JP7081606B2 - Methods, systems, and computer programs to determine a subject's fall response - Google Patents

Description

The present invention relates, but not exclusively, to methods, systems and storage media for determining a subject's tipping response.

Falling is one of the main causes of injury to the elderly, especially those living alone. In addition, it is difficult for elderly people living alone to call for help, so if they are injured, they are likely to be in a serious situation. Therefore, monitoring the behavior of the elderly and detecting fall events is important for promptly coping with such events.

Currently, several methods are used to detect a fall event. However, with the general-purpose method of detecting a fall event, it is unavoidable that an erroneous alert, that is, an alert is sent to the care service provider even though the fall has not occurred. Such false alerts are an obstacle for both users and long-term care service providers.

Therefore, in order to deal with the above problems, there is a need for a method and a system for determining a fall response and thereby detecting a fall event while reducing false alerts.

The method according to one aspect of the present invention
A sensor attached to the subject receives a signal indicating the state of the subject from one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Compare with the state,
Based on the comparison between the first state and the second state, the fall response of the subject is determined.

The system according to one aspect of the present invention is
A receiving means that receives a signal indicating a state of the subject from a sensor attached to the subject regarding one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. A means of comparison to compare states,
It has a determination means for determining the fall response of the subject based on the comparison between the first state and the second state.

The computer-readable medium according to one aspect of the present invention is
Receiving processing that receives a signal indicating the state of the target from a sensor attached to the target regarding one of three or more actions of the target.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Comparison processing to compare states and
It includes a program that enables a computer to perform a determination process for determining a fall response of the target based on a comparison between the first state and the second state.

The computer system according to one aspect of the present invention is
With a memory device
It has at least one processor connected to the memory device, and the processor is
A sensor attached to the subject receives a signal indicating the state of the subject from one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Compare with the state,
Based on the comparison between the first state and the second state, the fall response of the subject is determined.

Embodiments of the present invention will be appropriately and easily understood by those skilled in the art in combination with the drawings based on the description which is merely exemplary.
FIG. 1 is a flowchart showing a method of determining a fall response of a target according to the first embodiment of the present embodiment. FIG. 2 is a schematic diagram showing a system for determining a fall response of a target according to the first embodiment of the present embodiment. FIG. 3 shows an exemplary system for determining a subject's fall response according to this embodiment. FIG. 4 shows an object to which the sensor according to the present embodiment is attached. FIG. 5 is a schematic diagram showing a system for determining a fall response of a target according to the present embodiment. FIG. 6 is a schematic diagram showing processing between the FD engine and the ADL engine in the system according to the present embodiment. FIG. 7 is a schematic diagram showing that the behavior of daily life is detected based on the approach by the binary tree structure according to the present embodiment according to the embodiment. FIG. 8 is a schematic diagram showing that the parameters of the fall detection algorithm are adjusted by using the behavior of daily life according to the present embodiment. FIG. 9 is a schematic diagram of a system for determining a fall response of a target according to the present embodiment. FIG. 10 shows an example of a computer suitable for performing the method of determining a subject's fall response according to this embodiment.

Unless the context has different meanings, the following terms shall have the following meanings.
"Subject" refers to a person, animal, robot, or anything else that is fitted with a sensor to determine a fall response and thus confirm a fall event.
The "fall" and "fall event" refer to an event in which the target state is in a fall state. In other words, it refers to the state in which the subject has been confirmed to have fallen.
"Possible falls" and "suspected falls" refer to situations where the condition is undecided and more information is needed to determine whether the subject's condition has been confirmed to be in a fall condition. .. In other words, it refers to a state in which it cannot be confirmed that the subject has fallen.
"Fall response" refers to the state of the subject after a fall event. The fall response is detected by a sensor after it is confirmed that the subject has fallen.
"Alert" includes any action (eg, message, notification, signal, etc.) to inform the intended party (eg, a user of the system) that the subject has fallen. Alerts (whether visual or auditory alerts) may be sent by any means of communication as appropriate. The "alert" may be sent with the likelihood (likelihood) of a fall event, for example 95%. This possibility can be estimated based on historical data on fall event detection.
"Nursing caregiver", "nursing care service provider" and "user" are supposed to take care of the target and receive an emergency call (or alert) when the target's fall event is confirmed. Refers to a person or organization.
"One or more databases" refers to one or more databases stored in a computing system, a remote server such as a computer in a hospital, a cloud server, or the like. The one or more databases may be a cloud database running on a cloud platform.

An embodiment of the present invention will be described with reference to the drawings as a mere example. Similar symbols in the figure refer to similar elements or their equivalents.

