Method for detecting whether someone is on sensor based on ballistocardiogram signal

Abstract

The invention discloses a method for detecting whether a person is on a sensor based on a ballistocardiogram signal. The method comprises the following steps: collecting a ballistocardiogram original signal; according to the original signal, obtaining a slope value of a discrete power spectrum after straight line fitting in unit time; judging whether a person is on the sensor or not according to the discrete power spectrum slope value and a person slope threshold value; if all the discrete power spectrum slope values within the specified time are larger than a human-person slope threshold value, it is judged that a person is on the sensor, and if all the discrete power spectrum slope values within the specified time are smaller than the human-person slope threshold value, it is judged that no person is on the sensor; wherein if the discrete power spectrum slope values are all smaller than the third specified value and the second variable coefficient is smaller than the fourth specified value, the human slope threshold value is refreshed. By collecting the frequency domain original signal of the ballistocardiogram signal and dynamically refreshing the judgment threshold value, the robustness of the algorithm is improved, the situation that the sensor is prone to high-frequency interference of other equipment is effectively restrained, the problem that the signal amplitude and the signal standard deviation are affected by the high-frequency interference is solved, and the accuracy is improved.

Description

[0001] Cross - reference to related applications

[0002] This invention is a divisional application of the invention patent application with the application number 202211079637.5, the application date of September 5, 2022, and the invention title of "Method for Detecting Whether There Is a Person on a Sensor Based on Ballistocardiogram Signals". Technical field

[0003] The present invention relates to the technical field of sensor detection, and particularly relates to a method and system for detecting whether there is a person on a sensor based on ballistocardiogram signals. Background art

[0004] Ballistocardiogram signals, as physiological signals reflecting the heart, respiration, and body movement mechanical activities, have a signal frequency of 0.05 to 20 Hz. Ballistocardiogram signals can be detected by a variety of different sensors, including piezoelectric sensors, fiber optic sensors, and radar sensors. During the acquisition process of ballistocardiogram signals, they contain a lot of information, including respiratory signals, heartbeat signals, body movement signals, and high - frequency interference generated by electronic devices.

[0005] In the actual use of detecting human body signals through piezoelectric sensors, fiber optic sensors, and radar sensors, how to determine whether there is a person on the sensor is a prerequisite for detecting human body signals. The common methods currently used are setting thresholds through signal amplitude and setting thresholds through signal standard deviation. However, in common usage scenarios such as homes and hospitals, there is usually high - frequency interference from other devices. When the high - frequency interference is severe, the signal amplitude and signal standard deviation will be affected, resulting in incorrect judgments by the algorithm. At the same time, even without the influence of high - frequency interference, different sleeping postures, sitting postures, and different sensor placement positions will have different signal amplitudes and signal standard deviations, and the possibility of errors with fixed thresholds is relatively high. Summary of the invention

[0006] In the existing technology for using sensors to determine whether there is a person on a bed, the sensors are easily affected by high - frequency interference from other devices, resulting in the signal amplitude and signal standard deviation being affected, thus causing incorrect judgments by the method.

[0007] In view of the above problems, a method for detecting whether there is a person on a sensor based on ballistocardiogram signals is proposed. By collecting the time - series / frequency - domain original signals of ballistocardiogram signals and dynamically refreshing the judgment threshold, the robustness of the algorithm is increased, effectively suppressing the high - frequency interference that the sensor is easily affected by other devices, solving the problem that the signal amplitude and signal standard deviation are affected, and improving the accuracy.

[0008] In the first aspect, a method for detecting whether there is a person on a sensor based on ballistocardiogram signals includes:

[0009] Step 100: Collect the original ballistocardiogram signal;

[0010] Step 200: Obtain the high-frequency signal based on the original signal;

[0011] Step 300: Calculate the amplitude ratio between the original signal and the high-frequency signal;

[0012] Step 400: If all the amplitude ratios within a specified time are greater than the human presence amplitude threshold, it is determined that there is a person on the sensor; if all the amplitude ratios within the specified time are less than the human presence amplitude threshold, it is determined that there is no person on the sensor;

[0013] Among them, Step 400 includes:

[0014] Step 410: If the amplitude ratios are all less than a first specified value and the first coefficient of variation is less than a second specified value, refresh the human presence amplitude threshold.

