WO2022057071A1 - 一种跌倒检测方法、系统、终端以及存储介质 - Google Patents

一种跌倒检测方法、系统、终端以及存储介质 Download PDF

Info

Publication number
WO2022057071A1
WO2022057071A1 PCT/CN2020/129534 CN2020129534W WO2022057071A1 WO 2022057071 A1 WO2022057071 A1 WO 2022057071A1 CN 2020129534 W CN2020129534 W CN 2020129534W WO 2022057071 A1 WO2022057071 A1 WO 2022057071A1
Authority
WO
WIPO (PCT)
Prior art keywords
fall
action
resultant acceleration
limit threshold
monitored person
Prior art date
Application number
PCT/CN2020/129534
Other languages
English (en)
French (fr)
Inventor
赵国如
王光辉
宁运琨
郑凯
张宇
蔡凌峰
Original Assignee
中国科学院深圳先进技术研究院
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 中国科学院深圳先进技术研究院 filed Critical 中国科学院深圳先进技术研究院
Publication of WO2022057071A1 publication Critical patent/WO2022057071A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/02—Alarms for ensuring the safety of persons
    • G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
    • G08B21/0407—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis
    • G08B21/043—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis detecting an emergency event, e.g. a fall
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C9/00—Measuring inclination, e.g. by clinometers, by levels
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
    • G01P15/00—Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration
    • G01P15/18—Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration in two or more dimensions
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/02—Alarms for ensuring the safety of persons
    • G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
    • G08B21/0438—Sensor means for detecting
    • G08B21/0446—Sensor means for detecting worn on the body to detect changes of posture, e.g. a fall, inclination, acceleration, gait

