WO2017113653A1 - 一种人体运动状态的识别方法和装置 - Google Patents

一种人体运动状态的识别方法和装置 Download PDF

Info

Publication number
WO2017113653A1
WO2017113653A1 PCT/CN2016/086933 CN2016086933W WO2017113653A1 WO 2017113653 A1 WO2017113653 A1 WO 2017113653A1 CN 2016086933 W CN2016086933 W CN 2016086933W WO 2017113653 A1 WO2017113653 A1 WO 2017113653A1
Authority
WO
WIPO (PCT)
Prior art keywords
sleep
walking
state
human body
sensor
Prior art date
Application number
PCT/CN2016/086933
Other languages
English (en)
French (fr)
Inventor
李波
李娜
Original Assignee
歌尔股份有限公司
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 歌尔股份有限公司 filed Critical 歌尔股份有限公司
Priority to EP16876980.0A priority Critical patent/EP3369375A4/en
Priority to US15/541,313 priority patent/US10856777B2/en
Publication of WO2017113653A1 publication Critical patent/WO2017113653A1/zh

Links

Images

Classifications

    • 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
    • A61B5/1123—Discriminating type of movement, e.g. walking or running
    • 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
    • A61B5/1118—Determining activity level
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
    • 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
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4806—Sleep evaluation
    • A61B5/4809—Sleep detection, i.e. determining whether a subject is asleep or not
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4806—Sleep evaluation
    • A61B5/4812—Detecting sleep stages or cycles
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4806—Sleep evaluation
    • A61B5/4815—Sleep quality
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7271—Specific aspects of physiological measurement analysis
    • A61B5/7278—Artificial waveform generation or derivation, e.g. synthesising signals from measured signals
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C22/00—Measuring distance traversed on the ground by vehicles, persons, animals or other moving solid bodies, e.g. using odometers, using pedometers
    • G01C22/006—Pedometers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2562/00—Details of sensors; Constructional details of sensor housings or probes; Accessories for sensors
    • A61B2562/02—Details of sensors specially adapted for in-vivo measurements
    • A61B2562/0219—Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A61B5/681—Wristwatch-type devices
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6843—Monitoring or controlling sensor contact pressure
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7246—Details of waveform analysis using correlation, e.g. template matching or determination of similarity

