Action recommendation device, action recommendation method, and program

Abstract

To mitigate the problem of physical or mental stress that may be experienced by disaster victims who are sheltering in vehicles.SOLUTION: An action recommendation device according to the present disclosure comprises acquisition means for acquiring information on daily life of disaster victims sheltering in a vehicle as obtained based on detection results of sensors installed in the vehicle, determination means for determining actions to be recommended to the disaster victims to maintain the health of the disaster victims on the basis of the information on daily life of the disaster victims, and output means for outputting the information on the actions. The action recommendation device can be used to support decision making of users, etc.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to an action recommendation device, an action recommendation method, and a program. [Background technology]

[0002] When a disaster occurs, victims may evacuate in vehicles, and may end up living in those vehicles.

[0003] Patent Document 1 describes a disaster evacuation support device that acquires disaster situation information and disaster victim situation information and generates an evacuation plan. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2023-168077 Summary of the Invention [Problem to be solved by the invention]

[0005] When disaster victims take refuge in their vehicles, they may be subjected to physical or mental stress.

[0006] One of the purposes of the present disclosure is to provide an action recommendation device etc. that can promote the health of disaster victims who are living in evacuation centers in vehicles during a disaster. [Means for solving the problem]

[0007] An action recommendation device in one aspect of the present disclosure includes an acquisition means for acquiring information about the lives of disaster victims living in evacuation shelters in vehicles, obtained based on the detection results of sensors installed in the vehicles; a decision means for deciding actions to recommend to the disaster victims for maintaining their health based on the information about the disaster victims' lives; and an output means for outputting the action information.

[0008] A method for recommending actions in one aspect of the present disclosure acquires information about the lives of disaster victims living in evacuation shelters in vehicles based on the detection results of sensors installed in the vehicles, determines actions to be recommended to the disaster victims to maintain their health based on the information about the disaster victims' lives, and outputs information about the actions to maintain their health.

[0009] A program in one aspect of the present disclosure causes a computer to acquire information about the lives of disaster victims living in evacuation shelters in vehicles, based on the detection results of sensors installed in the vehicles, determine recommended actions for maintaining the health of the disaster victims based on the information about the disaster victims' lives, and output information about the actions for maintaining health.

[0010] Each program may be stored in a non-transitory computer-readable recording medium. [Effects of the Invention]

[0011] One example of the effect of the present disclosure is to provide an action recommendation device or the like that can promote the health of disaster victims living in evacuation shelters in vehicles during a disaster. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of an action recommendation system including an action recommendation device. [Figure 2] FIG. 1 is a block diagram showing an example of the configuration of an action recommendation device. [Figure 3] FIG. 10 is a block diagram showing another example of the configuration of the action recommendation device. [Figure 4] 10 is a flowchart showing the operation of the action recommendation device. [Figure 5] FIG. 1 is a block diagram showing an example of the configuration of an action recommendation device. [Figure 6] 10 is a flowchart showing the operation of the action recommendation device. [Figure 7] FIG. 2 is a diagram illustrating an example of a hardware configuration of an action recommendation device. DETAILED DESCRIPTION OF THE INVENTION

[0013] Embodiments of the present disclosure will be described in detail with reference to the drawings.

[0014] [First embodiment] The configuration of an action recommendation system including an action recommendation device 10 will be described with reference to Fig. 1. Fig. 1 is a diagram showing an example of the configuration of an action recommendation system including an action recommendation device 10. Referring to Fig. 1, the action recommendation device 10 is connected to a sensor 901 installed in a vehicle 90 via a wired or wireless network.

[0015] The vehicle 90 is a vehicle used by a disaster victim. An example of the vehicle 90 is a passenger car. In particular, the vehicle 90 is a vehicle in which a disaster victim lives as an evacuee. The vehicle 90 is, for example, a private car owned by the disaster victim. The vehicle 90 may also be a vehicle prepared for use by a disaster victim after a disaster occurs. Referring to FIG. 1 , the action recommendation device 10 is connected to one vehicle 90. The number of vehicles 90 connected to the action recommendation device 10 is not limited to one. The action recommendation device 10 may be connected to multiple vehicles 90. When the action recommendation device 10 is connected to multiple vehicles 90, the multiple vehicles 90 may be owned by different disaster victims. Furthermore, the multiple vehicles 90 may be owned by the same disaster victim.

[0016] In the following description, the period after a disaster occurs may be referred to as a disaster time. Furthermore, normal times refer to, for example, the period before a disaster occurs or when a disaster has not yet occurred. Furthermore, victims may use the vehicle 90 even in normal times. Therefore, a victim in normal times refers to the state of a person who becomes a victim due to the occurrence of a disaster before the disaster occurs. Furthermore, in the following description, the occupants of the vehicle 90 include the victim.

[0017] The sensor 901 installed in the vehicle 90 is a sensor used for operating the vehicle 90 under normal circumstances. The sensor 901 is a sensor used, for example, for the safe running of the vehicle 90 before a disaster occurs. The sensor 901 may also be a sensor used, for example, to ensure the safety of occupants aboard the vehicle 90 before a disaster occurs. Specific examples of the sensor 901 are a load sensor and a capacitance sensor. The load sensor is provided, for example, on a frame of a seat of the vehicle 90. The capacitance sensor is provided on a seat cushion of the vehicle 90. The sensor 901 may also be a pressure sensor that measures the internal pressure of an air bag provided inside a seat cover that constitutes the seating surface of the seat of the vehicle 90. These sensors 901 are used for the purpose of detecting occupants of the vehicle 90 under normal circumstances. The sensor 901 is not limited to these. The sensor 901 may also be a radar sensor that is used for the purpose of detecting occupants. The radar sensor is, for example, a millimeter-wave radar.

[0018] In the example shown in Fig. 1, the sensor 901 is a sensor installed in a seat of the vehicle 90. The sensor 901 is represented by a rectangle in Fig. 1. The sensor installed in the seat of the vehicle 90 may be built into the seat. Furthermore, examples of the sensor 901 are not limited to these.

[0019] Sensor 901 may be a camera that records the state of vehicle 90 while it is traveling or when an accident occurs. The camera may be an exterior camera that records the outside of vehicle 90. The camera may be an interior camera that records the inside of vehicle 90. The camera may be, for example, a drive recorder. Sensor 901 may be, for example, a sensor installed on the steering wheel of vehicle 90. The sensor installed on the steering wheel may be, for example, a sensor that measures the pulse rate of the driver of vehicle 90.

[0020] In the example shown in FIG. 1, the number of sensors 901 is one. The number of sensors 901 installed in the vehicle 90 is not limited to one. There may be multiple sensors 901. When there are multiple sensors 901, the multiple sensors 901 may include sensors 901 of the same type. Furthermore, the multiple sensors 901 may include sensors 901 of different types.

[0021] The action recommendation device 10 may be connected to an in-vehicle device (not shown) mounted on a vehicle 90. The in-vehicle device is connected to a sensor 901 via a wired or wireless network. Therefore, the in-vehicle device can acquire the detection results of the sensor 901. Furthermore, the in-vehicle device and the action recommendation device 10 are connected via a wired or wireless network. The number of in-vehicle devices is not particularly limited. Here, the action recommendation device 10 may be included in the in-vehicle device as part of the functions of the in-vehicle device.

[0022] Referring to FIG. 1 , the action recommendation device 10 may be connected to a communication terminal 91 via a wired or wireless network. The communication terminal 91 is a terminal that can be used by a disaster victim. The communication terminal 91 is also a terminal that the disaster victim uses under normal circumstances. The communication terminal 91 used by the disaster victim during a disaster and the communication terminal 91 used by the disaster victim under normal circumstances may be the same or different. The communication terminal 91 is, for example, a smartphone, a tablet terminal, a wearable terminal, or a personal computer. The communication terminal 91 is not limited to these. The communication terminal 91 is not limited to one unit. The action recommendation device 10 may be connected to multiple communication terminals 91. The action recommendation device 10 may be connected to the communication terminals 91 of multiple disaster victims. The action recommendation device 10 may also be connected to multiple communication terminals 91 used by a single disaster victim.

[0023] A sensor 911 is installed in the communication terminal 91. The sensor 911 is, for example, an acceleration sensor. The sensor 911 installed in the communication terminal 91 is not limited to this. There may be multiple sensors 911. When there are multiple sensors 911, the multiple sensors 911 may include sensors 911 of the same type. Furthermore, the multiple sensors 911 may include sensors 911 of different types.

[0024] The configuration of the action recommendation device 10 will be described with reference to FIG. 2. FIG. 2 is a block diagram showing an example of the configuration of the action recommendation device 10. The action recommendation device 10 includes an acquisition unit 101, a determination unit 102, and an output unit 103. The action recommendation device 10 may also include an estimation unit 104. FIG. 3 is a block diagram showing another example of the configuration of the action recommendation device 10. In the example of FIG. 3, the action recommendation device 10 includes an acquisition unit 101, a determination unit 102, an output unit 103, and an estimation unit 104. In the following description, the action recommendation device 10 will be described as including the acquisition unit 101, the determination unit 102, and the output unit 103.

[0025] The acquisition unit 101 is one aspect of an acquisition means for acquiring information about the lives of disaster victims living as evacuees in the vehicle 90, which information is obtained based on the detection results of the sensor 901 installed in the vehicle 90. The acquisition unit 101 analyzes the detection results of the sensor 901 installed in the vehicle 90 during a disaster. The acquisition unit 101 then acquires information about the lives of the disaster victims by analyzing the detection results. Here, the analysis of the detection results of the sensor 901 may be performed by a component different from the acquisition unit 101. For example, an analysis unit may analyze the detection results of the sensor 901. In this case, the acquisition unit 101 can acquire information about the lives of the disaster victims by acquiring the analysis results by the analysis unit. Furthermore, the acquisition unit 101 may acquire the detection results of the sensor 901 as information about the lives of the disaster victims. The acquisition unit 101 does not have to analyze the detection results of the sensor 901.

[0026] Information about the lives of disaster victims is information about the lives of disaster victims in evacuation shelters. In other words, information about the lives of disaster victims is information that indicates how disaster victims spend their time in evacuation shelters. Information about the lives of disaster victims includes behavioral history inside vehicle 90. The behavioral history inside vehicle 90 is a record of the behavior of disaster victims inside vehicle 90. The behavioral history inside vehicle 90 indicates how disaster victims spend their time inside vehicle 90 while living as evacuees.

[0027] The behavioral history inside the vehicle 90 includes, for example, posture inside the vehicle. The posture inside the vehicle is, for example, the seated or lying position of the occupant on the seat of the vehicle 90. The behavioral history inside the vehicle 90 may also include postural habits inside the vehicle. Postural habits inside the vehicle are the tendency of the occupant's posture when assuming the same posture inside the vehicle. One example of a postural habit is when the victim sits in a hunched position. Another example of a postural habit is when the victim sits in a seated position, leaning to the left or right. Yet another example of a postural habit is when the victim sleeps facing to the left or right when lying down.

[0028] The acquisition unit 101 can acquire, for example, posture or posture habits inside the vehicle obtained based on the detection results of a sensor intended for occupant detection. That is, the acquisition unit 101 can acquire information about posture or posture habits inside the vehicle. Here, a sensor intended for occupant detection may be used during normal operation, for example, to prevent occupants from falling asleep while driving. For example, a load sensor, capacitance sensor, or pressure sensor included in the sensor 901 can detect load distribution or pressure distribution on the seat surface, backrest, or frame. Then, posture or posture habits inside the vehicle are estimated by a trained model that uses the load distribution or pressure distribution detected by the sensor 901 as input. The acquisition unit 101 can acquire the estimation result of posture or posture habits inside the vehicle using the trained model.

[0029] The acquisition unit 101 may also acquire the posture or posture habits inside the vehicle from the detection results of the radar sensor. The posture or posture habits inside the vehicle are estimated by a trained model that uses point cloud data collected by the radar sensor as input. Then, the acquisition unit 101 can acquire the estimation result of the posture or posture habits inside the vehicle using the trained model.

