013-09-30 2018-01-23 吴家宝 人体摔倒检测模型建立方法及模型系统
CN104408876A (zh) * 2014-12-10 2015-03-11 范子成 一种从高处摔落的检测及报警方法和装置
CN104574441B (zh) * 2014-12-31 2017-07-28 浙江工业大学 一种基于gmm和时序模型的跌倒实时检测方法
US11120679B2 (en) 2015-04-30 2021-09-14 Honeywell International Inc. System for integrating multiple sensor data to predict a fall risk
WO2016186160A1 (ja) * 2015-05-21 2016-11-24 コニカミノルタ株式会社 画像処理システム、画像処理装置、画像処理方法、および画像処理プログラム
CN105823479B (zh) * 2016-03-04 2018-08-17 上海畅星软件有限公司 驾驶行为分析方法
CN106384480A (zh) * 2016-09-30 2017-02-08 福建工程学院 一种在运动状态下的摔倒监测方法、系统及终端
CN106228753A (zh) * 2016-09-30 2016-12-14 福建工程学院 一种意外摔倒监控方法、系统及终端
CN106327799A (zh) * 2016-09-30 2017-01-11 福建工程学院 一种在行走状态下的摔倒监测方法及系统
CN106373338A (zh) * 2016-09-30 2017-02-01 福建工程学院 一种摔倒监测方法及系统
CN107126198B (zh) * 2017-06-05 2020-06-30 河池学院 穿戴式摔倒检测设备
CN107126217A (zh) * 2017-06-05 2017-09-05 河池学院 穿戴式摔倒检测设备
CN109407504B (zh) * 2018-11-30 2021-05-14 华南理工大学 一种基于智能手表的人身安全检测系统及方法
CN110638459B (zh) * 2019-09-03 2023-05-05 宁波路晟电器科技有限公司 一种基于加速度传感器的人体运动跌倒检测装置及方法
CN110738822B (zh) * 2019-10-24 2022-09-20 合肥盛东信息科技有限公司 一种摔倒预警系统
CN113520306B (zh) * 2020-04-17 2024-03-22 青岛海尔空调器有限总公司 一种人体睡眠状态的监测方法和智能家居装置
CN112438726A (zh) * 2020-11-20 2021-03-05 深圳市卓翼科技股份有限公司 基于可穿戴设备的跌倒报警方法、可穿戴设备及存储介质

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0816986A2 (en) * 1996-07-03 1998-01-07 Hitachi, Ltd. Method, apparatus and system for recognizing motions
US20050093709A1 (en) * 2003-07-31 2005-05-05 Wellcare Systems Inc. Comprehensive monitoring system
US20070107068A1 (en) * 2005-10-14 2007-05-10 Oqo, Inc. Hybrid hardware/firmware multi-axis accelerometers for drop detect and tumble detect
WO2009151412A1 (en) * 2008-06-12 2009-12-17 Valerii Kanevskyi Method for detecting the critical position in orientation and indicating the change of the position of a human being body
CN101650869A (zh) * 2009-09-23 2010-02-17 中国科学院合肥物质科学研究院 一种人体摔倒自动检测并报警的装置及其信息处理方法
JP2010207488A (ja) * 2009-03-12 2010-09-24 Gifu Univ 行動解析装置及びプログラム
US20110025493A1 (en) * 2005-03-11 2011-02-03 Aframe Digital, Inc. Mobile wireless customizable health and condition monitor
US20120059927A1 (en) * 2006-12-21 2012-03-08 24/8 Llc Apparatus and method for wireless autonomous infant mobility detection, monitoring, analysis and alarm event generation
CN103577836A (zh) * 2013-09-30 2014-02-12 吴家宝 人体摔倒检测模型建立方法及模型系统
CN103581852A (zh) * 2013-09-30 2014-02-12 吴家宝 人体摔倒检测的方法、装置及移动终端系统