[0015] In the first possible implementation manner in combination with the method described in the first aspect, Step 200 includes:

[0016] Step 210: Filter the original signal of the ballistocardiogram signal within a specified time using finite impulse response filtering or zero-phase shift infinite impulse response filtering or wavelet filter filtering;

[0017] Step 220: Retain the original signal above a specified frequency value within a specified time to obtain the high-frequency signal.

[0018] In the third possible implementation manner in combination with the first or second possible implementation manner of the first aspect, Step 410 includes:

[0019] Step 411: Obtain the first average value of all amplitude ratios less than the first specified value;

[0020] Step 412: Multiply the first average value by a first adjustment coefficient to obtain the refreshed human presence amplitude threshold;

[0021] Step 413: Determine whether there is a person on the sensor according to the refreshed human presence amplitude threshold:

[0022] If the sensor is determined to be in the no-person state, continuously refresh the human presence amplitude threshold using the first adjustment coefficient.

[0023] In the fourth possible implementation manner in combination with the third possible implementation manner of the first aspect, the adjustable range of the human presence amplitude is 1.5 - 4.5.

[0024] In a second aspect, a method for detecting whether there is a person on a sensor based on a ballistocardiogram signal includes:

[0025] Step 500: Collect the original ballistocardiogram signal;

[0026] Step 600: Obtain the slope value after fitting a straight line to the discrete power spectrum within a unit time according to the original signal;

[0027] Step 700: Determine whether there is a person on the sensor according to the discrete power spectrum slope value and the slope threshold for a person:

[0028] If all the discrete power spectrum slope values within a specified time are greater than the slope threshold for a person, it is determined that there is a person on the sensor; if all the discrete power spectrum slope values within a specified time are less than the slope threshold for a person, it is determined that there is no person on the sensor;

[0029] Among them, Step 700 includes:

[0030] Step 710: If the discrete power spectrum slope values are all less than a third specified value and the second coefficient of variation is less than a fourth specified value, refresh the slope threshold for a person.

[0031] Combined with the method described in the second aspect, in the first possible implementation manner, Step 600 includes:

[0032] Step 610: Calculate the power spectrum of the original signal within a unit time;

[0033] Step 620: Retain the power spectrum of the original signal less than a second frequency value.

[0034] Combined with the first possible implementation manner of the second aspect, in the second possible implementation manner, Step 600 further includes:

[0035] Step 630: Fit a straight line to the power spectrum of the original signal;

[0036] Step 640: Calculate the slope value of the straight line fitted to all the discrete power spectra within a specified time.

[0037] Combined with the first or second possible implementation manner of the second aspect, in the third possible implementation manner, Step 710 includes:

[0038] Step 711: Obtain the second average value of all the discrete power spectrum slope values greater than a third specified value;

[0039] Step 712: Multiply the second average value by a second adjustment coefficient to obtain the refreshed slope threshold for a person;

[0040] Step 713: Determine whether there is a person on the sensor according to the refreshed slope threshold for a person:

[0041] If the sensor is determined to be in an unmanned state, the second adjustment coefficient is continuously used to refresh the slope threshold for the presence of a person.

[0042] Combined with the method described in the second aspect, in the fourth possible implementation manner, the adjustable range of the slope threshold for the presence of a person is -0.75 to -3.

[0043] Implementing the method for detecting whether there is a person on the sensor based on the ballistocardiogram signal of the present invention, by collecting the time-domain / frequency-domain raw signals of the ballistocardiogram signal, dynamically refreshing the judgment threshold, increases the robustness of the algorithm, effectively suppresses the high-frequency interference that the sensor is susceptible to from other devices, solves the problem that the signal amplitude and signal standard deviation will be affected by it, and improves the accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0045] Figure 1 It is the first schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0046] Figure 2 It is the second schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0047] Figure 3 It is the third schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0048] Figure 4 It is the fourth schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0049] Figure 5 It is the fifth schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0050] Figure 6 It is the sixth schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0051] Figure 7 It is the seventh schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention;

[0052] Figure 8 It is the eighth schematic diagram of a method for detecting whether there is a person on the sensor based on the ballistocardiogram signal in the present invention. Detailed implementation manners

[0053] The technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0054] In the existing technology of using sensors to determine whether there is someone in the bed, the sensors are easily affected by high-frequency interference from other devices, resulting in the signal amplitude and signal standard deviation being affected, thus causing the method to make incorrect judgments.

[0055] In view of the above problems, a method and system for detecting whether there is someone on a sensor based on ballistocardiogram signals are proposed.