[0030] Furthermore, the acquisition unit 101 may acquire posture or posture habits inside the vehicle using images captured by an interior camera. Here, the images include moving images. The images may also be a set of images captured at a predetermined time interval. The predetermined time is a preset time. Joint keypoints of the occupant's body are estimated by a trained model that inputs images captured by the interior camera. Furthermore, posture or posture habits inside the vehicle are estimated by a trained model that inputs the joint keypoints of the occupant's body. The acquisition unit 101 can acquire an estimation result of posture or posture habits inside the vehicle using the trained model. The trained model for estimating joint keypoints and the trained model for estimating posture or posture habits inside the vehicle may be different trained models or may be the same trained model.

[0031] The above-mentioned trained model may be included in the acquisition unit 101. Furthermore, analysis of the detection results of the sensor 901 using the trained model may be performed by the above-mentioned analysis unit. Furthermore, the trained model may be included in an information processing device (not shown) different from the action recommendation device 10. In this case, the acquisition unit 101 can acquire the analysis results performed in the information processing device. The action recommendation device 10 is connected to the information processing device via a wired or wireless network.

[0032] The behavior history inside the vehicle 90 may include the time spent seated inside the vehicle. The time spent seated inside the vehicle is the time during which the occupant is detected as being seated. For example, at least one of a load sensor, a capacitance sensor, and a pressure sensor is used to detect the occupant's seating. A radar sensor or an in-vehicle camera may also be used to detect the occupant's seating. The acquisition unit 101 acquires the time during which the occupant is detected as being seated by these sensors as the time spent seated.

[0033] The behavior history in the vehicle 90 may include the amount of physical activity while maintaining a seated position. The amount of physical activity while maintaining a seated position is, for example, a change in the position of the center of gravity of the occupant while sitting. The acquisition unit 101 can use, for example, detection results from at least one of a load sensor, a capacitance sensor, a pressure sensor, a radar sensor, and an in-vehicle camera. The acquisition unit 101 acquires the change in the position of the center of gravity of the occupant obtained based on the detection results as the amount of physical activity while maintaining a seated position.

[0034] The behavioral history in the vehicle 90 may include vital signs. One example of a vital sign is a heart rate. Another example of a vital sign is a respiratory rate. Vital signs are not limited to these. The acquisition unit 101 can acquire vital signs obtained based on the detection results of a sensor 901 installed in the vehicle 90. If the sensor 901 is a load sensor, the load sensor detects a ballistocardiogram. The acquisition unit 101 can acquire a heart rate estimated from the detection results of the ballistocardiogram by the load sensor. The acquisition unit 101 can also acquire a respiratory rate estimated from the detection results of the ballistocardiogram by the load sensor. The heart rate or respiratory rate may be estimated by the analysis unit. For example, the analysis unit applies a Fourier transform to the signal detected by the load sensor. Then, the analysis unit can estimate the respiratory rate by considering a peak of a frequency close to the respiratory cycle as a peak of fluctuations associated with breathing.

[0035] The acquiring unit 101 may also acquire vital signs from the detection results of a capacitance sensor, a pressure sensor, or a radar sensor. The heart rate or the respiratory rate is estimated based on periodic fluctuations in the signal from the sensor 901 that accompany heartbeats or breathing.

[0036] The behavioral history in the vehicle 90 may include heart rate variability or pulse rate variability. Heart rate variability is the variation in the interval between heartbeats. Pulse rate variability is the variation in the interval between pulses. Heart rate variability and pulse rate variability may be used as indicators of autonomic nervous function or stress. The acquisition unit 101 can acquire heart rate variability or pulse rate variability obtained based on the detection results of a sensor 901 installed in the vehicle 90. The signal of the sensor 901 includes peaks of heartbeats or pulse rates due to variations caused by heartbeats. Therefore, the heart rate variability or pulse rate variability is estimated based on the interval between peaks. The acquisition unit 101 can acquire the heart rate variability or pulse rate variability estimated based on the interval between peaks. The estimation of the heart rate variability or pulse rate variability may be performed by an analysis unit.

[0037] The behavior history in the vehicle 90 is not limited to these. Furthermore, the method of acquiring the behavior history in the vehicle 90 is not limited to these.

[0038] The information about the life of the disaster victim may include, for example, a behavior history outside the vehicle 90. The behavior history outside the vehicle 90 is, for example, a record of the disaster victim's behavior around the vehicle. An example of the disaster victim's behavior around the vehicle is the disaster victim getting out of the vehicle 90 and stretching. The acquisition unit 101 may acquire the behavior history outside the vehicle 90 using, for example, images captured by an exterior camera as the sensor 901 installed in the vehicle 90. The exterior camera captures the disaster victim's behavior outside the vehicle within the range that the camera can capture. Here, as described above, the images include moving images. The images may also be a set of images captured at a predetermined time interval. The exterior camera may be used during normal times to capture images of vehicles in front of and behind the vehicle 90.

[0039] The acquisition unit 101 may acquire, for example, the victim's movement posture or postural habits outside the vehicle. An example of the victim's movement posture outside the vehicle is a hunched posture when walking. The method for estimating the victim's movement posture or postural habits is the same as the method for estimating the posture or postural habits inside the vehicle. Furthermore, the estimation of the movement posture or postural habits may be performed by the analysis unit.

[0040] Information about the victim's life is not limited to these. Information about the victim's life may also include the victim's physical condition. An example of the physical condition is the condition of the victim's joints or muscles. The physical condition may particularly be the condition of the victim's joints or muscles when the victim assumes a specific posture or movement. The joint condition is, for example, the range of motion of the joint. The specific posture or movement may be predetermined. The acquisition unit 101 can acquire the victim's physical condition using, for example, images captured by an exterior camera as the sensor 901 installed in the vehicle 90. The exterior camera captures images of the victim when the victim assumes a specific posture or movement. The condition of the victim's joints or muscles is then estimated using the images captured by the exterior camera. The images include moving images. The method for estimating the condition of the joints or muscles is performed using known technology. For example, the condition of the victim's joints or muscles is estimated from the shape of the victim's spine included in the moving images. The estimation of the condition of the joints or muscles may be performed by an analysis unit.

[0041] The information about the life of the victim may include the mental state of the victim. The mental state may be, for example, the emotion or stress of the victim. The emotion or stress of the victim may be estimated, for example, based on the victim's vital signs. The emotion or stress of the victim may be estimated, for example, based on the victim's heart rate variability or pulse rate variability. Alternatively, the emotion or stress of the victim may be estimated from the facial expression of the victim captured by an in-car camera. Publicly known techniques are used to estimate the emotion or stress of the victim. The emotion or stress of the victim may be estimated by an analysis unit.

[0042] The information about the lives of the disaster victims may include information about the in-vehicle environment. The in-vehicle environment includes information about the temperature or humidity inside the vehicle 90. The acquisition unit 101 can acquire the information about the in-vehicle environment using a thermometer or hygrometer of the vehicle 90 as the sensor 901 installed in the vehicle 90. The thermometer or hygrometer of the vehicle 90 may be a sensor that is used to keep the in-vehicle environment comfortable under normal circumstances.

[0043] The acquisition unit 101 may acquire information about the life of the disaster victim under normal circumstances, obtained based on the detection results of the sensor 901 installed in the vehicle 90. The information about the life of the disaster victim under normal circumstances is information about the life inside the vehicle when the disaster victim uses the vehicle 90 under normal circumstances. The information about the life of the disaster victim under normal circumstances includes at least one of the history of actions inside the vehicle 90, the history of actions outside the vehicle 90, the physical condition, the mental state, and the in-vehicle environment.

[0044] The behavioral history inside the vehicle 90 under normal circumstances is a record of the behavior inside the vehicle 90 when the victim uses the vehicle 90 under normal circumstances. Furthermore, the behavioral history outside the vehicle 90 under normal circumstances is a record of the behavior around the vehicle when the victim uses the vehicle 90 under normal circumstances. When the victim uses the vehicle 90 under normal circumstances, this includes when the victim is in the vehicle 90 under normal circumstances. Furthermore, when the victim uses the vehicle 90 under normal circumstances, this may include, for example, when the victim performs maintenance or cleaning on the vehicle 90 under normal circumstances. The contents of the physical condition, mental state, and in-vehicle environment are the same as those in the methods described above. The methods for acquiring the behavioral history inside the vehicle 90, the behavioral history outside the vehicle 90, the physical condition, mental state, and in-vehicle environment are the same as those in the methods described above.

[0045] The acquisition unit 101 can acquire information related to the lifestyles of multiple disaster victims. Multiple disaster victims may be living in evacuation shelters in the vehicle 90. Furthermore, multiple disaster victims may use the vehicle 90 even during normal times. In addition, the seats in which multiple occupants sit may not be the same. Therefore, the acquisition unit 101 can distinguish and acquire information related to multiple occupants based on the detection results of the sensor 901 installed in the vehicle 90.

[0046] A case will be described in which the acquisition unit 101 uses, for example, a load sensor or a pressure sensor as the sensor installed in the vehicle 90. The acquisition unit 101 can distinguish between occupants using their weights. The acquisition unit 101 can distinguish between occupants based on differences in their weights. Furthermore, the acquisition unit 101 can associate the behavior history of each occupant even when the seating position of the occupant changes.

[0047] Furthermore, when the acquisition unit 101 acquires the physical condition, it can distinguish the occupant from an image of the occupant outside the vehicle. Publicly known technology can be used to distinguish the occupant included in the image. For example, the weight of the occupant is estimated by a machine learning model that uses an image of the occupant outside the vehicle 90 as input. Then, the acquisition unit 101 can associate the behavior history of the occupant with the physical condition.

[0048] Detection by the sensor 901 installed in the vehicle 90 may be performed continuously. Furthermore, detection by the sensor 901 installed in the vehicle 90 may be performed periodically. The timing of detection by the sensor 901 may be different or the same during a disaster and during normal times. For example, detection may be performed continuously during a disaster, while detection may be performed at predetermined intervals during normal times. Furthermore, if detection is performed at predetermined intervals during both a disaster and normal times, the detection intervals may be different during a disaster and during normal times. During a disaster, the vehicle 90 may run out of fuel. Therefore, the detection interval during a disaster may be longer than the detection interval during normal times.

[0049] There is no particular limitation on the timing at which the detection results by the sensor 901 are analyzed. For example, the detection results by the sensor 901 may be analyzed in real time. Also, for example, the detection results by the sensor 901 for one day may be analyzed all at once.

[0050] The acquisition unit 101 may store information obtained from the detection results of the sensor 901 installed in the vehicle 90, for example, in a database. Furthermore, when an analysis unit analyzes the detection results of the sensor 901 installed in the vehicle 90, the analysis unit may store the analysis results in a database. Then, the determination unit 102 or the estimation unit 104, which will be described later, can refer to the information stored in the database as necessary.

[0051] The acquisition unit 101 may acquire information about the lives of disaster victims obtained from detection results of a sensor other than the sensor installed in the vehicle 90. In the event of a disaster, the acquisition unit 101 can acquire information about the lives of disaster victims obtained based on detection results of a sensor 911 installed in a communication terminal 91 used by the disaster victim.

[0052] The information about the life of the disaster victim includes, for example, a behavioral history outside the vehicle 90. Examples of the behavioral history outside the vehicle 90 include the location information and the number of steps taken by the disaster victim. The acquisition unit 101 can acquire the behavioral history of the disaster victim by analyzing the detection results of the sensor 911 installed in the communication terminal 91. The analysis of the detection results may be performed by the analysis unit described above. The acquisition unit 101 may also acquire the behavioral history of the disaster victim, which is the result of the analysis by the communication terminal 91, from the communication terminal 91. The acquisition unit 101 can acquire the behavioral history of the disaster victim obtained based on the detection results of the sensor 911 installed in the communication terminal 91 during disasters and normal times.

