JP2014067082A - Watching reporting apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a mechanism for sure detection while casually watching in a simple method when a failure occurs in a life of an elderly solitary person with an increase in elderly solitary person households.SOLUTION: A person's actions in a daily life emit unique sounds along with the respective actions. A series of actions show characteristics that these sounds are emitted in a constant order. For example, when a door chime rings, normally noise accompanying walking is heard without a lapse. When an abnormality is recognized in this pattern, some failure is predicted. Here, a mechanism is obtained for watching in a method for detection by specifying types of respective actions and occurrences from characteristics of many sounds (spectrums) emitted by the life actions and detecting a proper transition of these series of actions and occurrences.

Description

  The present invention relates to a method for detecting a failure of a person being watched at an early stage while watching the daily life of a single person living alone.

An increasing number of single-person households have become conspicuous as the aging society arrives. There is a demand for an efficient mechanism for watching the situation where these people's lives continue without failure.
For example, a method of constantly monitoring with a surveillance camera can be considered, but this is not a desirable method from the viewpoint of privacy protection. It is expected to establish an effective mechanism without putting an excessive burden on the people who are watched over.

Patent Publication 2012-095796 Patent Publication 2012-014598 Patent Publication 2008-210093 Patent Publication 2005-301960 Patent Publication 2004-174168 Japanese Patent Application No. 2011-114333

Several methods have been disclosed in the past. For example, if the opening / closing operation is not permitted within a certain period while monitoring information indicating the opening / closing status of the entrance and toilet doors performed by the person being watched over, a method for reporting that there has been a failure is presented. ing. Instead of door opening / closing information, a method for detecting that a person to be watched enters a specific space region using electromagnetic waves or infrared rays is also presented.

However, for example, the method of monitoring the opening and closing of the door is not useful when the person to be watched does not move between rooms or go out. Even a method for detecting that a specific space area is entered does not necessarily enter the setting area, and just because there is no such information regarding the person being watched over does not mean that a failure can be recognized.
Here, realization of a mechanism with higher accuracy for detection is an issue.

The present invention is a method for detecting operations such as opening and closing of doors, for example, from specific sounds generated in the daily life of a person to be watched, and grasping whether these sequences and combinations are in a normal state, The purpose is to detect the failure of the person being watched over. For example, when there is a visitor, a sound accompanying the walking of the person to be watched should normally be generated along with the ringing of the chime at the entrance.

If the door of the toilet is opened and closed, after a while the flushing sound of the toilet is generated. When the phone rings, pick up the handset and start talking. Therefore, it is considered that it is possible to monitor that normally captured events are observed, and when these greatly deviate from a certain pattern, it is assumed that some failure has occurred in the person being watched over.

For this reason, the sound generated with the life of the person being watched over is used here. For convenience, the operation of the person being watched over will be described in two types. In other words, it consists of events and sounds associated with individual actions (hereinafter referred to as partial actions) such as door opening and closing and walking sounds, and actions consisting of a series of partial actions with a certain purpose (hereinafter referred to as target actions). is there. This is because it is considered that the specific partial motion or the target action is accompanied by the generation of a specific sound, and the behavior of the watching target person can be estimated from the characteristics of the sound.

If there is a specific partial action or objective action at one or more indoor locations where the watched person resides and it can be determined that this is an action that accompanies normal life, this is the normal life of the watched person It can be estimated. In addition to events that are notified by sensors that detect opening and closing of doors and entering the space, a method of determining the living situation of the person being watched over is also conceivable.

In other words, if the target behavior is estimated based on partial movements detected from specific sounds at multiple locations in the room where the person being watched over lives and events obtained from sensors such as door opening and closing, normal if certain conditions are met Although it is determined that there is a comfortable life, it is also considered effective to report that it has been determined that some failure has occurred if it deviates from the conditions. Since the correspondence between specific life sounds and partial actions and the contents of the target behavior vary depending on the lifestyle of the person being watched over, tools that can be easily set for each person being watched over are provided here.

In this way, by monitoring the characteristics that appear in the behavior of the person being watched over, it is possible to quickly grasp the occurrence of a serious failure or a minor failure.

Monitoring device configuration diagram Configuration example of detection target sound data part Configuration example of action description data part

The configuration of the present invention will be described with reference to FIG.
FIG. 1 includes a sound collection sensor 1, a sensor 4 that senses an operation, a notification device 10, an IP network 20, and a terminal device 30. However, here, the configuration may be such that the reporting device and the terminal device are directly connected. In this case, the IP network 20 may not be provided.

The sound collection sensor 1 includes a sound collection unit 2 and a signal transmission unit 3 for notifying a notification device of a sound signal. The sensor 4 that detects the operation is a sensor that detects opening / closing of the door or entering the space. On the other hand, the notification device 10 includes an audio signal receiving unit 11 that receives a sound signal from the sound collecting sensor 1, an audio analysis processing unit 12 that detects a partial motion from a specific living sound emitted by the person being watched over, and events such as opening and closing of a door. An input / output processing unit 13 that receives an event occurrence notification from the sensing sensor 4, a voice analysis processing unit 12, a notification determination processing unit 14 that discriminates a target action based on partial actions and events notified from the input / output processing unit 13; A detection target sound data unit 15 that describes characteristics for detecting partial motions from daily sounds emitted by the watching target person, an action description data unit 16 that describes a target action behavior that the watching target person normally performs, an IP network connection unit 17, and The setting unit 18 is configured.

Here, the setting unit 18 writes and registers the association data of the characteristics of the living sound and the partial motion generated in accordance with the specific motion of the person being watched over with respect to the detection target sound data unit 15, and the behavior description data unit. 16 is a function for writing and registering the content data of the target action.