[0056] The first implementation manner

[0057] Embodiment 1

[0058] A method for detecting whether there is someone on a sensor based on ballistocardiogram signals, as Figure 1 , Figure 1 is the first schematic diagram of a method for detecting whether there is someone on a sensor based on ballistocardiogram signals in the present invention, including:

[0059] Step 100: Collect the original ballistocardiogram signals; Step 200: Obtain the high-frequency signals according to the original signals; Step 300: Calculate the amplitude ratio between the original signals and the high-frequency signals; Step 400: If all the amplitude ratios within a specified time are greater than the human presence amplitude threshold, it is determined that there is someone on the sensor. If all the amplitude ratios within a specified time are less than the human presence amplitude threshold, it is determined that there is no one on the sensor. Among them, Step 400 includes: Step 410: If the amplitude ratios are all less than the first specified value and the first coefficient of variation is less than the second specified value, the human presence amplitude threshold is refreshed. By collecting the sequential original signals of the ballistocardiogram signals and dynamically refreshing the judgment threshold, the robustness of the algorithm is increased, the high-frequency interference from other devices that the sensor is easily affected by is effectively suppressed, the problem that the signal amplitude and signal standard deviation are affected by it is solved, and the accuracy is improved.

[0060] In this embodiment, a signal acquisition unit can be used to collect the ballistocardiogram signals to obtain the original ballistocardiogram signal data sequence, that is, the original signals.

[0061] The high-frequency signals can be obtained by filtering the collected ballistocardiogram signals using a signal filtering unit. Preferably, the original signals above 15HZ are retained, that is, the original signals are subjected to high-pass filtering.

[0062] Preferably, asFigure 2 , Figure 2 It is the second schematic diagram of a method for detecting whether there is a person on a ballistocardiogram signal detection sensor in the present invention; Step 200 includes: Step 210, filtering the original signal of the ballistocardiogram signal within a specified time using finite impulse response filtering or zero-phase shift infinite impulse response filtering or wavelet filter filtering; Step 220, retaining the original signal above a specified frequency value within a specified time to obtain a high-frequency signal.

[0063] Finite impulse response filtering can be used for high-pass filtering to obtain a high-frequency signal.

[0064] In this embodiment, preferably, 10 seconds of original signal is collected, and 10 seconds of high-frequency signal is also obtained, and the original signal and the high-frequency signal are time-aligned and equal.

[0065] When calculating the amplitude, preferably, 10 discrete amplitudes of the 10-second original signal and 10 discrete amplitudes of the 10-second high-frequency signal can be selected, and the amplitude ratio between the original signal and the high-frequency signal is compared with the human presence amplitude threshold to determine whether there is a person on the sensor.

[0066] It should be noted that the amplitude can be calculated for every second of the 10 seconds, or the discrete amplitude within other time periods can be calculated. For example, it is calculated in units of 100 milliseconds for 1 second, and it is not limited here.

[0067] The human presence amplitude threshold is the human presence threshold regarding the signal amplitude. When making a determination, if all the amplitude ratios of the 10 seconds are greater than the human presence amplitude threshold, it is determined that there is a person on the sensor; if all the amplitude ratios of the 10 seconds are less than the human presence amplitude threshold, it is determined that there is no person on the sensor.

[0068] However, since during the sensor detection process, it is easily affected by high-frequency interference from other devices in the environment, which often leads to misjudgment. To solve such problems, the human presence amplitude threshold needs to be continuously updated.

[0069] Preferably, as Figure 4 , Figure 4 It is the fourth schematic diagram of a method for detecting whether there is a person on a ballistocardiogram signal detection sensor in the present invention; Step 410 includes: Step 411, obtaining the first average value of all amplitude ratios less than the first specified value; Step 412, multiplying the first average value by the first adjustment coefficient to obtain the refreshed human presence amplitude threshold; Step 413, judging whether there is a person on the sensor according to the refreshed human presence amplitude threshold: if the sensor is determined to be in the no-person state, the human presence amplitude threshold is continuously refreshed using the first adjustment coefficient.

[0070] When updating the human presence amplitude value, specifically, it can be assumed that the human presence amplitude threshold is 2. If there are 10 discrete amplitude ratios in the range [2, 2.5], and the coefficient of variation of these 10 amplitude ratios is less than the second specified value, and the second specified value is preferably 0.2, then it is necessary to adjust the human presence amplitude threshold using the first adjustment coefficient. Preferably, the first adjustment coefficient is 1.5.

