Device and method for detecting fall

Abstract

To provide a fall detection device capable of accurately detecting a signal relevant to a fall from radar signals.SOLUTION: A fall detection device according to an embodiment of the present invention comprises: a transceiver configured to transmit a radar signal toward a subject and receive radar signals reflected from the subject; an observation determination unit configured to determine whether or not to enter an observation mode for the subject by analyzing the received radar signals; and a fall determination unit configured to derive a range value between the transceiver and the subject when entering in the observation mode and determine whether the subject has fallen or not based on the range value.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an apparatus and method for detecting a fall injury to a target object using a radar signal. [Background technology]

[0002] Falling accidents are common accidents that can happen to anyone, regardless of age or gender, and as we move into an aging society, the incidence of falling accidents among the elderly is increasing. In particular, elderly people living alone may suffer serious injuries or even death from falling. Therefore, it is important to quickly detect such falling accidents and respond appropriately.

[0003] Sensor technologies used in conventional fall detection methods include acceleration sensors, tilt sensors, geomagnetic sensors, and cameras. Of these, acceleration sensors and tilt sensors are inconvenient because they must be attached to or carried by the user to detect sudden movements or changes in posture at the moment of a fall, which can be a burden to the user. Geomagnetic sensors operate in a similar manner, but both of these sensors have the limitation that they must be worn or carried by the user at all times. Furthermore, camera-based detection methods pose a major problem in protecting personal privacy. While cameras can detect fall injuries relatively accurately, they can infringe on personal privacy, and their use is particularly limited in private spaces.

[0004] While these existing sensors provide a certain level of functionality for fall detection, there is room for improvement in terms of ease of use, protection of the user's privacy, detection accuracy, and reliability. For elderly users and users with mobility issues, constantly wearing or carrying a sensor can be a significant burden, which can hinder the continued use of fall detection systems. Therefore, there is a need to develop a new fall detection method that overcomes the limitations of existing methods and is more user-friendly, respects the user's privacy, and provides high detection accuracy.

[0005] Radar technology can detect targets' distance and speed and can be used to detect falls. However, because radar sensors detect all moving objects, technology is needed to distinguish between falls and other related movements in environments such as toilets, where water movement can be confused with falls. Existing radar-based fall detection methods have limited ability to distinguish between these movements in such environments, which is why more sophisticated processing methods and algorithms are needed. Summary of the Invention [Problem to be solved by the invention]

[0006] SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a fall detection device that can accurately detect a fall-related signal among radar signals.

[0007] The present invention provides a device that can detect whether a subject has fallen, protecting the subject's privacy by detecting a fall using a radar signal.

[0008] The present invention provides a device that can detect whether a fall has occurred by detecting a fall using a radar signal, regardless of whether the user is wearing or carrying a detection device.

[0009] However, the technical objectives to be achieved by this embodiment are not limited to the above-mentioned technical objectives, and other technical objectives may also exist. [Means for solving the problem]

[0010] As a means for achieving the above-mentioned technical object, one embodiment of the present invention provides an apparatus for detecting a fall using a radar signal, the apparatus including: a transceiver unit that transmits a radar signal toward a target object and receives the radar signal reflected from the target object; an observation determination unit that analyzes the received radar signal and determines whether to enter an observation state for the target object; and a fall determination unit that derives a distance value between the transceiver unit and the target object when the observation state is entered, and determines whether the target object has fallen based on the distance value.

[0011] The above-described solutions are merely exemplary and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, there may be additional embodiments as described in the drawings and detailed description of the invention. [Effects of the Invention]

[0012] According to any one of the above-mentioned means for solving the problem of the present invention, the present invention can improve the accuracy of detecting a fall by distinguishing and detecting a signal related to a fall among radar signals.

[0013] The present invention has the effect of detecting whether or not a fall has occurred while protecting privacy by detecting a fall using a radar signal without directly photographing the object or recording sound.

[0014] The present invention detects a fall using a radar signal, so that whether or not a fall has occurred can be detected regardless of whether the user is wearing a detection device or not, thereby increasing the convenience of detecting a fall and reducing the burden on the user.

[0015] However, the technical objectives to be achieved by this embodiment are not limited to the above-mentioned technical objectives, and other technical objectives may exist. [Brief explanation of the drawings]