Some parts of the description below are presented, either explicitly or implicitly, as a functional or symbolic representation of the algorithm and operation on the data in computer memory. These algorithmic descriptions and functional or symbolic representations are the means used by those skilled in the art of data processing to most effectively convey the essence of the technique to others. The algorithm is also understood here and generally as a self-consistent sequence of steps leading to the desired result. A step is one that requires physical treatment of a physical quantity such as an electrical, magnetic or optical signal that can be treated by storage, transfer, coupling, comparison and other methods.

Unless otherwise stated, and as will be apparent from the description below, throughout this specification, "receive," "detect," "apply," "train," and "determine." Descriptions using terms such as "compare," "extract," "specify," or similar shall refer to the operation and processing of a computer system or similar device. The computer system or similar device may display data represented as a physical quantity in the computer system as a physical quantity in the computer system or in another device that stores, transmits or displays information. Treatment and transformation so that

The present specification also discloses an apparatus for performing an operation according to the above method. This device may be configured specifically for the required purpose, or may have a computer or other device that is selectively started or reconfigured by a computer program stored in the computer. .. The algorithms and indications described herein are not inherently associated with any particular computer or other device. Various devices can be used for the program in accordance with the disclosure herein. Alternatively, it may be appropriate to configure a more specialized device to perform the required method steps. The structure of the computer will be described later.

Further, the specification implicitly discloses a computer program, whereby those skilled in the art may perform individual steps of the methods described herein by computer code. Computer programs are not limited to any particular programming language and are not intended to be implemented. The disclosure content of this specification can be realized by various programming languages and coding using them. Moreover, computer programs are not limited to any particular control flow. There are various variants of computer programs, and it is possible to use various control flows without departing from the gist and scope of the present invention.

Moreover, one or more steps in a computer program can be performed in parallel rather than sequentially. Such computer programs can be stored on any computer-readable medium. Examples of computer-readable media include magnetic or optical discs, storage devices such as memory chips, and other storage devices capable of interfacing with computers. Further, the computer-readable medium includes a wired medium represented by an Internet system and a wireless medium represented by a GSM (Global Mobile Communication System) mobile phone system. By mounting and executing a computer program on such a computer, it is possible to effectively realize a device capable of executing the steps of a suitable method.

FIG. 1 is a flowchart showing a method 100 for determining a fall response of a target according to an embodiment of the present invention. Method 100 may be performed by a computer coupled with one or more databases. Further, the method 100 may be executed by an arithmetic unit such as a server system, a mobile device (for example, a smartphone or a tablet computer), or a personal computer. Further details about the computer and database will be described later with reference to FIGS. 9 and 10.

Method 100 generally includes the following steps.
Step 102: The processor receives a signal regarding the behavior of the subject. The state of the signal represents one of three or more actions.
Step 104: In response to detecting the fall of the target based on the received signal, the state of the signal after the first predetermined time has elapsed after the fall is compared with the state of the signal after the second predetermined time has elapsed after the fall. ..
Step 106: In the processor, the fall response of the subject is determined according to the result of the comparison.

Step 102 includes receiving a signal regarding the behavior of the subject. For example, the subject may be an elderly person, especially an elderly person living alone. Further, or instead, the subject may be an infant, an inpatient, or other people in need of a fall detection monitor. Further, or instead, the subject may be a pet dog, a pet cat, or some other animal. Further, the subject may be a bipedal robot, a quadrupedal robot, or some other footed robot.

The state of the received signal relates to one of three or more actions of interest. For example, the subject's actions may be sitting, standing, lying down, walking, running, or jumping. That is, the state of the received signal relates to one of the above three or more target actions. It will be appreciated that three or more behaviors of the subject are identified to improve the accuracy of the fall response determination. That is, actions of types other than the above may be included in the actions to be detected. The behavior of the subject may depend on each subject. For example, if the subject is an infant or elderly person who cannot jump, the "jump" may be excluded from the subject's behavior.

A sensor attached to the subject, such as an accelerometer, collects information about the behavior of the subject and sends a signal about the collected behavior of the subject. FIG. 4 shows an example of the position of the sensor attached to the object. A receiver connected to the processor may receive a signal about the behavior of the subject. Signals may be continuously received over a predetermined period of time. Further, or instead, a signal may be received when some action is detected. The signal may be transmitted to the receiver via Bluetooth, WiFi, NFC (Near Field Communication), or other types of wireless communication. The receiver may be a receiving module in a tablet computer, a smartphone, or a personal computer.