Definitions

  • the present application belongs to the technical field of human fall detection, and in particular, relates to a fall detection method, system, terminal and storage medium.
  • fall warning algorithms include:
  • Patent No. CN105632101A proposes a human body fall prevention early warning method and system. According to the resultant acceleration mean value obtained from the resultant accelerations of all sampling points and the preset resultant acceleration mean value threshold, combined with the Euler angle difference sequence and the preset The Euler angle difference threshold determines whether the monitored object has a tendency to fall; alarm information is generated when the monitored object has a tendency to fall, so as to give an alarm before the monitored object falls.
  • Patent No. CN106683342A proposes an anti-fall early warning system, which is worn on the user, small in size and easy to carry, does not affect the normal movements of people during activities, monitors the user's movement characteristics in real time, and provides real-time protection and protection to people who are prone to falls. Treat in time.
  • the shortcomings of the fall early warning algorithm are: it needs to rely on high-performance computing platforms and machine learning algorithms, the calculation process is complex, and the amount of calculation is large. Generally, offline methods are used for research, and real-time monitoring and real-time motion data cannot be performed Fall warning.
  • the present application provides a fall detection method, system, terminal, and storage medium, aiming to solve one of the above-mentioned technical problems in the prior art at least to a certain extent.
  • a fall detection method comprising:
  • the fall direction of the fall-like action is identified according to the posture angle data, and whether the fall-like action is a real fall action is determined according to the posture angle limit threshold of each fall direction.
  • the technical solution adopted in the embodiment of the present application further includes: before the extraction of the resultant acceleration and attitude angle data from the motion data of the monitored person includes:
  • the movement data of the monitored person is acquired in real time through the nine-axis inertial sensor in the wearable fall early warning device; the wearable fall early warning device is worn on the monitored person.
  • the technical solution adopted in the embodiment of the present application further includes: performing sampling analysis on the resultant acceleration of the monitored person during the fall process, and identifying the monitored person according to the waveform change rate and the trough value of the resultant acceleration during the fall process.
  • the monitor's fall-like maneuvers include:
  • the resultant acceleration is sampled at a set frequency, and the set number of sampling points is taken as a unit, and when the resultant acceleration is detected to be less than the first set threshold within a unit time, the number of sampling points T1 is accumulated. , and start accumulating the number of sampling points T 2 when the combined acceleration is less than the second set threshold, and when both T 1 and T 2 reach the set threshold of the number of sampling points, it is determined that the current action is a fall-like action; Otherwise, it is determined that the current action is a daily action.
  • the technical solutions adopted in the embodiments of the present application further include: the first set threshold is 0.85g, the second set threshold is 0.75g, the threshold for the number of sampling points for T 1 is 5, and the T 2 threshold is 5. The threshold for the number of sampling points is 2.
  • the technical solution adopted in the embodiment of the present application further includes: the attitude angle data includes a pitch angle and a roll angle.
  • the technical solution adopted in the embodiment of the present application further includes: identifying the fall direction of the fall-like action according to the attitude angle data, and determining whether the fall-like action belongs to a real fall according to the attitude angle limit thresholds of each fall direction Actions include:
  • the first roll angle limit threshold and the first pitch angle limit threshold are respectively set in the front and rear inversion process; the second pitch angle limit threshold and the second roll angle limit threshold are respectively set in the left and right inversion process; wherein, the first roll angle The limit threshold is greater than the second roll angle limit threshold, and the second pitch angle limit threshold is greater than the first pitch angle limit threshold;
  • both the roll angle and the pitch angle of the fall-like action are greater than the first roll angle limit threshold and the first pitch angle limit threshold, then determine that the fall-like action is a forward or backward fall;
  • both the pitch angle and the roll angle of the fall-like action are greater than the second pitch angle limit threshold and the second roll angle limit threshold, it is determined that the fall-like action is a left or right fall.
  • the technical solution adopted in the embodiment of the present application further includes: after it is determined that the fall-like action is a real fall action, the following further includes:
  • the air bag is inflated.
  • a fall detection system comprising:
  • Data extraction module used to extract the resultant acceleration and attitude angle data from the motion data of the monitored person
  • the first fall identification module used to sample and analyze the resultant acceleration during the fall process of the monitored person, and identify the fall-like fall of the monitored person according to the waveform change rate and the trough value of the resultant acceleration during the fall process. action;
  • the second fall recognition module is used to identify the fall direction of the fall-like action according to the posture angle data, and determine whether the fall-like action belongs to a real fall according to the posture angle limit thresholds of each fall direction.
  • a terminal includes a processor and a memory coupled to the processor, wherein,
  • the memory stores program instructions for implementing the fall detection method
  • the processor is configured to execute the program instructions stored in the memory to control fall detection.
  • a storage medium storing program instructions executable by a processor, where the program instructions are used to execute the fall detection method.
  • the beneficial effects of the embodiments of the present application are: the fall detection method of the embodiments of the present application extracts the resultant acceleration and attitude angle data in real time through the motion data of the nine-axis inertial sensor, and the resultant acceleration and attitude angle are extracted in real time.
  • the data is analyzed in real time, and the falling movement of the monitored person is identified within the fall warning lead time, and real-time warning of the falling movement is carried out.
  • the warning time is short and the accuracy is high.
  • Figure 2 is a waveform diagram of the combined acceleration of the fall action
  • FIG. 3 is a schematic diagram of the resultant acceleration sampling statistics according to an embodiment of the present application.
  • FIG. 4 is a schematic structural diagram of a fall detection system according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of a terminal according to an embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of a storage medium according to an embodiment of the present application.
  • the fall detection method of the embodiment of the present application extracts the resultant acceleration and attitude angle data in real time from the motion data received by the nine-axis inertial sensor in the wearable fall warning device. Carry out real-time analysis to identify the falling action of the monitored person, and complete the early warning of the falling action within the fall warning lead time. At the same time, the fall signal is converted into an electrical signal to inflate the airbag in the wearable fall warning device. In order to effectively protect the human body when the human body touches the ground.
  • FIG. 1 is a flowchart of a fall detection method according to an embodiment of the present application.
  • the fall detection method of the embodiment of the present application includes the following steps:
  • the wearable fall warning device is worn on the person being monitored, and it includes a nine-axis inertial sensor, an embedded fall warning module, a fall warning device and at least one airbag.
  • the nine-axis inertial sensor is used to obtain the movement data of the monitored person.
  • the embedded fall early warning module is used to extract the combined acceleration and attitude angle data from the motion data, and identify the fall action and fall direction according to the combined acceleration and attitude angle data; the fall early warning device is used for the fall warning before the monitored person falls.
  • the alarm information is generated within the set time to give an alarm before the monitored object falls; the airbag is used to convert the fall signal into an electrical signal within the fall warning lead time when the monitored person falls, so that the steering gear can be rotated and punctured.
  • the compressed gas cylinder releases a large amount of gas to inflate the airbag device, so as to effectively protect the human body when the human body touches the ground.
  • attitude angle data includes but is not limited to pitch angle and roll angle, etc.
  • S30 Sampling and analyzing the resultant acceleration during the fall process, and according to the waveform change rate and the trough value of the resultant acceleration during the fall process, determine whether the current movement of the monitored person belongs to a daily movement or a fall-like movement. If it belongs to a fall-like movement, execute S40 ; otherwise, execute S20 again;
  • the fall process means that the overall waveform shows a downward trend in the fall warning lead time.
  • the resultant acceleration includes two important characteristics, that is, the speed of the waveform falling from the normal value to the trough is very fast (the rate of change), and the trough is very fast.
  • the valley value is small, and based on the above two features, the fall-like actions can be distinguished from other daily actions.
  • area I is the daily action area before the fall
  • area II is the fall warning area
  • area III is the fall protection area