Family Cites Families (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5982573A (en) * 1993-12-15 1999-11-09 Hewlett-Packard Company Disk drive and method for minimizing shock-induced damage
JP2005114402A (ja) * 2003-10-03 2005-04-28 Sony Corp 落下検出方法および装置
KR100618866B1 (ko) * 2004-10-02 2006-08-31 삼성전자주식회사 전자기기의 자유 낙하 검출방법 및 장치
US7369345B1 (en) * 2004-12-03 2008-05-06 Maxtor Corporation Mobile hard disk drive free fall detection and protection
CN101111770B (zh) * 2005-01-31 2010-07-14 日立金属株式会社 掉落检测方法和掉落检测设备
WO2009037970A1 (ja) * 2007-09-21 2009-03-26 Murata Manufacturing Co., Ltd. 落下検知装置、磁気ディスク装置および携帯電子機器
US8206325B1 (en) * 2007-10-12 2012-06-26 Biosensics, L.L.C. Ambulatory system for measuring and monitoring physical activity and risk of falling and for automatic fall detection
EP2274734B1 (en) * 2008-05-12 2012-07-11 Koninklijke Philips Electronics N.V. Displacement measurement in a fall detection system
CN102099859B (zh) * 2008-07-23 2015-04-15 株式会社村田制作所 下落检测装置、磁盘装置、以及便携式电子设备
US8990041B2 (en) * 2009-04-27 2015-03-24 Mayo Foundation For Medical Education And Research Fall detection
BR112012012229A2 (pt) * 2009-11-25 2017-12-26 Koninl Philips Electronics Nv detector de queda para uso na detecção de quedas por um usuário, método para a estimativa de uma velocidade vertical e/ou um deslocamento vertical de um objeto que compreende um acelerômetro, método para uso na detecção de quedas por um usuário de um detector de queda que compreende um acelerômetro e produto de programa de computador
JP6023809B2 (ja) * 2011-08-18 2016-11-09 コーニンクレッカ フィリップス エヌ ヴェＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｎ．Ｖ． 加速度測定に基づく水平又は垂直方向の速度推定
US9402568B2 (en) * 2011-08-29 2016-08-02 Verizon Telematics Inc. Method and system for detecting a fall based on comparing data to criteria derived from multiple fall data sets
US9020476B2 (en) * 2011-09-12 2015-04-28 Leipzig Technology, Llc System and method for remote care and monitoring using a mobile device
CN202210337U (zh) * 2011-09-29 2012-05-02 西安中星测控有限公司 人体跌倒检测报警器
US9818281B2 (en) * 2011-11-14 2017-11-14 Vital Connect, Inc. Method and system for fall detection of a user
CN102542724B (zh) * 2011-12-15 2014-01-01 上海移为通信技术有限公司 跌落检测报警方法
CN102657533B (zh) * 2012-04-28 2013-11-13 浙江大学城市学院 一种跌倒检测方法、跌倒检测装置和手腕式设备
ES2482591B1 (es) * 2012-08-29 2015-05-29 Vodafone España, S.A.U. Detección de caídas
CN103177530B (zh) * 2013-02-25 2015-07-08 宁波中科集成电路设计中心有限公司 一种基于三轴加速度的人体摔倒检测方法
US9589442B2 (en) * 2013-09-03 2017-03-07 Verizon Telematics Inc. Adaptive classification of fall detection for personal emergency response systems
CA2875843A1 (en) * 2013-12-20 2015-06-20 Solutions Novika Activity, posture and heart monitoring system and method

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0816986A2 (en) * 1996-07-03 1998-01-07 Hitachi, Ltd. Method, apparatus and system for recognizing motions
US20050093709A1 (en) * 2003-07-31 2005-05-05 Wellcare Systems Inc. Comprehensive monitoring system
US20110025493A1 (en) * 2005-03-11 2011-02-03 Aframe Digital, Inc. Mobile wireless customizable health and condition monitor
US20070107068A1 (en) * 2005-10-14 2007-05-10 Oqo, Inc. Hybrid hardware/firmware multi-axis accelerometers for drop detect and tumble detect
US20120059927A1 (en) * 2006-12-21 2012-03-08 24/8 Llc Apparatus and method for wireless autonomous infant mobility detection, monitoring, analysis and alarm event generation
WO2009151412A1 (en) * 2008-06-12 2009-12-17 Valerii Kanevskyi Method for detecting the critical position in orientation and indicating the change of the position of a human being body
JP2010207488A (ja) * 2009-03-12 2010-09-24 Gifu Univ 行動解析装置及びプログラム
CN101650869A (zh) * 2009-09-23 2010-02-17 中国科学院合肥物质科学研究院 一种人体摔倒自动检测并报警的装置及其信息处理方法
CN103577836A (zh) * 2013-09-30 2014-02-12 吴家宝 人体摔倒检测模型建立方法及模型系统
CN103581852A (zh) * 2013-09-30 2014-02-12 吴家宝 人体摔倒检测的方法、装置及移动终端系统

