Apparatus and method for predicting risk for using collected information of non-contact type sensing device and energy consumption information

Abstract

The risk prediction method of the present invention comprises the steps of: collecting energy usage and behavioral pattern information of a risk monitoring target and storing them by time period; selecting a comparison target group including at least one comparison target having the same age group, gender, and housing type as the risk monitoring target; collecting the energy usage of each of the comparison targets and calculating an average value by time period to derive the energy usage by time period of the comparison target group; comparing and analyzing the energy usage by time period of the risk monitoring target and the energy usage by time period of the comparison target group; and monitoring the behavioral pattern information by time period of the risk monitoring target when the deviation between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation tolerance range. The method includes a step of comparing and analyzing the behavioral patterns of the aforementioned risk monitoring subject by date and time period, and predicting the risk of the said risk monitoring subject when the difference between the current behavioral pattern and the behavioral pattern of a past date / same time period exceeds a preset allowable range for behavioral pattern deviation, thereby predicting risk occurrence situations that reflect age group, gender, residential conditions, or health status, and thus having the advantage of obtaining accurate prediction results.

Description

Technology Field

[0001] The present invention relates to a risk prediction device and a method thereof, and more specifically, to a risk prediction device and a method thereof that predicts risk using environmental information and energy usage information collected using a non-contact type detection device (e.g., IoT equipment, etc.). Background Technology

[0002] With the advancement of IoT technology, consumers have been able to use various wearable devices, and through the convergence of IoT and wearable technology, services based on smart home environments have expanded to include safety management.

[0003] Representative safety management technologies for protecting home occupants include technologies that utilize physical security equipment.

[0004] Home physical security refers to security technologies for personal identification, video detection, and prevention of disasters and emergencies aimed at ensuring the personal safety of residents and establishing a safe management environment for home facilities. Representative technologies include video surveillance solutions, biometric recognition, object recognition unmanned electronic security, and intelligent cameras.

[0005] Home safety systems based on home physical security technology provide services such as internal home monitoring and safety accident notifications to protect residents.

[0006] As such prior art, Korean published patent No. 10-2020-0042604 discloses a method for comprehensive residential management using environmental information, comprising the steps of: connecting a smart plug equipped with an environmental sensor to a comprehensive home appliance and a user terminal through authentication; the smart plug collecting environmental information from the environmental sensor and collecting operation status information of the home appliance in real time from the connected comprehensive home appliance and transmitting the environmental information and operation status information of the home appliance to the user terminal; and a user possessing the user terminal receiving the environmental information and operation status information of the home appliance from the smart plug and identifying the environment and the state of the home appliance within the residence based on the environmental information and operation status information of the home appliance.

[0007] According to the above patent, by establishing a comprehensive residential management system using indoor environmental information, it is possible to go beyond existing safety management services focused on one-off monitoring and post-accident response, and to expect prevention of safety accidents and continuous safety management for residents.

[0008] However, these conventional technologies had a problem in that they could not reflect the individual characteristics of the risk monitoring targets by considering only indoor environment information or the operating status information of home appliances. Prior art literature

[0009] Korean Published Patent No. 10-2020-0042604 The problem to be solved

[0010] Accordingly, the present invention aims to provide a risk prediction device and a method capable of solving the above problem by generating a comparison target group comprising comparison targets that are similar to a risk monitoring target in at least one of age group, gender, residential conditions, and health status, and by comparing the energy consumption by season / time of day between the comparison target group and the risk monitoring target, thereby predicting a risk occurrence situation reflecting age group, gender, residential conditions, or health status and obtaining an accurate prediction result.