  • area IV is the area after the fall;
  • the rate of resultant acceleration in time zone II varies greatly, and the rate of change can be represented by the derivation of the resultant acceleration with respect to time t, that is, the acceleration of the waveform graph, or by the decreasing value per unit time.
  • the trough is usually less than 0.35g, but if the time axis in region II is elongated, it will be seen that there are still many peaks and troughs within a small time period.
  • the embodiment of the present application samples the resultant acceleration during the fall process, and counts the sampling points of the resultant acceleration a 1 g and a 2 g per unit time respectively. According to the set threshold of the number of sampling points, it is judged whether the current action is a fall-like action.
  • the resultant acceleration sampling process specifically includes: by sampling the resultant acceleration at a frequency of 100Hz, that is, the time interval between two sampling points is 10ms, and taking 50 sampling points as a unit, when the resultant acceleration is detected in a unit time When it is less than the first set threshold a 1 g, start to accumulate the number of sampling points T 1 , and start to accumulate the number of sampling points when it is detected that the resultant acceleration is less than the second set threshold a 2 g (a 2 g ⁇ a 1 g).
  • a 2 g and a 1 g are real-time accelerations.
  • both T 1 and T 2 do not reach the set threshold for the number of sampling points, it means that the acceleration of the current action has not yet reached the level of a fall-like action, and it is determined that the current The action is a daily action; on the contrary, if both T 1 and T 2 reach the set threshold for the number of sampling points, it means that the acceleration of the current action has reached the level of a fall-like action, that is, the current action is determined to be a fall-like action.
  • the optimal thresholds for the number of sampling points for T 1 and T 2 are obtained, specifically: the threshold for the number of sampling points for T 1 is 5, the threshold for the number of sampling points for T 2 is 2, and the threshold for the number of sampling points for T 2 is 2.
  • the value can also be set on a case-by-case basis.
  • a 2 g is used as the threshold to distinguish fall-like actions from daily actions such as walking, sitting up, bending over, etc.
  • the peaks and troughs within the range of the cells between sampling save the analysis time of fall warning.
  • the embodiment of the present application collects daily movements such as walking, sitting and standing up, slowly lying down and standing up, going up and down stairs, jumping, squatting up, jogging, etc.
  • the duration of the area II and the area III in FIG. 2 is the fall warning lead time.
  • the embodiment of the present application mainly performs the sampling and statistical analysis of a 2 g and a 1 g for the area II. The earlier the fall warning in the area II is, the longer the time left for fall protection will be.
  • the specific fall warning lead time is as follows: According to statistics, the human body starts to tilt until the human body touches the ground, and then the lead time of the fall warning is 428ms, the lead time of the left fall warning is 463ms, and the right fall warning lead time is 463ms.
  • the time threshold can be set according to the specific situation.
  • S40 Determine the falling direction of the fall-like action according to the attitude angle data, and determine whether the current action belongs to a real falling action or a daily action according to the attitude angle limit thresholds in different directions;
  • the fall-like actions also include many non-fall daily actions, such as sitting down, lying down, going down stairs, etc., in order to reduce the misjudgment of daily actions, it is necessary to further conduct the fall-like actions according to the posture angle data. identification.
  • the fall direction usually includes front and rear and left and right.
  • the pitch angle varies greatly.
  • the embodiment of the present application while judging the fall direction by the attitude angle, sets a limit threshold for the attitude angle in different directions, and identifies the real fall movement in the fall-like movement according to the threshold value.
  • the attitude angle limit thresholds in different directions are set as follows: the first roll angle limit threshold is set during the forward and backward inversion process. At the same time, since there will be a certain angle of left and right inclination during the forward and backward inversion process, it is also necessary to set a small range of the first roll angle limit threshold. A pitch angle limit threshold to improve the early warning accuracy of the front and rear inversion; a second pitch angle limit threshold is set during the left and right inversion process.
  • the second roll angle limit threshold is set to improve the early warning accuracy of the left-right reverse action; wherein, the first roll angle limit threshold is greater than the second roll angle limit threshold, and the second pitch angle limit threshold is greater than the first pitch angle limit threshold. If the current roll angle and pitch angle are both greater than the set first roll angle limit threshold and first pitch angle limit threshold, it is determined that the current action is falling forward or backward; accordingly, if the current pitch angle and roll angle are If both are greater than the set second pitch angle limit threshold and second roll angle limit threshold, it is determined that the current action is falling to the left or right.
  • the embodiment of the present application collects daily movements such as walking, sitting and standing up, slowly lying down and standing up, going up and down stairs, jumping, squatting up, jogging, etc. , and then perform statistical analysis on these actions, and set the values of pitch angle and roll angle.
  • the real fall action can be accurately identified, the false alarm rate in the fall warning is reduced, and the accuracy of the fall warning is improved.
  • the compressed gas cylinder is punctured to inflate the right airbag, so as to provide more accurate and effective protection for the monitored person.
  • the fall detection method of the embodiment of the present application recognizes the falling action of the monitored person by extracting the resultant acceleration and attitude angle data in real time from the motion data of the nine-axis inertial sensor, and performing real-time analysis on the resultant acceleration and attitude angle data. Real-time early warning of falls within the fall warning lead time, with short warning time and high accuracy.
  • the rate of change in the waveform diagram of the resultant acceleration is represented by a statistical method, the calculation process and the amount of calculation are reduced, the influence of the peaks and valleys on the rate of change in a small period of time is avoided, and the fall action can be warned earlier.
  • the embodiment of the present application has a small amount of calculation and low power consumption, and is suitable for an embedded platform.
  • the following examples use the fall warning algorithm of the embodiments of the present application to respectively perform the following steps: walking, sitting and standing up, slowly lying down and standing up, going up and down stairs, jumping, squatting up and standing up, jogging, etc.
  • Experiments were carried out on daily activities and fall movements such as backward fall, left fall, and right fall, and the number of experiments was 100 times. Among them, the results of the false alarm rate of daily actions are shown in Table 1, and the accuracy rate of falling actions is shown in Table 2.
  • FIG. 4 is a schematic structural diagram of a fall detection system according to an embodiment of the present application.
  • the fall detection system 40 of the embodiment of the present application includes:
  • Data extraction module 41 used to extract the resultant acceleration and attitude angle data from the motion data of the monitored person
  • the first fall identification module 42 is used to sample and analyze the resultant acceleration during the fall process of the monitored person, and identify the type of the monitored person according to the waveform change rate and the trough value of the resultant acceleration during the fall process. fall action;
  • the second fall recognition module 43 is configured to identify the fall direction of the fall-like action according to the attitude angle data, and determine whether the fall-like movement belongs to a real fall according to the attitude angle limit thresholds of each fall direction.
  • FIG. 5 is a schematic structural diagram of a terminal according to an embodiment of the present application.
  • the terminal 50 includes a processor 51 and a memory 52 coupled to the processor 51 .
  • the memory 52 stores program instructions for implementing the above-described fall detection method.
  • the processor 51 is configured to execute program instructions stored in the memory 52 to control fall detection.
  • the processor 51 may also be referred to as a CPU (Central Processing Unit, central processing unit).
  • the processor 51 may be an integrated circuit chip with signal processing capability.
  • the processor 51 may also be a general purpose processor, digital signal processor (DSP), application specific integrated circuit (ASIC), off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component .
  • DSP digital signal processor
  • ASIC application specific integrated circuit
  • FPGA off-the-shelf programmable gate array
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • FIG. 6 is a schematic structural diagram of a storage medium according to an embodiment of the present application.
  • the storage medium of this embodiment of the present application stores a program file 61 capable of implementing all the above methods, wherein the program file 61 may be stored in the above-mentioned storage medium in the form of a software product, and includes several instructions to make a computer device (which may It is a personal computer, a server, or a network device, etc.) or a processor that executes all or part of the steps of the methods of the various embodiments of the present invention.
  • a computer device which may It is a personal computer, a server, or a network device, etc.
  • a processor that executes all or part of the steps of the methods of the various embodiments of the present invention.
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes , or terminal devices such as computers, servers, mobile phones, and tablets.