[0053] The acquisition unit 101 may acquire basic information of the disaster victim. The basic information of the disaster victim includes attribute information or medical history of the disaster victim. The attribute information of the disaster victim includes, for example, at least one of the disaster victim's age, gender, family composition, and place of residence. The disaster victim's medical history includes information on illnesses the disaster victim is currently suffering from and illnesses the disaster victim has suffered from in the past. The disaster victim's medical history may also include a medication history. The acquisition unit 101 can acquire the basic information of the disaster victim from the communication terminal 91. The disaster victim can input their own basic information into the communication terminal 91. The acquisition unit 101 can then acquire the basic information input by the disaster victim from the communication terminal 91. The acquisition of the basic information by the acquisition unit 101 may be performed during normal times or during a disaster.

[0054] The acquisition unit 101 may acquire vehicle exterior environment information during a disaster and under normal circumstances from the communication terminal 91. The vehicle exterior environment during a disaster refers to the environment outside the vehicle 90 where disaster victims are taking refuge. The vehicle exterior environment during normal circumstances refers to the environment outside the vehicle 90 used by disaster victims under normal circumstances. The vehicle exterior environment includes at least one of weather, temperature, and humidity. The vehicle exterior environment during a disaster includes at least one of disaster occurrence information and the damage situation caused by the disaster. The disaster occurrence information is information about a disaster that occurred around the vehicle. The disaster information includes at least one of the type of disaster that occurred, the date and time of occurrence, and the extent of the disaster. The damage situation caused by the disaster indicates the type and extent of damage caused by the disaster. The damage caused by the disaster includes damage to transportation infrastructure, damage to living infrastructure, damage to housing, etc. The extent of the damage includes the extent of the damage. The acquisition unit 101 can acquire the vehicle exterior environment information from the communication terminal 91. The acquisition unit 101 may also acquire the vehicle exterior environment information from, for example, a database. Here, the database manages, for example, information on weather, temperature, and humidity, and the database may also store information on the occurrence of disasters or information on the damage caused by disasters.

[0055] Detection by the sensor 911 installed in the communication terminal 91 may be performed continuously. Furthermore, detection by the sensor 911 installed in the communication terminal 91 may be performed periodically. The timing of detection by the sensor 911 installed in the communication terminal 91 may be different or the same during a disaster and during normal times. For example, detection may be performed continuously during a disaster, while detection may be performed at predetermined intervals during normal times.

[0056] Furthermore, there is no particular restriction on the timing at which the detection results by the sensor 911 installed in the communication terminal 91 are analyzed. For example, the detection results by the sensor 911 installed in the communication terminal 91 may be analyzed in real time. Also, for example, one day's worth of detection results by the sensor 911 installed in the communication terminal 91 may be analyzed all at once.

[0057] The acquisition unit 101 may store information obtained from the detection results of the sensor 911 installed in the communication terminal 91, for example, in a database. Furthermore, when an analysis unit analyzes the detection results of the sensor 911 installed in the communication terminal 91, the analysis unit may store the analysis results in a database. Then, the determination unit 102 or the estimation unit 104, which will be described later, can refer to the information stored in the database as necessary.

[0058] The acquisition unit 101 may acquire information other than information about the lives of disaster victims, in addition to information about the lives of disaster victims.

[0059] The acquisition unit 101 may further acquire information regarding the usage status of the vehicle 90 under normal circumstances. The usage status of the vehicle 90 under normal circumstances includes, for example, the driving time or driving frequency under normal circumstances. Furthermore, the usage status of the vehicle 90 under normal circumstances includes the rest time or intervals between rests taken while driving. The usage status of the vehicle 90 under normal circumstances may represent the range of activities that a disaster victim can tolerate in the vehicle under normal circumstances. For example, a disaster victim who drives for long periods under normal circumstances may feel less stressed living as an evacuation center in the vehicle 90 than a disaster victim who rarely drives under normal circumstances. Furthermore, a disaster victim who takes frequent breaks while driving under normal circumstances may want to get out of the vehicle frequently to exercise even when living as an evacuation center in the vehicle 90.

[0060] The acquisition unit 101 can acquire the usage status of the vehicle 90 under normal circumstances using the sensor 901 installed in the vehicle 90. For example, the acquisition unit 101 analyzes the usage status of the vehicle 90 under normal circumstances based on images captured by an interior camera of the vehicle 90. Then, the acquisition unit 101 can acquire the analysis result of the usage status of the vehicle 90. Publicly known technology can be used to analyze the usage status of the vehicle 90. Furthermore, the acquisition unit 101 can analyze the usage status of the vehicle 90 under normal circumstances based on the operating status of the engine of the vehicle 90. Then, the acquisition unit 101 can acquire the analysis result of the usage status of the vehicle 90. For example, the acquisition unit 101 may determine the time during which the engine is running as the normal driving time. Alternatively, the acquisition unit 101 may analyze the usage status of the vehicle 90 based on location information of the vehicle 90. Then, the acquisition unit 101 can acquire the analysis result of the usage status of the vehicle 90. For example, when the vehicle 90 is stopped at a service area on a highway for a predetermined time, the acquisition unit 101 can analyze the stopping time as a rest time. A publicly known technique can be used to analyze the usage status of the vehicle 90. The analysis of the usage status of the vehicle 90 may also be performed by the above-mentioned analysis unit.

[0061] The acquisition unit 101 may acquire the usage status of the vehicle 90 under normal circumstances using a sensor 911 installed in the communication terminal 91. The acquisition unit 101 may also acquire the usage status of the vehicle 90 under normal circumstances from input by a disaster victim to the communication terminal 91.

[0062] The acquisition unit 101 may further acquire information on items that can be used in the event of a disaster. The acquisition unit 101 can particularly acquire information on items that can be used by the disaster victim when the disaster victim performs the recommended action. The items that can be used in the event of a disaster are, for example, items that can assist the disaster victim in exercising. The information on the items that can be used in the event of a disaster includes the contents of the items and the locations of the items.

[0063] The acquisition unit 101 acquires information about items inside and around the vehicle using a sensor 901 installed in the vehicle 90. The acquisition unit 101 can acquire information about items that can be used in the event of a disaster by analyzing images captured by a camera installed in the vehicle 90. The cameras installed in the vehicle 90 include at least one of an exterior camera and an interior camera. The acquisition unit 101 can acquire information about items around the vehicle from the analysis results of images captured by the exterior camera. Furthermore, the acquisition unit 101 can acquire information about at least items inside the vehicle from the analysis results of images captured by the interior camera. The analysis of images captured by the camera installed in the vehicle 90 can be performed using known technology. Furthermore, the analysis of images captured by the camera installed in the vehicle 90 may be performed by the analysis unit described above.

[0064] The acquisition unit 101 can further acquire information on items that can be used in the event of a disaster from the communication terminal 91. The acquisition unit 101 may acquire information on items that can be used in the event of a disaster, obtained based on images taken by a camera installed in the communication terminal 91. The acquisition unit 101 can acquire information on items that can be used in the event of a disaster by analyzing the images taken by the camera installed in the communication terminal 91. Publicly known technology can be used to analyze the images taken by the camera installed in the communication terminal 91. Furthermore, the analysis of the images taken by the camera installed in the communication terminal 91 may be performed by the analysis unit described above.

[0065] The acquisition unit 101 may acquire information about items that can be used in the event of a disaster through input by the disaster victim to the communication terminal 91. For example, the disaster victim can input information about items that can be used in the event of a disaster to the communication terminal 91. Then, the acquisition unit 101 acquires the information about items that can be used in the event of a disaster that has been input by the disaster victim.

[0066] The information on items that can be used in the event of a disaster acquired by the acquisition unit 101 is stored in a database, for example, in the form of a list. The list stored in the database is then referred to by the determination unit 102 when the determination unit 102, which will be described later, determines an action to be taken.

[0067] The acquisition unit 101 can also acquire community information related to the community to which the disaster victim belongs from the communication terminal 91. For example, the disaster victim may belong to a local community under normal circumstances. Therefore, the acquisition unit 101 acquires community information related to the community to which the disaster victim belongs. The community information is information on the frequency of the disaster victim's participation in the community and information on members belonging to the same community. The acquisition unit 101 may acquire information on items that can be used in the event of a disaster, obtained based on location information of the communication terminal 91. Furthermore, the acquisition unit 101 may acquire community information based on input by the disaster victim to the communication terminal 91. In this case, the community information may include details of community activities.

[0068] The decision unit 102 is one aspect of a decision means that decides on an action to be recommended to a disaster victim for maintaining the health of the disaster victim, based on information about the disaster victim's life.

[0069] The actions taken by disaster victims to maintain their health are actions taken by disaster victims living in evacuation centers inside the vehicle 90 to prevent damage to their health. In addition, the actions taken by disaster victims to maintain their health are actions taken to improve the health of disaster victims living in evacuation centers inside the vehicle 90. Specific examples of actions taken by disaster victims to maintain their health include stretching their arms and legs, moving from inside the vehicle to outside, eating and drinking, operating the vehicle to improve the environment inside the vehicle, and exercising.

[0070] Here, a description will be given of a method for determining an action to maintain the health of a disaster victim by the determination unit 102. The action recommendation device 10 may include an estimation unit 104.

[0071] The estimation unit 104 is one aspect of estimation means that estimates the health state of the disaster victim based on information about the disaster victim's life. The determination unit 102 can then determine an action depending on the health state estimated by the estimation unit 104. Therefore, the determination unit 102 will be described after the description of the estimation unit 104.

[0072] The estimation unit 104 estimates the health condition of a disaster victim. The health condition of a disaster victim represents the physical condition of the disaster victim living as an evacuee in the vehicle 90. Specific examples of health conditions include the risk of musculoskeletal pain, the risk of musculoskeletal disorders, joint and other movement disorders, and mental stress. Examples of musculoskeletal pain risk include the probability of lower back pain, shoulder pain, and neck pain. Examples of health conditions are not limited to these. The health condition may also be the risk of a worsening of a chronic illness or the risk of contracting a disease.

[0073] The health condition may be expressed, for example, as a probability. For example, the degree of risk of a musculoskeletal disorder may be expressed as a probability. The health condition may also be expressed as a rank or score divided into multiple stages. For example, mental stress may be expressed by a rank divided into stages according to the intensity of the stress. The health condition may also be expressed as the presence or absence of risk, disorder, or stress. Alternatively, the health condition may be expressed as the degree of change. For example, the risk of worsening a chronic illness may be expressed as the magnitude of change if the risk of worsening is increasing. The health condition may also be expressed as the rate of change within a specific period.

[0074] The estimation unit 104 may estimate the health condition of the disaster victim under normal circumstances based on information about the disaster victim's daily life.

[0075] The method of estimating the health condition of a disaster victim by the estimation unit 104 will be described. The estimation unit 104 estimates the health condition of a disaster victim using a trained model. Here, the trained model used to estimate the health condition of a disaster victim is called a health condition estimation model. The health condition estimation model is a model that learns the relationship between information about the life of a disaster victim and the health condition of the disaster victim through machine learning. The estimation unit 104 uses the information about the life of a disaster victim acquired by the acquisition unit 101 as input to the health condition estimation model. Then, the estimation unit 104 acquires an estimation result of the health condition of the disaster victim, which is the output of the health condition estimation model. The estimation unit 104 may estimate the health condition of the disaster victim using a predetermined algorithm.

[0076] When the estimation unit 104 estimates the health condition of a disaster victim during a disaster, the input information about the victim's lifestyle includes information during the disaster. When the estimation unit 104 estimates the health condition of a disaster victim during a disaster, the input information about the victim's lifestyle may include information during normal times. By using information about the victim's lifestyle during normal times to estimate the health condition of a disaster victim, the estimation unit 104 can estimate the health condition of a disaster victim during a disaster based on the victim's health condition during normal times. When the estimation unit 104 estimates the health condition of a disaster victim during normal times, the input information about the victim's lifestyle is information during normal times.