[0016] [Figure 1]1 is a block diagram of a fall detection system according to an embodiment of the present invention; [Figure 2] 1 is a diagram illustrating the configuration of a fall detection device according to an embodiment of the present invention. [Figure 3] 10 is an exemplary diagram illustrating a process of generating distance-Doppler map information according to an embodiment of the present invention. [Figure 4a] 4 is an exemplary diagram illustrating a process of deriving a peak signal according to an embodiment of the present invention; [Figure 4b] 4 is an exemplary diagram illustrating a process of deriving a peak signal according to an embodiment of the present invention; [Figure 5a] 4 is an exemplary diagram illustrating a process of deriving an operational peak signal according to an embodiment of the present invention; [Figure 5b] 4 is an exemplary diagram illustrating a process of deriving an operational peak signal according to an embodiment of the present invention; [Figure 6] 10 is an exemplary diagram illustrating a process of analyzing a pattern of an operational peak signal according to an embodiment of the present invention; [Figure 7a] 10 is an exemplary diagram illustrating a process of deriving a trend line according to an embodiment of the present invention. FIG. [Figure 7b] 10 is an exemplary diagram illustrating a process of deriving a trend line according to an embodiment of the present invention. FIG. [Figure 8] 10 is an exemplary diagram illustrating a process of deriving a distance value according to an embodiment of the present invention; [Figure 9] 10 is an exemplary diagram illustrating a process for determining whether or not there is a fall according to an embodiment of the present invention; [Figure 10] 10 is an exemplary diagram illustrating a process for determining whether or not there is a fall according to an embodiment of the present invention; [Figure 11] 1 is a flowchart illustrating a method for detecting a fall according to an embodiment of the present invention. [Figure 12] 1 is a flowchart illustrating a method for detecting a fall according to an embodiment of the present invention. [Figure 13] 1 is a flowchart illustrating a method for detecting a fall according to an embodiment of the present invention. [Figure 14] 10 is an exemplary diagram illustrating parameters used in a process of determining whether or not there is a fall according to an embodiment of the present invention. FIG. [Figure 15] 10 is an exemplary diagram illustrating parameters used in a process of determining whether or not there is a fall according to an embodiment of the present invention. FIG. [Figure 16] 1 is a flowchart illustrating a method for detecting a fall according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0017] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] The present invention will now be described in detail with reference to the accompanying drawings so that those skilled in the art can easily understand and practice the present invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. In the drawings, parts that are not relevant to the description are omitted in order to clearly explain the present invention, and like reference numerals are used to refer to like parts throughout the specification.

[0018] Throughout the specification, when a part is said to be "connected" to another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. Furthermore, when a part is said to "comprise" a certain component, this does not mean excluding other components, but may further include other components, unless otherwise specified, and should be understood as not precluding the presence or possibility of addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0019] In the present specification, the term "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both hardware and software. Also, one unit may be realized using two or more pieces of hardware, and two or more units may be realized by one piece of hardware.

[0020] In the present specification, some of the operations and functions described as being performed by a terminal or device may instead be performed by a server connected to the terminal or device, and similarly, some of the operations and functions described as being performed by a server may instead be performed by a terminal or device connected to the server.

[0021] The functionality provided by the components described herein may be embodied in processing circuitry, including general-purpose processors, special-purpose processors, integrated circuits, ASICs (Application Specific Integrated Circuits), CPUs (Central Processing Units), circuits, and / or combinations thereof, programmed to provide the functionality. A processor includes transistors and other circuitry and is considered a circuit or processing circuit. A processor may also be a programmed processor that executes a program stored in a memory.

[0022] In the present specification, a circuit, a part, a unit, or a means is hardware that is programmed to realize or executes the described functions, which may be all hardware disclosed in the present specification or any hardware known to be programmed to realize or execute the functions.

[0023] When the hardware is a processor considered to be a circuit type, the circuit, the part, means or unit is a combination of hardware and software used to configure the hardware and / or processor.

[0024] Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[0025] FIG. 1 is a diagram showing the configuration of a fall detection system according to an embodiment of the present invention.

[0026] As shown in FIG. 1, the disease diagnosis system 1 may include a fall injury detection device 100 and a radar 110 .

[0027] The components of the fall detection system 1 in Fig. 1 are generally connected via a network. For example, as shown in Fig. 1, the fall detection device 100 and the radar 110 may be connected simultaneously or at intervals.

[0028] A network refers to a connection structure that enables information exchange between nodes such as terminals and servers, and includes a local area network (LAN), a wide area network (WAN), the Internet (WWW), a wired and wireless data communication network, a telephone network, a wired and wireless television communication network, etc. Examples of wireless data communication networks include, but are not limited to, 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi, Bluetooth communication, infrared communication, ultrasonic communication, visible light communication (VLC), LiFi, etc.

[0029] The fall detection device 100 can analyze a radar signal reflected from a target object 111 using a radar 110. If the target object 111 falls while active in an indoor space (e.g., a shower room, a toilet, a workshop, or a hallway), the fall detection device 100 can detect whether the target object 111 has fallen or not based on the changed radar signal. For example, the fall detection device 100 is placed on the ceiling of a shower room to be spaced apart from the target object 111 in the shower, and can transmit a radar signal toward the target object 111 through the radar 110 and receive the radar signal reflected from the target object 111.

[0030] The fall detection device 100 can determine whether the target object 111 is falling or not using the radar 110.

[0031] Therefore, by using the fall detection device 100, it is possible to accurately determine whether the target object 111 is falling or has fallen, even if the target object 111 carries or does not wear a separate device for detecting falls.

[0032] The fall detection device 100 may have a radar 110 attached thereto or may have a radar built therein to detect a fall of the target object 111. In addition, at least some elements of the fall detection device 100 may be disposed in a space separated from the radar 110 and may communicate with the radar 110 wirelessly or via a wired connection via a network to detect a fall of the target object 111.

[0033] Each component of the fall injury detection device 100 will now be described.

[0034] FIG. 2 is a diagram showing the configuration of a fall detection device according to an embodiment of the present invention.

[0035] 2, the fall detection device 100 may include a transceiver unit 210, a peak signal deriving unit 220, an observation determining unit 230, and a fall determination unit 240. However, the above components 210 to 240 are merely illustrative examples of components that may be controlled by the fall detection device 100.

[0036] The transceiver 210 transmits a radar signal toward a target object and receives the radar signal reflected from the target object.

[0037] The peak signal derivation unit 220 can derive the peak signal by filtering the radar signal.

[0038] The observation determining unit 230 may analyze the received radar signal and determine whether or not to enter an observation state for the target object.

[0039] When entering the observation state, the damage determining unit 240 derives a distance value between the transceiver unit 210 and the target object, and determines whether the target object has fallen based on the distance value.