Landscapes

  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Gerontology & Geriatric Medicine (AREA)
  • Emergency Management (AREA)
  • Social Psychology (AREA)
  • Psychology (AREA)
  • Psychiatry (AREA)
  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Emergency Alarm Devices (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

一种跌倒检测方法、系统、终端以及存储介质。所述方法包括：从被监测者的运动数据中提取合加速度和姿态角数据；对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。本方法能够在跌倒预警前置时间内快速识别出被监测者的跌倒动作，并对跌倒动作进行实时预警，预警时间短，且准确性高。

Description

一种跌倒检测方法、系统、终端以及存储介质 技术领域
本申请属于人体跌倒检测技术领域，特别涉及一种跌倒检测方法、系统、终端以及存储介质。
背景技术
人体在走路或站立时为了保持微妙的平衡，会自主通过脚踝、臀部、躯干及跨步策略来调整姿态，这种平衡的控制会受到运动系统和中枢神经系统的影响。当平衡被打破、无法通过自身调整体位来维持稳定，或者平衡能力、神经认知功能严重受损无法自主调整，平衡无法保持就会出现跌倒的情况。随着年龄的增长，平衡能力逐步退化，跌倒发生的风险也会随之增大，跌倒后身体自主采取补救措施的能力也越缓慢。髋部骨折是老年人常见的严重损伤，90％以上的老人髋部骨折是由跌倒所致，发生髋部骨折后，一般需要长时间的卧床，常伴随肌肉萎缩、坠积性肺炎、泌尿系统感染和长褥疮等系统并发症，造成生活不能自理，甚至危及生命。有研究报告，老年人发生髋部骨折后，死亡率高达50％，而五年存活率只有20％。跌倒预警作为跌倒预防的有效手段，研究易跌人群的运动信息，在其跌倒过程给予髋部缓冲保护，可有效的减轻老年人跌倒后受得伤害，对解决人口老年化带来的医疗和社会问题具有重大意义。
现有技术中，跌倒预警算法包括：
专利号CN105632101A中提出一种人体防跌倒预警方法及系统，根据由所有采样点的合加速度得到的合加速度均值及预设的合加速度均值阈值，结合所述欧拉角差值序列及预设的欧拉角差值阈值判断被监测对象是否有跌倒倾向；在被监测对象有跌倒倾向时生成报警信息，以在被监测对象跌倒前进行报警。
专利号CN106683342A中提出的一种防跌倒预警系统,穿戴在使用者身上，体积小便于携带，不影响人在活动时的正常动作，实时监测使用者的运动特征，对易跌倒人群进行实时防护和及时救治。
专利号CN206697009U中提出的一种防跌倒预警系统，微控制主板植入基于阈值的跌倒预警算法，负责计算出人体所处的大致姿态，判断姿态角是否超出预设阈值，若超出阈值符合摔倒条件，则发出警报信息。
综上所述的跌倒预警算法存在的不足在于：需要依靠高性能的计算平台和机器学习算法，计算过程复杂，计算量较大，一般使用离线方法进行研究，无法对运动数据进行实时监测以及实时跌倒预警。
发明内容
本申请提供了一种跌倒检测方法、系统、终端以及存储介质，旨在至少在一定程度上解决现有技术中的上述技术问题之一。
为了解决上述问题，本申请提供了如下技术方案：
一种跌倒检测方法，包括：
从被监测者的运动数据中提取合加速度和姿态角数据；
对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；
根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。
本申请实施例采取的技术方案还包括：所述从被监测者的运动数据中提取合加速度和姿态角数据前包括：
通过穿戴式跌倒预警设备中的九轴惯性传感器实时获取被监测者的运动数据；所述穿戴式跌倒预警设备佩戴于被监测者身上。
本申请实施例采取的技术方案还包括：所述对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作包括：
以设定频率对所述合加速度进行采样，并以设定数量的采样点为一个单位，在单位时间内当检测到合加速度小于第一设定阈值时开始对采样点个数T 1进行累计，并在合加速度小于第二设定阈值时开始对采样点个数T 2进行累计，当T 1、T 2均达到设定的采样点个数阈值时，即判定当前动作为类跌倒动作；否则，判定当前动作为日常动作。
本申请实施例采取的技术方案还包括：所述第一设定阈值为0.85g，所述第二设定阈值为0.75g，所述T 1的采样点个数阈值为5，所述T 2的采样点个数阈值为2。
本申请实施例采取的技术方案还包括：所述姿态角数据包括pitch角与roll角。
本申请实施例采取的技术方案还包括：所述根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作包括：
在前后倒过程中分别设置第一roll角限制阈值以及第一pitch角限制阈值；在左右倒过程中分别设置第二pitch角限制阈值以及第二roll角限制阈值；其中，所述第一roll角限制阈值大于第二roll角限制阈值，所述第二pitch角限制阈值大于第一pitch角限制阈值；
如果所述类跌倒动作的roll角和pitch角均大于所述第一roll角限制阈值和第一pitch角限制阈值，则判定所述类跌倒动作为向前或向后跌倒；
如果所述类跌倒动作的pitch角和roll角均大于所述第二pitch角限制阈值和第二roll角限制阈值，则判定所述类跌倒动作为向左或向右跌倒。
本申请实施例采取的技术方案还包括：当判定所述类跌倒动作属于真正的跌倒动作后还包括：
触发所述穿戴式跌倒预警设备中的跌倒预警器在跌倒预警前置时间内生成报警信息，同时，将跌倒信号转化成电信号，控制舵机转动并刺破压缩气瓶对对应跌倒方向的安全气囊进行充气。
本申请实施例采取的另一技术方案为：一种跌倒检测系统，包括：
数据提取模块：用于从被监测者的运动数据中提取合加速度和姿态角数据；
第一跌倒识别模块：用于对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；