Definitions

  • the present invention relates to the field of motion state recognition technology, and in particular, to a method and device for identifying a human motion state.
  • the existing motion state monitoring devices are mostly based on acceleration sensors to monitor the motion state of the human body.
  • an accelerometer-based pedometer is used to count the number of steps.
  • This kind of step-by-step scheme mainly utilizes people in the process of walking or running, and various parts of the human body are moving, thereby generating corresponding acceleration and utilizing acceleration.
  • Features such as quasi-periodicity to count the number of motion steps.
  • this step-by-step approach does not effectively distinguish between walking and running. Therefore, a motion state recognition scheme that can effectively distinguish between walking and running is desired.
  • the sleep situation will also reflect people's physical health to a certain extent. People also expect to have a certain understanding and grasp of sleep conditions, so the recording of sleep status is also necessary.
  • the current solution is mainly to understand the quality of sleep by counting the duration of people's sleep state.
  • the existing sleep statistics method introduces a problem of false sleep, such as a person who has a long time of motion during sleep, which is similar to the case where the device is not worn, and the device is placed away from the human body, which is sometimes difficult. Sleep is mistaken for sleep when the device is placed at rest. Therefore, how to solve the problem of fake sleep is another problem that needs to be solved in the statistics of sleep state.
  • the present invention provides an identification method capable of effectively distinguishing a human body motion state of walking and running, and an identification method capable of effectively solving a human body motion state in a sleep state statistical sleepiness problem.
  • an embodiment of the present invention provides a method for identifying a motion state of a human body, for effectively distinguishing between a walking state and a running state, the method comprising:
  • a three-axis acceleration sensor and a body sign sensor are disposed in the wearable device
  • the acceleration signal collected by the three-axis acceleration sensor it is determined that the human body is in a walking state, the walking step of the human body is calculated, and the walking step frequency is calculated according to the walking steps;
  • the body motion state is determined to be the running state, and the calculated walking step number is recorded as the running step number; otherwise, the human body motion state is determined to be the walking state. And count the calculated number of walking steps as the number of walking steps.
  • an embodiment of the present invention further provides an apparatus for recognizing a state of motion of a human body, which is configured to effectively distinguish between a walking state and a running state, and the identifying device includes:
  • a walking step frequency calculation unit configured to calculate a walking step of the human body according to an acceleration signal collected by the three-axis acceleration sensor when determining that the human body is in a walking state according to an acceleration signal collected by the three-axis acceleration sensor of the wearable device, and according to the walking step Calculate the walking frequency of the number;
  • a sign frequency calculation unit configured to calculate a corresponding sign frequency during the walking process according to the vital sign signal collected by the body sign sensor in the wearable device
  • a comparing unit configured to compare the calculated walking pitch frequency and the physical frequency with the step frequency threshold and the physical frequency threshold
  • a motion state identifying unit configured to determine that the human body motion state is a running state when the walking step frequency is greater than the step frequency threshold, and the body sign frequency is greater than the body sign frequency threshold value, and record the calculated walking step number as the running step number; otherwise, determine the human body
  • the motion state is the walking state, and the calculated number of walking steps is recorded as the walking step.
  • the above technical solution provided by the embodiment of the present invention is based on different characteristics of a step frequency and a human body sign of a running state and a walking state, and a plurality of sensors, such as an acceleration sensor and a human body sign sensor, are used in the wearable device, and the acceleration sensor and the human body are utilized.
  • the vital signs sensor collects the acceleration signal and the human body sign signal during the movement of the human body respectively, calculates the stride frequency and the corresponding human body sign frequency based on the acceleration signal and the human body sign signal respectively, and distinguishes the walking state and running by combining the stride frequency and the human body sign frequency. status.
  • the present invention provides a method for recognizing a state of motion of a human body, which is used to effectively solve the problem of slumber in sleep state statistics, and the method includes:
  • a three-axis acceleration sensor and a body sign sensor are disposed in the wearable device
  • the present invention provides an apparatus for recognizing a state of motion of a human body for effectively solving a problem of slumbering in sleep state statistics, the identification apparatus comprising:
  • a momentary movement detecting unit configured to detect a momentary movement of the human body according to an acceleration signal collected by the triaxial acceleration sensor of the wearable device;
  • a wearing determining unit configured to determine, according to the vital sign signal collected by the human body vitality sensor of the wearable device, whether to wear the wearable device;
  • a motion state identifying unit configured to determine that the body motion state is a sleep state when the number of instantaneous motions detected during the sleep state statistical time is less than or equal to the set first number threshold, and determine that the human body wears the wearable device; , to determine the state of motion of the human body is awake.
  • the foregoing technical solution provided by the embodiment of the present invention provides a plurality of sensors, such as an acceleration sensor and a human body sign sensor, in the wearable device, and respectively uses an acceleration sensor and a body sign sensor to separately collect an acceleration signal and a body sign signal during the movement of the human body, and combine the acceleration signal.
  • the sign signal identifies the sleep state of the human body to improve the accuracy of the recognition result, and avoids counting the state in which the wearable device is placed away from the human body to be in a sleep state.
  • Embodiment 1 is a flow chart of a method for identifying a motion state of a human body according to Embodiment 1;
  • FIG. 2 is a schematic diagram of an acceleration signal generated by a three-axis acceleration sensor in three directions during the running according to the first embodiment
  • FIG. 3 is a flow chart showing a method for identifying a running state and a walking state according to Embodiment 1;
  • FIG. 4 is a flow chart of a method for identifying a motion state of a human body according to Embodiment 2;
  • FIG. 5 is a schematic diagram of an acceleration signal generated by a three-axis acceleration sensor in three directions during a sleep process according to Embodiment 2;
  • FIG. 6 is a flowchart of a sleep state identification method provided by Embodiment 2;
  • FIG. 7 is a schematic diagram of an apparatus for identifying a motion state of a human body according to Embodiment 3;
  • FIG. 8 is a schematic diagram of an apparatus for identifying a motion state of a human body according to Embodiment 4.
  • Embodiment 1 is a diagrammatic representation of Embodiment 1:
  • This embodiment adopts a multi-sensor combined with human body features (such as heart rate detection) to achieve more accurate effects of walking and running state of the human body.
  • the present embodiment uses a wearable device with an acceleration sensor and a vital sign sensor to monitor the movement of the human body in real time, and combines the motion feature and the biological feature to identify the walking state of the human body.
  • the walking state of the embodiment includes a walking state and a running state, and the walking steps include the number of walking steps and the number of running steps.
  • FIG. 1 is a flowchart of a method for identifying a motion state of a human body according to the embodiment. As shown in FIG. 1 , the method in FIG. 1 includes:
  • S110 a three-axis acceleration sensor and a human body sign sensor are disposed in the wearable device.
  • the present embodiment preferably uses a heart rate sensor to acquire a heart rate characteristic of the human body.
  • S120 Determine, according to the acceleration signal collected by the three-axis acceleration sensor, that the human body is in a walking state, calculate a walking step of the human body, and calculate a walking step frequency according to the walking step.
  • S130 Calculate a corresponding physical frequency of the walking process according to the vital sign signal collected by the human body sign sensor.
  • the human body sign sensor set in the wearable device is a heart rate sensor
  • the corresponding body sign frequency during the walking process is calculated according to the vital sign signal collected by the body sign sensor: the running process is calculated according to the periodic fluctuation of the heart rate signal collected by the heart rate sensor. Heart rate.
  • S140 comparing the calculated walking frequency and the body frequency with the step frequency threshold and the body frequency threshold, respectively, determining that the human body motion state is a running state when the walking frequency is greater than the step frequency threshold, and the body frequency is greater than the body frequency threshold.
  • the calculated number of walking steps is recorded as the running step number; otherwise, the human body motion state is determined to be the walking state, and the calculated walking step number is recorded as the walking step number.
  • the specific content of the above steps S120 to S140 may be performed by a wearable device such as a smart watch or a smart bracelet.
  • a three-axis acceleration sensor and a human body sign sensor are set, and the acceleration signal of the human body motion is collected by using the three-axis acceleration sensor and the human body sign sensor respectively.
  • the human body sign signals based on the acceleration signal and the human body sign signal, respectively calculate the stride frequency and the corresponding human body sign frequency, and combine the stride frequency and the human body sign frequency to distinguish the walking state and the running state.
  • FIG. 2 shows an acceleration signal generated by three-axis acceleration sensors in three directions during walking, and ax/g, ay/g, and az/g in FIG. 2 are respectively three-axis acceleration sensors.
  • the normalized acceleration signal produced on the x-axis, the y-axis, and the z-axis, g represents the gravitational acceleration. It can be seen from Fig. 2 that during the walking or running of the person, the acceleration signal collected by the three-axis acceleration sensor exhibits obvious periodic characteristics on at least one coordinate axis. Therefore, the number of walking steps can be determined by counting the extreme points of the acceleration signal, such as a maximum point (or a minimum point) corresponding to each step of the left leg and the right leg, that is, one extreme point corresponds to two steps. .
  • FIG. 3 is a flowchart of a method for identifying a running state and a walking state according to the embodiment.
  • FIG. 3 illustrates a method for identifying a running state and a walking state, including:
  • S321 Acquire a heart rate signal having a set sampling duration from an output of the heart rate sensor of the wearable device.
  • the acceleration signal of step S311 and the heart rate signal of step S321 correspond to the same motion state, and the sampling duration is the same.
  • the sampling duration in this embodiment should not be too long, preferably less than 5 minutes. If the sampling duration is too long, the walking state within the sampling duration may include the walking state and the running state, which is not conducive to distinguishing between the walking state and the running state.
  • the acceleration signal output by the three-axis acceleration sensor usually contains a DC component, and the presence of the DC component interferes with the analysis of the acceleration signal. Therefore, the present embodiment filters out the DC component in the acceleration signal by high-pass filtering.
  • the high-pass filtered acceleration signal may also contain multiple frequency components corresponding to different body rhythms, such as fundamental frequency components, frequency doubling components, and other high frequency components.
  • fundamental frequency component is associated with the most basic rhythm, and the quasi-periodicity of the signal is more accurate according to the fundamental frequency component.
  • the high-pass filtered acceleration signal is attenuated, that is, the signal is attenuated by a filter that attenuates the signal energy from a low frequency to a high frequency to suppress high-frequency components in the acceleration signal, thereby The fundamental frequency component in the acceleration signal is highlighted, and the error of the fundamental frequency to be solved is reduced.
  • the autocorrelation function ⁇ ( ⁇ ) of the attenuated signal is obtained by the following equation, and the reciprocal of the ⁇ value corresponding to the maximum value of the autocorrelation function ⁇ ( ⁇ ) is the fundamental frequency of the signal.
  • N is the predetermined length of the signal