[0040] FIG. 3 is an exemplary diagram illustrating a process of generating range-Doppler map information according to an embodiment of the present invention.

[0041] 3 shows a process of generating distance-Doppler map information from a radar signal in order to derive a peak signal from the radar signal by the peak signal deriving unit 220. The peak signal deriving unit 220 can generate distance-Doppler map information through a pre-processing process of processing the radar signal, which is raw data acquired from the radar.

[0042] The peak signal deriving unit 220 receives a radar signal reflected from a target object and samples the received radar signal for each chirp using an ADC (Analog-to-Digital Converter) to generate a digital signal 310. In this process, the sampled data for each chirp forms the structure of the original radar signal.

[0043] Then, the peak signal deriving unit 220 can apply a two-dimensional fast Fourier transform (2D-FFT) to the digital signal. Through this Fourier transform, the digital signal is converted from the time domain to the range-Doppler domain. The first Fourier transform is a process of converting the time domain signal to the range domain 320 to extract distance information to each target object, and the second Fourier transform is a process of analyzing changes in Doppler frequency for each distance to obtain velocity information 330 of the target object.

[0044] In a range-Doppler map, signal strength with distance can be displayed on the horizontal axis and Doppler frequency on the vertical axis.

[0045] The converted distance-Doppler map information shows the distance and speed information of the object that reflected the radar signal as a 2D image, and can provide important information for detecting specific events such as falls.Furthermore, the distance-Doppler map information is analyzed by a fall detection algorithm and used to determine whether a fall accident has occurred.

[0046] 4a and 4b are exemplary diagrams illustrating a process of deriving a peak signal according to an embodiment of the present invention.

[0047] The peak signal deriving unit 220 may generate range-Doppler map information based on the radar signal and derive the peak signal by filtering the range-Doppler map information. Here, the range-Doppler map information may be generated for each frame corresponding to the time point when the transceiver unit 210 receives the radar signal. That is, the range-Doppler map information may include at least one range-Doppler map corresponding to each frame.

[0048] The peak signal deriving unit 220 can remove noise from the range-Doppler map and derive a signal having an intensity equal to or greater than a predetermined threshold value as a peak signal.

[0049] In this case, the peak signal derivation unit 220 can derive a peak signal by applying a Constant False Alarm Rate (CFAR) algorithm. The peak signal derivation unit 220 can analyze the level of background noise around the target object 111 through the CFAR algorithm. Furthermore, through this, it can select a surrounding area that is not the target object and calculate the average signal strength of the selected area. Then, the peak signal derivation unit 220 can dynamically determine a threshold value for distinguishing the signal for the target object from noise based on the calculated average noise level. Here, the threshold value is adjusted in proportion to the surrounding noise level, so that consistent signal detection performance can be maintained even in various environments.

[0050] In addition, the peak signal derivation unit 220 can derive a peak signal from within the range-Doppler map using a thresholding method utilizing a histogram, in addition to the above-mentioned CFAR algorithm.

[0051] The peak signal derived by the peak signal deriving unit 220 can be filtered for power, range, and Doppler to obtain a desired peak signal, which can be used as feature information for pattern recognition.

[0052] Thereafter, the peak signal deriving unit 220 scans the entire range-Doppler map and regards signals exceeding the set threshold as peak signals. Through this process, peak signals corresponding to the actual position and velocity information of the target object 111 can be derived.

[0053] FIG. 4a illustrates an example of distance-Doppler map information generated when a target object falls, and FIG. 4b illustrates an example of a distance-Doppler map generated when a certain movement other than a fall (e.g., water sprayed from a shower falling to the floor of a room) occurs within a space where a fall detection device is installed.

[0054] 4a or 4b illustrates one frame among a plurality of distance-Doppler maps generated for each frame. For example, FIG. 4a may represent a distance-Doppler map corresponding to one frame among frames corresponding to a time when a target object falls and is injured. FIG. 4b may represent a distance-Doppler map corresponding to one frame among frames corresponding to a time when water falls while the target object is taking a shower without falling. In FIG. 4a or 4b, peak signals may be indicated by bold dots.

[0055] In Figure 4a, when a crash occurs, a peak signal 410 appears in a relatively narrow Doppler bin range and a wide range bin range. On the other hand, in Figure 4b, when a crash does not occur but another movement occurs, a peak signal 420 appears in a relatively wide Doppler bin range and a narrow range bin range.

[0056] FIG. 5a or 5b is an exemplary diagram illustrating a process of deriving an operational peak signal according to an embodiment of the present invention.

[0057] The observation determining unit 230 derives a motion peak signal related to the motion of the target object 111 from the radar signal, analyzes the pattern of the motion peak signal, and determines whether to enter the observation state based on the pattern analysis result.

[0058] Here, the operational peak signal may be derived by filtering at least some Doppler bin regions of the radar signal or based on the distribution of the radar signal.

[0059] As shown in FIG. 4a or 4b, the distribution pattern of the peak signals changes depending on whether the target object 111 has fallen or been injured. FIG. 5a illustrates a peak signal distribution 510 in the situation shown in FIG. 4a. FIG. 5b illustrates a peak signal distribution 530 in the situation shown in FIG. 4b. The observation decision unit 230 can derive, from these peak signals, a motion peak signal that is more closely related to the motion of the target object 111. To this end, the observation decision unit 230 can derive, as the motion peak signal, a peak signal that is more closely related to the motion of the target object 111, excluding peak signals corresponding to a certain Doppler bin range 520 in the distance-Doppler map. That is, the observation decision unit 230 can derive only peak signals related to the motion of the target object 111 from among many motions that occur while the target object 111 is taking a shower in a shower room. To this end, the observation decision unit 230 can derive, as the motion peak signal of the target object, a peak signal that is more closely related to the motion of the target object, excluding a Doppler bin range 520 related to the motion of water. In this case, the Doppler bin range 520 to be excluded can be arbitrarily determined based on the designer's intention or an experimentally derived value.