第二跌倒识别模块：用于根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。
本申请实施例采取的又一技术方案为：一种终端，所述终端包括处理器、与所述处理器耦接的存储器，其中，
所述存储器存储有用于实现所述跌倒检测方法的程序指令；
所述处理器用于执行所述存储器存储的所述程序指令以控制跌倒检测。
本申请实施例采取的又一技术方案为：一种存储介质，存储有处理器可运行的程序指令，所述程序指令用于执行所述跌倒检测方法。
相对于现有技术，本申请实施例产生的有益效果在于：本申请实施例的跌倒检测方法通过对九轴惯性传感器的运动数据进行合加速度以及姿态角数据的实时提取，对合加速度以及姿态角数据进行实时性分析，在跌倒预警前置时间内识别出被监测者的跌倒动作，并对跌倒动作进行实时预警，预警时间短，且准确性高。
附图说明
图1是本申请实施例的跌倒检测方法的流程图；
图2为跌倒动作合加速度波形图；
图3为本申请实施例的合加速度采样统计原理图；
图4为本申请实施例的跌倒检测系统结构示意图；
图5为本申请实施例的终端结构示意图；
图6为本申请实施例的存储介质的结构示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。
针对现有技术的不足，本申请实施例的跌倒检测方法通过对穿戴式跌倒预警设备中九轴惯性传感器接收到的运动数据进行合加速度以及姿态角数据的实时提取，对合加速度以及姿态角数据进行实时性分析，识别出被监测者的跌倒动作，并在跌倒预警前置时间内完成对跌倒动作的预警，同时将跌倒信号转化成电信号，使穿戴式跌倒预警设备中的安全气囊膨胀，以在人体接触地面时对人体进行有效保护。
具体的，请参阅图1，是本申请实施例的跌倒检测方法的流程图。本申请实施例的跌倒检测方法包括以下步骤：
S10：通过穿戴式跌倒预警设备实时获取被监测者的运动数据；
其中，穿戴式跌倒预警设备佩戴于被监测者身上，其包括九轴惯性传感器、嵌入式跌倒预警模块、跌倒预警器以及至少一个安全气囊，九轴惯性传感器用于获取被监测者的运动数据，嵌入式跌倒预警模块用于从运动数据中提取出合加速度以及姿态角数据，根据合加速度以及姿态角数据进行跌倒动作以及跌倒方向的识别；跌倒预警器用于在被监测者发生跌倒动作的跌倒预警前置时间内生成报警信息，以在被监测对象跌倒前进行报警；安全气囊用于在被监测者发生跌倒动作的跌倒预警前置时间内将跌倒信号转化成电信 号，使舵机转动，刺破压缩气瓶，释放出大量气体对安全气囊装置进行充气，以在人体接触地面时对人体进行有效保护。
S20：从运动数据中提取合加速度和姿态角数据；
其中，由于跌倒过程中人体处于失重状态，失重过程中的合加速度下降很快，且人体运动过程中姿态角数据也会不断变化，因此，本申请实施例通过提取合加速度和姿态角数据对跌倒动作和日常动作进行准确区分。其中姿态角数据包括但不限于pitch角与roll角等。
S30：对跌倒过程中的合加速度进行采样分析，根据跌倒过程中合加速度的波形变化速率和波谷谷值判断被监测者的当前动作属于日常动作还是类跌倒动作，如果属于类跌倒动作，执行S40；否则，重新执行S20；
本步骤中，如图2所示，为跌倒动作合加速度波形图。跌倒过程即在跌倒预警前置时间内波形图整体呈下降趋势，在此时间段内合加速度包括两个重要特征，即波形从正常值下降到波谷的速度很快(变化速率)，且波谷的谷值很小，基于上述两个特征可以对类似于跌倒的类跌倒动作与其他日常动作进行区分。假设将整个跌倒动作分为四个区域，区域Ⅰ为跌倒前的日常动作区域，区域Ⅱ为跌倒预警区域，区域Ⅲ为跌倒防护区域，区域Ⅳ为倒后区域；如图2所示，跌倒的时候区域Ⅱ内合加速度的速率变化很大，变化速率可以通过对合加速度关于时间t求导表示，即波形图的加速度，也可以通过单位时间内的下降值表示。跌倒过程中，波谷通常小于0.35g，但如果将区域Ⅱ中的时间轴拉长就会看到还存在很多小时间段内的波峰波谷，求导一方面需要花费更长时间进行分析，计算量较大功耗变大，会延长跌倒预警的时间；另一方面小时间段内的波峰波谷对合加速度曲线求导的影响较大。为了避免 波形图中短时间的波峰波谷对变化速率的影响，本申请实施例通过对跌倒过程中的合加速度进行采样，并分别统计单位时间内合加速度a 1g、a 2g的采样点个数，根据设定的采样点个数阈值判断当前动作是否属于类跌倒动作。
请一并参阅图3，为本申请实施例的合加速度采样统计原理图。合加速度采样过程具体包括：通过以100Hz的频率对合加速度进行采样，即两个采样点之间的时间间隔为10ms，并以50个采样点为一个单位，在单位时间内当检测到合加速度小于第一设定阈值a 1g时开始对采样点个数T 1进行累计，当检测到合加速度小于第二设定阈值a 2g(a 2g<a 1g)时开始对采样点个数T 2进行累计(由于a 2g<a 1g，因此当合加速度小于a 2g时开始对其采样点个数T 2进行累计的同时，合加速肯定小于a 1g，T 1也在累计)；a 2g和a 1g是实时的加速度，如果T 1、T 2均没有达到设定的采样点个数阈值，表示当前动作的加速度还未达到类跌倒动作的程度，则判定当前动作是日常动作；反之，如果T 1、T 2均达到设定的采样点个数阈值，表示当前动作的加速度已经达类跌倒的程度，即判定当前动作为类跌倒动作。本申请实施例中，经过实验，得到T 1和T 2的最优采样点个数阈值，具体为：T 1的采样点个数阈值为5，T 2的采样点个数阈值为2，该值也可根据具体情况进行设定。
通过上述采样统计方法，只用统计单位时间内合加速度小于a 1g、a 2g的采样点个数，由于合加速度小于a 1g的采样点个数T 1和合加速度小于a 2g的采样点个数T 2是很小的(T 2是包含在T 1里面的)，相当于T 1时间内检测到合加速度不仅小于a 1g而且小于a 2g的采样点个数为T 2，表示T 1时间内合加速度曲线的斜率。对于T 2个合加速度小于a 2g的采样点，以a 2g为阈值对类跌倒动作与失重较轻的走路、坐下起立、弯腰等日常动作进行区分，从而可以忽略开始 采样与结束采样之间小区间范围内的波峰波谷，节约跌倒预警的分析时间。同时，由于a 2g在跌倒动作合加速度波形图中对应的时间点属于跌倒预警前置时间较为提前的，因此可以较早的对跌倒动作进行预警。其中，本申请实施例通过采集多人的走路、坐下站起、缓慢躺下站起、上下楼梯、跳跃、蹲下站起、慢跑等日常动作以及后倒、左倒、右倒等跌倒动作，然后对这些动作进行统计分析，分别得到a 2g和a 1g以及T 1和T 2的最优值，具体为：a 1g＝0.85g，a 2g＝0.75g，该值也可根据具体情况进行设定。