[0011] In addition, the present invention aims to provide a risk prediction device and a method that, in order to solve the above problem, predict a risk occurrence situation of a risk monitoring target by reflecting at least one of the past behavioral pattern of the risk monitoring target and energy consumption by season / time of day, thereby predicting a risk occurrence situation that reflects an individual environment or behavioral pattern, and thereby obtaining a prediction result that reflects the individual characteristics of the risk monitoring target. means of solving the problem

[0012] To achieve the above objective, the risk prediction device provided in the present invention comprises: a first storage unit for storing time-based energy usage and behavioral pattern information of a risk monitoring target; a second storage unit for storing time-based energy usage of a comparison target group including at least one comparison target; a first information collection unit for collecting the energy usage and behavioral pattern information of the risk monitoring target and storing it in the first storage unit by time; a comparison target group selection unit for selecting a comparison target group including at least one comparison target that has the same age group, gender, and housing type as the risk monitoring target; a second information collection unit for collecting energy usage from each of the comparison targets; a comparison target group information generation unit for calculating time-based average values ​​for the energy usage of each of the comparison targets to derive the time-based energy usage of the comparison target group and storing it in the second storage unit; and a comparison analysis unit for comparing and analyzing the time-based energy usage of the risk monitoring target and the time-based energy usage of the comparison target group. The method is characterized by including: a behavioral pattern monitoring unit that monitors the time-based behavioral pattern information of the risk monitoring target when, as a result of the above comparative analysis, the deviation between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowable range; and a risk prediction unit that predicts the risk of the risk monitoring target when, by comparing and analyzing the date and time-based behavioral patterns of the risk monitoring target, the difference between the current behavioral pattern and the behavioral pattern of a past date / same time period exceeds a preset behavioral pattern deviation allowable range.

[0013] Preferably, each of the first and second storage units can be implemented as a Long Short-Term Memory (LSTM).

[0014] Preferably, the first information collection unit may include: a measuring instrument installed in a space where the risk monitoring target resides to collect energy usage of the risk monitoring target, including gas, electricity, and water usage; at least one non-contact type detection device installed in a space where the risk monitoring target resides; a behavior pattern derivation unit that collects non-contact data from the non-contact type detection device and derives a behavior pattern of the risk monitoring target; a time counter that counts time to determine the energy usage and the time of collection of the non-contact data; and an information generation unit that generates the energy usage and behavior pattern of the risk monitoring target by time and season based on the time of collection for each energy usage, the non-contact data, and the corresponding time of collection.

[0015] Preferably, the comparison target group selection unit may select at least one target among the comparison targets that has the same disease history as the risk monitoring target that affects at least one of energy consumption and behavioral patterns as the comparison target group.

[0016] Preferably, the device further includes a future prediction unit that predicts future energy usage and behavioral patterns based on energy usage and behavioral pattern information of the risk monitoring target collected / stored during a past predetermined period, and the risk prediction unit can predict the risk of the risk monitoring target when the future energy usage and behavioral patterns deviate from a preset normal range.

[0017] Preferably, the risk prediction unit may adopt a recall evaluation method that determines whether the risk monitored object is at risk based on an error matrix.

[0018] Meanwhile, to achieve the above objective, the risk prediction method provided in the present invention comprises: a risk monitoring target information storage step for collecting and storing the energy usage and behavioral pattern information of the risk monitoring target by time period; a comparison target group selection step for selecting a comparison target group including at least one comparison target that has the same age group, gender, and housing type as the risk monitoring target; a comparison target group information collection step for collecting the energy usage of each of the comparison targets and calculating the average value by time period to derive the energy usage by time period of the comparison target group; a comparison analysis step for comparing and analyzing the energy usage by time period of the risk monitoring target and the energy usage by time period of the comparison target group; and a behavioral pattern monitoring step for monitoring the behavioral pattern information by time period of the risk monitoring target when the deviation between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation tolerance range. The method is characterized by including a risk prediction step that predicts the risk of the said risk monitoring target by comparing and analyzing the behavioral patterns of the said risk monitoring target by date and time period, and when the difference between the current behavioral pattern and the behavioral pattern of a past date / same time period exceeds a preset allowable range for behavioral pattern deviation.