[0077] The estimation unit 104 receives, for example, the posture in the vehicle, postural habits, the time spent sitting in the vehicle, and the amount of physical activity while sitting. The estimation unit 104 can estimate the risk of musculoskeletal pain using a health state estimation model or a predetermined algorithm. The estimation unit 104 also receives, for example, vital signs, heart rate variability, emotions, and stress as input. The estimation unit 104 can estimate mental stress using a health state estimation model or a predetermined algorithm. The heart rate variability may be pulse rate variability. The estimation unit 104 also receives, for example, the state of a joint or muscle when a specific posture or movement is assumed. The estimation unit 104 can estimate a malfunction of the joint movement using a health state estimation model or a predetermined algorithm. The estimation unit 104 can also estimate the risk of musculoskeletal pain using, for example, the posture in the vehicle, postural habits, the time spent sitting in the vehicle, and the amount of physical activity while sitting, as input. This is because the way victims perceive pain varies depending on factors such as the state of their musculoskeletal system as well as mental stress.

[0078] The health state estimation model or the predetermined algorithm may be a known technique. The health state estimation model may be an additionally trained model for estimating a health state.

[0079] The health state estimation model may be an additionally trained model for estimating the importance of information used to estimate a health state. The additionally trained health state estimation model estimates, for example, the importance of information used to estimate a health state. The information used to estimate a health state is information related to the lives of disaster victims. The information related to the lives of disaster victims includes at least one of the victim's behavioral history inside vehicle 90, the victim's behavioral history outside vehicle 90, and the victim's physical condition or mental state. The information used to estimate a health state may also include information related to the victim's life under normal circumstances. The importance of each of the above information is estimated by the additionally trained health state estimation model. The importance is determined by repeatedly estimating a health state using the health state estimation model.

[0080] An example of a method for estimating importance will now be described. Estimating the importance of data X will be described as an example. A health state estimated by the estimation unit 104 using data X is compared with a health state S' estimated by the estimation unit 104 without using data X. If S and S' are close in value, the importance of data X is estimated to be low relative to the estimated health state. For example, the absolute error, squared error, or cross-entropy error between S and S' is used as an index of importance.

[0081] Another example of a method for estimating importance will be described. The importance of information can be estimated using the health condition input by the disaster victim. The estimation unit 104 estimates the health condition of the disaster victim using information related to the life of the disaster victim acquired by the acquisition unit 101. The importance of information can be estimated by comparing the estimated health condition with the health condition input by the disaster victim. Here, the input of the health condition by the disaster victim can be acquired by the acquisition unit 101 via, for example, a device installed in the vehicle 90 or a communication terminal 91.

[0082] The information to be acquired by the acquisition unit 101 is determined by a health state estimation model that has been additionally trained on estimating the importance of information. The data to be acquired by the acquisition unit 101 is determined according to the importance of the information. For example, the acquisition unit 101 may acquire information with high importance and not acquire information with low importance.

[0083] The health state estimation model may be a health state estimation model that has been additionally trained on the correlation between information obtained from the detection results of sensor 901 installed in vehicle 90 and information obtained from the detection results of sensor 911 installed in communication terminal 91 under normal circumstances. Under normal circumstances, estimation unit 104 acquires the occupant's posture, postural habits, time spent sitting in the vehicle, and amount of physical activity while sitting, based on the detection results of sensor 901 installed in vehicle 90. Furthermore, acquisition unit 101 acquires the occupant's posture, postural habits, time spent sitting in the vehicle, and amount of physical activity while sitting, based on the detection results of sensor 911 installed in communication terminal 91. Then, a correlation coefficient is calculated between the information obtained from the detection results of sensor 901 installed in vehicle 90 and information obtained from the detection results of an acceleration sensor or the like installed in communication terminal 91.

[0084] Then, in the event of a disaster, the information to be acquired by the acquisition unit 101 is determined by the health state estimation model that has been additionally trained on the correlation. The acquisition unit 101 acquires, for example, highly correlated information from either the sensor installed in the vehicle 90 or the sensor 911 installed in the communication terminal 91. This is because highly correlated information is information for which similar information can be obtained from either the sensor installed in the vehicle 90 or the sensor 911 installed in the communication terminal 91. For highly correlated information, the acquisition unit 101 may select a sensor from which to acquire information depending on the remaining battery capacity or power consumption. Furthermore, the acquisition unit 101 may switch the sensor from which to acquire information depending on the remaining battery capacity or power consumption. By acquiring multiple pieces of highly correlated information from either the sensor installed in the vehicle 90 or the sensor 911 installed in the communication terminal 91, it is possible to reduce power consumption and calculation costs.

[0085] The determination unit 102 will be described again. The determination unit 102 determines the actions to be taken by the disaster victim to maintain their health based on the health status of the disaster victim at the time of the disaster estimated by the estimation unit 104 using information related to the life of the disaster victim. The actions to be taken by the disaster victim to maintain their health are as described above.

[0086] The determination unit 102 determines an action to maintain the health of the disaster victim using a trained model that has been machine-learned. Here, the trained model used to determine an action to maintain the health of the disaster victim is called a behavioral decision model. The behavioral decision model is a model that has learned the relationship between the disaster victim's health state and the action to maintain the health of the disaster victim through machine learning. The determination unit 102 uses the disaster victim's health state estimated by the estimation unit 104 as input to the behavioral decision model. Then, the determination unit 102 obtains the result of the determination of the action to maintain the health of the disaster victim, which is the output of the behavioral decision model. The estimation unit 104 may estimate the disaster victim's health state using a predetermined algorithm.

[0087] A specific example of determining an action to maintain the health of a disaster victim will be described. For example, a case will be described in which the estimation unit 104 estimates a health condition indicating a high risk of lower back pain based on the state of the joints or muscles. In this case, the determination unit 102 can determine, as the action to maintain the health of the disaster victim, an exercise that is effective for the joints or muscles that are considered to be the main cause of the lower back pain. Furthermore, if the estimation unit 104 estimates that the disaster victim is experiencing high mental stress, the determination unit 102 may determine, as the action to maintain the health of the disaster victim, stretching of the muscles involved in breathing. In this case, the determination unit 102 may determine, as the action to maintain the health of the disaster victim, stretching that is considered to be effective in relieving anxiety and stress.

[0088] When the estimation unit 104 estimates that multiple risks or stresses are high, the determination unit 102 can determine appropriate actions for maintaining the health of the disaster victim using the behavior determination model. At this time, the scores of each of the multiple risks or stresses estimated by the estimation unit 104 are input to the behavior determination model. Then, the behavior determination model outputs appropriate actions for maintaining the health of the disaster victim.

[0089] Here, the determining unit 102 may determine the timing and location of the health maintenance behavior in addition to the content of the health maintenance behavior. The determining unit 102 may also determine the frequency of performing the health maintenance behavior. The determining unit 102 can determine the timing, location, and frequency of performing the health maintenance behavior to be appropriate depending on the content of the health maintenance behavior.

[0090] The determining unit 102 may determine one action for maintaining the health of the disaster victim. Alternatively, the determining unit 102 may determine a recommended program including a plurality of actions. For example, the recommended program may determine the order in which the plurality of actions are to be performed. Alternatively, the recommended program may determine the timing in which the plurality of actions are to be performed. The action for maintaining the health of the disaster victim determined by the determining unit 102 may be candidate actions to be performed by the disaster victim. The determining unit 102 can determine candidate actions for maintaining the health of the disaster victim based on the disaster victim's health condition at the time of the disaster. At this time, the disaster victim can select the action they wish to perform from the candidate actions.

[0091] The decision unit 102 can decide on an action to maintain the health of the disaster victim by using information other than the disaster victim's health condition in addition to the disaster victim's health condition.

[0092] The determination unit 102 may determine an action to maintain the health of the disaster victim using, for example, the in-vehicle environment. The in-vehicle environment includes information on the temperature or humidity inside the vehicle 90. Depending on the in-vehicle environment, it may be difficult for the disaster victim to perform an action to maintain the health of the disaster victim inside the vehicle 90. On the other hand, the in-vehicle environment may be the optimum in-vehicle environment for performing the health maintenance action. Therefore, the determination unit 102 may determine an action to maintain the health of the disaster victim using the in-vehicle environment.

[0093] The determination unit 102 can input, for example, information about the in-vehicle environment into the behavioral decision model. Then, the determination unit 102 can acquire, as an output of the behavioral decision model, a behavior for maintaining health determined based on the environment outside the vehicle. Also, an appropriate environment may be determined in advance for each behavior for maintaining health. The determination unit 102 may determine, as a behavior for maintaining health, a behavior for which the in-vehicle environment matches the appropriate environment.

[0094] The determination unit 102 may determine an action for maintaining the health of the disaster victim using, for example, the outside-vehicle environment. The outside-vehicle environment includes at least one of weather, temperature, and humidity. Depending on the outside-vehicle environment, it may be difficult for the disaster victim to perform an action for maintaining the health of the disaster victim inside the vehicle 90. For example, if the temperature outside the vehicle is too high, it may be difficult for the disaster victim to exercise outside the vehicle. Therefore, the determination unit 102 may determine an action for maintaining the health of the disaster victim using the outside-vehicle environment.

[0095] The determination unit 102 may determine a health maintenance action for a disaster victim using, for example, information about the occurrence of a disaster or the damage situation caused by the disaster. As described above, the outside-vehicle environment includes at least one of information about the occurrence of a disaster or the damage situation caused by the disaster. Depending on the disaster situation, some actions may be difficult for a disaster victim to perform. For example, in the case of a heavy rain disaster, if the rain continues to fall, it may be difficult for a disaster victim to perform a health maintenance action outside the vehicle. Therefore, the determination unit 102 determines a health maintenance action that can be performed inside the vehicle 90 using information about the occurrence of a disaster or the damage situation caused by the disaster. Furthermore, damage caused by a disaster may include poor road and soil conditions around evacuation sites. In such cases, for example, it may be difficult for a disaster victim to walk away from the vehicle 90. Therefore, the determination unit 102 determines a health maintenance action that can be performed inside the vehicle 90 or around the vehicle 90 using information about the occurrence of a disaster or the damage situation caused by the disaster.

[0096] The determination unit 102 can input, for example, information about the outside-vehicle environment to the behavioral decision model. Then, the determination unit 102 can acquire, as an output of the behavioral decision model, a health maintenance behavior determined based on the outside-vehicle environment. Also, an appropriate environment may be determined in advance for each health maintenance behavior. The determination unit 102 may determine, as a health maintenance behavior, a behavior in which the outside-vehicle environment matches the appropriate environment. Also, the determination unit 102 may determine a health maintenance behavior for the disaster victim using the inside-vehicle environment and the outside-vehicle environment.

[0097] The determination unit 102 may determine an action for maintaining the health of a disaster victim using information regarding the usage of the vehicle 90 under normal circumstances. The usage of the vehicle 90 under normal circumstances includes, as described above, the driving time or driving frequency under normal circumstances. Furthermore, the usage of the vehicle 90 under normal circumstances includes the rest time or the interval between rests taken while driving. A person who drives infrequently under normal circumstances may be more likely to feel stressed while living as an evacuee in the vehicle 90 than a person who drives frequently under normal circumstances. Therefore, the determination unit 102 may determine an action for relieving stress as an action for maintaining health for a person who drives infrequently under normal circumstances. Furthermore, the determination unit 102 may determine, as a recommended program for maintaining health, an action in which the disaster victim gets out of the vehicle and stretches at the same interval as the interval between rests taken under normal circumstances. For a disaster victim whose physical condition deteriorates if he or she continues driving for a predetermined period of time under normal circumstances, the determination unit 102 may determine, as an action for maintaining health, a recommended program in which the disaster victim repeatedly stretches at intervals shorter than the predetermined time.