本申请实施例中，图2中区域Ⅱ和区域Ⅲ的持续时间即为跌倒预警前置时间。本申请实施例主要针对区域Ⅱ进行a 2g和a 1g的采样统计分析，在区域Ⅱ内的预警跌倒越靠前，则留给跌倒防护的时间也会越长。跌倒预警前置时间具体为：据统计，人体从身体开始倾斜到人体接触地面，其后倒预警前置时间为428ms，左倒预警前置时间为463ms，右倒预警前置时间为463ms，具体时间阈值可根据具体情况进行设置。
S40：根据姿态角数据判定类跌倒动作的跌倒方向，并根据不同方向的姿态角限制阈值判定当前动作属于真正的跌倒动作还是日常动作；
本步骤中，由于类跌倒动作中还包含很多非跌倒的日常动作，例如坐下、躺下、下楼梯等，为了减少对日常动作的误判，还需要根据姿态角数据对类跌倒动作进行进一步的识别。人体运动过程中，不同的运动状态其姿态角变化规律也不一样，具体体现为：跌倒方向通常包括前后倒和左右倒，人体在前后倒过程中的roll角变化较大，左右倒过程中的pitch角变化较大。基于上述姿态角特征，本申请实施例在通过姿态角判别跌倒方向的同时，通过对 不同方向的姿态角设置限制阈值，根据该阈值识别类跌倒动作中真正的跌倒动作。
具体的，不同方向的姿态角限制阈值设置方式为：在前后倒过程中设置第一roll角限制阈值，同时由于前后倒过程中会有一定角度的左右倾斜，因此，还需要设置小范围的第一pitch角限制阈值，以提高前后倒动作的预警准确性；在左右倒过程中设置第二pitch角限制阈值，同时由于左右倒过程中会有一定角度的前后倾斜，因此，还需要设置小范围的第二roll角限制阈值，以提高左右倒动作的预警准确性；其中，第一roll角限制阈值大于第二roll角限制阈值，第二pitch角限制阈值大于第一pitch角限制阈值。如果当前的roll角和pitch角均大于设定的第一roll角限制阈值和第一pitch角限制阈值，则判定当前动作为向前或向后跌倒；相应地，如果当前的pitch角和roll角均大于设定的第二pitch角限制阈值和第二roll角限制阈值，则判定当前动作为向左或向右跌倒。其中，本申请实施例通过采集多人的走路、坐下站起、缓慢躺下站起、上下楼梯、跳跃、蹲下站起、慢跑等日常动作以及后倒、左倒、右倒等跌倒动作，然后对这些动作进行统计分析，并设定pitch角和roll角的数值。
基于上述，通过对合加速度的统计分析和姿态角的判断，可以准确识别出真实的跌倒动作，降低跌倒预警中的误报率，提高跌倒动作的预警准确性。
S50：当判定到真正的跌倒动作时，触发跌倒预警器在跌倒预警前置时间内生成报警信息，以在被监测对象跌倒前进行报警；同时，将跌倒信号转化成电信号对对应跌倒方向的安全气囊进行充气，已在被监测者跌倒时对其进行保护；
其中，假设识别到被监测者的跌倒方向为向右跌倒，则刺破压缩气瓶对右侧安全气囊进行充气，以对被监测者进行更加精准有效的防护。
本申请实施例的跌倒检测方法通过对九轴惯性传感器的运动数据进行合加速度以及姿态角数据的实时提取，对合加速度以及姿态角数据进行实时性分析，识别出被监测者的跌倒动作，并在跌倒预警前置时间内对跌倒动作进行实时预警，预警时间短，且准确性高。本申请通过将合加速度的波形图中的变化速率用统计方法进行表示，减少计算过程与计算量，避免小时间段内波峰波谷对变化速率的影响，可以更早的对跌倒动作进行预警。本申请实施例的计算量较小，功耗低，适用于嵌入式平台。
为了验证本申请的可行性和有效性，以下实施例采用本申请实施例的跌倒预警算法分别对走路、坐下站起、缓慢躺下站起、上下楼梯、跳跃、蹲下站起、慢跑等日常活动和后倒、左倒、右倒等跌倒动作进行实验，实验次数为100次。其中，日常动作的误报率结果如表一所示，跌倒动作准确率如表二所示。
表一日常动作误报率
Figure PCTCN2020129534-appb-000001
表二跌倒动作准确率
动作类型 后倒 左倒 右倒
准确性 100％ 100％ 100％
请参阅图4，是本申请实施例的跌倒检测系统的结构示意图。本申请实施例的跌倒检测系统40包括：
数据提取模块41：用于从被监测者的运动数据中提取合加速度和姿态角数据；
第一跌倒识别模块42：用于对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；
第二跌倒识别模块43：用于根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。
请参阅图5，为本申请实施例的终端结构示意图。该终端50包括处理器51、与处理器51耦接的存储器52。
存储器52存储有用于实现上述跌倒检测方法的程序指令。
处理器51用于执行存储器52存储的程序指令以控制跌倒检测。
其中，处理器51还可以称为CPU(Central Processing Unit，中央处理单元)。处理器51可能是一种集成电路芯片，具有信号的处理能力。处理器51还可以是通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现成可编程门阵列(FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
请参阅图6，为本申请实施例的存储介质的结构示意图。本申请实施例的存储介质存储有能够实现上述所有方法的程序文件61，其中，该程序文件61可以以软件产品的形式存储在上述存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)或处理器(processor)执行本发明各个实施方式方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质，或者是计算机、服务器、手机、平板等终端设备。
对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本申请。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本申请中所定义的一般原理可以在不脱离本申请的精神或范围的情况下，在其它实施例中实现。因此，本申请将不会被限制于本申请所示的这些实施例，而是要符合与本申请所公开的原理和新颖特点相一致的最宽的范围。