[0060] Meanwhile, the observation determining unit 230 can determine the distribution pattern of the peak signals according to the situation and derive the operational peak signals by filtering out peak signals corresponding to noise from among the peak signals.

[0061] FIG. 6 is an exemplary diagram illustrating a process of analyzing a pattern of an operational peak signal according to an embodiment of the present invention.

[0062] Figure 6 illustrates range-Doppler maps generated continuously over time. Range-Doppler maps 610, 620, 630, and 640 in Figure 6 illustrate the distribution changes of motion peak signals 611, 621, 631, and 641 over time when a fall injury occurs. More specifically, the examples in Figure 6 illustrate range-Doppler map 610 corresponding to the 124th frame, range-Doppler map 620 corresponding to the 131st frame, range-Doppler map 630 corresponding to the 140th frame, and range-Doppler map 640 corresponding to the 143rd frame, respectively.

[0063] As can be seen from each distance-Doppler map shown in Fig. 6, as time passes, the main distribution of the motion peak signal moves from a small distance bin range to a large distance bin range. In other words, within a certain range of Doppler bin range, the distance bins tend to increase over time.

[0064] FIG. 7a or 7b is an exemplary diagram illustrating a process of deriving a trend line according to an embodiment of the present invention.

[0065] The observation determination unit 230 derives a distance index of the motion peak signal over time and derives a trend line based on the amount of change in the distance index, thereby analyzing the pattern of the motion peak signal. In this case, the trend line can be derived by applying the least squares method to the amount of change in the distance index.

[0066] Furthermore, the observation determination unit 230 can determine to enter the observation state if the slope of the trend line is equal to or greater than a predetermined value.

[0067] 7a or 7b illustrates a two-dimensional coordinate system in which the horizontal axis represents time and the vertical axis represents a range index, and in this coordinate system, each point 710 represents a motion peak signal. That is, FIG. 7a or 7b shows the distribution of motion peak signals over time. In FIG. 7a, each point 710 may correspond to the result of displaying all range bins of a moving object over time in a range-Doppler map.

[0068] As shown in Figure 7a, it can be seen that the motion peak signal is distributed in an increasing manner from the point where the time value is greater than about 33. In other words, the point where the time value is greater than 33 can be determined as a fall-suspected section 720. In Figures 7a and 7b, the time axis may be in units of seconds or frames.

[0069] 7b is a diagram illustrating an example of trend lines derived based on the distance index values ​​of each motion peak signal in a suspected fall section 720. A first trend line 730 is a trend line derived based on all motion peak signals, and a second trend line 740 is an example of a trend line derived using only the motion peak signal with the smallest distance index value among the motion peak signals.

[0070] Due to the characteristics of radar, even a single target may contain multiple peak signals in the received data due to multiple reflections of electromagnetic waves. For more accurate analysis, the following explanation will be made assuming the use of the second trend line.

[0071] The observation decision unit 230 may derive a trend line by considering the distribution of each motion peak signal from the time-distance index coordinate system. Such a trend line may be derived based on the fact that when a fall occurs on the target object, the distance between the radar and the target object increases. Using this characteristic, the observation decision unit 230 may suspect that a fall on the target object has occurred and may decide to enter an observation state, which will be described later.

[0072] The slope of the trend line can be calculated using the following formula (1).

number

[0073] In the above, a means the slope of the trend line, x means the time value, and y means the distance index value.

[0074] The more the slope value of the trend line increases in the positive direction, the faster the target object moves away from the radar, which means that the possibility of a fall is high. Therefore, if the slope value of the trend line exceeds a predetermined value, the observation decision unit 230 may determine to enter an observation state, which will be described later, because a fall is suspected. Here, the standard of the slope value of the trend line, which is the criterion for entering the observation state, may be arbitrarily determined based on the designer's intention or an experimentally derived value.

[0075] FIG. 8 is an exemplary diagram illustrating a process of deriving a distance value according to an embodiment of the present invention.

[0076] The fall damage determination unit 240 derives the distance value between the radar and the target object in the observation state, derives the fall damage range according to the height of the space in which the radar is installed, and can determine whether the target object has fallen based on the distance value and the fall damage range.

[0077] Here, the distance value may refer to a value derived from the RMS (Root Mean Square) power obtained by filtering a signal corresponding to the movement of the target object from the radar signal, moving a window corresponding to the time domain from the filtered signal.

[0078] Furthermore, the damage range may refer to the range between a long-distance critical point corresponding to the distance from the radar position to the bottom of the space and a short-distance critical point corresponding to a position a predetermined distance away from the long-distance critical point in the direction in which the radar is installed.

[0079] The fall damage determination unit 240 determines whether or not a fall damage has occurred if the distance value is smaller than the critical point for re-entry or the average distance is greater than the critical point for absence, and the critical point for re-entry can be derived based on the distance value derived when the target object is re-entering the space, and the critical point for absence can be derived based on the distance value derived when the target object is absent from the space.