  • 1 ⁇ n ⁇ N is the delay time
  • ⁇ ( ⁇ ) is the normalized autocorrelation function of the signal.
  • the fundamental frequency obtained by the fundamental frequency detection is used as a cutoff frequency to set a low pass or band pass filter, and the low pass or band pass filter is used to perform low pass or band pass filtering on the corresponding high pass filtered acceleration signal.
  • a smoother signal can be obtained, which facilitates accurate statistics of the extreme points of the acceleration signal.
  • This embodiment can perform extreme point removal processing based on the time interval of adjacent extreme points. Specifically, the extreme point of the acceleration signal extreme point and the time interval of the previous acceleration signal extreme point is less than a predetermined threshold. Wherein the predetermined threshold is much smaller than the period of the fundamental component of the uniaxial acceleration signal.
  • the extreme point removal method among the extreme points that are closer to each group, only the leftmost extreme point is retained, and the remaining extreme points are regarded as interference extreme points and removed. In this way, the interference extreme point in the extreme point of the acceleration signal is removed by the time interval between the extreme points of the acceleration signal.
  • interference extreme points for example, by using an acceleration signal extreme point in which the amplitude of the acceleration signal extreme points in each set of time intervals is less than a predetermined threshold is not the maximum.
  • the extreme point removal method among the extreme points that are closer to each group, only the extreme value of the acceleration signal extreme point is retained, and the remaining extreme points are regarded as the interference extreme point and are removed.
  • the number of extreme points of the acceleration signal after removing the interference extreme point in the three uniaxial acceleration signals needs to be counted; and the number of walking steps is determined according to the number of effective extreme points corresponding to each uniaxial acceleration signal.
  • the number of walking steps can be determined by the following method: if the energy of each uniaxial acceleration signal is not much different, the number of extreme points of the acceleration signal after removing the interference extreme point corresponding to each axis can be averaged to The average number is used as the walking step obtained by the current step counting process; if the energy of each uniaxial acceleration signal has a large difference, the acceleration after removing the interference extreme point corresponding to the uniaxial acceleration signal with the largest energy may be obtained. The number of signal extreme points determines the number of walking steps obtained during the current round of the step.
  • step S330 comparing the walking step frequency and the heart rate with the step frequency threshold and the heart rate threshold respectively, determining whether the walking step frequency is greater than the step frequency threshold, and whether the heart rate is greater than the heart rate threshold; if the walking step frequency is greater than the step frequency threshold, and the heart rate is greater than the heart rate threshold Then, step S340 is performed; otherwise, step S350 is performed.
  • the step frequency threshold and the heart rate threshold in this embodiment may be set according to data statistics.
  • the step frequency threshold may be set to 2.5 steps/second; and the heart rate threshold is set to 120 times/minute.
  • the resting heart rate of a person is in the range of 60 to 100 beats per minute, while when the person is exercising, The heart rate will increase. Since the exercise is relatively strong in walking, the heart rate during running is higher than that when walking. Therefore, the heart rate threshold can be set according to whether the obtained heart rate satisfies the heart rate range at the time of running, for example, the heart rate threshold can be set to 120 times/minute. Since heart rate is related to factors such as age, physical state and resting heart rate, you can refer to these parameters when setting the heart rate threshold, so that the distinction between walking and running will be more accurate.
  • S340 Determine the current human exercise state as the running state, and record the calculated walking step number as the running step number.
  • S350 Determine the current human motion state as a walking state, and record the calculated walking step number as the walking step.
  • Embodiment 2 is a diagrammatic representation of Embodiment 1:
  • This embodiment adopts a multi-sensor combined with human body features (such as heart rate detection) to achieve a more accurate distinction between the human sleep state and the awake state.
  • the sleep state is divided into a light sleep state and a deep sleep state. Since the exercise energy corresponding to different degrees of sleep state is different, and the wearable device is placed at rest when the wearable device is in a sleep state, the human body sign sensor output is The signs of the human body are not the same. Therefore, the present embodiment utilizes a wearable device having an acceleration sensor and a human body sign sensor to recognize the sleep state of the human body in combination with the motion characteristics and the human body signs.
  • FIG. 4 is a flowchart of a method for identifying a motion state of a human body according to the embodiment. As shown in FIG. 4, the method in FIG. 4 includes:
  • S410 a three-axis acceleration sensor and a human body sign sensor are disposed in the wearable device.
  • This embodiment preferably utilizes a heart rate sensor to acquire heart rate signs, or utilizes a capacitive sensor to acquire capacitive signs of human skin.
  • the acceleration signal of the embodiment has the characteristics as shown in FIG. 5, and FIG. 5 is a schematic diagram of the acceleration signal generated by the triaxial acceleration sensor in three directions during the sleep process, as shown in FIG. 5, during the sleep process, Most of the time the acceleration signal is very small and gentle, and the time during the sleep process (such as the transient change such as turning over during the 600-610 second period in Figure 5) accounts for a small proportion of the total sleep time.
  • the present embodiment quantitatively analyzes the quality of sleep by analyzing the number of occurrences of transient disturbances in the statistical state of the sleep state.
  • S430 Determine whether to wear the wearable device according to the vital sign signal collected by the human body sign sensor.
  • determining whether to wear the wearable device according to the vital sign signal collected by the human body sign sensor in this step is specifically: determining whether to wear the above according to the energy or amplitude of the heart rate signal collected by the heart rate sensor. Wearable device.
  • determining whether to wear the wearable device according to the vital sign signal collected by the human body sign sensor in this step is specifically: determining whether to wear the above according to the energy or amplitude of the capacitance signal collected by the capacitive sensor. Wearable device.
  • the acceleration signal and the body sign sensor are respectively used to collect the acceleration signal and the body sign signal during the movement of the human body, and the acceleration signal and the body sign signal are combined to identify the sleep state of the human body to improve the accuracy of the recognition result, and the wearable device is prevented from being detached from the human body.
  • the situation of static placement is counted in the sleep state.
  • detecting the instantaneous change of the human body according to the acceleration signal collected by the triaxial acceleration sensor includes:
  • the plurality of instantaneous energies are respectively compared with the first energy threshold, and when the plurality of instantaneous energies are less than the first energy threshold, determining that no instantaneous change occurs in the unit time; otherwise, determining that the unit time has a transient transaction.
  • determining that the human body motion state is the sleep state includes:
  • the sleep state statistical time is evenly divided into a plurality of exercise amount statistical periods, and the number of instantaneous movements occurring in each exercise amount statistical period is counted;
  • the body motion state is a sleep state.
  • the motion state recognition method further includes:
  • the sleep state statistical time is divided into a plurality of deep and shallow sleep statistical periods according to certain conditions, and the number of instantaneous transaction occurrences in each of the deep and shallow sleep statistical periods is counted; the length of the deep and shallow sleep statistical period is greater than the exercise amount statistical period;
  • the number of instantaneous movements in each of the shallow and light sleep statistical periods is sequentially compared with the set second number threshold; since the person is in a deep sleep state, the number of movements such as turning over, scratching, etc. is less than that of the person being in a shallow sleep state, thereby Further subdividing the deep sleep state and the light sleep state in the sleep state, the second number threshold is required to be smaller than the first number threshold;
  • the sleep state in the period is a deep sleep state; otherwise, the sleep state in the period is determined to be a mild sleep state.
  • the sleep state statistical time is divided into a plurality of deep and shallow sleep statistical periods according to certain conditions, and the acceleration signal energy after the filtering process in each of the deep and light sleep statistical periods is calculated; the length of the deep and shallow sleep statistical period is greater than the exercise amount statistical period;
  • the acceleration signal energy of each of the shallow and light sleep statistical periods is sequentially compared with the set second energy threshold; since the person is in a deep sleep state, the number of movements such as turning over and scratching is less than that of the person being in a shallow sleep state, thereby further Subdividing the deep sleep state and the light sleep state in the sleep state, the second energy threshold is required to be smaller than the first energy threshold;
  • the above-mentioned sleep state statistical time is divided into a plurality of deep and shallow sleep statistical periods, and a static division method may be adopted, for example, the sleep state statistical time is evenly divided into a plurality of deep and shallow sleep statistical periods according to a fixed duration; and a dynamic division method may also be adopted, such as During the sleep state statistical time, the window of the set length is moved according to the set step size, and the moving overlap window is divided into a plurality of deep and shallow sleep statistical periods.
  • FIG. 6 is a flowchart of a method for identifying a sleep state according to an embodiment of the present invention. As shown in FIG. 6, the method for identifying a sleep state includes:
  • the acceleration signal during most of the sleep process is very small and gentle, and the time of turning over, scratching, etc. during sleep is a small proportion of the total sleep time. Therefore, when analyzing the acceleration signal generated by sleep, the sampling length of the signal should be large to ensure that the transaction signal during sleep can be collected.
  • S621 Acquire a human body sign signal from an output of a human body vitality sensor of the wearable device.
  • acceleration signal of step S611 and the human body sign signal of step S621 correspond to the same motion state, and the sampling duration is the same.
  • the body sign sensor in this embodiment is preferably a heart rate sensor or a capacitance sensor.
  • the sleep state statistical time is evenly divided into a plurality of exercise amount statistical periods, and the number of instantaneous transaction occurrences in each exercise amount statistical period is counted.
  • the number of instantaneous transactions that occur during each statistical period of the exercise amount is calculated by the following method:
  • the acceleration signal is filtered to filter out the DC signal.
  • the acceleration signal output by the triaxial acceleration sensor usually contains a direct current component, and the presence of the direct current component interferes with the analysis of the acceleration signal. Therefore, this embodiment The DC component of the acceleration signal is filtered by high pass filtering.
  • the instantaneous energy is the acceleration signal energy of each sub-period after the unit time is divided.
  • the unit time in this embodiment should not be too long, and should not be too short.
  • the unit time is too long to be convenient for statistical changes. If it is too short, it is easy to count one instantaneous transaction behavior as two or count as multiple; the unit time is preferably 1 second. bell.
  • the unit time of 1 second is equally divided into two sub-periods, and the duration of each sub-period is 0.5 seconds, and the instantaneous energy STD of each sub-period is calculated by the following formula:
  • N is the length of the acceleration signal
  • N 0.5
  • s 0 is the average value of the acceleration signal.
  • the plurality of instantaneous energies are respectively compared with the first energy threshold.
  • the plurality of instantaneous energies are less than the first energy threshold, it is determined that no instantaneous change occurs in the unit time; otherwise, the instantaneous change occurs in the unit time.
  • the first energy threshold of the present embodiment needs to be able to distinguish between transient and respiratory motions that characterize turning over, scratching.
  • the first energy threshold is a value close to zero.
  • the instantaneous energy of one of the instantaneous energies of the two sub-periods is greater than the first energy threshold, it is determined that the instantaneous transaction occurs within the unit time.