[0098] The determination unit 102 can input, for example, information on the usage status of the vehicle 90 under normal circumstances to the behavioral decision model. Then, the determination unit 102 can acquire, as an output from the behavioral decision model, an action for maintaining health that is determined based on the information on the usage status of the vehicle 90 under normal circumstances. Furthermore, an appropriate action for maintaining health may be determined in advance for the information on the usage status of the vehicle 90 under normal circumstances. The determination unit 102 may determine, as the action for maintaining health, an action that corresponds to the information on the usage status of the vehicle 90 under normal circumstances.

[0099] The determining unit 102 may also determine an action for maintaining the health of a disaster victim using information about items that can be used in the event of a disaster. Some health maintenance actions can be more effective when performed using items. Therefore, the determining unit 102 may determine, as the health maintenance action, an action performed using items that can be used in the event of a disaster. As described above, the acquiring unit 101 stores the acquired information about items that can be used in the event of a disaster in a database, for example, in list format. Furthermore, for each health maintenance action, necessary items and substitutes for the items are determined in advance. That is, information associating the health maintenance action with the necessary items and substitutes is stored, for example, in a database. Therefore, the determining unit 102 can determine, as the health maintenance action, an action for which an item or substitute item required for the health maintenance action is included in a list of information about items that can be used in the event of a disaster.

[0100] A specific example of determining an action to maintain the health of a disaster victim using information about items that can be used in the event of a disaster will be described. An example of an exercise that can be more effective when performed using a stretching pole will be described. If a disaster victim performs this exercise, it is unlikely that the stretching pole will be stored in the vehicle 90 during a disaster. However, a simple substitute for a stretching pole can be made by rolling a bath towel into a cylindrical shape. A simple substitute for a stretching pole can also be made by wrapping a towel around the core of a wrap. For an exercise using a stretching pole, it is predetermined that a substitute for the stretching pole is a towel or a wrap core. Therefore, if a towel or a wrap core is included in the list of information about items that can be used in the event of a disaster, an exercise that can be more effective when performed using a stretching pole can be determined as an action to maintain health.

[0101] The determination unit 102 may use the community information to determine the health maintenance behavior of the disaster victim. As described above, the community information includes the frequency of community participation. Therefore, the determination unit 102 may determine, as the health maintenance behavior, a recommended program in which the disaster victim exercises at the same frequency as the frequency with which the disaster victim participated in the community. Furthermore, the community information includes information on participating members of the community to which the disaster victim belongs. For example, the determination unit 102 may determine, as the health maintenance behavior, exercising with participating members of the community even at an evacuation destination based on location information of the participating members. For example, disaster victims who participate in the same community may evacuate to the same evacuation destination or nearby evacuation destinations. Therefore, even during a disaster, disaster victims may be able to refresh themselves by exercising with participating members of the community they normally participate in.

[0102] Alternatively, the determining unit 102 may determine that disaster victims who normally belong to different communities exercise together during a disaster, as a health maintenance behavior. When evacuating to a location far from their homes during a disaster, it may be difficult for disaster victims to participate in the communities they normally belong to. However, even different communities may have the same activities. For example, among disaster victims evacuated to the same evacuation site, there may be disaster victims who belong to communities with the same activities. In this case, the determining unit 102 may determine that disaster victims who belong to communities with the same activities get together to exercise, as a health maintenance behavior.

[0103] Here, the action recommendation device 10 does not need to include the estimation unit 104. For example, the estimation unit 104 may be included in the determination unit 102. Furthermore, the health state estimation model and the action determination model may be the same model.

[0104] The output unit 103 is one aspect of output means that outputs information about actions. The output unit 103 outputs information about actions recommended to disaster victims to maintain their health. The information about actions to maintain the health of disaster victims is information necessary for disaster victims to perform the actions. The information about actions to maintain the health of disaster victims is a description of the actions. For example, if the action to maintain the health of disaster victims is exercise, the description of the action may include an explanation of the body parts to be moved, the direction of movement, and tips for the exercise. The description of the action may also include the number of times the action is performed, the time required for the action, the name of the action, and the timing for performing the action.

[0105] The information on the actions for maintaining the health of the disaster victim is not limited to these. If the actions for maintaining the health of the disaster victim are a recommended program, the information on the actions for maintaining the health of the disaster victim may include the order in which the actions are to be performed. Furthermore, the information on the actions for maintaining the health of the disaster victim may include the reason for recommending the actions. The reason for recommendation is the reason why the determination unit 102 decided on the actions. The reason for recommendation may be, for example, the health condition of the disaster victim that the determination unit 102 used to decide on the actions.

[0106] The output unit 103 outputs information on the disaster victim's health maintenance actions in a format that can be confirmed by the disaster victim. By outputting information on the disaster victim's health maintenance actions in a format that can be confirmed by the disaster victim, the disaster victim can obtain information on the recommended health maintenance actions. The output unit 103 outputs the information on the health maintenance actions to a device (not shown) installed in the vehicle 90. The device installed in the vehicle 90 may be a device used for operating the vehicle 90 under normal circumstances.

[0107] The devices installed in the vehicle 90 include a display device. An example of a display device is a display. The vehicle 90 may be equipped with a display. The display may be used as a car navigation system under normal circumstances. The output unit 103 may output information on actions to be taken by the disaster victim to maintain their health to the display. The display can display the information on actions to be taken by the disaster victim on the screen. The disaster victim can then check the information on the actions to be taken by the disaster victim to maintain their health by looking at the display on the screen. An explanation of the actions can be displayed on the display screen using video.

[0108] The devices installed in the vehicle 90 also include audio output devices. An example of an audio output device is a speaker. The audio output device can output information about health maintenance actions by audio. The audio output devices include an in-vehicle audio output device and an outside-vehicle audio output device. The in-vehicle audio output device may be used as a speaker for a car audio system under normal circumstances. The outside-vehicle audio output device may be used as a speaker for an alarm device under normal circumstances.

[0109] The output unit 103 may output information about health maintenance behavior to the communication terminal 91. The communication terminal 91 can display the information about health maintenance behavior on a screen of the communication terminal 91. The communication terminal 91 can also output the information about health maintenance behavior as sound from a speaker provided in the communication terminal 91.

[0110] The output destination of the output unit 103 may differ depending on the content of the action. When the health maintenance action determined by the determination unit 102 is an action to be performed outside the vehicle 90, the output unit 103 can output information about the health maintenance action to an external audio output device. On the other hand, when the health maintenance action determined by the determination unit 102 is an action to be performed in the vehicle 90, the output unit 103 can output information about the health maintenance action to an internal audio output or display device.

[0111] The output destination of the output unit 103 may differ depending on the location of the victim. When the victim is inside the vehicle 90, the output unit 103 can output information about health maintenance actions to an in-vehicle audio output or display device. On the other hand, when the victim is outside the vehicle, the output unit 103 can output information about health maintenance actions to an outside-vehicle audio output device. Whether the victim is inside the vehicle can be determined, for example, from the detection result of a sensor 901 installed in the vehicle 90.

[0112] The output destinations of the output unit 103 are not limited to these. Furthermore, the output destinations of the output unit 103 may be one or more. For example, the output unit 103 may output information on health maintenance behaviors to a display device and an in-car audio output device. The output destination of the output unit 103 can be determined by, for example, the determination unit 102.

[0113] The timing at which the output unit 103 outputs the information on health maintenance actions may be set in advance. For example, the output unit 103 may output the information on health maintenance actions periodically. For example, the output unit 103 may output the information on health maintenance actions at 12:00 every day. Furthermore, the timing at which the output unit 103 outputs the information on health maintenance actions may be set in advance by the disaster victim. The disaster victim can set the timing at which they wish to receive the information on health maintenance actions. Then, the output unit 103 may output the information on health maintenance actions at the timing set by the disaster victim.

[0114] The timing at which the output unit 103 outputs information about the health maintenance behavior may be determined based on the content of the behavior. As described above, the determination unit 102 determines the timing at which to perform the health maintenance behavior. Therefore, the output unit 203 can output information about the health maintenance behavior in accordance with the timing at which to perform the health maintenance behavior. The output unit 103 may, for example, output information about the health maintenance behavior a predetermined time before the timing at which to perform the health maintenance behavior. The predetermined time is, for example, a time that allows the disaster victim to prepare for the health maintenance behavior. Furthermore, the predetermined time is a time that allows the disaster victim to remember to perform the health maintenance behavior. For example, if the output unit 103 outputs information too early compared to the timing at which to perform the behavior, the disaster victim may forget to perform the health maintenance behavior. An example of the predetermined time is 30 minutes before the timing at which to perform the behavior, but is not limited to this.

[0115] Furthermore, the timing at which the output unit 103 outputs information about the health maintenance action may be determined according to the urgency of the health maintenance action. For example, if the disaster victim's health condition indicates a high risk of economy class syndrome, the determination unit 102 may determine that walking outside the vehicle is the health maintenance action. In this case, if the output unit 103 outputs the information slowly, it may be too late. Therefore, if the urgency of the health maintenance action is high, the output unit 103 may output the information immediately after the determination by the determination unit 102. On the other hand, if the urgency of the health maintenance action is low, for example, the output unit 103 may output the information at a timing set by the disaster victim.

[0116] Furthermore, the timing at which the output unit 103 outputs the information on the health maintenance action may be determined according to the content of the health maintenance action and the location of the disaster victim. If the health maintenance action is an action performed outside the vehicle, the output unit 103 may output the information on the health maintenance action when the disaster victim is outside the vehicle. Also, if the health maintenance action is an action performed inside the vehicle 90, the output unit 103 may output the information on the health maintenance action when the disaster victim is inside the vehicle 90. Whether the disaster victim is inside the vehicle 90 can be determined, for example, from the detection result of a sensor 901 installed in the vehicle 90.

[0117] The timing at which the output unit 103 outputs the information on the behavior for maintaining health may be determined by the determination unit 102, for example.

[0118] The timing at which the output unit 103 outputs the information on the behavior for maintaining health is not limited to these. Furthermore, the output by the output unit 103 may be one time or multiple times. When the output unit 103 outputs the information on the behavior for maintaining health multiple times, the output unit 103 may output the information on the behavior for maintaining health to a different output destination for each output.

[0119] 4, the operation of the action recommendation device 10 including the acquisition unit 101, the determination unit 102, the output unit 103, the acquisition unit 101, and the estimation unit 104 will be described. FIG. 4 is a flowchart showing the operation of the action recommendation device 10.

[0120] In step S101, the acquisition unit 101 acquires information about the lives of disaster victims living as evacuees in the vehicle 90, obtained based on the detection results of sensors installed in the vehicle 90. In step S102, the estimation unit 104 estimates the health state of the disaster victims based on the information about the disaster victims' lives. In step S103, the determination unit 102 determines an action to be recommended to the disaster victims for maintaining their health, based on their health state. In step S104, the output unit 103 outputs information about the action to maintain their health. Then, the action recommendation device 10 ends its operation.

[0121] A case will be described in which the action recommendation device 10 includes an acquisition unit 101, a determination unit 102, and an output unit 103. In this case, after step S101, the determination unit 102 determines an action to be recommended to the disaster victim for maintaining the health of the disaster victim based on information about the disaster victim's life. Then, the output unit 103 outputs information about the action for maintaining the health, and the action recommendation device 10 ends its operation.

[0122] In this embodiment, the action recommendation device 10 includes an acquisition unit 101 that acquires information about the lives of disaster victims living as evacuees in the vehicles 90, the information being obtained based on detection results from sensors installed in the vehicles 90. The determination unit 102 then determines an action to be recommended to the disaster victims for maintaining their health, based on the information about the disaster victims' lives. When a disaster occurs, the disaster victims may evacuate. The disaster victims may also live as evacuees in the vehicles 90. This may put physical or mental strain on the disaster victims. For example, the disaster victims may become ill. Therefore, the determination unit 102 determines an action to be recommended to the disaster victims for maintaining their health, based on the information about the disaster victims' lives. The output unit 103 then outputs the information about the action to maintain health, thereby promoting the health of disaster victims living as evacuees in the vehicles 90 during a disaster. The output unit 103 then outputs the information about the action to maintain health, thereby supporting decision-making regarding the action to maintain health of disaster victims.