[0080] 8 is a diagram illustrating a process in which the damage determination unit 240 derives a distance value when entering an observation state. The damage determination unit 240 can derive distance bin values ​​810 derived by time from the radar signal. The distance bin values ​​810 may correspond to the result of displaying signal strength on the distance axis of a zero Doppler region where a stationary object exists in chronological order. A stationary object should be measured at a constant signal strength over time. However, if another object with slight movement exists, the zero Doppler region will also change due to the movement of the object.

[0081] Furthermore, the damage determination unit 240 extracts only the magnitude component of the zero Doppler signal for the derived range bin value 810, and then performs primary filtering in a manner that removes meaningless components in a virtual sensing system due to stationary targets or clutter in the time domain, resulting in the result 820. By removing the clutter, signal changes over time in the zero Doppler domain can be detected.

[0082] Thereafter, the injury determination unit 240 can output a result 830 of deriving Root Mean Squared (RMS) power in the time domain for the primary filtering result 820. Furthermore, for the derived result 830, it can output a result 840 of deriving an average distance, i.e., a distance value, based on the distance bins.

[0083] The steps 810 to 840 in which the injury determining unit 240 derives the distance value will be described in more detail below.

[0084] The damage determination unit 240 can extract the magnitude component of the signal from a zero Doppler signal, which is a signal returned when the target object is not moving relative to the radar, i.e., when the velocity is 0. Furthermore, the damage determination unit 240 can remove clutter to filter components corresponding to background signals and noise in order to analyze only the changing parts of the radar signal. Through this, the damage determination unit 240 can remove background signals from the derived distance bin value 810, leaving only objects that vibrate or move with the body. In this way, the process of removing clutter can be defined by the following (Equation 2).

number

[0085] Here, y[i] means a value from which clutter has been removed at time point i, x[i] means the magnitude of the input signal at time point i, and α means a tuning parameter of the system. More specifically, x[i] means the input signal at the current time point, and x[i-1] means the input signal at the previous time point. The output signal from which clutter has been removed can be obtained through Equation 2 above.

[0086] Thereafter, the damage determination unit 240 sets a window size corresponding to a certain period in the time domain and calculates the root mean squared power (RMS power) within each window while moving the set window. The RMS power indicates the average power of the radar signal and can reflect the fluctuation of the signal over time.

[0087] The process for determining the RMS power can be defined by the following (Equation 3).

number

[0088] Here, y[i] denotes the RMS power output at time i, and x[i] denotes the value input at time i. In this case, x[i] denotes the value from which clutter has been removed (y[i] in Equation 2) derived through Equation 2. Furthermore, N denotes a value corresponding to the window size. Using the window size (N) set in the time domain, the average of the squared input signal values ​​can be calculated. Thereafter, the RMS power value can be calculated by calculating the square of the calculated average value. That is, the RMS power value, which is a measure of the average size of a signal, can be derived through Equation 3.

[0089] Thereafter, the damage determining unit 240 can calculate the average RMS power in the distance bins to calculate the average signal strength for each distance interval.

[0090] The process of calculating the average RMS power can be defined by the following (Equation 4).

number

[0091] Here, y[i] denotes the average RMS power at time point i, and x[i, n] denotes the RMS power corresponding to the nth distance bin at time point i.

[0092] Finally, the damage determination unit 240 may calculate a distance value using a Gaussian distribution on the average of the derived RMS power. Through this process, the damage determination unit 240 may obtain a distance value from the target object. By deriving the distance value using the average value of the normal distribution, the average distance between the target object and the radar may be more accurately estimated.

[0093] The process of deriving the distance value can be defined by the following (Equation 5).

number

[0094] Here, y[i] is the distance value at time point i, and x[i, n] means the RMS power corresponding to each n-th distance bin at time point i.

[0095] 9 and 10 are exemplary diagrams illustrating a process of determining whether or not there is a fall according to an embodiment of the present invention.

[0096] 9 is a diagram illustrating an example in which radar 920 of fall detection device 100 is installed on a ceiling 910 of an indoor space. Radar 920 of fall detection device 100 transmits a radar signal toward floor 940 of the indoor space and receives the reflected radar signal. In this case, the distance between ceiling 910 where radar 920 is installed and floor 940 can be defined as height 930 of the indoor space. Furthermore, fall range 960 can be defined as the range between the floor and a position 950 that is a predetermined distance away from floor 940. Fall detection device 100 can determine that a target object has fallen if the position of the target object, derived from the distance value to the target object, remains within fall range 960 for a predetermined time or more.

[0097] In the process of deriving the above-described damage range 960, the bottom 940 may be defined as a far-distance critical point, and the height of a position 950, which is a predetermined distance away from the bottom 940, may be defined as a near-distance critical point. In this case, the height of the near-distance critical point may be arbitrarily determined based on the designer's intention or an experimentally derived value.

[0098] FIG. 10 is a graph 1010 illustrating a distance value 1020 between a radar and a target object calculated by the fall damage determination unit 240 over time. FIG. 10 shows that the distance value 1020 has a value between 300 and 350 at a first time point 1030, then decreases to a value between 100 and 150 at a second time point 1040. Here, the first time point 1030 represents the time when the target object enters the indoor space where the radar is installed, and the second time point 1040 represents the time when the target object falls and is injured. FIG. 10 illustrates that the distance value 1020 to the target object changes within a certain range after the second time point 1040 when the target object falls and is injured. The fall damage determination unit 240 can determine that the target object is in a fall damage state if the distance value 1020 to the target object has a value within a fall damage range 1050 and this state continues for a predetermined time or longer.