  • the sleep state statistical time is evenly divided into a plurality of exercise amount statistical periods, and the number of instantaneous transaction occurrences in each exercise amount statistical period is counted.
  • the sleep state statistical time in the embodiment ranges from 20 minutes to 60 minutes, preferably 30 minutes; the duration of the exercise amount statistical period ranges from 1 minute to 3 minutes.
  • this step specifically divides the sleep state statistical time 30 minutes into 30 exercise amount statistical periods according to the length of 1 minute, and counts each time. The number of transient changes that occur within minutes.
  • the body sign sensor provided in the wearable device is a heart rate sensor, whether the wearable device is worn according to the energy or amplitude of the heart rate signal; when the body sign sensor provided in the wearable device is a capacitive sensor, according to the capacitance signal The energy or amplitude determines whether the wearable device is worn.
  • step S630 determining whether the number of instantaneous transactions of each exercise amount statistical period is less than or equal to the first number threshold, and determining whether to wear the wearable device; if each motion quantity statistical period is instantaneously changed If the number is less than or equal to the first number threshold, and it is determined that the human body wears the wearable device, step S640 is performed; otherwise, step S650 is performed.
  • the first number threshold in this embodiment ranges from 0 to 60, preferably 30.
  • the first number threshold in this embodiment ranges from 0 to 80, preferably 30.
  • step S640 determining that the current human motion state is a sleep state, and performing step S660.
  • S660 The sleep state statistical time is divided into a plurality of deep and shallow sleep statistical periods, and the number of instantaneous transaction occurrences in each time period is counted; the length of the dark and shallow sleep statistical period is greater than the exercise amount statistical period.
  • the length of the deep and light sleep statistical period in this embodiment ranges from 4 minutes to 10 minutes.
  • the static division method can be used to divide the dark and shallow sleep statistical time into multiple deep and shallow sleep statistical periods, and the length of the deep and shallow sleep statistical period is 5 minutes.
  • the sleep state statistical time 30 minutes can be evenly divided into 6 deep and shallow sleep statistical periods according to the length of 5 minutes.
  • the dynamic method of windowing movement may also be used to divide the sleep state statistical time. If the window length of the window is 5 minutes and the moving step is 1 minute, the sleep state statistical time may be 5 minutes according to 5 minutes. The length of time is evenly divided into 29 deep and light sleep statistical periods.
  • step S670 sequentially comparing the number of instantaneous movements of each of the deep and light sleep statistical periods with the set second number of thresholds, when the number of instantaneous changes in the dark and light sleep statistical period is less than or equal to the second number of thresholds, step S680; Otherwise, step S690 is performed.
  • the second number threshold in this step is smaller than the first number threshold, and the second number threshold is in the range of 0-15.
  • step S660 of the embodiment the acceleration signal energy after the filtering process in each of the deep and light sleep statistical periods may also be calculated
  • the step S670 is performed to: compare the acceleration signal energy of each of the deep and light sleep statistical periods with the second energy threshold in sequence, and when the acceleration signal energy of the deep and light sleep statistical period is less than or equal to the second energy threshold, perform step S680; otherwise Step S690 is performed.
  • the second energy threshold is less than the first energy threshold.
  • Embodiment 3 is a diagrammatic representation of Embodiment 3
  • the present embodiment provides an apparatus for recognizing a human body motion state.
  • FIG. 7 is a schematic diagram of a device for recognizing a state of motion of a human body according to the embodiment of the present invention. As shown in FIG. 7 , the device is applicable to a wearable device having multiple sensors, and may be a specific module in the wearable device.
  • the identification device includes:
  • Walking step frequency calculation unit 71 for collecting on a three-axis acceleration sensor according to the wearable device
  • the acceleration signal determines that the human body is in a walking state
  • the number of walking steps of the human body is calculated according to the acceleration signal collected by the three-axis acceleration sensor
  • the walking step frequency is calculated according to the walking step.
  • the sign frequency calculation unit 72 is configured to calculate a corresponding sign frequency during the walking process according to the vital sign signal collected by the body sign sensor of the wearable device.
  • the comparing unit 73 is configured to compare the calculated walking step frequency and the body frequency with the step frequency threshold and the body frequency threshold, respectively.
  • the motion state identifying unit 74 is configured to determine that the body motion state is the running state when the walking step frequency is greater than the step frequency threshold, and the body sign frequency is greater than the body sign frequency threshold, and record the calculated walking step number as the running step number; otherwise, determine The state of motion of the human body is the walking state, and the calculated number of walking steps is recorded as the number of walking steps.
  • the walking step frequency calculation unit 71 includes:
  • a filtering module for filtering the acceleration signal
  • An interference extreme point removal module is configured to remove an interference extreme point in the acceleration signal after the filtering process
  • the extreme point counting module is configured to calculate the number of effective extreme points in the acceleration signal after removing the interference extreme point, and the number of the effective extreme points is the walking step;
  • the step frequency calculation module is configured to count the number of signal sampling points between two adjacent steps, and multiply the number of signal sampling points by the signal sampling time to obtain a walking period; and use the reciprocal of the walking period to obtain a walking step frequency.
  • the vital frequency calculation unit 72 is specifically configured to calculate the heart rate during the walking according to the periodic fluctuation of the heart rate signal collected by the heart rate sensor.
  • Embodiment 4 is a diagrammatic representation of Embodiment 4:
  • the present embodiment provides an apparatus for recognizing a human body motion state.
  • FIG. 8 is a schematic diagram of a device for recognizing a state of motion of a human body according to an embodiment of the present invention, which is applicable to a wearable device having multiple sensors, and may be a specific module in the wearable device, and the identification device includes:
  • the instantaneous change detecting unit 81 is configured to detect the instantaneous change of the human body according to the acceleration signal collected by the triaxial acceleration sensor of the wearable device.
  • the wearing determining unit 82 is configured to determine whether to wear the wearable device according to the vital sign signal collected by the human body vitals sensor in the wearable device.
  • the body sign sensor may be a heart rate sensor or a capacitive sensor; when the body sign sensor is a heart rate sensor, the wearing determining unit is specifically configured to determine whether to wear the wearable device according to the energy or amplitude of the heart rate signal collected by the heart rate sensor; In the case of a capacitive sensor, the wearing determining unit is specifically configured to determine whether to wear the wearable device according to the energy or amplitude of the capacitive signal collected by the capacitive sensor.
  • the motion state identifying unit 83 is configured to determine that the human body motion state is a sleep state when the number of transient motions detected in the sleep state statistical time is less than or equal to the set first number threshold value, and when the human body wears the wearable device Otherwise, it is determined that the human body motion state is awake.
  • the instantaneous transaction detecting unit 81 includes:
  • a filtering module configured to filter and process the acceleration signal collected by the three-axis acceleration sensor, and filter the DC signal
  • the instantaneous energy calculation module is configured to calculate a plurality of instantaneous energy per unit time of the filtered acceleration signal, wherein the instantaneous energy is an acceleration signal energy of each sub-period after equalizing the unit time;
  • the instantaneous energy judgment processing module is configured to compare the plurality of instantaneous energies with the first energy threshold respectively, and when the plurality of instantaneous energies are less than the first energy threshold, determine that no instantaneous change occurs in the unit time; otherwise, determine the Instantaneous changes occur per unit time.
  • the motion state recognizing unit 83 includes:
  • the instantaneous transaction number statistics module is configured to evenly divide the sleep state statistical time into a plurality of exercise quantity statistical periods, and count the number of instantaneous movements occurring in each exercise quantity statistical period;
  • the instantaneous transaction number comparison judging module is configured to compare the instantaneous transaction number of each exercise quantity statistical period with the set first number threshold, and the instantaneous transaction number in each exercise quantity statistical period is less than or equal to the first
  • the threshold value is determined, and when the human body wears the wearable device, it is determined that the human body motion state is a sleep state.
  • the motion state recognition unit 83 further includes a depth and light sleep determination module
  • the depth and sleep judgment module is configured to divide the sleep state statistical time into a plurality of deep and shallow sleep statistical periods, and count the number of instantaneous movements occurring in each of the deep and shallow sleep statistical periods; and sequentially change the instantaneous change of each of the deep and shallow sleep statistical periods Comparing the number with the set second number threshold, determining that the sleep state in the deep sleep statistics period is a deep sleep state when the number of transient movements in the deep sleep statistics period is less than or equal to the second number threshold; otherwise determining the The sleep state in the dark sleep statistics period is a light sleep state; wherein, the length of the depth sleep statistics period is greater than the exercise amount statistics period, and the second number threshold is less than the first number threshold;
  • the depth sleep determination module is configured to divide the sleep state statistical time into a plurality of deep and shallow sleep statistical periods, and calculate the acceleration signal energy after the filtering process in each of the deep and light sleep statistical periods; and sequentially accelerate the acceleration time of each of the deep and shallow sleep statistical periods
  • the signal energy is compared with the set second energy threshold.
  • the acceleration signal energy of the deep sleep statistics period is less than or equal to the second energy threshold, determining that the sleep state in the deep sleep statistical period is a deep sleep state; otherwise determining the The sleep state in the dark sleep statistics period is a mild sleep state; wherein the length of the depth sleep statistics period is greater than the exercise amount statistics period, and the second energy threshold is less than the first energy threshold.
  • the present invention provides an identification method and apparatus capable of effectively distinguishing a human body motion state of walking and running, and the technical solution of the present invention is based on a step frequency of a person in a running state and a walking state.
  • the acceleration sensor and the human body sign sensor are arranged in the wearable device, and the acceleration signal and the human body sign signal are respectively collected by the acceleration sensor and the human body sign sensor, and are respectively calculated based on the acceleration signal and the human body sign signal.
  • the walking frequency and the corresponding human body sign frequency during the movement are combined with the walking frequency and the human body frequency to distinguish the walking state and the running state.
  • the present invention also provides a method and apparatus for identifying a human motion state that can effectively solve the problem of slumber in sleep state statistics.
  • the technical solution of the present invention utilizes an acceleration sensor and a human body sensor in a wearable device.
  • the acceleration sensor and the body sign sensor respectively collect the acceleration signal and the body sign signal during the movement of the human body, and combine the acceleration signal and the body sign signal to recognize the sleep state of the human body to improve the accuracy of the recognition result, and avoid the situation that the wearable device is detached from the human body. Counted to sleep state.
  • the words “first” and “second” are used to distinguish the same items or similar items whose functions and functions are substantially the same. Personnel can understand that the words “first” and “second” do not limit the quantity and order of execution.