Claims (10)

  1. 一种跌倒检测方法，其特征在于，包括：
    从被监测者的运动数据中提取合加速度和姿态角数据；
    对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；
    根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。
  2. 根据权利要求1所述的跌倒检测方法，其特征在于，所述从被监测者的运动数据中提取合加速度和姿态角数据前包括：
    通过穿戴式跌倒预警设备中的九轴惯性传感器实时获取被监测者的运动数据；所述穿戴式跌倒预警设备佩戴于被监测者身上。
  3. 根据权利要求2所述的跌倒检测方法，其特征在于，所述对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作包括：
    以设定频率对所述合加速度进行采样，并以设定数量的采样点为一个单位，在单位时间内当检测到合加速度小于第一设定阈值时开始对采样点个数T 1进行累计，并在合加速度小于第二设定阈值时开始对采样点个数T 2进行累计，当T 1、T 2均达到设定的采样点个数阈值时，即判定当前动作为类跌倒动作；否则，判定当前动作为日常动作。
  4. 根据权利要求3所述的跌倒检测方法，其特征在于，所述第一设定阈值为0.85g，所述第二设定阈值为0.75g，所述T 1的采样点个数阈值为5，所述T 2的采样点个数阈值为2。
  5. 根据权利要求1所述的跌倒检测方法，其特征在于，所述姿态角数据包括pitch角与roll角。
  6. 根据权利要求5所述的跌倒检测方法，其特征在于，所述根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作包括：
    在前后倒过程中分别设置第一roll角限制阈值以及第一pitch角限制阈值；在左右倒过程中分别设置第二pitch角限制阈值以及第二roll角限制阈值；其中，所述第一roll角限制阈值大于第二roll角限制阈值，所述第二pitch角限制阈值大于第一pitch角限制阈值；
    如果所述类跌倒动作的roll角和pitch角均大于所述第一roll角限制阈值和第一pitch角限制阈值，则判定所述类跌倒动作为向前或向后跌倒；
    如果所述类跌倒动作的pitch角和roll角均大于所述第二pitch角限制阈值和第二roll角限制阈值，则判定所述类跌倒动作为向左或向右跌倒。
  7. 根据权利要求2至6任一项所述的跌倒检测方法，其特征在于，当判定所述类跌倒动作属于真正的跌倒动作后还包括：
    触发所述穿戴式跌倒预警设备中的跌倒预警器在跌倒预警前置时间内生成报警信息，同时，将跌倒信号转化成电信号，控制舵机转动并刺破压缩气瓶对对应跌倒方向的安全气囊进行充气。
  8. 一种跌倒检测系统，其特征在于，包括：
    数据提取模块：用于从被监测者的运动数据中提取合加速度和姿态角数据；
    第一跌倒识别模块：用于对所述被监测者跌倒过程中的合加速度进行采样分析，根据所述跌倒过程中合加速度的波形变化速率和波谷谷值识别出所述被监测者的类跌倒动作；
    第二跌倒识别模块：用于根据所述姿态角数据识别所述类跌倒动作的跌倒方向，并根据各个跌倒方向的姿态角限制阈值判定所述类跌倒动作是否属于真正的跌倒动作。
  9. 一种终端，其特征在于，所述终端包括处理器、与所述处理器耦接的存储器，其中，
    所述存储器存储有用于实现权利要求1-7任一项所述的跌倒检测方法的程序指令；
    所述处理器用于执行所述存储器存储的所述程序指令以控制跌倒检测。
  10. 一种存储介质，其特征在于，存储有处理器可运行的程序指令，所述程序指令用于执行权利要求1至7任一项所述跌倒检测方法。
PCT/CN2020/129534 2020-09-21 2020-11-17 一种跌倒检测方法、系统、终端以及存储介质 WO2022057071A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010994202.8A CN112233374B (zh) 2020-09-21 2020-09-21 一种跌倒检测方法、系统、终端以及存储介质
CN202010994202.8 2020-09-21