[0099] 11 to 13 are flow charts illustrating a method for detecting a fall according to an embodiment of the present invention.

[0100] 11 to 13 are flow charts for more specifically explaining the process of detecting a fall injury to a target object by the fall injury detection device 100. For convenience of explanation, Figs. 11 to 13 are explained using predefined parameters.

[0101] 11 and 12 illustrate the operations performed by the peak signal deriving unit 220 and the observation determining unit 230 based on the radar signal received from the transceiver unit 210 of the fall detection device 100. FIG.

[0102] In step 1110, the fall detection device 100 can initialize variables and execute a main function (Sp_vMain) based on the radar signal received from the transceiver unit 210. At this time, the variables are initialized as follows: State_Observation=0, BufCount=0, Fall_Timer=FALLDET_FALL_TIMER, Fall_Count=0.

[0103] Here, State_Observation is a parameter that indicates the observation state, and indicates whether the state is entering the observation state through a value of 0 or 1. BufCount is a variable used in the fall detection algorithm, and is a value that increases if a movement that is suspicious of a fall occurs with each radar scan, and decreases in the opposite case. Fall_Timer is a parameter for the time during which the observation state is executed, and in the observation state, its value decreases with each radar scan. If FallCount does not reach Th_FallCount until the value reaches 0, it is determined that there is no fall, and the observation state can end. In the observation state, FallCount can increase or decrease depending on the situation.

[0104] In step 1121, the fall detection device 100 generates and updates distance-Doppler map information from the radar signal acquired from the transceiver unit 210 for each radar scan, and can derive a peak signal.

[0105] The fall detection device 100 may filter the updated distance-Doppler map information and derive an operational peak signal from the peak signal in step 1122. In this case, the process of deriving the operational peak signal may be performed by checking whether the value of the Doppler index (Dop_Index) is within the range between Th_lower_Didx and Th_upper_Didx and whether the power (PeakPower) of the peak signal exceeds Th_PeakPower.

[0106] Here, Th_lower_Didx and Th_upper_Didx are parameters for removing movements unrelated to falls (e.g., water movements) from the algorithm for detecting fall movements, and can be defined by their respective pre-set Doppler index values.

[0107] Th_Peak Power refers to a parameter for removing the movement of objects other than the target object, and can be defined based on a preset decibel (dB).

[0108] If the above conditions are not met, the process returns to the previous initial step 1121, and if the conditions are met, the process proceeds to the subsequent step 1123.

[0109] In step 1123, the fall detection device 100 checks whether the number of operational peak signals (filtered_PeakCount) is greater than the previously set critical number of operational peak signals (Th_PeakCount), and if the condition is met, it proceeds to the next step 1124; if not, it decrements Buf_Count and then returns to the initial step 1121.

[0110] In step 1124, the fall injury detection device 100 may select the minimum value from the filtered distance-index (R-index) data, store the distance-index data having the minimum value, and increase the value of BufCount (1124).

[0111] In step 1125, the fall injury detection device 100 checks whether the value of BufCount has reached the previously set value of Th_BufCount, and if so, proceeds to the next step 1130; if not, it decrements BufCount and returns to the initial step 1121.

[0112] Step 1130 and subsequent steps will be explained subsequently with reference to FIG.

[0113] Following step 1130, the fall injury detection device 100 can derive a trend line using the least squares method based on the minimum distance index stored in step 1210.

[0114] The fall detection device 100 may compare the slope (LeastSquares_slope) of the trendline derived in step 1220 with the magnitude of Th_slope. Here, Th_slope is a parameter related to the slope of the trendline derived through the least squares method, and if a trendline with a slope greater than this parameter is derived, it may be used as a critical point for entering the observation state. The larger the value of Th_slope, the lower the sensitivity for fall detection. If the above conditions are met, the fall detection device 100 may proceed to the next step 1230; otherwise, it may proceed to step 1240.

[0115] In step 1230, the fall detection device 100 can adjust the value of State_Observation to set the conditions for entering the observation state, and can initialize the values ​​of FallCount and Fall_Timer, which are parameters used in the observation state.

[0116] The fall detection device 100 can check the value of State_Observation to determine whether to enter the observation state in step 1240. If the value of State_Observation has a value for entering the observation state (e.g., 1), the device proceeds to the next step 1260; otherwise, the device can decrement the value of BufCount and then proceed to step 1250. In this case, step 1250 refers to step 1121 in FIG. 11.

[0117] Step 1260 and subsequent steps will be explained subsequently with reference to FIG.

[0118] FIG. 13 is a diagram illustrating an example of the operation performed by the fall determination unit 240 of the fall detection device 100 when entering the observation state.

[0119] Following step 1270, the fall detection device 100 may update related data for each radar scan in step 1310. At this time, the updated data is the distance value (AvgDistance) from the target object, and the value of Fall_Timer may be decreased each time the distance value (AvgDistance) is updated.

[0120] In step 1320, the fall detection device 100 determines whether the distance value (AvgDistance) is less than the re-entry critical point (Th_Pres) or greater than the absence critical point (Th_Absence). Here, the re-entry critical point (Th_Pres) refers to a critical point when a target object is present in the indoor space where the radar is installed and is engaged in normal activity, not a fall injury. The re-entry critical point (Th_Pres) can be defined based on the average of distance values ​​measured when a target object is present in the indoor space. The absence critical point (Th_Absence) refers to a critical point in an absence situation, where no one is present in the indoor space where the radar is installed. The absence critical point (Th_Absence) can be defined based on the average of distance values ​​measured when no target object is present in the indoor space.