[0123] The acquisition unit 101 also acquires information about the lives of disaster victims living as evacuees in the vehicle 90, obtained based on the detection results of sensors installed in the vehicle 90. By having disaster victims living as evacuees in the vehicle 90, the acquisition unit 101 can acquire information about the lives of disaster victims. In other words, the disaster victims may not need to perform tasks such as inputting information about the lives of disaster victims. This reduces the effort required of the disaster victims when deciding on recommended actions to maintain the health of the disaster victims. After a disaster occurs, the disaster victims may have many tasks to perform, such as responding to damage caused by the disaster and procuring food. In such a situation, it may be difficult for the disaster victims to perform tasks such as inputting information about the lives of the disaster victims. Therefore, reducing the effort required of the disaster victims makes it possible to easily promote the health of the disaster victims.

[0124] In the action recommendation device 10 according to this embodiment, the sensors are sensors used for operating the vehicle 90 before a disaster occurs. The vehicle 90 is equipped with sensors that are used for operating the vehicle 90 under normal circumstances. Examples of sensors used for operating the vehicle 90 under normal circumstances include a load sensor, a capacitance sensor, a pressure sensor, or a radar sensor that detects occupants of the vehicle 90. The sensors used for operating the vehicle 90 are not limited to these. After a disaster occurs, the acquisition unit 101 acquires information about the lives of disaster victims obtained based on the detection results of these sensors. In other words, there is no need to prepare new sensors to acquire information about the lives of disaster victims. This reduces the effort required to promote the health of disaster victims.

[0125] Furthermore, since the acquisition unit 101 uses sensors installed in the vehicle 90 during normal times, it can acquire information about the lives of disaster victims immediately after a disaster occurs. For example, if a new sensor is prepared to acquire information about the lives of disaster victims, it may be difficult for the acquisition unit 101 to acquire information about the lives of disaster victims until the new sensor is prepared. However, by using the sensors installed in the vehicle 90, the acquisition unit 101 can acquire information about the lives of disaster victims immediately after a disaster occurs. Therefore, the determination unit 102 can determine actions for maintaining the health of disaster victims based on information about the lives of more disaster victims. As a result, it is possible to promote the health of disaster victims who are living as evacuees in the vehicle 90 during a disaster.

[0126] In this embodiment, the action recommendation device 10 has an estimation unit 104 that estimates the health state of a disaster victim based on information about the disaster victim's life. The determination unit 102 then determines an action to maintain health based on the health state. A disaster victim living as an evacuee in a vehicle 90 may experience changes in health. The health state may include at least one of the following: risk of musculoskeletal pain, risk of musculoskeletal disorder, joint or other movement disorder, and mental stress. During a disaster, a disaster victim may experience, for example, an increased risk of pain or disorder, or increased mental stress. Depending on the disaster victim's health state, the action that the disaster victim should take to maintain their health may differ. Therefore, the determination unit 102 determines an action to maintain health based on the disaster victim's health state estimated by the estimation unit 104. Therefore, the action recommendation device 10 can recommend appropriate actions based on the disaster victim's health state. By having the disaster victim take the recommended action, the disaster victim's health can be improved during a disaster.

[0127] In this embodiment, the output unit 203 of the action recommendation device 10 outputs information about health maintenance actions to a device installed in the vehicle 90. A disaster victim may be living as an evacuee in the vehicle 90. Therefore, by having the output unit 103 output information about health maintenance actions to a device installed in the vehicle 90, the disaster victim can obtain information about health maintenance actions while living as an evacuee. In other words, the disaster victim can reduce the effort required to obtain information about health maintenance actions. The effort required to obtain information about health maintenance actions may involve, for example, accessing a specific website or visiting a specific location in an evacuation shelter to obtain information about health maintenance actions. Reducing the effort required to obtain information about health maintenance actions may make it easier for the disaster victim to perform health maintenance actions. As a result, it is possible to promote the health of disaster victims during a disaster.

[0128] Furthermore, by the output unit 103 outputting information on health maintenance actions to a device installed in the vehicle 90, a new output device may not be necessary. For example, if a new output device is needed to output information on health maintenance actions, the output unit 103 may have difficulty outputting information on health maintenance actions until the new output device is ready. However, by the output unit 103 outputting information on health maintenance actions to a device installed in the vehicle 90, the information on health maintenance actions can be output immediately after the occurrence of a disaster. This allows the output unit 103 to output information on health maintenance actions at an appropriate time. As a result, it is possible to promote the health of disaster victims living as evacuees in the vehicle 90 during a disaster.

[0129] In the action recommendation device 10 according to this embodiment, information about the life of a disaster victim includes at least one of the victim's behavior history inside the vehicle 90, the victim's behavior history outside the vehicle 90, and the victim's physical or mental state. A disaster victim living as an evacuee in the vehicle 90 may spend a lot of time inside the vehicle 90. After a disaster occurs, depending on the damage situation, it may be difficult for the victim to leave the evacuation site. Therefore, the disaster victim may spend a lot of time around the vehicle. Therefore, the acquisition unit 101 acquires information about the life of the disaster victim, including the behavior history inside the vehicle 90 or the behavior history outside the vehicle 90. Therefore, the estimation unit 104 may be able to accurately estimate the health state of the disaster victim. Then, the determination unit 102 can determine appropriate actions for maintaining health based on the accurately estimated health state of the disaster victim. As a result, it is possible to promote the health of a disaster victim living as an evacuee in the vehicle 90 during a disaster.

[0130] The behavioral history of the disaster victim in the vehicle 90 includes at least one of posture in the vehicle, postural habits, time spent sitting in the vehicle, amount of physical activity while sitting, vital signs, heart rate variability, and pulse rate variability. Since the space in the vehicle 90 where the disaster victim lives as an evacuee is limited, the postures they can assume in the vehicle may be limited, for example. Furthermore, since the disaster victim lives as an evacuee in the vehicle 90, they may spend a long time sitting in the vehicle's seat. Therefore, the behavioral history of the disaster victim in the vehicle 90 includes at least one of posture in the vehicle, postural habits, and time spent sitting in the vehicle. This allows the acquisition unit 101 to appropriately acquire how the disaster victim spends their time in the vehicle. Therefore, the estimation unit 104 may be able to accurately estimate the disaster victim's health condition. The determination unit 102 can then determine appropriate actions for maintaining health based on the accurately estimated health condition of the disaster victim. As a result, it is possible to promote the health of disaster victims living as evacuees in the vehicle 90 during a disaster.

[0131] In the action recommendation device 10 according to this embodiment, the estimation unit 104 estimates the health state of a disaster victim using a trained model that has been machine-learned. Then, the acquisition unit 101 acquires information about the life of the disaster victim determined according to the importance estimated by the trained model. For example, the acquisition unit 101 may acquire information with a high degree of importance and not acquire information with a low degree of importance. After a disaster occurs, the amount of power that can be consumed may be limited. Furthermore, communications may be restricted due to the disaster. In such an environment, the acquisition unit 101 acquires information about the life of the disaster victim determined according to the importance of the information. This makes it possible to reduce power consumption and communication traffic associated with information acquisition. Furthermore, it becomes possible to reduce the amount of information used by the estimation unit 104 to estimate the health state. As a result, it becomes possible to reduce the amount of calculation required for estimating the health state.

[0132] In the action recommendation device 10 according to this embodiment, the trained model is a model in which, before a disaster occurs, correlations between information obtained from the detection results of the sensors installed in the vehicle 90 and information obtained from the detection results of the sensors installed in the communication terminal are additionally learned. The acquisition unit 101 then acquires information about the lives of disaster victims determined according to the correlations. Information with a high correlation may be information for which similar information can be obtained from both the sensor installed in the vehicle 90 and the sensor 911 installed in the communication terminal 91. Therefore, the acquisition unit 101 can acquire information with a high correlation from either the sensor 901 installed in the vehicle 90 or the sensor 911 installed in the communication terminal 91. Furthermore, for information with a high correlation, the acquisition unit 101 may select a sensor from which to acquire information according to the remaining battery charge or power consumption. The acquisition unit 101 may switch the sensor from which to acquire information according to the remaining battery charge or power consumption. In this way, the acquisition unit 101 acquires information with a high correlation from either the sensor installed in the vehicle 90 or the sensor 911 installed in the communication terminal 91, thereby making it possible to reduce the power consumption and communication volume involved in acquiring the information. Also, it becomes possible to reduce the amount of information used by the estimation unit 104 to estimate the health state. As a result, it becomes possible to reduce the amount of calculation required for estimating the health state.

[0133] In this embodiment, when multiple disaster victims use the same vehicle 90, the acquisition unit 101 of the action recommendation device 10 distinguishes information about the lives of the disaster victims using the weights of the disaster victims detected by a sensor. In the event of a disaster, multiple disaster victims may evacuate to a single vehicle 90. In this case, the acquisition unit 101 needs to distinguish which disaster victim the acquired information about the lives of the disaster victims belongs to. Therefore, the acquisition unit 101 can distinguish which disaster victim the acquired information about the lives of the disaster victims belongs to using the weights of the disaster victims detected by a sensor. As a result, the estimation unit 104 may be able to accurately estimate the health status of each disaster victim. Then, the determination unit 102 can determine appropriate health maintenance actions for each disaster victim based on the accurately estimated health status of each disaster victim. As a result, it is possible for multiple disaster victims living as evacuees in the vehicle 90 to improve their health during a disaster.

[0134] In this embodiment, the action recommendation device 10 includes an acquisition unit 101 that acquires environmental information outside the vehicle, including disaster occurrence information or the damage caused by the disaster. A determination unit 102 determines health maintenance actions according to the occurrence information or the damage situation. Depending on the disaster occurrence situation, some actions may be difficult for disaster victims to perform. For example, in the case of a heavy rain disaster, rain may continue to fall. Furthermore, damage caused by the disaster may include poor road and soil conditions around evacuation sites. In such cases, it is difficult for disaster victims to perform health maintenance actions outside the vehicle. Therefore, the determination unit 102 determines health maintenance actions that can be performed inside the vehicle using the disaster occurrence information or the damage situation caused by the disaster. This allows disaster victims to perform health maintenance actions. As a result, it is possible to promote the health of disaster victims living as evacuees inside the vehicle 90 during a disaster.

[0135] In this embodiment, the action recommendation device 10 has an acquisition unit 101 that acquires information about items that can be used by disaster victims after a disaster occurs. Then, the determination unit 102 determines an action for maintaining health using the information about the items that can be used. Some health maintenance actions can be more effective when performed using items. Therefore, the determination unit 102 determines an action to be performed using items that can be used during a disaster as an action for maintaining health. A disaster victim who performs a health maintenance action using an item may be more effective than a disaster victim who performs a health maintenance action without using an item. As a result, it is possible to promote the health of disaster victims who are living as evacuees in vehicles 90 during a disaster.

[0136] Furthermore, the acquisition unit 101 acquires information about items that can be used by disaster victims after a disaster occurs. Then, the determination unit 102 can determine that exercise using the items is the health maintenance activity. In this case, the disaster victim may not need to search for or purchase new items to perform the health maintenance activity. In other words, the disaster victim's effort to perform the recommended health maintenance activity can be reduced. Furthermore, it may become easier for the disaster victim to perform the health maintenance activity. As a result, it is possible to promote the health of disaster victims who are evacuated in the vehicle 90 during a disaster.

[0137] [Second embodiment] The second embodiment will be described in detail with reference to the drawings. Below, the description of the second embodiment will be omitted to the extent that it does not make the description of the present embodiment unclear.