[0019] Preferably, the risk monitoring target information storage step may include: a first storage step of collecting non-contact data from at least one non-contact type detection device installed in a space where the risk monitoring target resides and storing it together with the date and time of collection; and a second storage step of generating and storing the time and seasonal behavioral patterns of the risk monitoring target based on the non-contact data and the corresponding date and time of collection.

[0020] Preferably, the comparison target group selection step may select at least one target among the comparison targets that has the same disease history as the risk monitoring target that affects at least one of energy usage and behavioral patterns as the comparison target group.

[0021] Preferably, the method further includes a future prediction step for predicting future energy usage and behavioral patterns based on energy usage and behavioral pattern information of the risk monitoring target collected / stored during a past predetermined period, and the risk prediction step can predict the risk of the risk monitoring target when the future energy usage and behavioral patterns deviate from a preset normal range.

[0022] Preferably, the risk prediction step may adopt a recall evaluation method that determines the risk status of the risk monitoring target based on an error matrix. Effects of the invention

[0023] As described above, the risk prediction device and method using information collected by a non-contact type detection device and energy usage information provided by the present invention have the advantage of generating a comparison target group that includes comparison targets similar to the risk monitoring target in at least one of age group, gender, living conditions, and health status, and predicting risk occurrence situations reflecting age group, gender, living conditions, or health status by comparing the energy usage by season / time of day between the comparison target group and the risk monitoring target, thereby obtaining accurate prediction results.

[0024] In addition, the present invention has the advantage of predicting a risk occurrence situation of a risk monitoring target by reflecting at least one of the past behavioral pattern of the risk monitoring target and energy consumption by season / time of day, thereby predicting a risk occurrence situation that reflects an individual environment or behavioral pattern, and thereby obtaining a prediction result that reflects the individual characteristics of the risk monitoring target. Brief explanation of the drawing

[0025] FIG. 1 is a schematic block diagram of a risk prediction device according to one embodiment of the present invention. FIG. 2 is a schematic block diagram of a risk monitoring target information collection unit according to one embodiment of the present invention. FIGS. 3 to 6 are schematic process flowcharts for a risk prediction method according to an embodiment of the present invention. FIGS. 7 and 8 are drawings illustrating risk criteria based on an individual's energy consumption or behavioral patterns by season and time of day, according to an embodiment of the present invention. Specific details for implementing the invention

[0026] Embodiments of the present invention are described below with reference to the attached drawings. The description is provided in detail to enable those skilled in the art to easily practice the present invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Meanwhile, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification have been given similar reference numerals. Furthermore, explanations of parts that can be easily understood by those skilled in the art even without detailed description have been omitted.

[0027] Throughout the specification and claims, when a part is described as including a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0028] FIG. 1 is a schematic block diagram of a risk prediction device according to an embodiment of the present invention. Referring to FIG. 1, a risk prediction device (100) according to an embodiment of the present invention includes a database unit (110), a risk monitoring target information collection unit (120), a comparison target group selection unit (130), a comparison target information collection unit (140), a comparison target group information generation unit (150), a comparison / analysis unit (160), a behavior pattern monitoring unit (170), a risk prediction unit (180), and a control unit (190).

[0029] The database unit (110) can store data that must be pre-set to operate the risk prediction device (100), or data generated during the operation of the risk prediction device (100). In particular, the database unit (110) can store a risk monitoring target information management DB (111) that stores the energy usage and behavioral pattern information of the risk monitoring target by time period, and a comparison target group information management DB (112) that stores the energy usage by time period of a comparison target group including at least one comparison target.

[0030] At this time, the risk monitoring target may be an elderly person, and the time-based energy consumption of the risk monitoring target may include gas consumption, electricity consumption, and water consumption, and the behavioral pattern information of the risk monitoring target is information collected through a non-contact type detection device installed in the space where the risk monitoring target resides (e.g., IoT sensors including door sensors, motion sensors, and sound sensors), and may include door opening or closing information, movement information, and sound information (noise due to movement, sound such as TV or music due to activity patterns).