Abstract

一种人体运动状态的识别方法和装置。所述方法包括：在可穿戴设备中设置三轴加速度传感器和人体体征传感器(S110)；根据三轴加速度传感器采集的加速度信号判断人体处于行走状态，计算人体的行走步数，并根据行走步数计算行走步频(S120)；根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率(S130)；将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较，在行走步频大于步频阈值，且体征频率大于体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则确定人体运动状态为走路状态，并将计算的行走步数记为走路步数(S140)。本方法和装置能够有效地区分行走状态中的走路状态和跑步状态。

Description

一种人体运动状态的识别方法和装置 技术领域
本发明涉及运动状态识别技术领域，特别涉及一种人体运动状态的识别方法和装置。
发明背景
随着社会经济的不断发展、物质生活水平的日渐提高，人们越来越关注自身健康，为自己制定各种运动方案来健身，并根据运动情况分析自身的健康状态。因此，出现了各种用于监测运动状态的设备。
现有运动状态监测设备多是基于加速度传感器来监测人体的运动状态。如利用基于加速度传感器的计步器来统计步数，该种计步方案主要是利用人在走路或跑步的过程中，人体的多处部位都在运动，从而会产生相应的加速度，利用加速度的准周期性等特征来统计运动步数。但是这种计步方案无法有效地区分走路和跑步。因此，人们期望一种能够有效地区分走路和跑步的运动状态识别方案。
此外，睡眠情况也会在一定程度上反映人们的身体健康状况，人们也期望能够对睡眠情况有一定的了解和把握，因此对睡眠状况的记录也是必要的。现有方案，主要是通过统计人们的睡眠状态所持续的时间来了解其睡眠质量。但是现有睡眠统计方法会引入假睡的问题，比如人在睡眠过程中会存在较长时间一动不动的情况，这种情况类似于未佩戴检测设备，设备脱离于人体静止地放置，从而有时难眠将设备静止放置的情况误判为睡眠。因此，如何解决假睡问题是睡眠状态统计中需要解决的另一个问题。
发明内容
鉴于上述问题，本发明提供一种能够有效地区分走路和跑步的人体运动状态的识别方法，以及一种能够有效地解决睡眠状态统计中假睡问题的人体运动状态的识别方法。
为达到上述目的，本发明的技术方案是这样实现的：
一方面，本发明实施例提供了一种人体运动状态的识别方法，用于有效地区分走路状态和跑步状态，该方法包括：
在可穿戴设备中设置三轴加速度传感器和人体体征传感器；
根据三轴加速度传感器采集的加速度信号判定人体处于行走状态，计算人体的行走步数，并根据行走步数计算行走步频；
根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率；
将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比 较，在行走步频大于步频阈值，且体征频率大于体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
另一方面，本发明实施例还提供了一种人体运动状态的识别装置，用于有效地区分走路状态和跑步状态，该识别装置包括：
行走步频计算单元，用于在根据可穿戴设备的三轴加速度传感器采集的加速度信号判定人体处于行走状态时，根据该三轴加速度传感器采集的加速度信号计算人体的行走步数，并根据行走步数计算行走步频；
体征频率计算单元，用于根据可穿戴设备中的人体体征传感器采集的体征信号计算行走过程中相应的体征频率；
比较单元，用于将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较；
运动状态识别单元，用于在行走步频大于步频阈值，且体征频率大于体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
本发明实施例提供的上述技术方案基于人在跑步状态和走路状态的步频和人体体征不同的特点，在可穿戴设备中设置多个传感器，如加速度传感器和人体体征传感器，利用加速度传感器和人体体征传感器分别采集人体运动过程中的加速度信号和人体体征信号，基于加速度信号和人体体征信号分别计算运动过程中的步频和相应的人体体征频率，结合步频和人体体征频率区分走路状态和跑步状态。
一方面，本发明提供了一种人体运动状态的识别方法，用于有效地解决睡眠状态统计中假睡问题，该方法包括：
在可穿戴设备中设置三轴加速度传感器和人体体征传感器；
根据三轴加速度传感器采集的加速度信号检测人体的瞬时异动；
根据人体体征传感器采集的体征信号判断是否佩戴可穿戴设备；
当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴上述可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
另一方面，本发明提供了一种人体运动状态的识别装置，用于有效地解决睡眠状态统计中假睡问题，该识别装置包括：
瞬时异动检测单元，用于根据可穿戴设备的三轴加速度传感器采集的加速度信号检测人体的瞬时异动；
佩戴判断单元，用于根据可穿戴设备的人体体征传感器采集的体征信号判断是否佩戴可穿戴设备；
运动状态识别单元，用于当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
本发明实施例提供的上述技术方案通过在可穿戴设备中设置多个传感器，如加速度传感器和人体体征传感器，利用加速度传感器和体征传感器分别采集人体运动过程中的加速度信号和体征信号，结合加速度信号和体征信号识别人体的睡眠状态来提高识别结果的准确性，避免了将可穿戴设备脱离于人体静止放置的情况统计到睡眠状态中。
附图简要说明
附图用来提供对本发明的进一步理解，并且构成说明书的一部分，与本发明实施例一起用于解释本发明，并不构成对本发明的限制。在附图中：
图1为实施例一提供的人体运动状态的识别方法流程图；
图2为实施例一提供的走跑过程中三轴加速度传感器在三个方向上产生的加速度信号示意图；
图3为实施例一提供的跑步状态和走路状态的识别方法流程图；
图4为实施例二提供的人体运动状态的识别方法流程图；
图5为实施例二提供的睡眠过程中三轴加速度传感器在三个方向上产生的加速度信号示意图；
图6为实施例二提供的睡眠状态识别方法流程图；
图7为实施例三提供的人体运动状态的识别装置示意图；
图8为实施例四提供的人体运动状态的识别装置示意图。
实施本发明的方式
为使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明实施方式作进一步地详细描述。
实施例一：
本实施例采用多传感器结合人体特征(如心率检测)达到更精确获得人体走路和跑步运动状态的效果。
由于不同的行走状态，对应的步频和体征不同，因此本实施例利用具有加速度传感器和体征传感器的可穿戴设备实时监测人体的运动，结合运动特征和生物特征来识别人体的行走状态。本实施例的行走状态包括走路状态和跑步状态，行走步数包括走路步数和跑步步数。
图1为本实施例提供的人体运动状态的识别方法流程图，如图1所示，图1中的方法包括：
S110，在可穿戴设备中设置三轴加速度传感器和人体体征传感器。
由于人体在走路和跑步时的心率会有明显的不同，因此本实施例优选地利用心率传感器获取人体的心率特征。
S120，根据三轴加速度传感器采集的加速度信号判定人体处于行走状态，计算人体的行走步数，并根据行走步数计算行走步频。
S130，根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率。
当可穿戴设备中设置的人体体征传感器为心率传感器，本步骤中根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率具体为：根据心率传感器采集的心率信号的周期性波动计算行走过程中的心率。
S140，将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较，在行走步频大于步频阈值，且体征频率大于体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
上述步骤S120至S140的具体内容可以由可穿戴设备(如智能手表、智能手环)执行。
本实施例基于人在跑步状态和走路状态的步频和人体体征不同的特点，设置三轴加速度传感器和人体体征传感器，利用三轴加速度传感器和人体体征传感器分别采集人体运动过程中的加速度信号和人体体征信号，基于加速度信号和人体体征信号分别计算运动过程中的步频和相应的人体体征频率，结合步频和人体体征频率区分走路状态和跑步状态。
为了更加详细地介绍本实施例识别人体运动状态的方法，本实施例通过一具体实现方案来说明。
由于人在行走过程中，身体的各部位都在运动，从而会产生一定的加速度，且由于人走路或跑步具有一定的周期性，因而由此产生的加速度也具有一定的周期性。如图2所示，图2示出了行走过程中三轴加速度传感器在三个方向上产生的加速度信号，图2中的ax/g、ay/g、az/g分别是三轴加速度传感器在x轴、y轴和z轴上产生的归一化的加速度信号，g表示重力加速度。从图2中可以看出，人在走路或跑步的行走过程中，三轴加速度传感器采集到的加速度信号至少在一个坐标轴上表现出明显的周期特征。因此，可以通过统计加速度信号的极值点来确定行走步数，如一个极大值点(或一个极小值点)对应左腿和右腿各迈一步，即一个极值点对应行走两步。
图3为本实施例提供的跑步状态和走路状态的识别方法流程图，图3所示，识别跑步状态和走路状态的方法包括：
S311，从可穿戴设备的三轴加速度传感器的输出中获取具有设定采样时长的加速度信号。
S321，从可穿戴设备的心率传感器的输出中获取具有设定采样时长的心率信号。
需要说明的是，步骤S311的加速度信号和步骤S321中的心率信号对应为同一运动状态，且采样时长相同。本实施例中采样时长不宜过长，优选地小于5分钟。若采样时长过长，采样时长内的行走状态可能即包括走路状态又包括跑步状态，不利于对走路状态和跑步状态的区分。
S312，对加速度信号进行滤波处理。
从图2中可以看出，三轴加速度传感器输出的加速度信号通常会包含直流分量，而该直流分量的存在会对加速度信号的分析产生干扰。因此，本实施例通过高通滤波来滤除加速度信号中的直流分量。
此外，高通滤波后的加速度信号也可能会包含与不同的身体律动相对应的多种频率分量，如基频分量、倍频分量以及其它高频分量。其中，基频分量与最基本的律动关联，根据基频分量判断信号的准周期性会更准确。为了能够获得只有基频分量的加速度信号，需要滤除加速度信号中的高频分量。而为了滤除高频分量，需要检测出基频分量的频率，以便构造合适的滤波器滤除基频分量之外的高频分量。
基频检测的方法很多，可以使用语音信号基音检测中常用的自相关函数方法，倒谱方法，线性预测编码方法，平均幅度差函数方法等经典方法。优选地，可以使用自相关函数方法。
具体说，首先，对于高通滤波后的加速度信号进行衰减处理，即利用对信号能量的衰减从低频到高频递增的滤波器对该信号进行衰减处理，以抑制加速度信号中的高频分量，从而突出加速度信号中的基频分量，减小待求解基频的误差。
然后，由下述公式求出衰减后的信号的自相关函数ρ(τ)，该自相关函数ρ(τ)的最大值所对应的τ值的倒数即为信号的基频。
Figure PCTCN2016086933-appb-000001
其中，a(n)为该信号的第n个值，N为该信号的预定长度，且1≤n≤N，τ为延迟时间，ρ(τ)为该信号的归一化自相关函数。
最后，利用基频检测所获得的基频作为截止频率设置低通或带通滤波器，并利用该低通或带通滤波器对相应的高通滤波后的加速度信号进行低通或带通滤波。低通或带通滤波后，可以获得较为平滑的信号，从而便于准确统计加速度信号的极值点。
S313，去除滤波处理后的加速度信号中的干扰极值点。
由于行走步数只与单轴加速度信号中的极值点的数目对应，而与这些极值点的准确位置关系不大。换言之，只要去除合适数目的极值点，以保证左腿和右腿各迈一步的运动周期与一个极大值点对应即可。因此，干扰极值点的去除方法并不唯一。
本实施例可以基于相邻极值点的时间间隔进行极值点去除处理。具体的，将加速度信号极值点与其前一个加速度信号极值点的时间间隔小于预定阈值的极值点去除。其中，该预定阈值远小于单轴加速度信号的基频分量的周期。在该极值点去除方法中，对于每一组靠得较近的极值点中，只保留最左边的一个极值点，其余极值点则视为干扰极值点而去除。这种方式下，通过加速度信号极值点之间的时间间隔，去除了加速度信号极值点中的干扰极值点。
当然，本实施例还可以采用其他方法去除干扰极值点，例如利用将每组时间间隔小于预定阈值的加速度信号极值点中的幅值非最大的加速度信号极值点去除。在该极值点去除方法中，对于每一组靠得较近的极值点中，只保留幅值最大的加速度信号极值点，其余的极值点则视为干扰极值点而去除。
S314，计算去除干扰极值点后的加速度信号中的有效极值点个数，该有效极值点个数为行走步数。
在本步骤中，需要统计三个单轴加速度信号中去除干扰极值点后的加速度信号极值点的数目；再根据每个单轴加速度信号对应的有效极值点数目确定行走步数。
本实施例可以通过如下方法确定行走步数：如果各单轴加速度信号的能量相差不大，则可以对各轴所对应的去除干扰极值点后的加速度信号极值点的数目进行平均，以该平均数作为本轮计步过程所获得的行走步数；如果各单轴加速度信号的能量相差较大，则可以根据其中能量最大的单轴加速度信号所对应的去除干扰极值点后的加速度信号极值点的数目来确定本轮计步过程所获得的行走步数。
S315，统计相邻两步数之间的信号采样点个数。
S316，将信号采样点个数乘以信号采样时间，得到行走周期，该行走周期的倒数即为行走步频。
S322，根据获取的心率信号的周期性波动计算行走过程中的心率。
S330，将行走步频和心率分别与步频阈值和心率阈值进行比较，判断行走步频是否大于步频阈值，且心率是否大于心率阈值；若行走步频大于步频阈值，且心率大于心率阈值，则执行步骤S340；否则执行步骤S350。
本实施例中的步频阈值和心率阈值可以根据数据统计设置，优选地，可以将步频阈值设定为2.5步/秒；将心率阈值设定为120次/分钟。
对于一般情况而言，人的静息心率在60～100次/分钟范围内，而人在运动时， 心率会增加。由于跑步相对走路的运动强度大，因而人在跑步时的心率较走路时的心率高。因此，可以根据获得的心率是否满足跑步时的心率范围设定心率阈值，例如可将心率阈值设定为120次/分钟。由于心率与年龄、身体状态和静息心率等因素有关，因此在设定心率阈值时可以参考这些方面的参数，使得走路和跑步的区分会更准确。
S340，确定当前人体运动状态为跑步状态，并将计算的行走步数记为跑步步数。
S350，确定当前人体运动状态为走路状态，并将计算的行走步数记为走路步数。
实施例二：
本实施例采用多传感器结合人体特征(如心率检测)达到更精确区分人体睡眠状态和清醒状态的效果。
本实施例将睡眠状态分为浅睡状态和深睡状态，由于不同程度的睡眠状态对应的运动能量不同，而且可穿戴设备静止放置时与人体佩戴可穿戴设备进入睡眠状态时，人体体征传感器输出的人体体征信号并不相同。因此，本实施例利用具有加速度传感器和人体体征传感器的可穿戴设备，结合运动特征和人体体征来识别人体的睡眠状态。
图4为本实施例提供的人体运动状态的识别方法流程图，如图4所示，图4中的方法包括：
S410，在可穿戴设备中设置三轴加速度传感器和人体体征传感器。
本实施例优选地利用心率传感器获取心率体征，或利用电容传感器获取人体皮肤的电容性体征。
S420，根据三轴加速度传感器采集的加速度信号检测人体的瞬时异动。