[0138] The configuration of the action recommendation device 20 will be described with reference to Fig. 5. Fig. 5 is a block diagram showing an example of the configuration of the action recommendation device 20. The action recommendation device 20 includes an acquisition unit 201, a determination unit 202, an output unit 203, and an evaluation unit 205. The action recommendation device 20 may also include an estimation unit 204.

[0139] The acquiring unit 201, the determining unit 202, the output unit 203, and the estimating unit 204 are the same as those in the first embodiment. That is, the acquiring unit 101 corresponds to the acquiring unit 201, the determining unit 102 corresponds to the determining unit 202, the output unit 103 corresponds to the output unit 203, and the estimating unit 104 corresponds to the estimating unit 204. The action recommendation device 20 further includes an evaluating unit 205 in addition to the components included in the action recommendation device 10.

[0140] The evaluation unit 205 is one aspect of evaluation means for evaluating the health maintenance actions taken by the disaster victim. The disaster victim may take the health maintenance action for which the output unit 203 has output information. In this case, the evaluation unit 205 can evaluate the health maintenance action taken by the disaster victim. Then, the output unit 203 can output content according to the evaluation of the health maintenance action taken by the disaster victim. In other words, the action recommendation device 20 can provide feedback, etc., for the action taken by the disaster victim.

[0141] First, the acquisition unit 201 acquires the state of the disaster victim performing the health maintenance behavior. The acquisition unit 201 acquires the state of the disaster victim performing the health maintenance behavior, which is captured by a camera installed in the vehicle 90. When the disaster victim performs the health maintenance behavior outside the vehicle 90, the state of the disaster victim is captured by an exterior camera. When the disaster victim performs the health maintenance behavior inside the vehicle 90, the state of the disaster victim is captured by an interior camera. The acquisition unit 201 may acquire the state of the disaster victim performing the health maintenance behavior from a sensor 911 installed in the communication terminal 91.

[0142] Then, the evaluation unit 205 evaluates the health maintenance behaviors taken by the disaster victim, acquired by the acquisition unit 201. The evaluation unit 205 may evaluate whether the disaster victim is performing the health maintenance behaviors. For example, if the health maintenance behaviors are programs, the evaluation of whether the disaster victim is performing the health maintenance behaviors may be expressed as a percentage of the programs that the disaster victim has performed.

[0143] The evaluation unit 205 may evaluate whether the disaster victim is appropriately performing the behavior for maintaining health. The evaluation unit 205 may express the appropriateness of the behavior performed by the disaster victim, for example, as a rank. For example, if the behavior is appropriately performed, it may be expressed as rank 5, and if it is not appropriately performed, it may be expressed as rank 1. The evaluation unit 205 may also express the appropriateness of the behavior performed by the disaster victim as a percentage. The evaluation unit 205 may also evaluate the behavior performed by the disaster victim for each body part. For example, if the behavior for maintaining health is exercises that move the hands and feet, the evaluation unit 205 may evaluate the hand movements and the foot movements, respectively.

[0144] The evaluation of whether the disaster victim is taking actions to maintain health and whether the disaster victim is appropriately taking actions to maintain health can be performed using known techniques. Furthermore, the evaluation unit 205 may evaluate the actions to maintain health taken by the disaster victim by acquiring an evaluation of the disaster victim himself / herself. Furthermore, the evaluation unit 205 may evaluate the actions to maintain health taken by the disaster victim by acquiring an evaluation of the disaster victim by a third party.

[0145] The evaluation unit 205 may evaluate whether the disaster victim's health maintenance behavior is effective. For example, the estimation unit 104 can estimate the disaster victim's health condition based on the state of the disaster victim performing the health maintenance behavior acquired by the acquisition unit 101. The method for estimating the health condition is as described above. Then, the evaluation unit 205 compares the health condition of the disaster victim before and after performing the health maintenance behavior. The evaluation unit 205 can evaluate the effect of the health maintenance behavior from the comparison result. The method for evaluating the effect of the health maintenance behavior of the disaster victim is not limited to this. Publicly known techniques can be used to evaluate the effect of the health maintenance behavior of the disaster victim.

[0146] The evaluation unit 205 may express the effect of the action taken by the disaster victim, for example, by a rank. For example, a large effect may be expressed by a rank of 5, and a small effect may be expressed by a rank of 1. The evaluation unit 205 may also express the effect of the action taken by the disaster victim as a percentage. The evaluation unit 205 may also evaluate the effect of the action taken by the disaster victim for each part of the body. For example, if the action for maintaining health is exercises that move the hands and feet, the evaluation unit 205 may evaluate the effect on the hands and the effect on the feet, respectively.

[0147] The evaluation content by the evaluation unit 205 is not limited to these. The evaluation unit 205 can perform evaluation according to the disaster victim or actions for maintaining health.

[0148] The output unit 203 outputs content according to the evaluation of the disaster victim's actions for maintaining health. The output unit 203 may output points for improvement regarding the actions for maintaining health taken by the disaster victim. The content of the points for improvement to be output to the disaster victim can be determined using publicly known technology.

[0149] The output unit 203 may output to the disaster victim a health maintenance action that is different from the health maintenance action already output by the output unit 103. There are cases where it is evaluated that the effect of the health maintenance action taken by the disaster victim is sufficient. In this case, the disaster victim may be able to further promote health by taking an action different from the action in question. On the other hand, if it is evaluated that the disaster victim cannot appropriately take the health maintenance action, the action in question may be too difficult for the disaster victim. In this case, it may be better for the output unit 203 to output a health maintenance action with a lower level of difficulty.

[0150] Therefore, the output unit 203 may output different health maintenance actions to the disaster victim. In this case, the different health maintenance actions output by the output unit 203 may be determined by the determination unit 202.

[0151] The output unit 203 may output a notification including the victim's status to a predetermined notification destination. The victim's status includes at least one of the victim's health condition, actions to maintain health, and the evaluation by the evaluation unit 205. The victim's status may also include information about the victim's evacuation location. The notification output by the output unit 203 may include a request for a visit from the victim, a request for medical treatment or care from the victim, etc. The output unit 203 may also output a notification to a predetermined notification destination according to the victim's health condition estimated by the estimation unit 204. Depending on the victim's appearance or health condition, for example, response by government or a medical institution may be necessary. Therefore, the output unit 203 may output a notification to a predetermined notification destination based on the victim's appearance acquired by the acquisition unit 101 or the health condition estimated by the estimation unit 204. The predetermined notification destination may be, for example, a responsible department of a city, ward, town, or village, a medical institution, a family doctor, a nurse, etc. The victim's status also includes the victim's health condition.

[0152] Here, the content corresponding to the evaluation to be output by the output unit 203 may be determined by the determination unit 202. For example, the content to be output by the output unit 203 is determined in advance according to the evaluation by the evaluation unit 205. Therefore, the determination unit 202 can determine the content corresponding to the evaluation to be output by the output unit 203 based on the evaluation by the evaluation unit 205.

[0153] The evaluation unit 205 may evaluate the disaster victim's behavior for maintaining health in normal times. At this time, the acquisition unit 201 acquires the state of the disaster victim's behavior for maintaining health in normal times. The evaluation content and evaluation method by the evaluation unit 205 are the same as those in the event of a disaster. Furthermore, the evaluation unit 205 may store the evaluation in normal times in, for example, a database.

[0154] In the event of a disaster, the determination unit 202 may determine an action for maintaining health using the evaluations under normal circumstances. For example, a case will be described in which the determination unit 202 determines exercise in a community as an action for maintaining health. In this case, the determination unit 202 may determine, for example, exercise in a community that includes disaster victims who have received high evaluations under normal circumstances as an action for maintaining health.

[0155] The operation of the action recommendation device 10, which includes an acquisition unit 201, a determination unit 202, an output unit 203, an estimation unit 104, and an evaluation unit 205, will be described with reference to Fig. 6. Fig. 6 is a flowchart showing the operation of the action recommendation device 20 ... when a disaster victim takes an action for maintaining health for which information has been output by the output unit 203. In other words, Fig. 6 is a flowchart showing the operation of the action recommendation device 10 after the operation shown in Fig. 4 has been performed.

[0156] In step S104 in FIG. 4, the output unit 103 outputs information about health maintenance actions. The disaster victim may then perform actions to maintain their health. Therefore, in step S201, the acquisition unit 101 acquires information about the disaster victim performing health maintenance actions. Then, in step S202, the evaluation unit 205 evaluates the health maintenance actions performed by the disaster victim. Then, the output unit 203 outputs content according to the evaluation of the disaster victim's health maintenance actions. The action recommendation device 10 then terminates its operation. Here, the content to be output by the output unit 203 may be determined by the determination unit 202. That is, after step S202, the determination unit 202 determines the content to be output according to the evaluation of the disaster victim's health maintenance actions. Then, in step S203, the output unit 203 may output content according to the evaluation of the disaster victim's health maintenance actions.

[0157] In this embodiment, the evaluation unit 205 of the action recommendation device 10 evaluates the health maintenance actions taken by the disaster victim. Then, the output unit 203 outputs content according to the evaluation of the health maintenance actions taken by the disaster victim. The disaster victim may take the health maintenance action for which the output unit 203 has output information. Therefore, the evaluation unit 205 evaluates the health maintenance actions taken by the disaster victim, and the output unit 203 outputs content according to the evaluation. This makes it possible to provide feedback on the health maintenance actions taken by the disaster victim.

[0158] An example of the content output in response to the evaluation is outputting points for improving behaviors for maintaining health for disaster victims. By identifying the points for improvement, disaster victims may be able to take appropriate actions for maintaining their health. As a result, it becomes possible to effectively promote the health of disaster victims living as evacuees in vehicles 90 during a disaster.

[0159] Another example of the content output in accordance with the evaluation is outputting information on different health maintenance actions. The evaluation unit 205 may evaluate that the health maintenance actions taken by the disaster victim have been sufficiently effective. In this case, the disaster victim may be able to further promote their health by taking an action different from the action in question. On the other hand, if the disaster victim is evaluated as being unable to appropriately take a health maintenance action, the action in question may be too difficult for the disaster victim. In this case, the output unit 203 may be better off outputting a health maintenance action with a lower level of difficulty. In other words, by the output unit 103 outputting different health maintenance actions in accordance with the evaluation, it becomes possible to efficiently promote the health of the disaster victim.

[0160] Another example of the content output in accordance with the evaluation is outputting a notification to a preset notification destination. Depending on the condition and health status of the victim, for example, assistance from the government or a medical institution may be required. Therefore, by the output unit 203 outputting a notification to a preset notification destination, the victim may be able to receive appropriate assistance. Also, for example, the government or medical institution that is the notification destination can grasp the condition of the victim. As a result, for example, the government or medical institution can visit the victim and provide medical care promptly.

[0161] In this embodiment, in the action recommendation device 10, the acquisition unit 201 acquires, after a disaster occurs, images of disaster victims performing health maintenance actions captured by a camera installed in the vehicle 90. The evaluation unit 205 then evaluates the acquired actions. Disaster victims living as evacuees in the vehicle 90 may perform health maintenance actions inside or near the vehicle 90. Therefore, the acquisition unit 201 can acquire images of disaster victims performing health maintenance actions using the camera installed in the vehicle 90. This reduces the effort required to capture images of disaster victims performing health maintenance actions. In addition, the camera is installed in the vehicle 90 during normal times. Therefore, disaster victims do not need to prepare a new device to capture images of disaster victims performing health maintenance actions. This reduces the effort required to promote the health of disaster victims.

[0162] [Hardware configuration example] 7 is a diagram illustrating an example of the hardware configuration of the action recommendation device 30 in the present disclosure. The action recommendation device 30 is realized by a computer. The action recommendation device 30 is an example in which the action recommendation device 10 or the action recommendation device 20 is realized by a computer.