[0121] If the conditions of step 1320 are met, the fall detection device 100 can proceed to step 1330. In step 1320, the fall detection device 100 initializes the FallTimer and FallCount, sets them to a non-observation state, and then proceeds to the next step 1340. On the other hand, if the conditions of step 1320 are not met, the fall detection device 100 can proceed to step 1350.

[0122] Here, step 1340 may refer to step 1121 in FIG.

[0123] In step 1350, the fall injury detection device 100 may check whether the distance value (AvgDistance) to the target object is greater than the near critical point (Th_Near) and less than the far critical point (Th_Far).

[0124] The near critical point (Th_Near) and the far critical point (Th_Far) will be described in more detail later.

[0125] If the condition of step 1350 is met, the fall detection device 100 increments FallCount and proceeds to step 1370, and if the condition is not met, the fall detection device 100 proceeds to step 1360.

[0126] The fall detection device 100 may check whether the value of Fall_Time matches a previously set value (e.g., 0) in step 1360. If the condition of step 1360 is met, the fall detection device 100 may proceed to step 1330, and if the condition is not met, the fall detection device 100 may proceed to step 1340.

[0127] The fall detection device 100 checks whether FallCount and Th_FallCount match in step 1370. Here, Th_FallCount is a parameter for making a final determination as to whether a fall has occurred. If FallCount reaches Th_FallCount, the fall detection device can finally determine that the target object has fallen.

[0128] If the condition in step 1370 is not met, the fall injury detection device 100 proceeds to step 1360, and if the condition is met, the fall injury detection device 100 proceeds to step 1380.

[0129] The fall detection device 100 may determine that the target object is in a fall state and output information related thereto in step 1380. For example, the fall detection device 100 may transmit a message or signal to an external device connected to the fall detection device 100 to warn that the target object is currently in a fall state.

[0130] Thereafter, the fall detection device 100 may initialize the Fall_Timer and FallCount and set them to a non-observation state in step 1390. Thereafter, the fall detection device 100 may proceed to step 1340.

[0131] 14 and 15 are exemplary diagrams illustrating parameters used in a process of determining whether or not there is a fall according to an embodiment of the present invention.

[0132] FIG. 14 is a diagram illustrating the near-distance critical point (Th_near), the far-distance critical point (Th_far), and the rear-distance critical point (Th_Pres) among the above-mentioned parameters.

[0133] 14, the radar 1410 is illustrated as being installed on the ceiling of an indoor space. In this case, the distance from the height at which the radar 1410 is installed to the bottom 1420 can be assumed to be the height 1430 of the indoor space. Furthermore, if a target object 1440 is returning to the indoor space, a distance 1450 between the top of the target object's head and the radar 1410 can be derived as a distance value by the fall detection device 100. In this case, the critical point (Th_Pres) of the return can be derived based on the average of the distance values ​​1450 measured when the target object 1440 is returning to the indoor space and performing general activities other than a fall.

[0134] Furthermore, if the distance from the position where the radar 1410 is installed to the floor 1420 is defined as the height 1430 of the indoor space, the distance value corresponding to the height 1430 of the indoor space can be defined as the far-distance critical point (Th_far). Furthermore, the near-distance critical point (Th_near) can be defined based on the distance 1480 from the position of the radar 1410 to the height 1470 of a position a predetermined distance away from the floor 1420 in the indoor space. A target object 1460 in a falling injury state can be located at a height between the far-distance critical point (Th_far) and the near-distance critical point (Th_near).

[0135] 15 is a graph illustrating an example of a change in distance value depending on whether the target object returns to the room or falls. More specifically, FIG. 15 illustrates an example of a change in distance value depending on a target object absence section 1510, a target object entering the room section 1520, a target object returning to the room section 1530, a target object falling section 1540, and a target object leaving section 1550.

[0136] In this case, the fall detection device 100 can determine whether the target object is in a fall state when a distance value lower than the fall critical point 1560 is derived. In addition, the fall detection device 100 can determine that the target object is in a fall state when the distance value of the target object does not leave the range between the close critical point 1570 and the far critical point 1580 for a certain period of time.

[0137] FIG. 16 is a flowchart illustrating a method for detecting a fall according to an embodiment of the present invention.

[0138] The fall detection method using radar shown in Figure 16 includes steps that are processed in time sequence according to the embodiments shown in Figures 1 to 15. Therefore, even if the content is omitted below, it also applies to the fall detection method using the fall detection device according to the embodiments shown in Figures 1 to 15.

[0139] As shown in FIG. 16, the fall detection method may include step S100 of transmitting a radar signal toward a target object, step S200 of receiving the radar signal reflected from the target object, step S300 of analyzing the received radar signal to determine whether the target object has entered an observation state, and step S400 of deriving a distance value between the transceiver unit and the target object and determining whether the target object has fallen or been injured based on the distance value if the target object has entered an observation state.

[0140] The above-described method for detecting a fall injury may be implemented in the form of a computer program stored in a computer-readable recording medium that is executed by a computer or a recording medium containing computer-executable instructions.The above-described method for detecting a fall injury may also be implemented in the form of a computer program stored in a computer-readable recording medium that is executed by a computer.

[0141] The computer-readable recording medium may be any solvent-compatible medium accessible by a computer, including both volatile and nonvolatile media, and both separable and non-separable media. The computer-readable recording medium may also include computer storage media. The computer storage media includes both volatile and nonvolatile, separable and non-separable media embodied in any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data.