[0071] When making the adjustment, first calculate the first average value of all amplitude ratios within the range [2, 2.5], multiply the first average value by the first adjustment coefficient, and use the result as the refreshed human presence amplitude threshold; determine whether there is a person on the sensor based on the refreshed human presence amplitude threshold. If the sensor is determined to be in the no-person state, continuously refresh the human presence amplitude threshold using the first adjustment coefficient. Preferably, the adjustable range of the human presence amplitude is 1.5 - 4.5.

[0072] Embodiment 2

[0073] Different from Embodiment 1, in this embodiment, as Figure 3 , Figure 3 is the third schematic diagram of a method for detecting whether there is a person on a sensor based on ballistocardiogram signals in the present invention; Step 200 includes: Step 230, filtering the original signal of the ballistocardiogram signal within a specified time using a zero-phase shift infinite impulse response; Step 240, retaining the original signal above the specified frequency value within the time period to obtain a high-frequency signal.

[0074] In this embodiment, a zero-phase shift infinite impulse response can be used for high-pass filtering to obtain a high-frequency signal.

[0075] Second method implementation

[0076] Embodiment 1

[0077] A method for detecting whether there is a person on a sensor based on ballistocardiogram signals, such as Figure 5 , Figure 5It is the fifth schematic diagram of a method for detecting whether there is a person on a ballistocardiogram signal detection sensor in the present invention; it includes: Step 500, collecting the original ballistocardiogram signal; Step 600, obtaining the slope value after fitting a straight line to the discrete power spectrum within a unit time according to the original signal; Step 700, determining whether there is a person on the sensor according to the discrete power spectrum slope value and the slope threshold for presence of a person: if all the discrete power spectrum slope values within the specified time are greater than the slope threshold for presence of a person, it is determined that there is a person on the sensor, and if all the discrete power spectrum slope values within the specified time are less than the slope threshold for presence of a person, it is determined that there is no person on the sensor; wherein, Step 700 includes: Step 710, if the discrete power spectrum slope values are all greater than a third specified value and the second coefficient of variation is less than a fourth specified value, then refresh the slope threshold for presence of a person. By collecting the frequency-domain original signal of the ballistocardiogram signal and dynamically refreshing the judgment threshold, the robustness of the algorithm is increased, the high-frequency interference of the sensor by other devices is effectively suppressed, the problem that the signal amplitude and the signal standard deviation are affected by it is solved, and the accuracy is improved.

[0078] In this embodiment, a method for detecting whether there is a person on a ballistocardiogram signal detection sensor is provided in the frequency domain. The discrete power spectrum slope value can be: Similarly, collect the original ballistocardiogram signal data sequence, calculate the power spectrum within 1 second, and retain the part less than 40HZ. Fit a straight line to the original signal power spectrum and calculate the slope value of the straight line. In this way, 10 slope values are collected respectively in units of 1 second within 10 seconds.

[0079] In this embodiment, the slope threshold for presence of a person is the threshold for presence of a person regarding the signal slope. When making a determination, if all the slope ratios within 10 seconds are greater than the slope threshold for presence of a person, it is determined that there is a person on the sensor, and if all the amplitude ratios within 10 seconds are less than the slope threshold for presence of a person, it is determined that there is no person on the sensor.

[0080] Preferably, as Figure 6 , Figure 6 It is the sixth schematic diagram of a method for detecting whether there is a person on a ballistocardiogram signal detection sensor in the present invention; Step 600 includes: Step 610, calculating the original signal power spectrum within a unit time; Step 620, retaining the original signal power spectrum less than the second frequency value.

[0081] Step 600 further includes: Step 630, fitting a straight line to the original signal power spectrum; Step 640, calculating the slope value of the straight line fitted to all the discrete power spectra within the specified time.

[0082] In this embodiment, the slope threshold for detecting a person is defaulted to -0.5. However, during the sensor detection process, it is vulnerable to high-frequency interference from other devices in the environment, often leading to misjudgment. To solve such problems, the slope threshold for detecting a person needs to be continuously updated.

[0083] Specifically, obtain the second average value of all discrete power spectrum slope values greater than the third specified value; multiply the second average value by the second adjustment coefficient to obtain the refreshed slope threshold for detecting a person; multiply the second average value by the second adjustment coefficient to obtain the refreshed slope threshold for detecting a person; determine whether there is a person on the sensor according to the refreshed slope threshold for detecting a person: if the sensor is determined to be in the state of no person, continuously refresh the slope threshold for detecting a person using the second adjustment coefficient.