[0163] The action recommendation device 30 includes a processor 301, a ROM (Read Only Memory) 302, a RAM (Random Access Memory) 303, a storage device 304 such as a hard disk for storing programs, an input / output interface 305 for inputting and outputting data, and a communication interface 306 for network connection. Each component is connected via a bus 307.

[0164] The processor 301 runs an operating system to control the entire computer. Examples of the processor 301 include a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a GPU (Graphics Processing Unit). The processor 301 loads programs stored in, for example, the ROM 302 or the storage device 304. The processor 301 then executes each process coded in the program. The processor 301 may execute processes or instructions in the illustrated flowchart based on the program.

[0165] The ROM 302 stores application programs, programs according to the embodiments, etc. The RAM 303 is used as a work area for the processor 301.

[0166] Examples of the storage device 304 include a semiconductor memory such as a flash memory, a hard disk drive (HDD), etc. The storage device 304 stores, for example, an operating system (OS) program, application programs, and programs according to each embodiment.

[0167] The input / output interface 305 is connected to peripheral devices (not shown) via a wired network or a wireless network.

[0168] The communication interface 306 is connected to a communication network (not shown), such as a LAN (Local Network) or WAN (Wide Area Network), via a wireless or wired network. The communication network may be configured with multiple communication networks. This allows the computer to be connected to external devices via the communication network. The action recommendation device 30 may have components other than those shown in FIG. 7. For example, the action recommendation device 30 may have a drive device or the like. For example, the processor 301 may be attached to a drive device or the like and read programs and data stored in a non-transitory tangible recording medium into the RAM 303.

[0169] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, the configurations in the respective embodiments can be combined with each other without departing from the scope of the present disclosure.

[0170] Some or all of the above-described embodiments can be described as, but are not limited to, the following supplementary notes.

[0171] (Appendix 1) an acquisition means for acquiring information on the lives of disaster victims living in evacuation shelters in the vehicles, the information being obtained based on the detection results of sensors installed in the vehicles; A decision-making means for deciding, based on information about the life of the disaster victim, an action to be recommended to the disaster victim for maintaining the health of the disaster victim; an output means for outputting information about the health maintenance behavior; An action recommendation device comprising:

[0172] (Appendix 2) The sensor is a sensor used for operating the vehicle before a disaster occurs. 1. A behavioral recommendation device as described in Appendix 1.

[0173] (Appendix 3) The sensor is a load sensor that detects an occupant of the vehicle before the disaster occurs. 1. The behavioral recommendation device described in Appendix 2.

[0174] (Appendix 4) The sensor is a capacitance sensor that detects an occupant of the vehicle before the disaster occurs. 4. The behavior recommendation device according to claim 2 or 3.

[0175] (Appendix 5) The sensor is a pressure sensor that detects an occupant of the vehicle before the disaster occurs. 5. The behavior recommendation device of any one of appendixes 2 to 4.

[0176] (Appendix 6) The sensor is a radar sensor that detects occupants of the vehicle before the disaster occurs. 10. The behavior recommendation device according to claim 2, wherein the behavior recommendation device is a device for recommending behaviors.

[0177] (Appendix 7) further comprising an estimation means for estimating the health status of the disaster victim based on information about the disaster victim's life; The decision means decides the action depending on the health condition. 7. The behavior recommendation device of any one of appendixes 1 to 6.

[0178] (Appendix 8) The health condition represents the physical condition of the disaster victim living in the evacuation vehicle. 10. The behavioral recommendation device described in Appendix 7.

[0179] (Appendix 9) The health condition includes at least one of the following: risk of musculoskeletal pain, risk of musculoskeletal disorder, joint or other movement disorder, and mental stress. 9. The behavior recommendation device according to claim 7 or 8.

[0180] (Appendix 10) the acquiring means acquires information about the lives of the disaster victims before the disaster occurred, The estimation means further estimates the health condition of the disaster victim based on information about the life of the disaster victim before the disaster occurred. 10. The behavior recommendation device of any one of appendixes 7 to 9.

[0181] (Appendix 11) the acquisition means acquires the vehicle usage status of the disaster victim before the disaster occurs, The decision means decides an action to be taken to maintain the health of the victim according to the usage situation. 11. The behavioral recommendation device of claim 10.

[0182] (Appendix 12) The output means outputs the information on the health maintenance behavior to a device installed in the vehicle. 12. The behavior recommendation device of any one of claims 1 to 11.

[0183] (Appendix 13) The output means outputs information about the health maintenance action at a timing according to the urgency of the health maintenance action. 13. The behavior recommendation device of any one of appendixes 1 to 12.

[0184] (Appendix 14) The information about the life of the victim includes at least one of the victim's behavior history in the vehicle, the victim's behavior history outside the vehicle, and the victim's physical condition or mental condition. 14. The behavior recommendation device of any one of claims 1 to 13.

[0185] (Appendix 15) The behavioral history of the victim in the vehicle includes at least one of posture in the vehicle, posture habits, time spent sitting in the vehicle, amount of physical activity while sitting, vital signs, heart rate variability, and pulse rate variability. 15. The behavioral recommendation device of claim 14.

[0186] (Appendix 16) The acquisition means acquires the physical condition based on an image captured by a camera for capturing an image of the outside of the vehicle. 16. The behavior recommendation device of claim 14 or 15.

[0187] (Appendix 17) The acquisition means acquires the mental state based on the behavior history of the victim in the vehicle. 17. The behavior recommendation device of any one of appendixes 14 to 16.

[0188] (Appendix 18) The method further includes an evaluation means for evaluating the health maintenance behaviors performed by the disaster victim. The output means outputs content according to the evaluation. 18. The behavior recommendation device of any one of appendixes 1 to 17.

[0189] (Appendix 19) the acquiring means acquires, after the occurrence of the disaster, a video of the disaster victim taking the health maintenance action, which is captured by a camera installed in the vehicle; The evaluation means evaluates the acquired behavior. 19. The behavioral recommendation device of claim 18.

[0190] (Appendix 20) The output means outputs points for improvement in accordance with the evaluation. 20. The behavior recommendation device of claim 18 or 19.

[0191] (Appendix 21) the determining means determines different actions for maintaining health depending on the evaluation; The output means outputs information about the different health maintenance behaviors. 21. The behavioral recommendation device of any one of clauses 18 to 20.

[0192] (Appendix 22) The output means outputs a notification to a preset notification destination in accordance with the evaluation. 22. The behavior recommendation device of any one of clauses 18 to 21.

[0193] (Appendix 23) The estimation means estimates the health status of the disaster victim using a trained model that has been machine-learned; The acquisition means acquires information about the lives of the disaster victims, the information being determined according to the importance estimated by the trained model. 10. The behavioral recommendation device described in Appendix 7.

[0194] (Appendix 24) The trained model is a model in which, before a disaster occurs, a correlation between information obtained from a detection result of a sensor installed in a vehicle and information obtained from a detection result of a sensor installed in a communication terminal is additionally trained, The acquiring means acquires information about the lives of disaster victims determined according to the correlation. 24. The behavioral recommendation device of claim 23.

[0195] (Appendix 25) When multiple victims use the same vehicle, The acquisition means uses the weight of the victim detected by the sensor to distinguish information about the life of the victim. 25. The behavior recommendation device of any one of claims 1 to 24.

[0196] (Appendix 26) The acquisition means acquires environmental information outside the vehicle. 26. The behavioral recommendation device of any one of clauses 1 to 25.

[0197] (Appendix 27) The environmental information outside the vehicle includes information on the occurrence of the disaster or a damage situation caused by the disaster after the occurrence of the disaster. 27. The behavioral recommendation device of claim 26.

[0198] (Appendix 28) The decision means decides on the action for maintaining health in accordance with the occurrence information or the damage situation. 28. The behavioral recommendation device of claim 27.

[0199] (Appendix 29) the acquiring means acquires information on items that can be used by the disaster victims after the disaster occurs, The decision means further decides on the health maintenance behavior using the information on the available items. 29. The behavior recommendation device of any one of appendices 1 to 28.

[0200] (Appendix 30) Acquire information about the lives of disaster victims living in evacuation shelters in the vehicles, which information is obtained based on the detection results of sensors installed in the vehicles; determining, based on information about the victim's lifestyle, recommended actions for the victim to maintain the victim's health; Outputting information on the health maintenance behavior Recommended action methods.

[0201] (Appendix 31) Acquire information about the lives of disaster victims living in evacuation shelters in the vehicles, which information is obtained based on the detection results of sensors installed in the vehicles; determining, based on information about the victim's lifestyle, recommended actions for the victim to maintain the victim's health; Outputting information on the health maintenance behavior A program that causes a computer to perform a process.

[0202] (Appendix 32) Acquire information about the lives of disaster victims living in evacuation shelters in the vehicles, which information is obtained based on the detection results of sensors installed in the vehicles; determining, based on information about the victim's lifestyle, recommended actions for the victim to maintain the victim's health; Outputting information on the health maintenance behavior A recording medium that stores a program that causes a computer to execute a process.

[0203] Some or all of the configurations described in Supplementary Notes 2-29, which are dependent on Supplementary Note 1, may also be dependent on Supplementary Notes 30-32 in the same manner as Supplementary Note 2-29. Not limited to Supplementary Notes 1 and 30-32, but also to various hardware, software, various recording devices for recording software, or systems, some or all of the configurations described as Supplements may be dependent on each other within the scope of the above-mentioned embodiments. [Explanation of symbols]

[0204] 10, 20, 30 Action recommendation device 101, 201 Acquisition Department 102, 202 Decision Section 103, 203 Output section 104, 204 Estimation part 205 Evaluation Department 301 processor 302 ROM 303 RAM 304 Storage device 305 Input / Output Interface 306 Communication Interface 307 Bus 90 vehicles 901 Sensor 91 Communication terminal 911 Sensor

Claims

1. an acquisition means for acquiring information on the lives of disaster victims living in evacuation shelters in the vehicles, the information being obtained based on the detection results of sensors installed in the vehicles; A decision-making means for deciding, based on information about the life of the disaster victim, an action to be recommended to the disaster victim for maintaining the health of the disaster victim; an output means for outputting information about the health maintenance behavior; An action recommendation device comprising:

2. The sensor is a sensor used for the operation of the vehicle under normal circumstances. The behavior recommendation device according to claim 1 .

3. further comprising an estimation means for estimating the health status of the disaster victim based on information about the disaster victim's life; The decision means decides the action depending on the health condition. The behavior recommendation device according to claim 1 .

4. The acquisition means acquires information about the lives of the disaster victims under normal circumstances, The estimation means further estimates the health condition of the disaster victim based on information about the disaster victim's daily life under normal circumstances. The behavior recommendation device according to claim 3 .

5. The output means outputs the information on the health maintenance behavior to a device installed in the vehicle. The behavior recommendation device according to claim 1 .

6. The information about the life of the victim includes at least one of the victim's behavior history in the vehicle, the victim's behavior history outside the vehicle, and the victim's physical condition or mental condition. The behavior recommendation device according to claim 1 .

7. The method further includes an evaluation means for evaluating the health maintenance behaviors performed by the disaster victim. The output means outputs content according to the evaluation. The behavior recommendation device according to claim 1 .

8. The estimation means estimates the health state of the disaster victim using a trained model that has been machine-learned; The acquisition means acquires information about the lives of the disaster victims, the information being determined according to the importance estimated by the trained model. The behavior recommendation device according to claim 3 .

9. Acquire information about the lives of disaster victims living in evacuation shelters in the vehicles, which information is obtained based on the detection results of sensors installed in the vehicles; determining, based on information about the victim's lifestyle, recommended actions for the victim to maintain the victim's health; Outputting information on the health maintenance behavior Recommended action methods.

10. Acquire information about the lives of disaster victims living in evacuation shelters in the vehicles, which information is obtained based on the detection results of sensors installed in the vehicles; determining, based on information about the victim's lifestyle, recommended actions for the victim to maintain the victim's health; Outputting information on the health maintenance behavior A program that causes a computer to perform a process.

Images