Publications (1)

Publication Number Publication Date
WO2022057071A1 true WO2022057071A1 (zh) 2022-03-24

Family

ID=74108097

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/129534 WO2022057071A1 (zh) 2020-09-21 2020-11-17 一种跌倒检测方法、系统、终端以及存储介质

Country Status (2)

Country Link
CN (1) CN112233374B (zh)
WO (1) WO2022057071A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AT526614A1 (de) * 2022-10-21 2024-05-15 Adaptive Regelsysteme Ges M B H Sicherheitsgerät und Verfahren zum Erkennen einer Notsituation einer Person

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112927475B (zh) * 2021-01-27 2022-06-10 浙江理工大学 一种基于深度学习的跌倒检测系统
CN114997245A (zh) * 2022-07-19 2022-09-02 广州帝隆科技股份有限公司 一种基于人工智能的防跌倒方法及装置
CN115024717B (zh) * 2022-08-09 2022-10-25 广东百年医疗健康科技发展有限公司 一种跌倒检测方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105336102A (zh) * 2015-11-25 2016-02-17 青岛理工大学 摔倒检测定位方法和装置
CN105632101A (zh) * 2015-12-31 2016-06-01 深圳先进技术研究院 一种人体防跌倒预警方法及系统
CN110992649A (zh) * 2019-12-18 2020-04-10 西南交通大学 一种用于判断行动中跌倒状态的方法
US20200250954A1 (en) * 2017-09-29 2020-08-06 Apple Inc. Detecting falls using a mobile device

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108392208B (zh) * 2018-02-23 2020-12-15 大同公元三九八智慧养老服务有限公司 一种基于多阶段模型的跌倒判断方法
CN110363959B (zh) * 2019-05-28 2021-01-22 哈尔滨理工大学 一种基于足底压力和三轴加速度传感器的摔倒判定方法
CN110876613B (zh) * 2019-09-27 2022-07-22 深圳先进技术研究院 一种人体运动状态识别方法、系统及电子设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105336102A (zh) * 2015-11-25 2016-02-17 青岛理工大学 摔倒检测定位方法和装置
CN105632101A (zh) * 2015-12-31 2016-06-01 深圳先进技术研究院 一种人体防跌倒预警方法及系统
US20200250954A1 (en) * 2017-09-29 2020-08-06 Apple Inc. Detecting falls using a mobile device
CN110992649A (zh) * 2019-12-18 2020-04-10 西南交通大学 一种用于判断行动中跌倒状态的方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AT526614A1 (de) * 2022-10-21 2024-05-15 Adaptive Regelsysteme Ges M B H Sicherheitsgerät und Verfahren zum Erkennen einer Notsituation einer Person

Also Published As

Publication number Publication date
CN112233374B (zh) 2022-06-21
CN112233374A (zh) 2021-01-15

Similar Documents

Publication Publication Date Title
WO2022057071A1 (zh) 一种跌倒检测方法、系统、终端以及存储介质
Vo et al. Fall detection based on movement and smart phone technology
Lai et al. Adaptive body posture analysis for elderly-falling detection with multisensors
Devi et al. Driver fatigue detection based on eye tracking
US9928717B2 (en) Human body tumbling detection method and device and mobile terminal system
Wang et al. Recognizing human daily activities from accelerometer signal
CN108257352B (zh) 一种基于智能穿戴设备的跌倒检测预警方法
CN113963192A (zh) 跌倒检测方法、装置和电子设备
CN110946585A (zh) 一种基于数据融合和bp神经网络的跌倒检测系统及方法
CN103177530A (zh) 一种基于三轴加速度的人体摔倒检测方法
Yang et al. Fall-prediction algorithm using a neural network for safety enhancement of elderly
CN113128379A (zh) 人体跌倒检测方法、计算机可读存储介质和电子装置
CN105551191B (zh) 一种跌倒检测方法
KR20210088151A (ko) 리드 타임을 고려한 낙상 감지 장치 및 그 방법
Xie et al. ART: adaptive and real-time fall detection using COTS smart watch
CN114469074A (zh) 一种跌倒预警方法、系统、设备及计算机存储介质
CN111887830B (zh) 睡眠监测方法、装置、设备及可读存储介质
Viet et al. Fall detection with smart phone sensor
CN109543762B (zh) 一种多特征融合姿态识别系统及方法
Ren et al. ALARM: A novel fall detection algorithm based on personalized threshold
CN112438721A (zh) 状态判断方法、电子设备以及计算机存储介质
Hnoohom et al. Pre-Impact and Impact Fall Detection Based on a Multimodal Sensor Using a Deep Residual Network.
Cai et al. Practical Fall Detection Algorithm based on Adaboost
CN110992649B (zh) 一种用于判断行动中跌倒状态的方法
Lee Determination of fall direction before impact using support vector machine

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: 20953928

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: 20953928

Country of ref document: EP

Kind code of ref document: A1

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 13-08-2023)