[0142] The fall detection method described above may be further divided into additional steps or combined into fewer steps in accordance with the embodiments described above with reference to Figures 1 to 15. Also, some steps may be omitted as necessary, and the sequence between steps may be switched.

[0143] The above description of the present invention is for illustrative purposes only, and those skilled in the art will appreciate that the present invention may be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. Therefore, it should be understood that the above-described embodiments are illustrative in all respects and are not limiting. For example, each component described as a single component may be implemented in a distributed form, and similarly, components described as distributed may be implemented in a combined form.

[0144] The scope of the present invention is indicated by the claims set forth below rather than by the above detailed description, and all modifications and variations derived from the meaning and scope of the claims, as well as from the concept of equivalents, should be interpreted as being included within the scope of the present invention. [Explanation of symbols]

[0145] 100: Fall detection device 210: Transmitter / receiver 220: Peak signal derivation unit 230: Observation Decision Department 240: Fall damage determination section

Claims

1. In a device for detecting falls using radar signals, a transceiver that transmits a radar signal toward a target object and receives the radar signal reflected from the target object; an observation determination unit that analyzes the received radar signal and determines whether to enter an observation state for the target object; a fall damage determination unit that derives a distance value between the transceiver unit and the target object when the observation state is entered, and determines whether the target object has fallen based on the distance value; A fall injury detection device including:

2. The fall detection device according to claim 1 , further comprising a peak signal deriving unit that derives a peak signal by filtering the radar signal.

3. The peak signal derivation unit 3. The fall injury detection device according to claim 2, further comprising: generating range-Doppler map information based on the radar signal; filtering the range-Doppler map information to derive the peak signal.

4. The distance-Doppler map information is The fall detection device of claim 3 , wherein the radar signal is generated for each frame corresponding to a time point when the radar signal is received.

5. The peak signal derivation unit 4. The fall injury detection device according to claim 3, wherein noise is removed from the distance-Doppler map, and a signal having an intensity equal to or greater than a predetermined threshold value is extracted as the peak signal.

6. The observation determination unit The fall detection device of claim 1, further comprising: deriving a motion peak signal associated with the motion of the target object from the radar signal; analyzing a pattern of the motion peak signal; and determining whether the target object has entered the observation state based on a pattern analysis result.

7. The fall detection device according to claim 6 , wherein the operational peak signal is derived by filtering at least a portion of a Doppler bin region of the radar signal or based on a distribution of the radar signal.

8. The observation determination unit The fall detection device according to claim 6, further comprising: deriving a distance index of the motion peak signal over time; and deriving a trend line based on the amount of change in the distance index to analyze a pattern of the motion peak signal.

9. The trend line is 9. The fall detection device according to claim 8, wherein the distance index is derived by applying a least squares method to the change in the distance index.

10. The observation determination unit 9. The fall detection device according to claim 8, wherein it is determined that the device is entering an observation state if the slope of the trend line is equal to or greater than a predetermined value.

11. The damage determination unit 2. The fall detection device of claim 1, wherein the distance value between the radar and the target object is derived in the observation state, the fall range is derived according to the height of the space in which the radar is installed, and whether or not the target object has fallen is determined based on the distance value and the fall range.

12. The distance value is The fall injury detection device of claim 11, wherein the fall injury detection value is derived from Root Mean Square (RMS) power derived by filtering a signal corresponding to the movement of the target object from the radar signal and moving a window corresponding to a time domain from the filtered signal.

13. The damage range is:

12. The fall detection device of claim 11, wherein the range is between a long-distance critical point corresponding to the distance from the position of the radar to the bottom of the space and a short-distance critical point corresponding to a position spaced a predetermined distance from the long-distance critical point in the direction in which the radar is installed.

14. The damage determination unit If the distance value is smaller than the critical point of the room, or if the distance value is larger than the critical point of the absence, determining whether or not there is a fall; The critical point of the re-entry is derived based on the distance value derived when the target object is re-entering the space, The fall detection device of claim 11 , wherein the critical point of absence is derived based on the distance value derived when the target object is absent from the space.

15. 1. A method for detecting a fall using a radar signal, comprising: transmitting a radar signal toward a target object; receiving a radar signal reflected from the target object; analyzing the received radar signal to determine whether the target object has entered an observation state; When entering the observation state, deriving a distance value between the radar and the target object, and determining whether the target object is damaged based on the distance value; A fall injury detection method comprising:

16. 16. The method of claim 15, further comprising the step of filtering the radar signal to derive a peak signal.

17. The step of deriving the peak signal comprises: The method for detecting a fall injury according to claim 16, further comprising generating range-Doppler map information based on the radar signal, filtering the range-Doppler map information, and deriving the peak signal.

18. The distance-Doppler map information is The method of claim 17, wherein the radar signal is generated for each frame corresponding to a time point when the radar signal is received.

19. The step of deriving the peak signal comprises: The method for detecting a fall injury according to claim 17, further comprising removing noise from the distance-Doppler map and deriving a signal having an intensity equal to or greater than a predetermined threshold value as the peak signal.

20. In a device that uses radar to detect falls, at least one processor; at least one memory containing computer program code; The at least one memory and the computer program code, through the at least one processor, cause the device to: Sends a radar signal to the target object, receiving a radar signal reflected from the target object; Analyzing the received radar signal to determine whether the target object has entered an observation state; The fall detection device is configured to derive a distance value between the radar and the target object when entering the observation state, and determine whether the target object has fallen based on the distance value.

Images