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CUI, YINGHUI ET AL.: "Fall detection system based on three-axis acceleration sensor for the elderly", MODERN ELECTRONICS TECHNIQUE, vol. 36, no. 3, 1 February 2013 (2013-02-01), pages 130 - 132 *
NI, YIYANG: "FALL DETECTION SOFTWARE FOR THE ELDERLY BASED ON ANDROID PLATFORM", COMPUTER CD SOFTWARE AND APPLICATIONS, 31 August 2013 (2013-08-31), pages 55 - 57 *
REN, ZHILING ET AL.: "Design of Fall Detection Technology and Alert System Based on Accelerometer", COMPUTER MEASUREMENT & CONTROL, vol. 21, no. 6, 30 June 2013 (2013-06-30), pages 1428 - 1430 AND 1433 *

Also Published As

Publication number Publication date
CN103581852B (zh) 2018-03-06
US9928717B2 (en) 2018-03-27
CN103581852A (zh) 2014-02-12
US20160210835A1 (en) 2016-07-21

Similar Documents

Publication Publication Date Title
WO2015043546A1 (zh) 人体摔倒检测的方法、装置及移动终端系统
CN103577836B (zh) 人体摔倒检测模型建立方法及模型系统
RU2584459C2 (ru) Жестовое управление для отслеживания показателей жизнедеятельности
Lee et al. Development of an enhanced threshold-based fall detection system using smartphones with built-in accelerometers
US20170372582A1 (en) Device for monitoring and preventing slump and processing method thereof
KR100988459B1 (ko) 낙상 감지 장치 및 방법
WO2017049957A1 (zh) 智能跌倒监护装置及其处理方法
WO2017125094A1 (zh) 一种人体摔倒智能控制系统及其方法
EP3188569B1 (en) Method for detecting fall-off of wearable device, and wearable device
CN107633657A (zh) 一种疲劳驾驶的提醒方法及终端
CN103503041A (zh) 用于跌倒检测器或跌倒检测系统中的装置及其操作方法
CN201691930U (zh) 一种人体运动状态监测装置
KR101609914B1 (ko) 물리적충격 및 심리적충격에 대응한 긴급상황감지방법과 이를 이용한 긴급상황감지장치
JP2007325842A (ja) 健康管理機能を備えた携帯端末
KR101259560B1 (ko) 손목 착용형 낙상 감지 장치 및 방법과 이를 이용한 낙상 감지 시스템
CN105528859A (zh) 基于人体跌倒检测技术的养老智能报警系统
CN102646319A (zh) 生物体跌倒检测装置及方法
KR20080050993A (ko) 낙상 감지 장치 및 그 방법과 그를 이용한 낙상 구조서비스 시스템 및 그 방법
KR101883008B1 (ko) 낙상 예측 또는 감지 장치 및 그를 위한 방법
KR20170004269A (ko) 낙상 감지 장치 및 방법
CN106953982A (zh) 一种低头族步行安全警报方法
Liang et al. Pre-impact alarm system for fall detection using MEMS sensors and HMM-based SVM classifier
CN203118166U (zh) 生物体跌倒检测装置
Ye et al. Research of fall detection and alarm applications for the elderly
Soaz et al. A new method to estimate the real upper limit of the false alarm rate in a 3 accelerometry-based fall detector for the elderly

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: 14847461

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14847461

Country of ref document: EP

Kind code of ref document: A1