[0031] Meanwhile, the risk monitoring target information management DB (111) and the comparison target group information management DB (112) can each be implemented as a Long Short-Term Memory (LSTM). This is because the information obtained in the previous step can be continuously reflected by using an LSTM memory block to store data, which adopts a method of selecting data to be stored and data to be discarded by adding a cell state inside the hidden layer, thereby solving the problem of long-term dependency (Gradient Vanishing) of the RNN.

[0032] The risk monitoring target information collection unit (120) collects energy usage and behavioral pattern information of the risk monitoring target and stores it in the risk monitoring target information management DB (111) by time period. FIG. 2 is a schematic block diagram of the risk monitoring target information collection unit (120). Referring to FIG. 2, the risk monitoring target information collection unit (120) includes a measuring instrument (121), an IoT sensor (122), a time counter (123), a behavioral pattern derivation unit (124), an information generation unit (125), and a control unit (126).

[0033] The measuring instrument (121) is a device (e.g., an electric meter, etc.) for measuring the energy usage of the risk monitoring target, and is installed in the space where the risk monitoring target resides to collect the energy usage of the risk monitoring target, including gas usage, electricity usage, and water usage.

[0034] At least one IoT sensor (122) is installed in the space where the risk monitoring target resides to detect behavioral information of the risk monitoring target in a non-contact manner. For example, the IoT sensor (122) includes a door sensor that detects information on the opening or closing of an external door or window, a motion sensor that detects whether there is indoor movement, and a sound sensor that measures the level of indoor noise (e.g., decibels), and can detect behavioral information of the risk monitoring target.

[0035] Additionally, the IoT sensor (122) further includes a sensing device capable of detecting surrounding environment information (e.g., temperature sensor, humidity sensor, solar radiation sensor, precipitation sensor, etc.) and can collect environmental information about the space where the risk monitoring target resides.

[0036] The time counter (123) counts the current time to determine the collection time of the data (i.e., energy usage and non-contact data) of the measuring instrument (121) and the IoT sensor (122), respectively.

[0037] The behavior pattern derivation unit (124) collects non-contact data from the IoT sensor (122) and derives the behavior pattern of the risk monitoring target. For example, the behavior pattern derivation unit (124) can derive the behavior pattern of the risk monitoring target by time by matching each of one or more non-contact data at the time of collection.

[0038] The information generation unit (125) generates the energy usage and behavior patterns of the risk monitoring target by time and season based on the energy usage collected from the measuring instrument (121) and the corresponding collection date and time, and the non-contact data collected from the IoT sensor (122) and the corresponding collection date and time. For example, the information generation unit (125) can learn the energy usage and behavior information (e.g., number of door openings, noise level, movement level, etc.) according to conditions such as specific time periods, seasons, or temperature / humidity, and generate information including the energy usage model and behavior patterns of the risk monitoring target by time period and season.

[0039] The control unit (126) controls the overall operation of the risk monitoring target information collection unit (120) based on a preset control algorithm. The control unit (126) controls the operation of the behavior pattern derivation unit (124) to derive the behavior pattern of the risk monitoring target based on the information collected from the measuring instrument (121) and the IoT sensor (122), and controls the operation of the information generation unit (125) to generate information for risk prediction of the risk monitoring target based on the collected information.

[0040] The comparison target group selection unit (130) selects a comparison target group that includes at least one comparison target serving as a comparison standard to predict whether the risk of the risk monitoring target is present. For example, the comparison target group selection unit (130) may select a comparison target group that includes at least one comparison target that is identical to the risk monitoring target in terms of age group, gender, and housing type. Additionally, the comparison target group selection unit (130) may select at least one of the comparison targets that has a disease history similar to or identical to the disease history of the risk monitoring target as the comparison target group. In this case, it is preferable that the disease history is a disease history that affects at least one of energy consumption and behavioral patterns.

[0041] The comparison target information collection unit (140) collects energy usage from each of the comparison targets. That is, the comparison target information collection unit (140) is a device (e.g., an electric meter, etc.) for measuring the energy usage of each of the comparison targets, and is installed in the space where each of the comparison targets resides, and can collect the energy usage of the risk monitoring target, including gas usage, electricity usage, and water usage.