In order to notify the sensor signal processing unit 11 of the sound signal from the transmission processing unit 3, a dedicated signal line is usually used, but here, in order to avoid the troublesomeness of laying the communication line, a communication method using electromagnetic waves is used. Alternatively, a communication method using infrared rays or other methods may be used, and any method may be used as long as the sound collected by the sound collection sensor 1 is correctly notified to the audio signal signal receiving unit 11 of the notification device 10. Absent.

The operation of the apparatus constituting the present invention will be described below. The operation sound emitted by the person being watched over is captured by the sound collection unit 2 of the sound collection sensor 1 and is notified from the audio signal transmission unit 3 to the audio signal reception unit 11 of the notification device 10. The sound signal received by the sound signal receiving unit 11 is notified to the sound analysis processing unit 12 and then subjected to fast Fourier expansion processing (hereinafter referred to as FFT processing), and the spectrum components constituting the sound signal are extracted. . Here, the partial motion of the person to be watched is estimated by a method of comparing with the feature data indicated by the spectrum for specifying the partial motion registered in the detection target sound data section 15.

For the identified partial action, a code indicating the type of the partial action is notified to the notification determination processing unit 13. The notification determination processing unit 13 determines whether or not the person to be watched behaves in accordance with the normal target behavior in accordance with the conditions registered in the behavior description data unit 16. Here, if there is an action deviating from the target action described in the action name, the failure of the person being watched over can be observed by the method using the IP network connection unit 17 and the watch device 30 via the IP network 20 by mail or telephone. Notify that it will be included.

Further, when a partial motion corresponding to the detection target sound data portion 15 is not recognized within a certain period, for example, 24 hours, or when the target behavior registered in the behavior description data portion 16 is not recognized, the IP The monitoring device 30 is notified via the network connection unit 17 and the IP network 20.

Here, each structure of the detection object sound data part 15 and the action description data part 16 is demonstrated easily. FIG. 2 shows the configuration of the detection target sound data unit 15. Here, the feature of the sound accompanying the partial action is registered. That is, it is well known that every sound signal can be viewed as a composite of sine waves composed of a plurality of frequency components.

Here, it is also widely known that the frequency component can be obtained by performing analog-digital conversion on the sound signal notified from the sound collection sensor 1 and then performing FFT on the obtained digital data. Therefore, here, the sound associated with the partial motion is registered in the detection target sound data section 15 with the frequency and the component value for the main component for each sound of the partial motion. The configuration example described in the detection target sound data unit 15 illustrated in FIG. 2 illustrates an example in which three main components are described, but this quantity may be any amount.

Next, the configuration of the action description data unit 16 will be briefly described. In FIG. 3, the structural example of the action description data part 16 is shown. Here, a plurality of partial actions and events constituting the target behavior of the watching target person are described. Also described is the transition time deemed necessary between multiple partial actions or events.

Finally, the operation of the setting unit will be described. This provides a support function for registering data related to the specific life of the watching target person in the detection target sound data unit 15 and the behavior description data unit 16. Usually, it is composed of a personal computer device called a maintenance tool. The setting unit 18 includes a command for capturing a spectrum feature (referred to as command M), a detection target data part call command (referred to as R1), a detection target data part write command (referred to as W1), and a behavior description data part call command (referred to as R2). ) And a behavior description data part write command (referred to as W2).

Here, using these commands, while a characteristic action sound produced by the person being watched over is ringing, an event is added to the analysis data of the voice analysis processing unit 12 and set and registered in the detection target sound data unit 15. . At the same time, one or more of the assigned events are combined and registered in the action description data unit 16. Once the setting in the detection target sound data unit 15 and the behavior description data unit 16 is completed, the setting unit 18 may be removed from the reporting device and operated.

In addition, in FIG. 1, although the structure which notifies a sound signal from a sound collection sensor to a notification apparatus was demonstrated, the structure which performs both sound collection and audio | voice analysis processing with a sound collection sensor, and notifies only the result to a notification apparatus may be sufficient. This is an easily imagined configuration.

According to the present invention, in a department that manages in a manner that does not put a mental burden on the person being watched over, it is possible to detect the state in a timely manner when a failure occurs while watching the living conditions of many persons. For this reason, for example, in a department that manages the health of elderly people living alone, it can be expected to reduce the burden of work.

DESCRIPTION OF SYMBOLS 1 Sound collection sensor 2 Sound collection part 3 Voice signal transmission part 4 Sensor which detects an event 10 Notification apparatus 11 Voice signal reception part 12 Voice analysis process part 13 Input / output process part 14 Report determination process part 15 Detection target sound data part 16 Action Description data part 17 IP network connection part 18 Setting part 20 IP network 30 Watching device

Claims (4)

With a configuration consisting of a sound collection sensor and a notification device, the operation is automatically estimated from the characteristics of the timbre of the sound generated with the specific operation of the person being watched over, and if the operation is not detected, the subject is in trouble. A method of watching over the person being watched, which has a function to report that it is expected. It consists of a sound collection sensor and a notification device, and automatically estimates the operation from the characteristics of the timbre of the sound generated by the specific operation of the person being watched over, and deviates from the sequence and elapsed time of multiple operations normally expected A method of watching over the target person having the function of notifying the target person that a failure is expected on the condition. It consists of a sound collection sensor, a sensor for detecting the motion, and a notification device. It automatically estimates the motion from the sensor that detects the timbre characteristics and motion of the sound that occurs with the specific motion of the person being watched over, and is normally expected. A method of watching over a monitoring target person having a function of notifying the target person that a failure is expected on the condition that it deviates from the order of the plurality of operations and the elapsed time. A monitoring device using the method according to claim 1.

Images