[0084] Preferably, as Figure 7 , Figure 7 is the seventh schematic diagram of a method for detecting whether there is a person on a sensor based on ballistocardiogram signals in the present invention; the step 600 further includes: step 630, fitting the original signal power spectrum with a straight line; step 640, calculating the discrete power spectrum slope values of all straight lines within a specified time period.

[0085] Preferably, as Figure 8 , Figure 8 is the eighth schematic diagram of a method for detecting whether there is a person on a sensor based on ballistocardiogram signals in the present invention; the step 710 includes: step 711, obtaining the second average value of all discrete power spectrum slope values greater than the third specified value; step 712, multiplying the second average value by the second adjustment coefficient to obtain the refreshed slope threshold for detecting a person; step 713, determining whether there is a person on the sensor according to the refreshed slope threshold for detecting a person: if the sensor is determined to be in the state of no person, continuously refresh the slope threshold for detecting a person using the second adjustment coefficient.

[0086] When updating the slope threshold for detecting a person, specifically, it can be assumed that the slope threshold for detecting a person is -0.5. If 10 discrete slope values are greater than -0.5 and the coefficient of variation of these 10 slope values is less than the third specified value, and the third specified value is preferably 0.2, then the second adjustment coefficient needs to be used to adjust the slope threshold for detecting a person. Preferably, the second adjustment coefficient is 1.5.

[0087] When making the adjustment, first obtain the second average value of all slope values greater than -0.5, multiply the second average value by the second adjustment coefficient, and the result is used as the refreshed slope threshold for detecting a person; determine whether there is a person on the sensor according to the refreshed slope threshold for detecting a person. If the sensor is determined to be in the state of no person, continuously refresh the slope threshold for detecting a person using the second adjustment coefficient. Preferably, the adjustable range of the slope threshold for detecting a person is -0.75 to -3.

[0088] Implement the method of detecting whether there is a person on the ballistocardiogram signal detection sensor of the present invention. By collecting the time-domain / frequency-domain original signals of the ballistocardiogram signals and dynamically refreshing the judgment threshold, the robustness of the algorithm is increased, the high-frequency interference that the sensor is susceptible to other devices is effectively suppressed, the problem that the signal amplitude and the signal standard deviation are affected by it is solved, and the accuracy is improved.

[0089] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for detecting whether there is a person on a ballistocardiogram signal detection sensor, characterized in that, Including: Step 500: Collect the original ballistocardiogram signal; Step 600: Obtain the slope value after fitting the discrete power spectrum within a unit time with a straight line according to the original signal; Step 700: Determine whether there is a person on the sensor according to the discrete power spectrum slope value and the slope threshold for a person: If all the discrete power spectrum slope values within a specified time are greater than the slope threshold for a person, it is determined that there is a person on the sensor; if all the discrete power spectrum slope values within a specified time are less than the slope threshold for a person, it is determined that there is no person on the sensor; Among them, the said Step 700 includes: Step 710: If all the discrete power spectrum slope values are less than a third specified value and the second coefficient of variation of the discrete power spectrum slope values is less than a fourth specified value, refresh the slope threshold for a person.

2. The method for detecting whether there is a person on the detection sensor according to claim 1, characterized in that, The said Step 600 includes: Step 610: Calculate the power spectrum of the original signal within a unit time; Step 620: Retain the power spectrum of the original signal that is less than a second frequency value.

3. The method for detecting whether there is a person on the detection sensor according to claim 2, wherein The said Step 600 further includes: Step 630: Fit the power spectrum of the original signal with a straight line; Step 640: Calculate the slope value of the straight line fitted by all the discrete power spectra within a specified time.

4. The method for detecting whether there is a person on the detection sensor according to claim 3, wherein, The said Step 710 includes: Step 711: Obtain the second average value of all the discrete power spectrum slope values greater than a third specified value; Step 712: Multiply the second average value by a second adjustment coefficient to obtain the refreshed slope threshold for a person; Step 713: Judge whether there is a person on the sensor according to the refreshed slope threshold for a person: If the sensor is determined to be in a state of no person, continuously refresh the slope threshold for a person by using the second adjustment coefficient.

5. The method for detecting whether there is a person on the detection sensor according to any one of claims 1-3, characterized in that, The adjustable range of the slope threshold for a person is -0.75 to -3.

Images