由于人在进入睡眠状态时，大部分时间时处于无动作的呼吸状态，只偶尔会出现翻身、惊吓、抽搐等瞬时异动行为。因此，本实施例的加速度信号具有如图5所示的特征，图5为睡眠过程中三轴加速度传感器在三个方向上产生的加速度信号示意图，如图5所示，人在睡眠过程中，大多数时间加速度信号是非常小而平缓的，睡眠过程中出现异动(例如在图5中600～610秒时间段出现的翻身等瞬时异动)的时间占整个睡眠时间的比例很小。
基于睡眠过程中加速度信号的特征，本实施例通过分析睡眠状态统计时间内瞬时异动发生的次数来定量地分析睡眠的质量。
S430，根据人体体征传感器采集的体征信号判断是否佩戴上述可穿戴设备。
当该可穿戴设备设置的人体体征传感器为心率传感器，本步骤中根据人体体征传感器采集的体征信号判断是否佩戴上述可穿戴设备具体为：根据心率传感器采集的心率信号的能量或幅度判断是否佩戴上述可穿戴设备。
当该可穿戴设备设置的人体体征传感器为电容传感器，本步骤中根据人体体征传感器采集的体征信号判断是否佩戴上述可穿戴设备具体为：根据电容传感器采集的电容信号的能量或幅度判断是否佩戴上述可穿戴设备。
S440，当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴上述可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
本实施例利用加速度传感器和体征传感器分别采集人体运动过程中的加速度信号和体征信号，结合加速度信号和体征信号识别人体的睡眠状态来提高识别结果的准确性，避免了将可穿戴设备脱离于人体静止放置的情况统计到睡眠状态中。
本实施例步骤S420中根据三轴加速度传感器采集的加速度信号检测人体的瞬时异动包括：
对三轴加速度传感器采集的加速度信号进行滤波处理，滤除直流信号；
计算滤波处理后的加速度信号单位时间内的多个瞬时能量，该瞬时能量为将单位时间均分后的每个子时段的加速度信号能量；
将多个瞬时能量分别与第一能量阈值进行比较，在该多个瞬时能量均小于第一能量阈值时，确定该单位时间未发生瞬时异动；否则，确定该单位时间发生瞬时异动。
则本实施例步骤S440中当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴可穿戴设备时，确定人体运动状态为睡眠状态包括：
将睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数；
将每个运动量统计时段的瞬时异动个数分别与设定的第一数目阈值进行比较，在每个运动量统计时段的瞬时异动个数均小于或等于第一数目阈值，且判断人体佩戴可穿戴设备时，确定人体运动状态为睡眠状态。
进一步地，在执行步骤S440后，该运动状态识别方法还包括：
将睡眠状态统计时间按照一定条件划分为多个深浅睡眠统计时段，并统计每个深浅睡眠统计时段内发生瞬时异动的个数；该深浅睡眠统计时段的时间长度大于运动量统计时段；
依次将每个深浅睡眠统计时段的瞬时异动个数与设定的第二数目阈值进行比较；由于人处于深睡状态时，翻身、挠痒等异动动作个数小于人处于浅睡状态，从而为了进一步细分睡眠状态中的深睡状态和浅睡状态，需要使第二数目阈值小于第一数目阈值；
在该深浅睡眠统计时段的瞬时异动个数小于或等于第二数目阈值时，确定 该时段内的睡眠状态为深度睡眠状态；否则确定该时段内的睡眠状态为轻度睡眠状态。
或者，将睡眠状态统计时间按照一定条件划分为多个深浅睡眠统计时段，并计算每个深浅睡眠统计时段内滤波处理后的加速度信号能量；该深浅睡眠统计时段的时间长度大于运动量统计时段；
依次将每个深浅睡眠统计时段的加速度信号能量与设定的第二能量阈值进行比较；由于人处于深睡状态时，翻身、挠痒等异动动作个数小于人处于浅睡状态，从而为了进一步细分睡眠状态中的深睡状态和浅睡状态，需要使该第二能量阈值小于第一能量阈值；
在该深浅睡眠统计时段的加速度信号能量小于或等于第二能量阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态。
其中，上述将睡眠状态统计时间划分为多个深浅睡眠统计时段，可以采用静态划分方法，如按固定时长将睡眠状态统计时间均匀划分为多个深浅睡眠统计时段；也可以采用动态划分方法，如在睡眠状态统计时间内将设定长度的窗口按照设定步长移动，移动重叠窗将划分为多个深浅睡眠统计时段。
为了更加详细地介绍本实施例识别人体运动状态的方法，本实施例通过一具体实现方案来说明。
图6为本实施例提供的睡眠状态识别方法流程图，如图6所示，该睡眠状态的识别方法包括：
S611，从可穿戴设备的三轴加速度传感器的输出中获取加速度信号。
如图5所示，睡眠过程中大多数时间的加速度信号是非常小而平缓的，睡眠过程中出现翻身、挠痒等异动的时间占整个睡眠时间的比例很小。因此，在分析睡眠产生的加速度信号时，信号的采样长度应该较大，以保证能够采集到睡眠过程中的异动信号。
S621，从可穿戴设备的人体体征传感器的输出中获取人体体征信号。
需要说明的是，步骤S611的加速度信号和步骤S621中的人体体征信号对应为同一运动状态，且采样时长相同。
本实施例中的人体体征传感器优选为心率传感器或电容传感器。
S612，将睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数。
本步骤中通过下述方法计算每个运动量统计时段内发生瞬时异动的个数：
首先，对加速度信号进行滤波处理，滤除直流信号。
从图5中可以看出，三轴加速度传感器输出的加速度信号通常会包含直流分量，而该直流分量的存在会对加速度信号的分析产生干扰。因此，本实施例 通过高通滤波来滤除加速度信号中的直流分量。
然后，计算滤波处理后的加速度信号单位时间内的多个瞬时能量，该瞬时能量为将单位时间均分后的每个子时段的加速度信号能量。
本实施例中的单位时间不宜过长，也不宜过短，单位时间过长不便于统计异动，过短则容易将一个瞬时异动行为统计为两个或者统计为多个；单位时间优选为1秒钟。
本实施例将1秒钟的单位时间均分为两个子时段，每个子时段的时长为0.5秒钟，通过下述公式计算每个子时段的瞬时能量STD：
Figure PCTCN2016086933-appb-000002
其中，si为加速度信号的第i个值，N为加速度信号的长度，N＝0.5，且1≤i≤N，s0为加速度信号的平均值。
接着，将该多个瞬时能量分别与第一能量阈值进行比较，在多个瞬时能量均小于第一能量阈值时，确定该单位时间未发生瞬时异动；否则，确定该单位时间发生瞬时异动。
由于呼吸运动所产生的能量很小，因此本实施例的第一能量阈值需要能够将表征翻身、挠痒的瞬时异动和呼吸运动区分开。优选地第一能量阈值为接近零的值。
当将1秒钟的单位时间均分为两个子时段时，若这两个子时段的瞬时能量中的一个子时段的瞬时能量大于第一能量阈值，则确定该单位时间内发生瞬时异动。
最后，将睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数。
本实施例中的睡眠状态统计时间范围为20分钟～60分钟，优选为30分钟；运动量统计时段的时间长度范围为1分钟～3分钟。
当睡眠状态统计时间为30分钟，运动量统计时段的时间长度为1分钟时，本步骤具体为，将睡眠状态统计时间30分钟按照1分钟的时间长度均匀划分为30个运动量统计时段，并统计每分钟内发生的瞬时异动个数。
S622，根据人体体征信号的能量或幅度判断是否佩戴可穿戴设备。
当可穿戴设备中设置的人体体征传感器为心率传感器时，则根据心率信号的能量或幅度判断是否佩戴该可穿戴设备；当可穿戴设备中设置的人体体征传感器为电容传感器时，则根据电容信号的能量或幅度判断是否佩戴该可穿戴设备。
S630，判断每个运动量统计时段的瞬时异动个数是否均小于或等于第一数目阈值，以及判断是否佩戴可穿戴设备；若每个运动量统计时段的瞬时异动个 数均小于或等于第一数目阈值，且判断人体佩戴可穿戴设备，则执行步骤S640；否则执行步骤S650。
本实施例中的第一数目阈值范围为0～60，优选为30。
本实施例中的第一数目阈值范围为0～80，优选为30。
S640，确定当前人体运动状态为睡眠状态，并执行步骤S660。
S650，确定当前人体运动状态为清醒状态。
S660，将睡眠状态统计时间划分为多个深浅睡眠统计时段，并统计每个时段内发生瞬时异动的个数；该深浅睡眠统计时段的时间长度大于运动量统计时段。
本实施例中的深浅睡眠统计时段的时间长度范围为4分钟～10分钟，可以采用静态划分方法将深浅睡眠统计时间划分为多个深浅睡眠统计时段，假设深浅睡眠统计时段的时间长度为5分钟，则可以将睡眠状态统计时间30分钟按照5分钟的时间长度均匀划分为6个深浅睡眠统计时段。
本实施例中还可以采用加窗移动的动态方法来划分睡眠状态统计时间，假设所加窗的窗口长度为5分钟、移动步长为1分钟，则可以将睡眠状态统计时间30分钟按照5分钟的时间长度均匀划分为29个深浅睡眠统计时段。
S670，依次将每个深浅睡眠统计时段的瞬时异动个数与设定的第二数目阈值进行比较，当该深浅睡眠统计时段的瞬时异动个数小于或等于第二数目阈值时，执行步骤S680；否则执行步骤S690。
本步骤中的第二数目阈值小于第一数目阈值，第二数目阈值范围为0～15。
需要说明的是，在本实施例的步骤S660中，还可以计算每个深浅睡眠统计时段内滤波处理后的加速度信号能量；
则步骤S670对应为：依次将每个深浅睡眠统计时段的加速度信号能量与第二能量阈值进行比较，当该深浅睡眠统计时段的加速度信号能量小于或等于第二能量阈值时，执行步骤S680；否则执行步骤S690。其中第二能量阈值小于第一能量阈值。
S680，确定该深浅睡眠统计时段内的睡眠状态为深睡状态。
S690，确定该深浅睡眠统计时段内的睡眠状态为浅睡状态。
实施例三：
基于与实施例一相同的技术构思，本实施例提供了一种人体运动状态的识别装置。
图7为本实施例提供的人体运动状态的识别装置示意图，如图7所示，适用于具有多传感器的可穿戴设备，可以是可穿戴设备中的具体模块，该识别装置包括：
行走步频计算单元71，用于在根据可穿戴设备的三轴加速度传感器采集的 加速度信号判定人体处于行走状态时，根据该三轴加速度传感器采集的加速度信号计算人体的行走步数，并根据该行走步数计算行走步频。
体征频率计算单元72，用于根据可穿戴设备的人体体征传感器采集的体征信号计算行走过程中相应的体征频率。
比较单元73，用于将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较。
运动状态识别单元74，用于在行走步频大于步频阈值，且体征频率大于体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
其中，行走步频计算单元71包括：
滤波模块，用于对加速度信号进行滤波处理；
干扰极值点去除模块，用于去除滤波处理后的加速度信号中的干扰极值点；
极值点统计模块，用于计算去除干扰极值点后的加速度信号中的有效极值点个数，该有效极值点个数为行走步数；
步频计算模块，用于统计相邻两步数之间的信号采样点个数，并将信号采样点个数乘以信号采样时间，得到行走周期；以及用于将行走周期取倒数得到行走步频。
在本实施例中，若可穿戴设备中设置的人体体征传感器为心率传感器，则体征频率计算单元72，具体用于根据心率传感器采集的心率信号的周期性波动计算行走过程中的心率。
实施例四：
基于与实施例二相同的技术构思，本实施例提供了一种人体运动状态的识别装置。
图8为本实施例提供的人体运动状态的识别装置示意图，适用于具有多传感器的可穿戴设备，可以是可穿戴设备中的具体模块，该识别装置包括：
瞬时异动检测单元81，用于根据可穿戴设备的三轴加速度传感器采集的加速度信号检测人体的瞬时异动。
佩戴判断单元82，用于根据可穿戴设备中的人体体征传感器采集的体征信号判断是否佩戴该可穿戴设备。
其中，人体体征传感器可以为心率传感器或电容传感器；当人体体征传感器为心率传感器时，佩戴判断单元具体用于根据心率传感器采集的心率信号的能量或幅度判断是否佩戴可穿戴设备；当人体体征传感器为电容传感器时，佩戴判断单元具体用于根据电容传感器采集的电容信号的能量或幅度判断是否佩戴可穿戴设备。
运动状态识别单元83，用于当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴该可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
其中，瞬时异动检测单元81包括：
滤波模块，用于对三轴加速度传感器采集的加速度信号进行滤波处理，滤除直流信号；
瞬时能量计算模块，用于计算滤波处理后的加速度信号单位时间内的多个瞬时能量，该瞬时能量为将单位时间均分后的每个子时段的加速度信号能量；
瞬时能量判断处理模块，用于将该多个瞬时能量分别与第一能量阈值进行比较，在该多个瞬时能量均小于第一能量阈值时，确定该单位时间未发生瞬时异动；否则，确定该单位时间发生瞬时异动。
则运动状态识别单元83包括：
瞬时异动个数统计模块，用于将睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数；
瞬时异动个数比较判断模块，用于将每个运动量统计时段的瞬时异动个数分别与设定的第一数目阈值进行比较，在每个运动量统计时段的瞬时异动个数均小于或等于第一数目阈值，且判断人体佩戴该可穿戴设备时，确定人体运动状态为睡眠状态。
进一步地，运动状态识别单元83还包括深浅睡眠判断模块；
该深浅睡眠判断模块，用于将睡眠状态统计时间划分为多个深浅睡眠统计时段，并统计每个深浅睡眠统计时段内发生瞬时异动的个数；依次将每个深浅睡眠统计时段的瞬时异动个数与设定的第二数目阈值进行比较，在该深浅睡眠统计时段的瞬时异动个数小于或等于第二数目阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态；其中，深浅睡眠统计时段的时间长度大于运动量统计时段，第二数目阈值小于第一数目阈值；
或者，深浅睡眠判断模块，用于将睡眠状态统计时间划分为多个深浅睡眠统计时段，并计算每个深浅睡眠统计时段内滤波处理后的加速度信号能量；依次将每个深浅睡眠统计时段的加速度信号能量与设定的第二能量阈值进行比较，在该深浅睡眠统计时段的加速度信号能量小于或等于第二能量阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态；其中，深浅睡眠统计时段的时间长度大于运动量统计时段，第二能量阈值小于第一能量阈值。
综上所述，本发明提供了一种能够有效地区分走路和跑步的人体运动状态的识别方法和装置，本发明的该技术方案基于人在跑步状态和走路状态的步频 和人体体征不同的特点，在可穿戴设备中设置加速度传感器和人体体征传感器，利用加速度传感器和人体体征传感器分别采集人体运动过程中的加速度信号和人体体征信号，基于加速度信号和人体体征信号分别计算运动过程中的步频和相应的人体体征频率，结合步频和人体体征频率区分走路状态和跑步状态。以及本发明还提供了一种能够有效地解决睡眠状态统计中假睡问题的人体运动状态的识别方法和装置，本发明的该技术方案通过在可穿戴设备中设置加速度传感器和人体体征传感器，利用加速度传感器和体征传感器分别采集人体运动过程中的加速度信号和体征信号，结合加速度信号和体征信号识别人体的睡眠状态来提高识别结果的准确性，避免了将可穿戴设备脱离于人体静止放置的情况统计到睡眠状态中。
为了便于清楚描述本发明实施例的技术方案，在发明的实施例中，采用了“第一”、“第二”等字样对功能和作用基本相同的相同项或相似项进行区分，本领域技术人员可以理解“第一”、“第二”等字样并不对数量和执行次序进行限定。
以上所述仅为本发明的较佳实施例而已，并非用于限定本发明的保护范围。凡在本发明的精神和原则之内所作的任何修改、等同替换、改进等，均包含在本发明的保护范围内。