In step 104, the state of the signal at the time when the first predetermined time has elapsed after the fall and the signal at the time when the second predetermined time has elapsed after the fall, in response to detecting the fall of the target (for example, a fall event). Includes comparing with the state of. For example, the first predetermined time may be 10 seconds and the second predetermined time may be 30 seconds. For example, an accelerometer worn on an elderly person may detect sudden acceleration beyond a predetermined threshold. Alerts have so far been sent to the caregiver or someone in the subject's family in response to the detection of a sudden acceleration when a fall is suspected. However, in the present embodiment, the transmission of an erroneous alert is reduced by using additional data for confirming whether or not the subject has fallen. For example, this additional data is collected twice from the accelerometer. For example, 10 seconds have passed since the suspected fall and 30 seconds have passed since the suspected fall. By comparing the data collected 10 seconds after the suspected fall and 30 seconds after the suspected fall, it is determined whether it is necessary to send an alert. For example, the additional data relates to the behavior of the subject 10 seconds after the suspected fall and 30 seconds after the suspected fall. The first predetermined time and the second predetermined time may be appropriately selected so as to reduce the transmission of erroneous alerts. By monitoring the signal status over a predetermined time range including the first predetermined timing and the second predetermined timing instead of checking the signal state at one timing, the accuracy of the fall alert can be improved.

Step 106 comprises determining the subject's tipping response in the processor based on the comparison in step 104. The result of the comparison in step 104 is processed to determine the subject's fall response. For example, the characteristics of the signal are extracted, and the signal is determined based on the extracted characteristics. In this way, one of the three or more behaviors of the subject is identified. For example, a classification algorithm is applied to the extracted features to identify one of three or more behaviors of interest.

For example, the classification algorithm includes a binary tree approach. The binary tree approach will be described later with reference to FIG. 7. Further, for example, an action represented by a signal state when a first predetermined time has elapsed after a fall and another action represented by a signal state when a second predetermined time has elapsed after a fall can be described as a fall response. Compare with the set of registered actions.

For example, when a fall response is determined, it is determined whether or not to send an alarm signal. The alarm may be sent to a family member living in the vicinity of the subject or a caregiver caring for the subject. If the target is an industrial robot, an alarm may be sent to the factory manager. The alarm may be a telephone, a message via SMS (Short Message Service), or an email. If the subject is in the hospital, the alarm may be a call to the nurse center.

FIG. 2 is a schematic diagram showing the system 200 according to the embodiment. The system 200 determines the subject's fall response. For example, the system 200 has an FD (fall detection) engine 202, an ADL (activity of daily living) engine 204, and an acceleration sensor 206. The acceleration sensor 206 can be attached, for example, in a patch type, and has a configuration for detecting acceleration data in three dimensions in the x, y, and z directions, for example. The acceleration data 208 detected by the acceleration sensor 206 is continuously transmitted to an external device such as a computer server, a tablet, or a mobile phone by wireless communication such as Bluetooth communication, as will be described later with reference to FIG. The patch-type accelerometer will be described later with reference to FIG.

In one example, the acceleration data 208 may be transmitted to a back-end device 216, such as a computer server running the ADL engine 204. Based on the acceleration data 208, the ADL engine 204 predicts the behavior of the subject in which the acceleration data 208 is observed by the living state, for example, the acceleration sensor 206. The predicted living state 210 of the target is transmitted to the FD engine 202 operating in the front-end device 214 such as a smartphone. The predicted living condition 210 of the subject can be used to adjust the threshold value of the fall event detection by the FD engine 202. When the front-end device 214 receives the acceleration data 208, the fall event is determined by comparing the received acceleration data 208 with the threshold value determined using the predicted living state 210 of the subject.

Upon receiving the predicted living condition 210 from the ADL engine 204, the FD engine 202 integrates the predicted living condition 210 with the acceleration data 208 received from the sensor 206 to estimate a possible fall event 212. .. The post-estimation mechanism may be activated as soon as an alert for a possible fall event is issued. The FD engine 202 estimates the scenario based on the living condition 210 predicted by the ADL engine 204, and detects whether or not the fall alert is a true alert. Post-estimation mechanisms can reduce false alerts for fall detection. Details of the functions of the ADL engine 204 and the FD engine 202 will be described later with reference to FIGS. 5 and 6.

FIG. 3 shows an example of the system 300 according to the embodiment. The system 300 determines the subject's fall response. The system 300 may include a front-end device 302, a back-end device 304, and an accelerometer 306. The acceleration sensor 306 transmits the detected acceleration data 308 to the front-end device 302 and the back-end device 304. For example, a back-end device 304 such as a computer server may predict the living condition of the target. The predicted living condition 310 may be transmitted to a front-end device 302 such as a smartphone. Therefore, the front-end device 302 receives information from the acceleration sensor 306 and the back-end device 304.

FIG. 4 shows an object 400 having a sensor 402 according to an embodiment. The sensor 402 may be a patch-type sensor attached to the central portion of the target, for example, the body portion. Preferably, a single sensor may be used instead of the plurality of sensors to detect the fall. This is because if two or more devices are attached, the subject may feel uncomfortable.

The position of the sensor 402 in the target 400 may be determined based on the accuracy of the acceleration data obtained at the position of the sensor 402. This accuracy may be determined based on how the detected acceleration data reflects the movement of the subject. For example, the sensor 402 is attached to the lumbar region of the target 400.