[0042] The comparison target group information generation unit (150) derives the energy usage amount of the comparison target group by time period and stores it in the comparison target group information management DB (112). At this time, the energy usage amount of the comparison target group by time period can be determined by calculating the average value of the energy usage amount of each of the comparison targets by time period.

[0043] The comparison / analysis unit (160) compares and analyzes the energy usage of the risk monitoring target by time period and the energy usage of the comparison target group by time period. To do this, the comparison / analysis unit (160) can read the energy usage of the risk monitoring target by time period and the energy usage of the comparison target group by time period from the risk monitoring target information management DB (111) and the comparison target group information management DB (112), respectively.

[0044] The behavior pattern monitoring unit (170) monitors the behavior pattern information of the risk monitoring target by time period when, as a result of the comparison analysis, the difference between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowable range.

[0045] The risk prediction unit (180) predicts the risk of the risk monitoring target based on the monitoring results of the behavior pattern monitoring unit (170), and predicts that a risk (e.g., accident, outbreak, etc.) has occurred to the risk monitoring target when there is a large deviation between the past behavior pattern and the current behavior pattern of the risk monitoring target by comparing the two. To this end, the risk prediction unit (180) can predict the risk of the risk monitoring target by comparing and analyzing the behavior patterns of the risk monitoring target by date and time, and when the difference between the current behavior pattern and the behavior pattern of past dates / same time periods exceeds a preset behavior pattern deviation allowance range. For example, the risk of the risk monitoring target can be predicted when the deviation between the current behavior pattern as of December 1, 2022, and the behavior pattern on December 1 of the past three years (i.e., 2021, 2020, 2019) exceeds a preset behavior pattern deviation allowance range. At this time, the reference information for determining the above-mentioned allowable range of behavioral pattern deviation may include movement information per unit time, the number of times the door sensor operates during a specific time period, noise information, etc.

[0046] In addition, the risk prediction device (100) of the present invention further includes a future prediction unit (not shown) that predicts future energy usage and behavioral patterns based on the energy usage and behavioral pattern information of the risk monitoring target collected / stored during a predetermined past period, and the risk prediction unit (180) can predict the risk of the risk monitoring target when the future energy usage and behavioral pattern deviate from a preset normal range. For example, the risk prediction device (100) can predict the energy usage and behavioral patterns of the risk monitoring target for the next year based on the energy usage and behavioral pattern information of the risk monitoring target collected / stored during the past 10 years, and predict whether there is a risk based on the result.

[0047] Meanwhile, the risk prediction unit (180) adopts a recall evaluation method that determines whether the risk monitoring target is at risk based on an error matrix, and may adopt accuracy, specificity, sensitivity, and precision as evaluation types. The reason for adopting such a recall evaluation method is to minimize errors in which the risk state of the risk monitoring target is judged as having no risk possibility even though it is an actual risk situation.

[0048] FIGS. 7 and FIGS. 8 are drawings for explaining risk criteria based on an individual's energy consumption or behavioral patterns by season and time of day according to an embodiment of the present invention. FIGS. 7 is a diagram illustrating the energy consumption by time of day of a risk monitoring target (i.e., individual model) and the energy consumption by time of day of a comparison target group (i.e., entire data model) as time-series data, and FIGS. 8 is a diagram illustrating information collected from each of the door sensor, motion sensor, and sound sensor, which are types of IoT sensors, as time-series data by season and time of day.

[0049] Referring to FIGS. 7 and FIGS. 8, the risk prediction unit (180) first analyzes time series data as exemplified in FIG. 7 to compare the energy usage of the risk monitoring target and the comparison target group. When the usage of a specific energy (i.e., electric energy) exceeds a preset energy deviation tolerance range (A) as exemplified in FIG. 7, the risk prediction unit (180) monitors the time-series behavioral pattern information of the risk monitoring target as exemplified in FIG. 8. That is, the risk prediction unit (180) analyzes time series data as exemplified in FIG. 8, and when a situation (B) in which the form of at least one non-contact data drops sharply at any time is detected, it predicts a drop in the activity score of the risk monitoring target and can predict the risk.

[0050] The control unit (190) controls the overall operation of the risk prediction device (100) based on a preset control algorithm. That is, the control unit (190) can control the operation of each of the risk monitoring target information collection unit (120), comparison target group selection unit (130), comparison target information collection unit (140), comparison target group information generation unit (150), comparison / analysis unit (160), behavior pattern monitoring unit (170), and risk prediction unit (180) to predict the risk of the risk monitoring target based on information collected from the risk monitoring target information collection unit (120) and the comparison target information collection unit (140).

[0051] FIGS. 3 to 6 are schematic flowcharts of a risk prediction method according to an embodiment of the present invention. A risk prediction method according to an embodiment of the present invention is described as follows with reference to FIGS. 1 to 6.

[0052] First, in step S110, the control unit (190) stores the risk monitoring target information (i.e., the energy usage and behavioral pattern information of the risk monitoring target) collected by the risk monitoring target information collection unit (120) in the risk monitoring target information management DB (111) by time period. To this end, in step S111, the IoT sensor (122) collects IoT data, and in step S112, the control unit (126) stores the IoT data in the risk monitoring target information management DB (111), while receiving the date and time of collection of the IoT data from the time counter (123) and storing them together. To this end, one or more IoT sensors (122) may be installed in the space where the risk monitoring target resides. In step S113, the behavior pattern derivation unit (124) generates the time-based and seasonal behavior patterns of the risk monitoring target based on the IoT data and the corresponding collection date and time, and in step S114, the control unit (126) stores the time-based and seasonal behavior patterns of the risk monitoring target in the risk monitoring target information management DB (111). Additionally, the control unit (126) can store the time-based and seasonal energy usage of the risk monitoring target collected from the measuring instrument (121) in the risk monitoring target information management DB (111).

[0053] In step S120, the comparison target group selection unit (130) selects a comparison target group that includes at least one comparison target serving as a comparison standard to predict the risk status of the risk monitoring target, wherein the comparison target group may include at least one comparison target that has the same age group, gender, and housing type as the risk monitoring target, or at least one target among the comparison targets that has a disease history similar to or identical to the disease history of the risk monitoring target. In this case, it is preferable that the disease history is a disease history that affects at least one of energy usage and behavioral patterns.

[0054] In step S130, the energy usage of each of the comparison targets is collected, and an average value is calculated for each time period to derive the energy usage of the comparison target group for each time period. To this end, in step S131, the comparison target information collection unit (140) collects energy usage from each of the comparison targets, including the gas usage, electricity usage, and water usage of each of the comparison targets. Meanwhile, in step S132, the comparison target group information generation unit (150) calculates an average value for each of the energy usage of the comparison targets for each time period, and in step S133, the comparison target group information generation unit (150) derives the energy usage of the comparison target group for each time period using the average value.

[0055] In steps S140 and S150, the comparison / analysis unit (160) compares and analyzes the time-based energy usage of the risk monitoring target and the time-based energy usage of the comparison target group to determine whether the deviation between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowance range.

[0056] In step S160, if the difference between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowance range, the behavior pattern monitoring unit (170) monitors the time-based behavior pattern information of the risk monitoring target.

[0057] In step S170, the risk prediction unit (180) predicts the risk of the risk monitoring target based on the monitoring results. To this end, in step S171, the risk prediction unit (180) compares and analyzes the behavioral patterns of the risk monitoring target by date and time period, comparing the current and past behavioral patterns, and in steps S172 and S173, the risk prediction unit (180) predicts the risk of the risk monitoring target if the difference between the current behavioral pattern and the behavioral pattern of a past date / same time period exceeds a preset behavioral pattern deviation allowance range.

[0058] Additionally, in step S174, a future prediction unit (not shown) predicts future energy usage and behavioral patterns based on the energy usage and behavioral pattern information of the risk monitoring target collected / stored during a past predetermined period, and in steps S175 and S176, the risk prediction unit (180) predicts the risk of the risk monitoring target when the future energy usage and behavioral patterns deviate from a preset normal range.

[0059] Meanwhile, in step S170, the risk prediction unit (180) predicts the risk by adopting a recall evaluation method that determines whether the risk monitoring target is at risk based on an error matrix.

[0060] In describing the risk prediction method of the present invention with reference to FIGS. 1 to 6, redundant descriptions regarding the contents mentioned in the description of the risk prediction device of the present invention with reference to FIGS. 1 and 2 have been omitted.

[0061] Thus, the present invention generates a comparison target group comprising comparison targets that are similar to the risk monitoring target in at least one of age group, gender, living conditions, and health status, and by comparing the seasonal / time-of-day energy consumption between the comparison target group and the risk monitoring target, it predicts risk occurrence situations reflecting age group, gender, living conditions, or health status, thereby having the advantage of obtaining accurate prediction results.

[0062] In addition, the present invention has the advantage of predicting a risk occurrence situation of a risk monitoring target by reflecting at least one of the past behavioral pattern of the risk monitoring target and energy consumption by season / time of day, thereby predicting a risk occurrence situation that reflects an individual environment or behavioral pattern, and thereby obtaining a prediction result that reflects the individual characteristics of the risk monitoring target.

[0063] Although embodiments of the present invention have been described above, the scope of the present invention is not limited thereto and includes all changes and modifications within the scope recognized as equivalents that can be easily changed by a person skilled in the art from the embodiments to which the present invention belongs. Explanation of the symbols

[0064] 100: Risk Prediction Device 111: Risk Monitoring Target Information Management DB 112: Comparison Target Group Information Management DB 120: Risk Monitoring Target Information Collection Department 121: Measuring Instrument 122: IoT Sensor 123: Time Counter 124: Behavior Pattern Derivation Unit 125: Information generation unit 126: Control unit 130: Comparison Group Selection Department 140: Comparison Information Collection Department 150: Comparison Target Group Information Generation Section 160: Comparison / Analysis Section 170: Behavioral Pattern Monitoring Department 180: Risk Prediction Department 190: Control unit

Claims

Claim 1 A risk prediction device for predicting risk to a specific risk monitoring target using information collected by a non-contact type detection device and energy usage information, comprising: a first storage unit for storing the time-based energy usage and behavioral pattern information of the risk monitoring target collected through a non-contact type detection device installed in the space where the risk monitoring target resides; a second storage unit for storing the time-based energy usage of a comparison target group including at least one comparison target; a first information collection unit for collecting the energy usage and behavioral pattern information of the risk monitoring target collected through the non-contact type detection device and storing it in the first storage unit by time; a comparison target group selection unit for selecting a comparison target group including at least one comparison target that has the same age group, gender, and residential type as the risk monitoring target, and has a disease history identical to that of the risk monitoring target, which affects at least one of the energy usage and behavioral pattern; a second information collection unit for collecting energy usage from each of the comparison targets; and after deriving the time-based energy usage of the comparison target group by calculating the time-based average value for the energy usage of each of the comparison targets, the A risk prediction device characterized by comprising: a comparison target group information generation unit that stores in a second storage unit; a comparison analysis unit that primarily compares and analyzes the time-based energy usage of the risk monitoring target and the time-based energy usage of the comparison target group; a behavior pattern monitoring unit that secondarily monitors the time-based behavior pattern information of the risk monitoring target only when, as a result of the comparison analysis, the deviation between the energy usage of the risk monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowable range; and a risk prediction unit that predicts the risk of the risk monitoring target based on the monitoring result of the behavior pattern monitoring unit, wherein the risk of the risk monitoring target is predicted when the difference between the current behavior pattern and the behavior pattern of a past date / same time period exceeds a preset behavior pattern deviation allowable range by comparing and analyzing the date and time-based behavior patterns of the risk monitoring target. Claim 2 A risk prediction device according to claim 1, characterized in that each of the first and second storage units is implemented as an LSTM (Long Short-Term Memory). Claim 3 A risk prediction device according to claim 1, wherein the first information collection unit comprises: a measuring instrument installed in a space in which the risk monitoring target resides, which collects the energy usage of the risk monitoring target including gas, electricity, and water usage; at least one non-contact type detection device installed in a space in which the risk monitoring target resides; a behavior pattern derivation unit that collects non-contact data from the non-contact type detection device and derives the behavior pattern of the risk monitoring target; a time counter that counts time to determine the energy usage and the time of collection of the non-contact data; and an information generation unit that generates the energy usage and behavior pattern of the risk monitoring target by time and season based on the time of collection for each energy usage, the non-contact data, and the corresponding time of collection. Claim 4 delete Claim 5 A risk prediction device according to claim 1, further comprising a future prediction unit that predicts future energy usage and behavioral patterns based on energy usage and behavioral pattern information of the risk monitoring target collected / stored during a past predetermined period, wherein the risk prediction unit predicts the risk of the risk monitoring target when the future energy usage and behavioral patterns deviate from a preset normal range. Claim 6 A risk prediction device according to claim 5, wherein the risk prediction unit adopts a recall evaluation method that determines whether the risk monitoring target is at risk based on an error matrix. Claim 7 A risk prediction method using a risk prediction device that predicts risk to a specific risk-monitoring target using information collected by a non-contact type detection device and energy usage information, wherein the risk prediction device comprises: a risk-monitoring target information storage step in which the risk prediction device collects energy usage and behavioral pattern information of the risk-monitoring target, collected through a non-contact type detection device installed in a space where the risk-monitoring target resides, and stores the information by time period; a comparison target group selection step in which the risk prediction device selects a comparison target group including at least one comparison target that has the same age group, gender, and residential type as the risk-monitoring target, and has a disease history identical to that of the risk-monitoring target, which influences at least one of energy usage and behavioral patterns; a comparison target group information collection step in which the risk prediction device collects the energy usage of each of the comparison targets and calculates an average value by time period to derive the energy usage by time period of the comparison target group; and a comparison analysis step in which the risk prediction device primarily compares and analyzes the energy usage by time period of the risk-monitoring target and the energy usage by time period of the comparison target group. A risk prediction method characterized by comprising: a behavior pattern monitoring step in which a risk prediction device secondarily monitors time-based behavior pattern information of the risk-monitoring target only when the deviation between the energy usage of the risk-monitoring target and the energy usage of the comparison target group exceeds a preset energy deviation allowable range as a result of processing the comparison analysis step; and a risk prediction step in which the risk prediction device compares and analyzes the date and time-based behavior patterns of the risk-monitoring target, and predicts the risk of the risk-monitoring target when the difference between the current behavior pattern and the behavior pattern of a past date / same time period exceeds a preset behavior pattern deviation allowable range. Claim 8 A risk prediction method according to claim 7, wherein the risk monitoring target information storage step comprises: a first storage step of collecting non-contact data from at least one non-contact type detection device installed in a space where the risk monitoring target resides and storing it together with the date and time of collection; and a second storage step of generating and storing the time-varying and seasonal behavioral patterns of the risk monitoring target based on the non-contact data and the corresponding date and time of collection. Claim 9 delete Claim 10 A risk prediction method according to claim 7, wherein the risk prediction device further includes a future prediction step for predicting future energy usage and behavioral patterns based on energy usage and behavioral pattern information of the risk monitoring target collected / stored during a past predetermined period, and wherein the risk prediction step predicts the risk of the risk monitoring target when the future energy usage and behavioral patterns deviate from a preset normal range. Claim 11 A risk prediction method according to claim 10, wherein the risk prediction step adopts a recall evaluation method that determines whether the risk monitoring target is at risk based on an error matrix.

Images