Claims (15)

  1. 一种人体运动状态的识别方法，其特征在于，所述方法包括：
    在可穿戴设备中设置三轴加速度传感器和人体体征传感器；
    根据三轴加速度传感器采集的加速度信号判定人体处于行走状态，计算人体的行走步数，并根据所述行走步数计算行走步频；
    根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率；
    将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较，在所述行走步频大于所述步频阈值，且所述体征频率大于所述体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
  2. 根据权利要求1所述的方法，其特征在于，所述根据的三轴加速度传感器采集的加速度信号判定人体处于行走状态，计算人体的行走步数，并根据所述行走步数计算行走步频包括：
    对所述加速度信号进行滤波处理，并去除滤波处理后的加速度信号中的干扰极值点；
    计算去除干扰极值点后的加速度信号中的有效极值点个数，该有效极值点个数为所述行走步数；
    统计相邻两步数之间的信号采样点个数；
    将所述信号采样点个数乘以信号采样时间，得到行走周期，所述行走周期的倒数即为行走步频。
  3. 根据权利要求1所述的方法，其特征在于，所述人体体征传感器为心率传感器，所述根据人体体征传感器采集的体征信号计算行走过程中相应的体征频率具体为：根据心率传感器采集的心率信号的周期性波动计算行走过程中的心率。
  4. 一种人体运动状态的识别方法，其特征在于，所述方法包括：
    在可穿戴设备中设置三轴加速度传感器和人体体征传感器；
    根据三轴加速度传感器采集的加速度信号检测人体的瞬时异动；
    以及根据人体体征传感器采集的体征信号判断是否佩戴所述可穿戴设备；
    当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数 目阈值，且判断人体佩戴所述可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
  5. 根据权利要求4所述的方法，其特征在于，所述根据三轴加速度传感器采集的加速度信号检测人体的瞬时异动包括：
    对所述三轴加速度传感器采集的加速度信号进行滤波处理，滤除直流信号；
    计算滤波处理后的加速度信号单位时间内的多个瞬时能量，所述瞬时能量为将单位时间均分后的每个子时段的加速度信号能量；
    将所述多个瞬时能量分别与第一能量阈值进行比较，在所述多个瞬时能量均小于第一能量阈值时，确定该单位时间未发生瞬时异动；否则，确定该单位时间发生瞬时异动。
  6. 根据权利要求5所述的方法，其特征在于，所述当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴所述可穿戴设备时，确定人体运动状态为睡眠状态包括：
    将所述睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数；
    将每个运动量统计时段的瞬时异动个数分别与设定的第一数目阈值进行比较，在每个运动量统计时段的瞬时异动个数均小于或等于第一数目阈值，且判断人体佩戴所述可穿戴设备时，确定人体运动状态为睡眠状态。
  7. 根据权利要求6所述的方法，其特征在于，所述确定人体运动状态为睡眠状态还包括：
    将所述睡眠状态统计时间划分为多个深浅睡眠统计时段，并统计每个时段内发生瞬时异动的个数；该深浅睡眠统计时段的时间长度大于运动量统计时段；
    依次将每个深浅睡眠统计时段的瞬时异动个数与设定的第二数目阈值进行比较，该第二数目阈值小于第一数目阈值；
    在该深浅睡眠统计时段的瞬时异动个数小于或等于所述第二数目阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态；
    或者，
    将所述睡眠状态统计时间划分为多个深浅睡眠统计时段，并计算每个深浅睡眠统计时段内滤波处理后的加速度信号能量；该深浅睡眠统计时段的时间长度大于运动量统计时段；
    依次将每个深浅睡眠统计时段的加速度信号能量与设定的第二能量阈值进行比较，该第二能量阈值小于第一能量阈值；
    在该深浅睡眠统计时段的加速度信号能量小于或等于所述第二能量阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态。
  8. 根据权利要求4所述的方法，其特征在于，所述人体体征传感器为心率传感器，所述根据人体体征传感器采集的体征信号判断是否佩戴所述可穿戴设备具体为：根据心率传感器采集的心率信号的能量或幅度判断是否佩戴所述可穿戴设备；
    或者，所述人体体征传感器为电容传感器，所述根据人体体征传感器采集的体征信号判断是否佩戴所述可穿戴设备具体为：根据电容传感器采集的电容信号的能量或幅度判断是否佩戴所述可穿戴设备。
  9. 一种人体运动状态的识别装置，适用于具有多传感器的可穿戴设备，其特征在于，所述识别装置包括：
    行走步频计算单元，用于在根据可穿戴设备的三轴加速度传感器采集的加速度信号判定人体处于行走状态时，根据该三轴加速度传感器采集的加速度信号计算人体的行走步数，并根据所述行走步数计算行走步频；
    体征频率计算单元，用于根据可穿戴设备中的人体体征传感器采集的体征信号计算行走过程中相应的体征频率；
    比较单元，用于将计算得到的行走步频和体征频率分别与步频阈值和体征频率阈值进行比较；
    运动状态识别单元，用于在所述行走步频大于所述步频阈值，且所述体征频率大于所述体征频率阈值时，确定人体运动状态为跑步状态，并将计算的行走步数记为跑步步数；否则，确定人体运动状态为走路状态，并将计算的行走步数记为走路步数。
  10. 根据权利要求9所述的识别装置，其特征在于，所述行走步频计算单元包括：
    滤波模块，用于对所述加速度信号进行滤波处理；
    干扰极值点去除模块，用于去除滤波处理后的加速度信号中的干扰极值点；
    极值点统计模块，用于计算去除干扰极值点后的加速度信号中的有效极值点个数，该有效极值点个数为所述行走步数；
    步频计算模块，用于统计相邻两步数之间的信号采样点个数，并将所述信号采样点个数乘以信号采样时间，得到行走周期；以及用于将所述行走周期取倒数得到行走步频；
    所述人体体征传感器为心率传感器，所述体征频率计算单元，具体用于根据心率传感器采集的心率信号的周期性波动计算行走过程中的心率。
  11. 一种人体运动状态的识别装置，适用于具有多传感器的可穿戴设备，其特征在于，所述识别装置包括：
    瞬时异动检测单元，用于根据可穿戴设备的三轴加速度传感器采集的加速度信号检测人体的瞬时异动；
    佩戴判断单元，用于根据可穿戴设备的人体体征传感器采集的体征信号判断是否佩戴所述可穿戴设备；
    运动状态识别单元，用于当睡眠状态统计时间内检测到的瞬时异动的个数小于或等于设定的第一数目阈值，且判断人体佩戴所述可穿戴设备时，确定人体运动状态为睡眠状态；否则，确定人体运动状态为清醒状态。
  12. 根据权利要求11所述的识别装置，其特征在于，所述瞬时异动检测单元包括：
    滤波模块，用于对所述三轴加速度传感器采集的加速度信号进行滤波处理，滤除直流信号；
    瞬时能量计算模块，用于计算滤波处理后的加速度信号单位时间内的多个瞬时能量，所述瞬时能量为将单位时间均分后的每个子时段的加速度信号能量；
    瞬时能量判断处理模块，用于将所述多个瞬时能量分别与第一能量阈值进行比较，在所述多个瞬时能量均小于第一能量阈值时，确定该单位时间未发生瞬时异动；否则，确定该单位时间发生瞬时异动。
  13. 根据权利要求12所述的识别装置，其特征在于，所述运动状态识别单元包括：
    瞬时异动个数统计模块，用于将所述睡眠状态统计时间均匀划分为多个运动量统计时段，并统计每个运动量统计时段内发生瞬时异动的个数；
    瞬时异动个数比较判断模块，用于将每个运动量统计时段的瞬时异动个数分别与设定的第一数目阈值进行比较，在每个运动量统计时段的瞬时异动个数均小于或等于第一数目阈值，且判断人体佩戴所述可穿戴设备时，确定人体运动状态为睡眠状态。
  14. 根据权利要求13所述的识别装置，其特征在于，所述运动状态识别单元还包括深浅睡眠判断模块；
    所述深浅睡眠判断模块，用于将所述睡眠状态统计时间均匀划分为多个深浅睡眠统计时段，并统计每个深浅睡眠统计时段内发生瞬时异动的个数；依次将每个深浅睡眠统计时段的瞬时异动个数与设定的第二数目阈值进行比较，在该深浅睡眠统计时段的瞬时异动个数小于或等于所述第二数目阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态；其中，深浅睡眠统计时段的时间长度大于运动量统计时段，第二数目阈值小于第一数目阈值；
    或者，
    所述深浅睡眠判断模块，用于将所述睡眠状态统计时间均匀划分为多个深浅睡眠统计时段，并计算每个深浅睡眠统计时段内滤波处理后的加速度信号能量；依次将每个深浅睡眠统计时段的加速度信号能量与设定的第二能量阈值进行比较，在该深浅睡眠统计时段的加速度信号能量小于或等于所述第二能量阈值时，确定该深浅睡眠统计时段内的睡眠状态为深度睡眠状态；否则确定该深浅睡眠统计时段内的睡眠状态为轻度睡眠状态；其中，深浅睡眠统计时段的时间长度大于运动量统计时段，第二能量阈值小于第一能量阈值。
  15. 根据权利要求11所述的识别装置，其特征在于，所述人体体征传感器为心率传感器，所述佩戴判断单元，具体用于根据心率传感器采集的心率信号的能量或幅度判断是否佩戴所述可穿戴设备；
    或者，所述人体体征传感器为电容传感器，所述佩戴判断单元，具体用于根据电容传感器采集的电容信号的能量或幅度判断是否佩戴所述可穿戴设备。
PCT/CN2016/086933 2015-12-28 2016-06-23 一种人体运动状态的识别方法和装置 WO2017113653A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP16876980.0A EP3369375A4 (en) 2015-12-28 2016-06-23 METHOD AND DEVICE FOR IDENTIFYING THE CONDITION OF THE MOVEMENT OF THE HUMAN BODY
US15/541,313 US10856777B2 (en) 2015-12-28 2016-06-23 Method and device for identifying human movement state

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201511003580.0 2015-12-28
CN201511003580.0A CN105496416B (zh) 2015-12-28 2015-12-28 一种人体运动状态的识别方法和装置

Publications (1)

Publication Number Publication Date
WO2017113653A1 true WO2017113653A1 (zh) 2017-07-06

Family

ID=55704988

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/086933 WO2017113653A1 (zh) 2015-12-28 2016-06-23 一种人体运动状态的识别方法和装置

Country Status (4)

Country Link
US (1) US10856777B2 (zh)
EP (1) EP3369375A4 (zh)
CN (1) CN105496416B (zh)
WO (1) WO2017113653A1 (zh)