FIG. 5 is a schematic diagram showing an ADL engine 502 and an FD engine 504 in the system 500 according to the embodiment. For example, the ADL engine 502 classifies the target behavior into one of the specific behaviors of the behavior classification module 506 in the ADL engine 502. The classified behavior 516 (that is, the specific behavior where the target behavior is classified) is set as a fall detection parameter under the specific behavior (that is, the classified behavior 516) as shown in module 508. It may be transmitted to the FD engine 504. In other words, the fall detection parameters are based on the behavior classified by the ADL engine. Therefore, different parameters may be applied to the fall detection due to different actions.

The FD engine 504 employs a threshold-based approach. If the threshold is exceeded in module 510, a suspicion of a fall is detected. By adjusting the thresholds according to the behavior, the FD engine 504 can dynamically set different thresholds based on the current behavior in which the fall occurs. For example, the thresholds for falls when running, sitting, and standing are different. According to this method, it is possible to distinguish various fall situations such as falling while walking or slipping off a chair while sitting. Reduced false alerts in module 510 fall detection by adjusting parameters based on the classified behavior 516 transmitted by the ADL engine 502 compared to using a unified threshold without considering behavior. can do.

A set of IF-THEN rules is predefined in the knowledge base to determine thresholds for acceleration values measured in the x, y and z directions. This will be described later with reference to FIG. The FD engine 504 needs to acquire information on current behaviors such as sitting, standing, and lying in daily life from the ADL engine 502 and select an appropriate threshold value based on the information.

Similarly, the classified actions 518 (same as the classified actions 516) may be transmitted from the ADL engine 502 to the FD engine 504 for knowledge estimation after the fall of the module 512. After a fall is detected, the classified behavior 518 is further used to reduce false fall detection. Knowledge estimation after a fall is activated when a suspicion of a fall is detected. The system does not send an alert if it is estimated that Module 512 has recovered from a suspected fall or was not a fall at all. Otherwise, the system sends an alert to a given contact, such as a member of the subject's family or a caregiver caring for the subject. Therefore, this method is useful in reducing false alerts for fall detection.

In the above process, a set of human knowledge can be predefined based on how long the subject should be monitored, what action to take at the same time as the fall, and so on. In order to carry out the knowledge estimation, the FD engine 504 also needs to acquire the current ADL information (that is, the classified action 518) detected by the ADL engine 502.

In one example, after a suspected fall is detected, the system confirms the fall event by continuous behavioral monitoring and knowledge estimation. An example of the knowledge estimation rule is shown in Table 1 below.

For example, if the system obtains a "potential recovery" as a result of a fall detection, the system will continue to monitor behavior for a predetermined threshold period (eg, one minute thereafter). The subject is recovering, or (ii) sitting or standing, if (i) is walking for the next minute, or is active, such as running or jumping. If it maintains a stationary state such as lying down or lying down, it can be considered that it has not recovered.

When it is confirmed that the subject has fallen, an alert is sent to the predetermined person in module 514 so that the predetermined person can take immediate action.

FIG. 6 is a schematic diagram showing processing between the FD engine 632 and the ADL engine 634 in the system 600 according to the embodiment. The two engines 632 and 634 operate in parallel in the front-end and back-end devices, respectively. For example, the sensor 602 transmits a signal of the target accelerometer data to the FD engine 632 and the ADL engine 634.

In the ADL engine 634, the acceleration data in the x, y and z directions from the sensor 602 is acquired as time series data in step 604 and stored in the memory 614. After that, the window size of the time series data is set, the data window is acquired, and the analysis is executed in step 606.

In step 608, the characteristics of the signal are extracted based on the time series data. In addition, the extracted features are intended to identify behaviors in daily living conditions such as sitting, standing, lying, walking, running, and jumping in step 610. It is used as an input to the behavior classification model. When the behavior of daily life is identified in step 612, the identified behavioral state is stored in the memory 614. The identified behavioral state is used in the FD engine 632.

In the FD engine 632, the acceleration data in the x, y and z directions by the sensor 602 is acquired from the memory 614. For example, the acquired acceleration data in the x, y and z directions is filtered using a high-pass filter in step 616.

In step 618, the current behavior in daily living conditions such as sitting, standing, lying down, walking, running, and jumping is read from the memory 614. Also, the threshold parameters are adjusted based on a predefined knowledge system.

If the acceleration data in the x, y, and z directions satisfy the fall condition adjusted based on the current behavior in the daily living state in step 618, the suspicion of falling is detected in step 620. If there is no suspicion of a fall, the FD engine 636 will continue to acquire new acceleration data in step 628.

In step 624, post-fall detection is performed by continuous behavior monitoring and knowledge estimation, as shown in Table 1. This step detects whether you have actually fallen or have recovered from a suspicion of a fall. If the fall event is confirmed in step 626, an alert is sent in step 636. If no fall is confirmed, or if the fall is recovered, monitoring ends and new acceleration data is detected in step 630. The FD engine 632 sends information about the above process to memory 614 for later reference.

FIG. 7 shows an example of the binary tree structure approach for detecting the behavior of daily life according to the embodiment. For example, the ADL engine implements a dichotomous tree-structured approach that identifies various activities of daily living (ADL) based on machine learning algorithms. This binary tree 700 example has five classification nodes 702, 704, 706, 708, and 710. Each of these nodes needs to be trained by a classification algorithm. When the ADL engine detects the acceleration data, the binary tree calculates and predicts the corresponding ADL states such as sitting, standing, lying down, and walking. In this example, a behavioral classification approach based on a binary tree is used. However, different targets have different classification algorithm designs.

For example, the ensemble SVM (Support Vector Machine) method is applied to the binary classification at each node of the binary tree. Upon receiving new sensor data, this approach extracts preprocessing, the statistical features of the ensemble SVM model. Statistical features may be, for example, mean, standard deviation, total value, percentile, skewness, kurtosis, correlation, coefficient of variation, inter-amplitude amplitude, zero intersection, and the like.

For example, upon receiving new sensor data, this approach first determines at node 702 "stationary or behavioral" whether the current behavior is a dynamic behavior or a static behavior. It is determined. If the current behavior is determined to be a static behavior, then at node 704 "" stationary + standing "or lying down", the current behavior is said to be stationary / standing. It is decided whether it is an action or an action of lying down. If the current behavior is not determined to be a lying behavior, then in node 708 "sitting or standing", the current behavior is sitting or standing. It is decided whether it is an action.

Similarly, in node 702, if the current action is determined to be a dynamic action, then in node 706 "whether walking or'running + jumping'", the current action is walking. It is determined whether it is an action of being, running or jumping. If the current action is not determined to be a walking action, then at node 710 "Running or Jumping", the current action is running or jumping. It is decided whether it is an action.

A single classification model always gives better results in the discrimination of two classes than in the discrimination of multiple classes. Therefore, the binary tree structure helps to improve the overall predictive power of behavior by integrating multiple binary classification models. This binary tree structure can also be expanded if it is necessary to expand the range of action. For example, the walking state may be distinguished by speed. Further binary nodes may be created as a substructure of node 708 "sitting or standing". Although described using the above approach, other approaches such as SVN (Support Vector Network), KNN (K-nearest Neighbor) algorithm, neural network, etc. may be applied to the prediction of ADL state.

FIG. 8 is a schematic diagram showing a process of setting parameters of a fall detection algorithm using actions in daily life according to an embodiment. In order to accurately detect a fall event, it is necessary to set different fall detection parameters according to various actions. If the current action is action 1 (YES in step 802), the threshold is set to parameter 1 (step 808). If the current behavior, i.e., behavior 1 is running, then there is an appropriate threshold for the fall in the running behavior. Based on this knowledge, the respective actions of steps 802, 804, 806 are associated with appropriate thresholds in steps 808, 810, 812. By selecting an appropriate threshold, it is possible to detect a fall event in each of these behaviors with high accuracy.

FIG. 9 is a schematic diagram of a network-based system 900 for determining a subject's fall response according to an embodiment of the present invention. The system 900 includes a computer 902, one or more databases 9041 to 904n, a user input module 906, and a user output module 908. Each of the one or more databases 9041 to 904n is communicably connected to the computer 902. The user input module 906 and the user output module 908 may be independent modules, and these modules may be communicably connected to the computer 902. Alternatively, the user input module 906 and the user output module 908 may be integrated as a single mobile electronic device (for example, a mobile phone, a tablet computer, etc.). The mobile electronic device may have an appropriate communication module for wireless communication with the computer 902 using existing communication protocols.

The computer 902 may have at least one processor and at least one memory in which the computer program code is stored. At least one memory and computer program code causes the computer to perform at least the following operations by at least one processor. (A) Receiving a signal relating to the behavior of the subject, the state of the signal relates to one of three or more behaviors of the subject. (B) Compare the signal state at the time when the first predetermined time elapses after the fall and the signal state at the time when the second predetermined time elapses after the fall according to the detection of the fall of the target. thing. (C) To determine the fall response of the target according to the comparison process.

Various types of data such as behavioral states, acceleration data in the x, y and z directions, knowledge estimation rules for estimating knowledge after a fall, etc. can be stored in one database (eg 9041) or multiple databases (eg x). , Y and z direction acceleration data is stored in the database 9041, knowledge estimation rules are stored in the database 904n, etc.). The databases 9041 to 904n may be configured using a cloud computing storage module and / or a dedicated server communicably connected to the computer 902.

FIG. 10 shows an exemplary computer / arithmetic unit 1000. Hereinafter, it is also referred to as a computer system 1000 in a compatible manner, and here, one or more arithmetic units 1000 may be used to facilitate the execution of the above-mentioned method for determining a fall response of a target. Also, one or more components of the computer system 1000 may be used to configure the computer 902. The following description of the arithmetic unit 1000 is for illustration purposes only and is not intended to be limiting.

As shown in FIG. 10, the exemplary arithmetic unit 1000 has a processor 1004 that executes software routines. For convenience, a single processor is shown, but the arithmetic unit 1000 may have a multiprocessor system. The processor 1004 is connected to the communication board 1006 for communication with other components of the arithmetic unit 1000. The communication infrastructure 1006 may be, for example, a communication bus, a crossbar, a network, or the like.

The arithmetic unit 1000 further has a main memory 1008 such as a RAM and a secondary memory 1010. The secondary memory 1010 is, for example, a storage device 1012 such as a hard disk drive, a solid state drive, a hybrid drive, and / or a magnetic tape drive, an optical disk drive, a solid state storage drive (for example, a USB flash drive, a flash memory device, a solid state drive, or the like. A removable storage device 1014 such as a memory card), or the same type may be included. The removable storage device 1014 reads and / or writes to the removable storage medium 1044 by a well-known method. The removable storage medium 1044, which is read and written by the removable storage device 1014, may be a magnetic tape, an optical disk, a non-volatile storage medium, or the same type. As will be appreciated by those skilled in the art, the removable storage medium 1044 comprises a computer-readable storage medium that stores computer-executable program code instructions and / or data.

Instead of the above configuration, the secondary memory 1010 may have additional or alternative means of allowing a computer program or other instruction to be mounted on the arithmetic unit 1000. Examples of such means include the removable storage unit 1022 and the interface 1040. Examples of the removable storage unit 1022 and the interface 1040 correspond to a program cartridge and a cartridge interface (like those found in a video game console), and a removable memory chip (EPROM (erasable programmable ROM), PROM (programmable ROM), etc.). Sockets, removable solid state storage drives (USB flash drives, flash memory devices, solid state drives, memory cards, etc.), and other removable storage units that can transfer software and data from the removable storage unit 1022 to the computer system 1000. 1022 and interface 1040 can be mentioned.

The arithmetic unit 1000 also has at least one communication interface 1024. The communication interface 1024 enables software and data to be transferred between the arithmetic unit 1000 and other external devices via the communication path 1026. In some embodiments of the present invention, the communication interface 1024 enables data transfer between the arithmetic unit 1000 and a data communication network such as public data or a private data communication network. The communication interface 1024 may be used to exchange data with some arithmetic units 1000 so that the arithmetic units 1000 form part of an interconnected computer network. Examples of communication interfaces 1024 include modems, network interfaces (eg Ethernet cards), communication ports (serial bus, parallel bus, printer bus, GPIB (general purpose interface bus), IEEE (American Electrical and Electronic Society) 1394, RJ (registered). Jacks) -45, USB, etc.), antennas mounted on the corresponding circuits, etc. The communication interface 1024 may be wired or wireless. The software or data transferred via the communication interface 1024 can be in the form of an electronic signal, an electromagnetic signal, an optical signal, or another signal that the communication interface 1024 can receive. These signals are transmitted to the communication interface 1024 via the communication path 1026.

As shown in FIG. 10, the arithmetic unit 1000 further includes a display interface 1002 that operates to provide an image to the corresponding display 1030 and an audio interface 1032 that operates to reproduce audio content through the corresponding speaker 1034.

As described herein, the term "computer program product", in part, refers to a removable storage medium 1044, a removable storage unit 1022, a hard disk installed in a storage device 1012, or software as a channel 1026 ( It may include carriers carried to the communication interface 1024 over a wireless link or wired). Computer-readable storage media include any non-volatile non-volatile tangible storage medium that provides recorded instructions and / or data to arithmetic unit 1000 for execution and / or processing. Examples of such storage media include magnetic tapes, CD-ROMs, DVDs, Blu-ray disks, hard disk drives, ROMs or integrated circuits, solid state storage drives (USB flash drives, flash memory devices, solid state drives, memory cards, etc.). ), Hybrid drives, magneto-optical disks, computer-readable cards such as SD memory cards, etc., regardless of whether they are inside or outside the arithmetic unit 1000. Temporary intangible computer-readable transmission media that may be involved in providing software, application programs, instructions and / or data to the compute device 1000 may be, for example, to a wireless or infrared transmission channel, other computer or networked device. It may be the Internet or an intranet containing information recorded on a network connection, e-mail transmission, website, or the like.

The computer program (also referred to as a computer program code) is stored in the main memory 1008 and / or the secondary memory 1010. The computer program can be received via the communication interface 1024. By executing such a computer program, the arithmetic unit 1000 can realize one or more characteristic embodiments of the embodiments described here. In certain embodiments, the execution of a computer program allows the processor 1004 to realize the characteristic embodiments of the embodiment. Therefore, such a computer program functions as a controller of the computer system 1000.

The software may be stored in a computer program product and read by the arithmetic unit 1000 using a removable storage device 1014, a storage device 1012, or an interface 1040. Alternatively, the computer program product may be downloaded to the computer system 1000 via the communication path 1026. By executing the software by the processor 1004, the arithmetic unit 1000 can realize the function described in the above embodiment.

The embodiment shown in FIG. 10 is merely an example. Therefore, depending on the embodiment, one or more features of the arithmetic unit 1000 may be omitted. Also, in certain embodiments, one or more features of the arithmetic unit 1000 may be combined. Further, in certain embodiments, one or more features of the arithmetic unit 1000 may be divided into one or more components.

A part or the whole of the above-described embodiment can be expressed as the following appendix. However, some or all of the embodiments as described above are not limited to the following description.

(Appendix 1)
A sensor attached to the subject receives a signal indicating the state of the subject from one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Compare with the state,
A method of determining the fall response of the subject based on a comparison between the first state and the second state.
(Appendix 2)
Extract the features in the signal
The method according to Supplementary Note 1, wherein the state of the signal is determined from the three or more actions based on the extracted characteristics.
(Appendix 3)
The method according to Appendix 2, wherein the process of determining the state of the signal includes a process of identifying one of the three or more actions by applying a classification algorithm to the extracted features.
(Appendix 4)
The method according to Appendix 3 for training the classification algorithm based on the extracted characteristics.
(Appendix 5)
The method according to Appendix 3 or Appendix 4, wherein the classification algorithm comprises a binary tree approach.
(Appendix 6)
The process of determining the fall response includes the process of comparing the first action represented by the first state and the second action represented by the second state with a series of actions registered for the fall response. The method according to any one of Appendix 5.
(Appendix 7)
The method according to any one of Supplementary note 1 to Supplementary note 6, which determines to transmit an alarm signal when a fall response is determined.
(Appendix 8)
The three or more actions are at least one of walking, lying down, standing, sitting, running, and jumping. The method according to any one of Supplementary note 1 to Supplementary note 7, including one.
(Appendix 9)
A receiving means that receives a signal indicating a state of the subject from a sensor attached to the subject regarding one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. A means of comparison to compare states,
A system comprising a determination means for determining a fall response of the subject based on a comparison between the first state and the second state.
(Appendix 10)
Further provided with a feature extraction means for extracting features in the signal,
The system according to Appendix 9, wherein the determination means determines the state of the signal from the three or more actions based on the extracted characteristics.
(Appendix 11)
The system according to Appendix 10, wherein the determination means identifies one of the three or more actions by applying a classification algorithm to the extracted features.
(Appendix 12)
The system according to Appendix 11, further comprising training means for training the classification algorithm based on the extracted features.
(Appendix 13)
The system according to Annex 11 or Annex 12, wherein the classification algorithm comprises a binary tree approach.
(Appendix 14)
The comparison means further any one of Supplementary Notes 9 to 13 for comparing the first action represented by the first state and the second action represented by the second state with a series of actions registered for the fall response. The system described in.
(Appendix 15)
The system according to any one of Supplementary note 9 to Supplementary note 14, wherein the determination means further determines to transmit an alarm signal when a fall response is determined.
(Appendix 16)
The three or more actions are at least one of walking, lying down, standing, sitting, running, and jumping. The system according to any one of Supplementary note 9 to Supplementary note 15, including one.
(Appendix 17)
Receiving processing that receives a signal indicating the state of the target from a sensor attached to the target regarding one of three or more actions of the target.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Comparison processing to compare states and
A computer-readable medium containing a computer program that enables a computer to perform a determination process for determining a fall response of the target based on a comparison between the first state and the second state.
(Appendix 18)
A computer program is stored that further realizes an extraction process for extracting a feature in the signal and a determination process for determining the state of the signal from the three or more actions based on the extracted feature. The medium according to Appendix 17.
(Appendix 19)
The medium according to Appendix 18, wherein the determination process determines one of the three or more actions by applying a classification algorithm to the extracted features in the process of determining the state of the signal.
(Appendix 20)
The medium according to Supplementary Note 19, wherein a computer program for further realizing a training process for training the classification algorithm based on the extracted characteristics is stored in the computer.
(Appendix 21)
The medium according to Appendix 19 or Appendix 20, wherein the classification algorithm comprises a binary tree approach.
(Appendix 22)
The process of determining the fall response includes the process of comparing the first action represented by the first state and the second action represented by the second state with a series of actions registered for the fall response. The medium according to any one of Supplementary note 17 to Supplementary note 21.
(Appendix 23)
The medium according to any one of Supplementary note 17 to Supplementary note 22, which contains a computer program for further realizing a second determination process for determining to transmit an alarm signal when a fall response is determined.
(Appendix 24)
The three or more actions are at least one of walking, lying down, standing, sitting, running, and jumping. The medium according to any one of Supplementary note 17 to Supplementary note 23, including the above.
(Appendix 25)
With a memory device
It comprises at least one processor connected to the memory device, the processor being
A sensor attached to the subject receives a signal indicating the state of the subject from one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Compare with the state,
A computer system that determines the fall response of the subject based on a comparison between the first state and the second state.
(Appendix 26)
The at least one processor
Extract the features in the signal
The computer system according to Appendix 25, which determines the state of the signal from the three or more actions based on the extracted characteristics.
(Appendix 27)
The at least one processor
The computer system according to Appendix 26, which identifies one of the three or more behaviors by applying a classification algorithm to the extracted features.
(Appendix 28)
The at least one processor
The computer system according to Appendix 27, which trains the classification algorithm based on the extracted features.
(Appendix 29)
The computer system according to Appendix 27 or Appendix 28, wherein the classification algorithm comprises a binary tree approach.
(Appendix 30)
The at least one processor
The computer system according to any one of Supplementary note 25 to Supplementary note 29, which compares the first action represented by the first state and the second action represented by the second state with a series of actions registered for a fall response.
(Appendix 31)
The at least one processor
The computer system according to any one of Supplementary note 25 to Supplementary note 30, which determines to transmit an alarm signal when a fall response is determined.
(Appendix 32)
The three or more actions are at least one of walking, lying down, standing, sitting, running, and jumping. The computer system according to any one of Supplementary Provisions 25 to 31 including the above.

As those skilled in the art will understand, the invention as shown in the embodiments described above may be modified and / or modified in various ways within the ideas or scope of the invention as described above. You may. Therefore, it can be considered that this embodiment is exemplary in all respects and is not limiting.

This application was filed on May 15, 2017 in Singapore Application No. Claim priority on the basis of 10201702112X and incorporate all of its disclosures here.

100 Method 200 System 202 Fall Detection Engine 204 Daily Life Behavior Engine 206 Acceleration Sensor 208 Acceleration Data 210 Predicted Living State 212 Possible Fall Events 214 Front End Device 216 Back End Device 300 System 302 Front End Device 304 Back End Device 306 Acceleration sensor 308 Acceleration data 310 Predicted living conditions 400 Target 402 Sensor 500 System 502 ADL engine 504 FD engine 506 Module 508 Module 510 Module 512 Module 514 Module 516 Classified behavior 518 Classified behavior 600 System 602 Sensor 614 Memory 632 FD engine 634 ADL engine 700 2 branch tree 702 node 704 node 706 node 708 node 710 node 900 system 902 computer 9041 database 904n database 906 user input module 908 user output module 1000 system 1002 display interface 1004 processor 1006 communication infrastructure 1008 main Storage 1010 Secondary storage 1012 Storage device 1014 Removable storage device 1022 Removable storage unit 1024 Communication interface 1026 Communication path 1030 Display 1032 Audio interface 1034 Speaker 1040 Interface 1044 Removable storage medium

Claims (10)

A sensor attached to the subject receives a signal indicating the state of the subject from one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Compare with the state,
A method of determining the fall response of the subject based on a comparison between the first state and the second state. A receiving means that receives a signal indicating a state of the subject from a sensor attached to the subject regarding one of three or more actions of the subject.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. A means of comparison to compare states,
A system comprising a determination means for determining a fall response of the subject based on a comparison between the first state and the second state. Further provided with a feature extraction means for extracting features in the signal,
The determination means determines the state of the signal from the three or more actions based on the extracted characteristics.
The system according to claim 2 . The determination means identifies one of the three or more actions by applying a classification algorithm to the extracted features.
The system according to claim 3 . The system according to claim 4 , further comprising a training means for training the classification algorithm based on the extracted features. The classification algorithm includes a binary tree approach.
The system according to claim 4 or 5 . The comparison means further compares the first action represented by the first state and the second action represented by the second state with a series of actions registered for the fall response.
The system according to any one of claims 2 to 6 . The determination means further determines to transmit an alarm signal when a fall response is determined.
The system according to any one of claims 2 to 7 . The three or more actions are at least one of walking, lying down, standing, sitting, running, and jumping. Including one
The system according to any one of claims 2 to 8 . Receiving processing that receives a signal indicating the state of the target from a sensor attached to the target regarding one of three or more actions of the target.
According to the detection of the fall of the target based on the above state, the first state of the signal when the first predetermined time has elapsed from the fall and the second state of the signal when the second predetermined time has elapsed from the fall. Comparison processing to compare states and
A computer program that enables a computer to perform a determination process for determining a fall response of an object based on a comparison between the first state and the second state.

Images