A user space based patient cognitive ability monitoring system and method

Abstract

The application relates to a user space-based patient cognitive ability monitoring system, comprising a first client, the first client being communicatively connected to a server, the server pre-storing a predetermined behavior sequence and a time length of a predetermined specific behavior in line with user behavior habits, being capable of generating a first early warning information, and obtaining a first confidence degree according to the accuracy of a behavior sequence of a user in a current period compared with a predetermined behavior sequence based on received behavior data of daily behaviors associated with the user; the application also relates to a user space-based patient cognitive ability monitoring method, wherein the server obtains a cognitive ability score of a user according to data obtained by the first client according to the following scoring formula: the server regularly compares the cognitive ability score of the user corresponding to a current period with the cognitive ability score of the user corresponding to a last period, and generates a first early warning information when the cognitive ability score of the user decreases by more than a preset trigger threshold.

Description

[0001] The original basis of this divisional application is a patent application with the application number 201910421469.5 and the application date of May 20, 2019, and the invention name of "a cognitive ability evaluation system and method based on daily behaviors of the elderly". TECHNICAL FIELD

[0002] The present application relates to the field of database construction, to machine learning in the field of artificial intelligence, and in particular to a cognitive ability evaluation system and method based on daily behaviors of the elderly. BACKGROUND

[0003] In the prior art, a method for monitoring activity ability of a subject as proposed in the patent document with the publication number CN108348194A includes determining, in at least one processing device, sensor data indicative of a sensor reading of each sensor of a plurality of sensors, the sensors being located within a living environment of the subject and the sensor data of each sensor being at least partially indicative of one or more activities performed by the subject; determining, for each activity domain of a plurality of activity domains, a domain score indicative of an activity level within the respective activity domain, the domain score being determined using respective combined sensor data from sensors associated with the respective domain; determining a reference activity level using a reference domain score measured during a reference time period; determining a current activity level using a current domain score measured during a monitoring time period; and generating an activity indicator at least partially in dependence on the current activity level and the reference activity level, the activity indicator being at least partially indicative of a difference between the current activity level and the reference activity level, thereby providing feedback regarding the activity ability of the subject. But this is a kind of existing rule-based method, in the case where the rules are not applicable to the specific environment, this method is very complex and prone to error.

[0004] In addition, on the one hand, due to the difference in understanding of those skilled in the art; on the other hand, because the inventors have studied a large number of literatures and patents when making the present application, but limited by the size and have not listed all the details and contents in detail, but this does not mean that the present application does not have the characteristics of these prior arts, on the contrary, the present application has all the characteristics of the prior art, and the applicant reserves the right to add relevant prior art in the background art. SUMMARY

[0005] To address the shortcomings of existing technologies, this invention provides a cognitive ability assessment system and method based on the daily behaviors of the elderly. Since cognitive decline is reflected in the daily behaviors of the elderly, this invention utilizes sensing technology to achieve long-term monitoring of the elderly's behavioral data, thereby enabling the assessment of their cognitive abilities. This eliminates the need for frequent assessments using cognitive impairment screening scales, reducing disruption to the user's daily life and lowering the workload of medical staff.

[0006] According to a preferred embodiment, a cognitive ability assessment system based on the daily behavior of the elderly is provided. The system includes a server and several behavioral sensors, wherein the server collects behavioral data of daily behaviors associated with the user in an implicit sensing manner through the several behavioral sensors deployed in the user's living space; the server assesses the user's cognitive ability based at least on the behavioral data of daily behaviors associated with the user to obtain the user's cognitive ability score.

[0007] According to a preferred embodiment, the server periodically compares the cognitive ability score of the user in the current period with the cognitive ability score of the user in the previous period, and the server generates a first warning message when the user's cognitive ability score decreases by more than a preset trigger threshold.

[0008] According to a preferred embodiment, the server assesses the user's cognitive ability according to the following scoring formula to obtain the user's cognitive ability score:

[0009] ;

[0010] Before assessing a user's cognitive ability to obtain a cognitive ability score, the server selects behavioral data corresponding to at most n daily triggered behaviors as the standard task X for evaluating the user's cognitive ability, denoted as... T is the set of standard times for an average person to complete each of the n tasks, denoted as: T i Let be the standard time for an average person to complete the i-th task out of the n tasks, and t be the set of average times for users to complete each of the n tasks, denoted as . , t i α is the average time a user takes to complete the i-th task out of the n tasks, and α is the set of task completion rates for each of the n tasks. i This represents the user's task completion rate for the i-th task out of the n tasks, where w is the set of weights for each task out of the n tasks. i This represents the weight of the i-th task among the n tasks.

[0011] According to a preferred embodiment, the behavioral data of daily behaviors associated with users collected by the server has a time attribute. After acquiring the behavioral data of daily behaviors associated with users, the server stores the behavioral data of daily behaviors associated with users into a time series database according to the time attribute. The server independently builds a hidden Markov model associated with each user to predict their behavior. The server uses the behavioral data related to users in the time series database to train the hidden Markov model associated with the user. The server predicts the user's predicted behavior based on the hidden Markov model associated with the user, and generates a second warning message when the user's actual behavior deviates from the predicted behavior and the deviation will lead to known risks and / or known losses.

[0012] According to a preferred embodiment, the system further includes a first client, which is communicatively connected to the server. The first client acts as a relay device to acquire behavioral data collected by the plurality of behavioral sensors deployed in the living space. After generating a second warning message, the server sends a risk verification request to the first client. The first client issues an alarm in response to the risk verification request, and the alarm can only be deactivated by the user on the first client after passing at least two biometric verifications. After deactivating the alarm, the first client sends a risk false alarm feedback or risk deactivation feedback to the server. The server deletes the second warning message in response to the risk false alarm feedback or risk deactivation feedback.

[0013] According to a preferred embodiment, the system further includes a camera device communicatively connected to a first client. The camera device can be positioned outside the entrance / exit of the living space. The camera device is used to capture a first set of photos and a second set of photos. The first set of photos is taken before the entrance / exit is opened, and the second set of photos is taken during a preset period after the entrance / exit is opened until it is closed. The first client acquires the first set of photos and the second set of photos from the camera device. The first client compares the first set of photos and the second set of photos to determine the changing status of people within the living space, and the first client responds to this determination. The living space includes several occupancy modes, comprising one of the following: a user-only mode with only the user present; a visitor mode with other people present or only other people present; and an empty mode with no one present. A first client analyzes the trigger of behavioral data collected by several behavioral sensors during the visitor mode. If the trigger of the corresponding behavioral data collected by the first client during the visitor mode cannot be determined, the client retains the behavioral data for which the trigger cannot be determined, and only sends the behavioral data collected by the several behavioral sensors during the visitor mode that can identify the user as the trigger to the server. Preferably, the preset time can be set by the manufacturer and / or the user. For example, the preset time can be 1-5 seconds. Specifically, for example, the preset time can be 2 seconds, 3 seconds, or 4 seconds. That is, assuming the preset time is 2 seconds, the second photo set includes photos taken within 2 seconds after the door is opened and closed.

[0014] According to a preferred embodiment, the plurality of sensors includes at least an angular displacement sensor mounted on the door / exit gate for identifying the door / exit gate's open / closed state. The angular displacement sensor for identifying the door / exit gate's open / closed state is communicatively connected to the camera device. The angular displacement sensor for identifying the door / exit gate's open / closed state is configured to immediately and actively send an electrical signal indicating that the door / exit gate is open to the camera device when the door / exit gate is rotated in the closing direction until the door / exit gate is completely closed. After receiving the electrical signal indicating that the door / exit gate is open, the camera device takes several photos within 1 second as a first set of photos. The angular displacement sensor for identifying the door / exit gate's open / closed state is configured to send an electrical signal indicating that the door / exit gate is closed only when the door / exit gate is rotated in the closing direction until the door / exit gate is completely closed. After the first set of photos is taken, the camera device continuously takes several photos until it receives the electrical signal indicating that the door / exit gate is closed, and the period of time that lasts for a preset time is used as a second set of photos.

[0015] According to a preferred embodiment, the system further includes a smart bracelet and an identity recognition module. The server assigns a unique identity to each user and stores it in the smart bracelet. The smart bracelet transmits electromagnetic waves carrying the identity within a limited communication range. Each behavior sensor is equipped with an identity recognition module to identify the electromagnetic waves carrying the identity, thereby detecting the smart bracelet worn by the user and identifying the trigger of the behavior data collected by the behavior sensor. When the first client determines that the occupancy mode of the living space is a user-only mode, the first client receives the behavior data collected by the behavior sensor and defaults to defining the trigger of the behavior data collected by the identity recognition module when the user is not wearing a smart bracelet as the user. When the first client determines that the occupancy mode of the living space changes from user-only mode to visitor mode in response to a change in the occupancy status of the living space, the first client instructs the user to wear the smart bracelet. The first client receives the behavior data collected by the behavior sensor and only identifies the behavior data collected by the identity recognition module when the user is wearing a smart bracelet as the trigger of the user's behavior data.

[0016] According to a preferred embodiment, the server pre-stores a predetermined sequence of behaviors conforming to user behavior habits and a predetermined duration of specific behaviors. After generating a first warning message, the server analyzes the accuracy of the user's behavior sequence within the current period compared to the predetermined behavior sequence based on received behavioral data related to the user's daily behavior, obtaining a first confidence level. The server then analyzes the second confidence level of the duration of the corresponding specific behavior within the current period compared to the predetermined duration of the specific behavior based on received behavioral data related to the user's daily behavior. The server stores preset first confidence level thresholds and second confidence level thresholds. If the first confidence level is greater than the second confidence level threshold, the server determines that the first warning message is invalid. The server directly confirms the validity of the first warning information when the first confidence level is less than or equal to the first confidence threshold and the second confidence level is less than or equal to the second confidence threshold. The server marks the first warning information as a first warning information to be verified when one of the conditions of the first confidence level being less than or equal to the first confidence threshold and the second confidence level being less than or equal to the second confidence threshold is met. The server waits for the behavioral data of the user's daily behavior in the next period sent by the first client to verify the first warning information to be verified. The server confirms the validity of the first warning information only when the cognitive ability score of the user in the next period decreases by more than a preset trigger threshold compared with the cognitive ability score of the user in the two periods before the next period. When the server confirms the validity of the first warning information, it sends the first warning information to a second client held by a person other than the user.

[0017] According to a preferred embodiment, a cognitive ability assessment method based on the daily behavior of the elderly is provided. The method uses the system described above to assess the user's cognitive ability. The system includes a server and several behavioral sensors, wherein the server implicitly collects behavioral data of the user's daily behavior in a perceptual manner through the several behavioral sensors deployed in the user's living space; the server assesses the user's cognitive ability based at least on the behavioral data of the user's daily behavior to obtain the user's cognitive ability score. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a preferred embodiment of the system of the present invention.

[0019] List of reference numerals

[0020] 100: First client; 200: Second client; 300: Server

[0021] 400: Behavior sensor; 500: Camera device; 600: Smart bracelet

[0022] 700: Identity Recognition Module Detailed Implementation

[0023] The following is a detailed explanation with reference to the accompanying drawings.

[0024] First, some of the terms used in the invention are defined:

[0025] Implicit sensing refers to the process by which behavioral sensors are coupled to objects in or around a living space to detect and collect behavioral data when a user touches or operates these objects. In other words, implicit sensing is the process of acquiring user behavioral data implicitly through behavioral sensors coupled to objects in the living space that are touched by the user's daily activities. This differs from active testing by the user; in implicit sensing, the user does not need to deliberately trigger the behavioral sensors, but rather triggers them incidentally while performing daily activities to obtain corresponding behavioral data. The placement and measurement methods of the behavioral sensors do not disturb the user's daily activities. That is, several behavioral sensors are placed in the user's living space in a way that does not disturb the user, collecting behavioral data related to the user's daily activities. Objects may include at least one of various types of doors, switches, and appliances. Specifically, objects may include at least one of interior doors, window doors, entrance doors, refrigerator doors, microwave oven doors, wardrobe doors, faucet switches, light switches, stove switches, power switches, wash cups, and washbasins.

[0026] A time-series database can refer to a database containing time series data. Time-series databases are primarily used to process data with time labels (data that changes sequentially over time, i.e., time-series data) or time attributes. Data with time labels or time attributes is also called time-series data.

[0027] Example 1

[0028] This embodiment discloses a cognitive ability assessment system based on the daily behavior of the elderly, or a cognitive ability assessment system based on the daily behavior of users, or a system for assessing cognitive ability based on a time-series database, or a system for assessing cognitive ability. This system is suitable for executing the various method steps described in this invention to achieve the desired technical effects. Where there is no conflict or contradiction, the whole and / or parts of the preferred embodiments of other embodiments can be used as supplements to this embodiment.

[0029] According to a preferred embodiment, see Appendix Figure 1The system may include a server 300 and / or several behavior sensors 400. The server 300 can collect behavioral data related to the user's daily behaviors through the behavior sensors 400 located or deployed in the user's living space. Alternatively, the server 300 can collect behavioral data related to the user's daily behaviors in an implicit sensing manner through the behavior sensors 400 deployed in the user's living space. Preferably, the behavior sensors 400 may include at least one of an angle sensor, a door magnetic switch sensor, a laser sensor, a water flow sensor, a micro switch sensor, a pressure sensor, a temperature sensor, a humidity sensor, a smart switch, and an angle sensor. The server 300 can assess the user's cognitive ability based on the behavioral data related to the user's daily behaviors to obtain a cognitive ability score. Preferably, the server 300 collects behavioral data related to the user's daily behaviors in an implicit sensing manner through the behavior sensors 400 located in the user's living space. Preferably, the living space refers to the user's living environment. For example, the living space may include at least one of a bedroom, a living room, a kitchen, and a toilet. The present invention can achieve at least the following beneficial technical effects by using this method: First, it can reduce the disturbance to users' daily behavior caused by frequent assessment of cognitive ability; second, it reduces the difficulty of continuously obtaining users' cognitive ability scores, so as to observe changes in users' cognitive ability over a long period of time.

[0030] According to a preferred embodiment, server 300 can periodically compare the cognitive ability score of a user in the current period with that of a user in the previous period. Server 300 can generate an early warning message when the user's cognitive ability score decreases by more than a preset trigger threshold. Preferably, server 300 can send the first early warning message to a second client. Preferably, the comparison period for server 300 to periodically compare the cognitive ability score of a user in the current period with that of a user in the previous period can be manually set. Preferably, the preset trigger threshold can be manually set. For example, the preset trigger threshold can be 5% to 60%. That is, assuming server 300 compares the user's cognitive ability score on a daily basis, the preset trigger threshold is set to 10%. If the user's cognitive ability score was 50 points the previous day and is 38 points today, then the user's cognitive ability score has decreased by 24% compared to the previous period, exceeding 10%, and server 300 generates an early warning message. Preferably, the comparison period for server 300 to periodically compare the cognitive ability score of a user in the current period with that of a user in the previous period can have multiple periods of different lengths. For example, server 300 can be configured to use at least one of the following periods: day, week, month, quarter, and year. For example, server 300 can be configured to compare the cognitive ability scores of corresponding users on a daily basis. Alternatively, server 300 can be configured to compare the cognitive ability scores of corresponding users on both daily and weekly basis. That is, server 300 compares the cognitive ability scores of users over two consecutive days, as well as over two consecutive weeks. Preferably, server 300 can periodically compare the cognitive ability scores of users in the current period with those of users in the previous period using at least two different comparison periods. Server 300 can generate an alert when a user's cognitive ability score decreases by more than a preset trigger threshold. Different comparison periods can correspond to different preset trigger thresholds. The longer the comparison period, the smaller the preset trigger threshold can be. For example, when server 300 compares the cognitive ability scores of users over two consecutive days, the corresponding preset trigger threshold can be 15%. When server 300 compares the cognitive ability scores of users over two consecutive weeks, the corresponding preset trigger threshold can be 10%. Preferably, the cycle length can refer to the duration contained in one cycle.This invention achieves at least the following beneficial technical effects: First, the server 300 periodically processes and compares user behavior data collected by sensors to provide feedback on changes in the user's cognitive abilities. Furthermore, the server 300 generates an early warning when the user's cognitive ability score decreases by more than a preset trigger threshold, thus indicating a potential rapid decline in the user's cognitive ability. Second, continuously evaluating a user's cognitive abilities daily using existing cognitive impairment screening scales is virtually impossible. This invention utilizes rich user behavior data collected daily by sensors to evaluate cognitive abilities. By periodically comparing data over multiple cycles, it can obtain changes in the user's cognitive abilities over periods of varying lengths and provide timely warnings. Third, generally, assuming a user's cognitive abilities decline slowly over time, the longer the time span, the greater the expected decline. Therefore, conventional methods are... While longer comparison periods generally correspond to larger preset trigger thresholds, the inventors of this invention believe that longer periods result in larger data volumes, leading to more accurate cognitive ability scores with smaller errors. Therefore, this invention breaks with convention by setting smaller preset trigger thresholds in server 300 for longer comparison periods. For example, when server 300 compares cognitive ability scores of users over two consecutive days, the preset trigger threshold is 19%; when comparing scores over two consecutive weeks, it is 16%; when comparing scores over two consecutive quarters, it is 13%; and when comparing scores over two consecutive quarters, it is 10%. This reduces false alarms and makes the warning information more accurate.

[0031] According to a preferred embodiment, the system may include at least one of a first client, a smart bracelet 600, and an identity recognition module 700. The server 300 may assign a unique identifier to each user and store the identifier in the smart bracelet 600. The smart bracelet 600 may be carried by the user. The smart bracelet 600 may transmit electromagnetic waves carrying the identifier within a limited communication range. Each behavior sensor 400 may be equipped with an identity recognition module 700 to identify the user associated with the behavior data collected by the behavior sensor 400. When the smart bracelet 600 moves with the user such that its communication range covers the corresponding identity recognition module 700, the corresponding identity recognition module 700 identifies the user associated with the behavior data collected by the behavior sensor 400 based on the acquired electromagnetic waves carrying the identifier and generates information with time attributes. The corresponding first client is associated with the user having the corresponding identifier.

[0032] According to a preferred embodiment, the system may include a plurality of cameras. These cameras may be positioned around the entrance to a user's living space. The system can identify the individuals within the living space based on images captured by the cameras. The system may ignore behavioral data that has been influenced by others other than the user.

[0033] According to a preferred embodiment, server 300 can assess a user's cognitive ability using the following scoring formula to obtain a user's cognitive ability score: Preferably, before assessing a user's cognitive ability to obtain a cognitive ability score, the server 300 can select behavioral data corresponding to up to n daily triggered behaviors as a standard task X for evaluating the user's cognitive ability, denoted as... T can be the set of standard times for an average person to complete each of the n tasks, denoted as... T i t can be the standard time for an average person to complete the i-th task out of the n tasks. t can be the set of average times for users to complete each of the n tasks, denoted as t = t / t. . t i α can be the average time a user takes to complete the i-th task out of the n tasks. α can be the set of the user's task completion rates for each of the n tasks. i This can represent the user's completion rate of the i-th task out of n tasks. w can be the set of weights for each of the n tasks. iThis can represent the weight of the i-th task among n tasks. For example, if a user opens a door and doesn't close it for a certain period, it's highly likely the user forgot to close it. The completion rate of the task completed by the door-opening / closing event decreases, and the cognitive ability score calculated using the cognitive ability scoring formula will also decrease. If the door-opening / closing event is one of the user's most frequently triggered n behaviors, it will be used to evaluate the user's cognitive ability; the more times the user forgets to open or close the door, the lower the cognitive ability score. Similarly, if a user takes a long time to turn off the water while washing vegetables, the time required to complete the vegetable-washing event will far exceed the normal time required, highly likely the user forgot to turn off the water. If the faucet-turning event is one of the user's most frequently triggered n behaviors, it will be used to evaluate the user's cognitive ability, and the calculated cognitive ability score will also decrease. This invention uses the user's most frequently triggered n behaviors to evaluate the user's cognitive ability, which can reduce the excessive interference from a few, accidental data points caused by insufficient data sources for evaluation. Preferably, different users have different habits; for example, some users may habitually leave doors and windows open for ventilation. For example, some users may have a habit of leaving the microwave oven door open for extended periods to air out odors when not in use. When using this invention, the cognitive ability scoring rules can be explained to the user beforehand, allowing them to avoid making mistakes due to personal habits that could lead to significant score deviations. Furthermore, whether users can consciously perform actions to achieve higher cognitive ability scores after understanding these rules also reflects their cognitive abilities. Preferably, n can be an integer greater than or equal to 1. n can be set manually, for example, to 1, 2, 3, or 4. The setting can be based on the number of behavioral sensors installed in the user's living space; the more behavioral sensors, the larger the value of n can be.

[0034] According to a preferred embodiment, the behavioral data of daily behaviors associated with users collected by server 300 may have a time attribute. After acquiring the behavioral data of daily behaviors associated with users, server 300 can store the behavioral data of daily behaviors associated with users into a time-series database according to the time attribute. Server 300 can independently build a Hidden Markov Model associated with each user to predict their behavior. Server 300 can use the behavioral data related to users in the time-series database to train the Hidden Markov Model associated with that user. Server 300 can predict the user's predicted behavior based on the Hidden Markov Model associated with that user. The server can generate a second warning message when the user's actual behavior deviates from the predicted behavior and this deviation will lead to known risks and / or known losses.

[0035] According to a preferred embodiment, the system may include a first client 100, which is communicatively connected to a server 300. The first client 100 can act as a relay device to acquire behavioral data collected by several behavioral sensors 400 deployed within the living space. The server 300 can generate a second warning message and send a risk verification request to the first client 100. The first client 100 can issue an alarm in response to the risk verification request. The alarm can only be deactivated after the user verifies at least two biometric identifications on the first client 100. After deactivating the alarm, the first client 100 can send a false alarm feedback or a risk deactivation feedback to the server 300. The server 300 can delete the second warning message in response to the false alarm feedback or the risk deactivation feedback.

[0036] According to a preferred embodiment, the system may include a camera device 500. The camera device 500 may be communicatively connected to a first client 100. The camera device 500 may be located outside the entrance / exit of the living space. The camera device 500 may be used to capture a first set of photos and a second set of photos. The first set of photos may be taken before the entrance / exit is opened. The second set of photos may be taken during a preset period after the entrance / exit is opened until it is closed. The first client 100 can acquire the first set of photos and the second set of photos from the camera device 500. The first client 100 can compare the first set of photos and the second set of photos to determine the changing status of people within the living space. The first client 100 can determine the occupancy mode of the living space in response to the changing status of people within the living space. The occupancy mode may include one of the following modes: a user-only mode where only the user is present in the living space; a visitor mode where other people besides the user are present or only other people are present in the living space; and an vacant mode where no one is present in the living space. The first client 100 can analyze the triggers of behavioral data collected by several behavior sensors 400 during the visitor mode. The first client 100 can intercept the corresponding behavioral data for which the trigger cannot be determined when the trigger of the corresponding behavioral data collected during the guest mode cannot be determined, and only send the behavioral data collected by several behavioral sensors 400 during the guest mode that can determine that the trigger is the user to the server 300.

[0037] Preferably, the length of the preset time can be set by the manufacturer and / or the user. For example, the preset time can be 1 to 5 seconds. Specifically, for example, the preset time can be 2 seconds, 3 seconds, or 4 seconds. That is, assuming the preset time is 2 seconds, the second photo group includes photos taken within 2 seconds after the door is opened and closed. This invention achieves at least the following beneficial technical effects: First, typically, existing technologies use a camera device 500 outside the door to collect features of people for feature recognition to determine whether the corresponding person has the right to enter, but this does not involve the present invention's determination of the occupancy pattern of people in the living space. This invention, by using this method, can both ensure the authenticity of behavioral data and reduce the inconvenience for the elderly. Although the elderly can also operate on the first client 100 to eliminate false behavioral data, due to the elderly's unfamiliarity with and difficulty in operating smart devices, this invention strives to reduce the number of steps the elderly need to take in order to simplify the system and minimize disruption to the elderly; Second, since the camera device 500 is located outside the door, it will not collect photos related to the elderly's privacy, thus avoiding any resistance from the elderly.

[0038] According to a preferred embodiment, the plurality of behavioral sensors may include angular displacement sensors mounted on the door / exit gate for identifying the opening / closing state of the door / exit gate. The angular displacement sensors for identifying the opening / exit gate state can be communicatively connected to the camera device 500. The angular displacement sensors for identifying the opening / exit gate state can be configured to immediately and actively send an electrical signal indicating that the door / exit gate is open to the camera device 500 when the door / exit gate is rotated 3° to 5° in the opening direction. Upon receiving the electrical signal indicating that the door / exit gate is open, the camera device 500 can take several photos within 1 second as a first set of photos. The angular displacement sensors for identifying the opening / exit gate state can be configured to send an electrical signal indicating that the door / exit gate is closed only when the door / exit gate is rotated in the closing direction until it is completely closed. The camera device 500 can continuously take several photos after the first set of photos is taken, and continue taking photos for a preset period of time until receiving the electrical signal indicating that the door / exit gate is closed, as a second set of photos. The present invention achieves at least the following beneficial technical effects by using this method: the angular displacement sensor of the present invention for identifying the opening and closing status of the entrance is communicatively connected to the camera device 500, and the time of the first and second photo sets collected by the camera device 500 is related to the opening and closing status of the entrance and exit, so that the first and second photo sets captured by the present invention are likely related to the changing status of people in the living space, and will not take too many useless photos, such as photos of passersby passing by the entrance and exit when the entrance and exit are not open. This can reduce the computational load of the first client 100 and enable the first client 100 to efficiently and accurately analyze the occupancy pattern of the people in the living space.

[0039] According to a preferred embodiment, the system may include a smart bracelet 600 and / or an identity recognition module 700. The server 300 may assign a unique identity to each user and store the identity in the smart bracelet 600. The smart bracelet 600 may transmit electromagnetic waves carrying the identity within a limited communication range. Each behavior sensor 400 may be equipped with an identity recognition module 700 for recognizing the electromagnetic waves carrying the identity, thereby detecting the user wearing the smart bracelet 600 and identifying the triggerer of the behavior data collected by the behavior sensor 400. When the first client 100 determines that the occupancy mode of the people in the living space is a user-only mode, the first client 100 may receive the behavior data collected by the behavior sensor 400 and default the triggerer of the behavior data collected by the identity recognition module 700 when no user is detected not wearing the smart bracelet 600 as the user.

[0040] When the first client 100 determines, in response to a change in the occupancy status of the living space, that the occupancy mode of the living space has changed from a user-only mode to a visitor mode, the first client 100 can instruct the user to wear the smart bracelet 600. When the first client 100 determines, in response to a change in the occupancy status of the living space, that the occupancy mode of the living space has changed from a user-only mode to a visitor mode, the first client 100 can receive behavioral data collected by the behavior sensor 400 and only identify the behavioral data collected by the identity recognition module 700 when it detects the user wearing the smart bracelet 600 as the triggering behavioral data of the user. The present invention achieves at least the following beneficial technical effects through this method: First, in the user-only living mode, the user does not need to wear the smart bracelet 600, giving the user more freedom; Second, in the user's daily life, due to various reasons such as relatives and friends visiting, maintenance personnel maintaining equipment, or water, electricity, and gas companies checking meters, it is inevitable that other people will enter the living space. However, in the visitor mode, the user can wear the smart bracelet 600 so that even when others are in the living space, only the behavioral data triggered by the user is collected to evaluate the user's cognitive ability, thereby avoiding the incorrect evaluation of the user's cognitive ability based on behavioral data triggered by others, and improving the accuracy of the cognitive ability evaluation of the present invention; Third, in the visitor mode, the false alarms of the first and / or second warning information caused by the user and visitors simultaneously triggering different behavioral sensors 400 are reduced.

[0041] According to a preferred embodiment, server 300 may pre-store a predetermined sequence of behaviors conforming to user habits and a predetermined duration of specific behaviors. After generating a first warning message, server 300 can analyze the accuracy of the user's behavior sequence in the current period compared to the predetermined behavior sequence based on received behavioral data related to the user's daily behavior to obtain a first confidence level. Server 300 can analyze the duration of the corresponding specific behavior in the current period compared to the predetermined duration of the specific behavior based on received behavioral data related to the user's daily behavior to obtain a second confidence level. Server 300 may store preset first confidence thresholds and second confidence thresholds. The first confidence threshold may be greater than the second confidence threshold. Server 300 may determine that the first warning message is invalid when the first confidence level is greater than the first confidence threshold and the second confidence level is greater than the second confidence threshold. Server 300 may directly confirm that the first warning message is valid when the first confidence level is less than or equal to the first confidence threshold and the second confidence level is less than or equal to the second confidence threshold. Server 300 may mark the first warning information as pending verification if either a first confidence level is less than or equal to a first confidence threshold or a second confidence threshold is met. Server 300 may wait for behavioral data on the user's daily behavior associated with the user in the next period from the first client 100 to verify the pending warning information. Server 300 confirms the validity of the first warning information only if the cognitive ability score of the user in the next period decreases by more than a preset trigger threshold compared to the cognitive ability score of the user assessed in the two periods prior to the next period. Server 300 may send the first warning information to a second client 200 held by someone other than the user when it confirms the validity of the first warning information. The present invention achieves at least the following beneficial technical effects by employing this method: First, the present invention reduces false alarms of the first warning information and improves the accuracy of the first warning information by setting a comparison between a first confidence level and a first confidence level threshold, and a comparison between a second confidence level and a second confidence level threshold; Second, because the duration of a user's behavior is more likely to fluctuate due to potential factors than the order of behavior, the first confidence level threshold preset in the server 300 is greater than the second confidence level threshold, thereby further reducing false alarms of the first warning information.

[0042] According to a preferred embodiment, the holder of the second client 200 may be, for example, a medical professional or the operator of the system of the present invention. The server 300 can send a first warning message to the second client 200 so that relevant personnel are aware of the change in the user's cognitive abilities. Preferably, after receiving the first warning message, the second client 200 can instruct its holder to go to the user's living space to evaluate the user's cognitive abilities using the MoCA scale and / or MMSE scale to obtain an on-site score reflecting the user's cognitive abilities. When the on-site score is lower than a corresponding preset score, the user is deemed no longer suitable for living alone. This allows for accurate on-site assessment of the cognitive abilities of users whose cognitive abilities have declined to the point of being unsuitable for living alone, and whose living alone would pose corresponding risks, and timely intervention in such users' behavior regarding living alone. For example, the user could be sent to a nursing home with centralized care by nursing staff, or their family members could be notified to care for the user, to prevent the risks arising from the user's severely declining cognitive abilities living alone.

[0043] Example 2

[0044] This embodiment discloses a method for assessing cognitive abilities based on the daily behaviors of the elderly, or a method for assessing cognitive abilities based on the daily behaviors of users, or a method for assessing cognitive abilities based on a time-series database, or a method for assessing cognitive abilities. This method can be implemented by the system of this invention and / or other alternative components. For example, the method of this invention can be implemented by using various components in the system of this invention. Where there is no conflict or contradiction, the whole and / or parts of the preferred embodiments of other embodiments can be used as supplements to this embodiment.

[0045] According to a preferred embodiment, the method can use the system of the present invention to assess a user's cognitive abilities.

[0046] According to a preferred embodiment, the method may include at least one of the following steps: 1, 2, 3, and 4. The method may include an information acquisition step, an expert knowledge base construction step, a data processing step, and a cognitive ability assessment and early warning step. The data acquisition step consists of sensors and a server; the sensors collect information about the elderly person's movements at each moment, and the server collects and integrates the sensor data. The information knowledge base construction step includes classifying existing features to a higher level and storing the classified data. The data processing step includes constructing a behavior prediction classifier based on the stored classification information, where a Hidden Markov Model is used as the classification model. The cognitive ability assessment step includes scoring the obtained behaviors and issuing an early warning for results where the score drops below a threshold.

[0047] Step 1: Collect basic information and construct a sensor information dataset, including a sensor device information table and a sensor data information table. The sensor device information table describes the basic information of various sensors, with attributes including: sensor name, MAC address, function description, sensor type, data unit, and installation location. The MAC address serves as a unique identifier for the sensor device information. The sensor data information table describes the information collected by the sensors, with attributes including: sensor device, data content, and upload time.

[0048] By deploying sensors in the room, data is uploaded to a cloud server when the sensors are triggered. The server stores the data and archives and processes it periodically based on the upload time.

[0049] Step 2: Construct an expert knowledge base. Expert knowledge provides prior knowledge about the relationship between sensors and behaviors, increasing the accuracy of behavior prediction and consequently, the accuracy of cognitive early warning. This is achieved by constructing a table showing the relationships between behaviors, sensors, and attributes, enabling behavior prediction that is not entirely rule-based. The expert knowledge base primarily includes a sensor behavior table, which contains attributes such as behavior name, behavior definition, triggering item, triggering sensor, trigger time, and location. Basic daily behaviors are defined using common sense, breaking them down into specific actions. The items triggered by these actions are associated with sensors, thus limiting the behavior to a few possible sensors. These rules improve the accuracy of behavior prediction. The sensor behavior table categorizes the underlying sensors into different behaviors, and the many-to-many relationships between these behaviors and sensors are stored in the expert knowledge base.

[0050] Step 3: Behavior Prediction Analysis. Taking advantage of the implicit temporal sequence in sensor and behavioral data, Hidden Markov Models (HMMs) are used to predict behavior based on sensor data. An HMM introduces a latent variable and assumes that changes in the state of the data are caused by and only by the latent variable at the previous time point, but this latent variable cannot be directly observed by the observer. In an HMM, the behavior at the time point to be predicted is solved from the transition matrix between the latent variables, the initial state probability matrix, and the emission matrix of the behavior corresponding to the latent variable. When using HMMs for behavior recognition, behavior is considered as a latent variable of the sensor; sensor data is observable, while behavioral data is hidden.

[0051] According to a preferred embodiment, the present invention trains a Hidden Markov Model (HMM) using pre-collected data. The HMM is then used to predict behavior based on sensor data collected each day, with a 24-hour cycle. The steps of the Markov model training and prediction process are described below:

[0052] Define variables for the Hidden Markov Model:

[0053]

[0054] Define matrix The Line 1 The element values ​​of the column are also determined by the first... each state Transfer to the each state The probability is :

[0055]

[0056] Define matrix The Line 1 The column value is also in the state of At that time, the observed data The probability of:

[0057]

[0058] Thus far, the Markov model The build is complete, but and All of these are unknown. The following will provide the process and formulas for solving Hidden Markov Models using the Viterbi algorithm.

[0059] The Viterbi algorithm aims to maximize the probability of the latent variable sequence given the observed data, i.e.:

[0060]

[0061] Viterbi Algorithm Definition

[0062]

[0063] The recursive formula can then be derived from the above formula:

[0064]

[0065] definition The Viterbi algorithm process is as follows:

[0066] Input: Model Observe the data sequence

[0067] Output: Latent variable sequence

[0068] Initialize local state:

[0069]

[0070]

[0071] Perform dynamic programming recursion Local state at time:

[0072]

[0073]

[0074] Calculation time The largest ,at this time That is The most likely hidden variable state at time 10:00

[0075]

[0076] use Backtracking :

[0077]

[0078] When using Hidden Markov Models for prediction, sensor data is treated as an observation sequence, i.e. Treating behavior as a latent variable, i.e. The behavior at each time step can be obtained using the Viterbi algorithm.

[0079] Step 4: By identifying the obtained behavioral information, assess cognitive abilities and issue early warnings. A decline in cognitive abilities manifests in daily life as behavioral errors, such as skipping or incorrectly executing important steps in completing an activity, rendering the activity meaningless. These behavioral errors might include forgetting to turn off the gas, leaving the refrigerator door open for an extended period, or spending too much time completing a simple task.

[0080] This invention selects the most n daily triggered behaviors as the standard task X, denoted as: The test obtains the time it takes for an average person to complete the n tasks as the standard time T for that behavior, denoted as: Assuming each person's task completion time is Task completion rate: The weight of each task is determined based on its importance. The formula for scoring cognitive ability is as follows:

[0081]

[0082] Cognitive abilities are standardized using scores to quantify the quality of activity completion. An alert is generated when a user's cognitive ability score drops by more than 10% based on daily ratings.

[0083] According to a preferred embodiment, the behavior sensor device information table describes basic information about various types of behavior sensors, and its attributes are shown in the table below:

[0084]

[0085] According to a preferred embodiment, the sensor device collects data and transmits it to a server. The server stores the sensor data information, the properties of which are shown in the table below:

[0086]

[0087] According to a preferred embodiment, the following deployment is made in a typical apartment building to detect basic human movements:

[0088]

[0089] According to a preferred embodiment, the behavior sensor device includes a communication gateway and a communication module. Different sensors transmit information to a first client via the communication module. The communication module can be a wired communication module or a wireless communication module. For example, the wireless communication module can be a Bluetooth module or a ZigBee module. The first client forwards all data to the server via Wi-Fi. The system is compatible with multiple behavior sensors; when a new behavior sensor is connected, the sensor's data only needs to be set to the communication protocol determined by the current system. Behavior sensors only transmit data when triggered by a behavior, reducing the amount of data transmitted.

[0090] According to a preferred embodiment, the communication protocol format is defined as shown in the table below:

[0091]

[0092] According to a preferred embodiment, after the data is stored on the server, due to the large amount of data, the data is archived at midnight every day, and the data collected that day is organized, packaged and stored.

[0093] According to a preferred embodiment, the behavior sensor-behavior mapping table constructed from expert knowledge is shown below:

[0094]

[0095] It should be noted that the specific embodiments described above are exemplary, and those skilled in the art can devise various solutions inspired by the disclosure of this invention. These solutions all fall within the scope of this invention and its protection. Those skilled in the art should understand that this specification and its accompanying drawings are illustrative and not intended to limit the scope of the claims. The scope of protection of this invention is defined by the claims and their equivalents.

Claims

1. A user-space-based patient cognitive ability monitoring system, characterized in that, The system includes: The first client (100) is used to acquire collected behavioral data; The server (300) assesses the user's cognitive ability based at least on behavioral data related to the user's daily behavior to obtain the user's cognitive ability score; The server (300) is used to periodically compare the cognitive ability score of the user in the current period with the cognitive ability score of the user in the previous period; The server (300) generates a first warning message when the user's cognitive ability score decreases by more than a preset trigger threshold. The server (300) can assess the user's cognitive ability according to the following scoring formula to obtain the user's cognitive ability score: ； Before assessing a user's cognitive ability to obtain a cognitive ability score, the server (300) selects behavioral data corresponding to at most n daily triggered behaviors as the standard task X for evaluating the user's cognitive ability, denoted as... T is the set of standard times for an average person to complete each of the n tasks, denoted as: T i Let be the standard time for an average person to complete the i-th task out of the n tasks, and t be the set of average times for users to complete each of the n tasks, denoted as . , t i α is the average time a user takes to complete the i-th task out of the n tasks, and α is the set of task completion rates for each of the n tasks. i This represents the user's task completion rate for the i-th task out of the n tasks, where w is the set of weights for each task out of the n tasks. i This represents the weight of the i-th task among the n tasks; The server (300) can obtain a first confidence level by analyzing the accuracy of the user's behavior sequence in the current period compared with a predetermined behavior sequence based on the received behavioral data of the user's daily behavior associated with the user; the server (300) can also obtain a second confidence level by analyzing the time length of the corresponding specific behavior of the user in the current period compared with the time length of a predetermined specific behavior based on the received behavioral data of the user's daily behavior associated with the user. The server (300) has a preset first confidence threshold and a second confidence threshold in its memory, wherein the first confidence threshold is greater than the second confidence threshold; when the first confidence is greater than the first confidence threshold and the second confidence is greater than the second confidence threshold, the first warning information is determined to be invalid; when the first confidence is less than or equal to the first confidence threshold and the second confidence is less than or equal to the second confidence threshold, the first warning information is directly confirmed to be valid.

2. The patient cognitive ability monitoring system based on user space according to claim 1, characterized in that, The server (300) can predict the user's behavior based on the hidden Markov model associated with the user, and generate a second warning message when the user's actual behavior deviates from the predicted behavior and such deviation will lead to known risks and / or known losses. The server (300) is able to send a risk verification request to the first client (100) after generating the second warning information.

3. The patient cognitive ability monitoring system based on user space according to claim 2, characterized in that, The first client (100) issues an alert in response to the risk verification request, and the alert can only be deactivated by the user on the first client after passing at least two biometric verifications. After the first client (100) can clear the alarm, it sends a risk false alarm feedback or risk clearance feedback to the server (300), and the server (300) deletes the second warning information in response to the risk false alarm feedback or risk clearance feedback.

4. A method for monitoring patient cognitive abilities based on user space using the monitoring system as described in any one of claims 1 to 3, characterized in that, Includes the following steps: Acquire collected behavioral data; At least based on behavioral data related to the user's daily behavior, the user's cognitive ability is assessed according to the following scoring formula to obtain the user's cognitive ability score: ； Before assessing a user's cognitive ability to obtain a cognitive ability score, the server (300) selects behavioral data corresponding to at most n daily triggered behaviors as the standard task X for evaluating the user's cognitive ability, denoted as... T is the set of standard times for an average person to complete each of the n tasks, denoted as: T i Let be the standard time for an average person to complete the i-th task out of the n tasks, and t be the set of average times for users to complete each of the n tasks, denoted as . , t i α is the average time a user takes to complete the i-th task out of the n tasks, and α is the set of task completion rates for each of the n tasks. i This represents the user's task completion rate for the i-th task out of the n tasks, where w is the set of weights for each task out of the n tasks. i This represents the weight of the i-th task among the n tasks; Periodically compare the cognitive ability scores of users in the current period with those of users in the previous period; The first warning message is generated when the user's cognitive ability score drops more than a preset trigger threshold.

5. The method for monitoring patient cognitive ability based on user space according to claim 4, characterized in that, The process includes the following steps: when one of the following conditions is met, the first warning information is marked as a first warning information to be verified; the first warning information to be verified is verified based on the behavioral data of the user's daily behavior associated with the next period.

6. The method for monitoring patient cognitive ability based on user space according to claim 5, characterized in that, It also includes the following steps: when the cognitive ability score of the user in the next cycle decreases by more than a preset trigger threshold compared with the cognitive ability score of the user in the two cycles prior to the next cycle, the first warning information is confirmed to be valid; when the first warning information is confirmed to be valid, the first warning information is sent to other persons other than the user.

Images