Families Citing this family (39)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105496416B (zh) 2015-12-28 2019-04-30 歌尔股份有限公司 一种人体运动状态的识别方法和装置
CN105943053A (zh) * 2016-06-01 2016-09-21 北京健康有益科技有限公司 健康检测方法及装置
CN106197470B (zh) * 2016-06-29 2019-11-26 联想(北京)有限公司 一种数据处理方法及电子设备
CN106388771B (zh) * 2016-08-16 2019-06-28 惠州市德赛工业研究院有限公司 一种自动检测人体生理状态的方法及运动手环
CN108778111B (zh) * 2016-09-12 2020-12-15 华为技术有限公司 一种心率检测方法及装置
CN106197639A (zh) * 2016-09-23 2016-12-07 初子超 一种具有唤醒功能的床、唤醒系统及唤醒方法
CN106510658B (zh) * 2016-10-25 2019-08-02 广东乐源数字技术有限公司 一种基于手环的人体情绪判断方法
CN106289309B (zh) * 2016-10-26 2019-08-16 深圳大学 基于三轴加速度传感器的计步方法及装置
CN106526619A (zh) * 2016-10-26 2017-03-22 广东小天才科技有限公司 自动开启gps定位功能的方法及装置
CN106333691A (zh) * 2016-10-27 2017-01-18 深圳市万机创意电子科技有限公司 判断人体睡眠状态、静止状态和运动状态的方法及装置
CN106473729A (zh) * 2016-11-14 2017-03-08 广东思派康电子科技有限公司 一种基于心率跳动的激光闪烁耳机及其实现方法
CN109222329B (zh) * 2017-04-12 2021-08-03 纵联汽车工业工程研究(天津)有限公司 一种步行长度计算方法及配置该方法的智能鞋垫
CN107314775B (zh) * 2017-05-17 2019-09-10 浙江利尔达物联网技术有限公司 一种基于三轴加速度传感器的动态切换计算轴的计步方法
CN107582061B (zh) * 2017-07-21 2020-03-27 青岛海信移动通信技术股份有限公司 一种识别人体运动状态的方法及智能移动设备
CN107703779B (zh) * 2017-07-25 2020-03-10 广东乐心医疗电子股份有限公司 通过识别可穿戴设备是否佩戴控制功能启闭的方法与装置
CN107545134B (zh) * 2017-07-25 2020-09-25 广东乐心医疗电子股份有限公司 用于可穿戴设备的与睡眠相关的特征数据处理方法与装置
US11487965B2 (en) 2017-08-23 2022-11-01 Huawei Technologies Co., Ltd. Method and apparatus for counting foot step based on stride frequency, and device
CN107515010A (zh) * 2017-08-28 2017-12-26 五邑大学 一种计步器的数据处理方法以及计步器装置
CN107860397A (zh) * 2017-10-25 2018-03-30 北京小米移动软件有限公司 统计运动信息的方法及装置
CN108008151A (zh) * 2017-11-09 2018-05-08 惠州市德赛工业研究院有限公司 一种基于三轴加速度传感器的运动状态识别方法及系统
CN108814618B (zh) * 2018-04-27 2021-08-31 歌尔科技有限公司 一种运动状态的识别方法、装置及终端设备
CN111374671A (zh) * 2018-12-28 2020-07-07 上海倍增智能科技有限公司 一种基于手机计步计数的运动指标管理系统
CN109949543A (zh) * 2019-04-18 2019-06-28 西安建筑科技大学 一种多功能鞋及基于压力感应的远程智能监控方法
CN109907736A (zh) * 2019-04-25 2019-06-21 蔡文贤 一种在计步软件上区分计步运动类型与强度的应用方法
CN111854737A (zh) * 2019-04-28 2020-10-30 百应科技（北京）有限公司 一种运动类型的判断方法及系统
KR20200134995A (ko) * 2019-05-24 2020-12-02 삼성전자주식회사 사용자의 이동 패턴 특징을 이용한 모드 제어 방법 및 장치
CN110180158B (zh) * 2019-07-02 2021-04-23 乐跑体育互联网（武汉）有限公司 一种跑步状态识别方法、系统及终端设备
CN110384505A (zh) * 2019-07-25 2019-10-29 四川云杉智途科技有限公司 一种自动检测人体运动类型的方法技术领域
CN110558990B (zh) * 2019-07-30 2022-04-12 福建省万物智联科技有限公司 一种步态分析方法及装置
CN110575176B (zh) * 2019-08-26 2024-03-26 南京理工大学 基于两层滑动窗口阈值的动作分割方法
CN111006683A (zh) * 2019-11-27 2020-04-14 青岛歌尔智能传感器有限公司 计步装置及其计步方法、控制器和可读存储介质
CN111166354B (zh) * 2020-01-23 2022-11-18 北京津发科技股份有限公司 影响情绪变化的因素的分析方法及电子设备
CN113285966B (zh) * 2020-02-19 2022-12-09 中国农业科学院农业信息研究所 一种智能猪行为异常监测方法及系统
CN113679340B (zh) * 2020-05-19 2024-01-19 瑞昱半导体股份有限公司 睡眠监测装置及方法
CN112130677B (zh) * 2020-09-23 2023-05-12 深圳市爱都科技有限公司 一种可穿戴终端及其举手识别方法
WO2022103335A1 (en) * 2020-11-12 2022-05-19 Kaha Pte. Ltd. Method, system and device for monitoring a sleep condition in user
CN112790752B (zh) * 2021-01-22 2022-09-27 维沃移动通信有限公司 心率值修正方法、装置及电子设备
CN114298105B (zh) * 2021-12-29 2023-08-22 东莞市猎声电子科技有限公司 一种跑步过程中快速响应抬腕动作并亮屏的信号处理方法
CN116725538B (zh) * 2023-08-11 2023-10-27 深圳市昊岳科技有限公司 一种基于深度学习的手环情绪识别方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090043531A1 (en) * 2007-08-08 2009-02-12 Philippe Kahn Human activity monitoring device with distance calculation
US20120083705A1 (en) * 2010-09-30 2012-04-05 Shelten Gee Jao Yuen Activity Monitoring Systems and Methods of Operating Same
EP2479966A1 (en) * 2011-01-19 2012-07-25 Vodafone Group PLC identifying personal context by correlating the output of multiple sensors
US20130053990A1 (en) * 2010-02-24 2013-02-28 Jonathan Edward Bell Ackland Classification System and Method
CN103767710A (zh) * 2013-12-31 2014-05-07 歌尔声学股份有限公司 人体运动状态监视方法和装置
US20140316305A1 (en) * 2012-06-22 2014-10-23 Fitbit, Inc. Gps accuracy refinement using external sensors
CN105496416A (zh) * 2015-12-28 2016-04-20 歌尔声学股份有限公司 一种人体运动状态的识别方法和装置

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6878121B2 (en) * 2002-11-01 2005-04-12 David T. Krausman Sleep scoring apparatus and method
US20090082994A1 (en) 2007-09-25 2009-03-26 Motorola, Inc. Headset With Integrated Pedometer and Corresponding Method
US10463300B2 (en) * 2011-09-19 2019-11-05 Dp Technologies, Inc. Body-worn monitor
US8948832B2 (en) * 2012-06-22 2015-02-03 Fitbit, Inc. Wearable heart rate monitor
US20140364770A1 (en) * 2013-06-06 2014-12-11 Motorola Mobility Llc Accelerometer-based sleep analysis
CN103727954A (zh) * 2013-12-27 2014-04-16 北京超思电子技术股份有限公司 一种计步器
CN103727959B (zh) * 2013-12-31 2016-09-14 歌尔声学股份有限公司 计步方法及装置
US10441212B2 (en) * 2014-04-11 2019-10-15 Withings Method to determine positions and states of an activity monitoring device

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090043531A1 (en) * 2007-08-08 2009-02-12 Philippe Kahn Human activity monitoring device with distance calculation
US20130053990A1 (en) * 2010-02-24 2013-02-28 Jonathan Edward Bell Ackland Classification System and Method
US20120083705A1 (en) * 2010-09-30 2012-04-05 Shelten Gee Jao Yuen Activity Monitoring Systems and Methods of Operating Same
EP2479966A1 (en) * 2011-01-19 2012-07-25 Vodafone Group PLC identifying personal context by correlating the output of multiple sensors
US20140316305A1 (en) * 2012-06-22 2014-10-23 Fitbit, Inc. Gps accuracy refinement using external sensors
CN103767710A (zh) * 2013-12-31 2014-05-07 歌尔声学股份有限公司 人体运动状态监视方法和装置
CN105496416A (zh) * 2015-12-28 2016-04-20 歌尔声学股份有限公司 一种人体运动状态的识别方法和装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3369375A4 *

Also Published As

Publication number Publication date
US20180020953A1 (en) 2018-01-25
EP3369375A4 (en) 2019-04-24
CN105496416A (zh) 2016-04-20
US10856777B2 (en) 2020-12-08
EP3369375A1 (en) 2018-09-05
CN105496416B (zh) 2019-04-30

Similar Documents

Publication Publication Date Title
WO2017113653A1 (zh) 一种人体运动状态的识别方法和装置
CN107106085B (zh) 用于睡眠监测的设备和方法
CN106971059B (zh) 一种基于神经网络自适应健康监测的可穿戴设备
WO2015100706A1 (zh) 人体运动状态监视方法和装置
CN105263403B (zh) 生物体信息处理装置以及生物体信息处理方法
KR101672609B1 (ko) 에너지 소비
WO2015100707A1 (zh) 计步方法及装置
CN104990562A (zh) 基于自相关运算的计步方法
US20200289047A1 (en) Fetal monitoring system and method
CN107469326A (zh) 一种用于可穿戴设备的游泳监测方法与装置及可穿戴设备
CN111643092A (zh) 一种癫痫报警装置及癫痫检测方法
Soltani et al. Real-world gait bout detection using a wrist sensor: An unsupervised real-life validation
Wong et al. Activity recognition and stress detection via wristband
JP5488135B2 (ja) 生体情報処理装置
Abbas et al. Characterizing peaks in acceleration signals–application to physical activity detection using wearable sensors
CN111278353A (zh) 一种生命体征信号噪声的检测方法与系统
WO2023205147A1 (en) System and method for assessing neuro muscular disorder by generating biomarkers from the analysis of gait
Lee et al. Mobile system design for scratch recognition
Tokmak et al. Unveiling the relationships between seismocardiogram signals, physical activity types and metabolic equivalent of task scores
CN115006824B (zh) 一种划船机动作计数方法、装置、介质和智能穿戴设备
Qi et al. Interference source-based quality assessment method for postauricular photoplethysmography signals
US20230380774A1 (en) Passive Breathing-Rate Determination
Li et al. Research on recognition of physical activity types based on a single triaxial acceleration sensor
Li et al. Research progress on wearable devices for daily human health management
Hou et al. Embedded system design for automatic screening of snore during nocturnal audio recording

Legal Events

Date Code Title Description
REEP Request for entry into the european phase

Ref document number: 2016876980

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 15541313

Country of ref document: US

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

Ref document number: 16876980

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE