An elderly care internet-of-things monitoring method and system

Abstract

The present application relates to the technical field of smart care for the aged and Internet of Things (IoT) care, and particularly relates to a method and system for IoT care for the aged. The method comprises the following steps: S1, generating a night-time getting-out-of-bed trigger event; S2, in response to the night-time getting-out-of-bed trigger event, starting a getting-out-of-bed return monitoring window; S3, mapping the path state data into a night-time getting-out-of-bed return path node sequence; and S4, forming an individualized path closure benchmark. The system comprises a getting-out-of-bed trigger module, a monitoring window starting module, a path data acquisition module, a path node sequence generation module, a path closure benchmark module, and a path closure judgment module. The present application no longer alarms only according to the total getting-out-of-bed duration, but judges whether the elderly has completed normal path closure according to the night-time getting-out-of-bed return path node sequence, thereby reducing false alarms caused by normal night-time toilet.

Description

Technical Field

[0001] This invention relates to the fields of smart elderly care and Internet of Things (IoT) care technology, and in particular to an IoT care method and system for elderly care. Background Technology

[0002] Existing IoT-based elderly care systems typically monitor the elderly using devices such as mattress sensors, infrared sensors, door magnetic sensors, millimeter-wave radar, smart bracelets, and emergency call buttons. When they detect that the elderly have been out of bed for too long, have not been active for an extended period, have fallen, have exceeded the toilet time limit, or have made an active call, they send alarm information to family members, caregivers, or the elderly care institution's management platform.

[0003] In nursing homes and home-based elderly care settings, getting out of bed at night to use the toilet is a common nighttime activity for the elderly, but it is also a high-risk process for falls, slips, dizziness, and sudden physical discomfort. Current technology typically handles nighttime bed-getting behavior by starting a timer when the mattress sensor detects the elderly person getting out of bed; if the duration of being out of bed exceeds a preset threshold, a timeout alarm is triggered.

[0004] The above method has the following drawbacks:

[0005] First, relying solely on the total time spent out of bed to trigger alarms can easily lead to false alarms. Different elderly people have significantly different nighttime toilet times. If a uniform threshold for time spent out of bed is used, elderly people who are slow to move or spend a long time in the toilet are more likely to trigger alarms frequently.

[0006] Second, to reduce false alarms, existing systems often set a relatively long timeout threshold for leaving the bed, but this can lead to a delay in detecting real anomalies. After an elderly person leaves the bed, anomalies may occur at the bedside, in the corridor, in the bathroom, or on the return route. If the system waits until the total time out of bed reaches a relatively long threshold before triggering an alarm, it will miss the opportunity for early intervention. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of existing technologies by proposing an IoT-based elderly care method and system.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A method for elderly care using IoT includes the following steps:

[0010] S1. During the nighttime care period, acquire bed status data generated by the bed detection device. When the bed status data changes from bed-lying state to bed-away state, generate a nighttime bed-away trigger event.

[0011] S2. In response to the nighttime bed leaving trigger event, open the bed leaving return monitoring window, and collect path status data generated by path detection devices in the bedside area, outbound passage area, bathroom area, return passage area and bed area within the bed leaving return monitoring window;

[0012] S3. Map the path status data into a nighttime bed-leaving return path node sequence. The nighttime bed-leaving return path node sequence includes multiple path nodes arranged according to the occurrence time. Each path node includes at least the node type, node trigger time, node region, node direction identifier, source device identifier, and node confidence level.

[0013] S4. Based on the historical normal nighttime bed-leaving and return samples of the elderly, establish individualized transfer duration thresholds between adjacent path nodes and path node dwell time thresholds to form individualized path closure benchmarks.

[0014] S5. Compare the real-time generated nighttime bed-leaving return path node sequence with the individualized path closure benchmark to determine whether the nighttime bed-leaving return path node sequence forms a closed path from the bed-leaving node to the bed-re-bed node.

[0015] S6. When the nighttime bed-leaving return path node sequence does not form a closed path, the abnormal interruption location and abnormality type are determined based on the last valid path node, the next expected path node, the path node dwell time, the adjacent path node transfer time, and the node confidence.

[0016] S7. Based on the location and type of the abnormal interruption, determine the recommended viewing area and output the hierarchical monitoring alarm information containing the recommended viewing area.

[0017] Preferably, the path nodes include a bed-leaving node, a bedside node, an outbound path node, a toilet node, a return path node, and a re-bed-up node;

[0018] The outbound passage node is a passage node generated when the elderly person moves from the bed area to the bathroom area;

[0019] The return route node is the route node generated when the elderly person moves from the bathroom area to the bed area.

[0020] Preferably, the path detection device includes one or more of the following: mattress sensor, bedside human body sensor, channel millimeter-wave radar, channel infrared sensor, bathroom door magnetic sensor, bathroom human body sensor, bathroom millimeter-wave radar, bedside lighting sensor, and channel lighting sensor;

[0021] When the bed detection device detects that an elderly person has left the bed, a bed departure node is generated;

[0022] When the bedside human body sensor or bedside millimeter-wave radar detects that an elderly person is in the bedside area, a bedside node is generated.

[0023] When the channel detection device detects that an elderly person is moving from the bed area to the bathroom area, a destination channel node is generated.

[0024] A bathroom node is generated when the bathroom door magnetic sensor, bathroom human body sensor, or bathroom millimeter-wave radar detects that an elderly person has entered or stayed in the bathroom area.

[0025] When the access detection device detects that an elderly person is moving from the toilet area to the bed area, a return access node is generated;

[0026] When the bed detection device detects that an elderly person has returned to a bedridden state, a "re-bedridden" node is generated.

[0027] Preferably, the node confidence level is determined based on the number of source devices, the status of source devices, the duration of data triggering, the matching relationship between adjacent path nodes, and the consistency of human movement direction;

[0028] When the same path node is triggered by at least two path detection devices within the same node confirmation time window, the confidence level of the node is increased.

[0029] When the source device is offline, has low battery, has delayed upload, or does not match with adjacent path nodes, the confidence level of the node is reduced.

[0030] Preferably, the individualized path closure benchmark is established in the following manner:

[0031] Obtain samples of nighttime bed-getting and returning to bed that were marked as normally completed within a preset historical period for elderly individuals;

[0032] Extract the transfer times of adjacent nodes between the bed-leaving node and the bedside node, the bedside node and the outbound path node, the outbound path node and the bathroom node, the bathroom node and the return path node, and the return path node and the re-bed-down node from the nighttime bed-leaving return sample;

[0033] Extract the node dwell time of bedside node, outbound route node, toilet node and return route node from the nighttime bed-leaving and return samples;

[0034] Individualized transfer duration thresholds and path node dwell duration thresholds are determined based on the adjacent node transfer duration and node dwell duration, respectively.

[0035] Preferably, the individualized transfer duration threshold is determined by the following formula:

[0036] ;

[0037] in, Indicates the first Individualized transfer time threshold for adjacent path node stages This indicates that the elderly person's historically normal nighttime bed-getting return sample is the first Average transition time of adjacent path node stages This represents the standard deviation of the corresponding transfer duration. This represents the threshold relaxation factor.

[0038] Preferably, the closed path refers to a nighttime bed-out return path node sequence that includes at least, in sequence, a bed-out node, a outbound path node, a toilet node, a return path node, and a bed-back node, and the actual transfer time between adjacent path nodes does not exceed the individualized transfer time threshold corresponding to the individualized path closure benchmark, and the actual stay time of the toilet node does not exceed the stay time threshold of the corresponding path node.

[0039] Preferably, when there are missing path nodes in the nighttime bed-leaving return path node sequence, a missing node compensation judgment is performed;

[0040] The missing node compensation judgment includes generating a missing path node or determining it as an untrusted path interruption based on the previous valid path node, the next valid path node, the direction of human movement, changes in bed status, and the presence status of human in the area.

[0041] Among them, when the bathroom door magnetic sensor is not triggered, but the bathroom human body sensor or bathroom millimeter-wave radar detects the presence of a human body, and the previous valid path node is the outbound channel node, a bathroom node is generated as compensation.

[0042] When the infrared sensor in the passage is not triggered, but the millimeter-wave radar in the passage detects a human body moving along the direction from the bed area to the bathroom area, a compensation is generated to create an outbound passage node.

[0043] When the return route node is missing, but the bed detection device detects a return to bed status and the toilet node has already been generated, a return route node will be generated as compensation.

[0044] Preferably, the location and type of the abnormal interruption are determined in the following manner:

[0045] When the last valid path node is a bedside node or a bedside node, and the next expected path node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the bedside area, and the abnormality type is determined to be the bedside getting up and staying abnormality.

[0046] When the last valid path node is the outbound channel node, and the next expected path node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the outbound channel area or the channel to the toilet entrance area, and the abnormality type is determined to be the outbound channel interruption abnormality.

[0047] When the last valid path node is the toilet node, and the actual dwell time of the toilet node exceeds the dwell time threshold of the corresponding path node, the abnormal interruption location is determined to be the toilet area, and the abnormal type is determined to be toilet stay abnormal.

[0048] When the last valid path node is the return route node, and the re-bed rest node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the return route area or the bedside area, and the abnormality type is determined to be the return return to bed interruption abnormality.

[0049] The suggested viewing area is determined based on the spatial connection relationship between the last valid path node and the next expected path node.

[0050] It also involves an IoT-based elderly care system, including:

[0051] The bed-leaving trigger module is used to generate nighttime bed-leaving trigger events based on bed status data generated by the bed detection device during nighttime care periods.

[0052] The monitoring window opening module is used to open the bed exit return monitoring window in response to the nighttime bed exit trigger event;

[0053] The path data acquisition module is used to collect path status data generated by path detection devices in the bedside area, outbound path area, toilet area, return path area and bed area within the bed departure and return monitoring window;

[0054] The path node sequence generation module is used to map the path status data into a nighttime bed leave and return path node sequence;

[0055] The path closure baseline module is used to establish individualized transfer duration thresholds between adjacent path nodes and path node dwell time thresholds based on the elderly’s historical normal nighttime bed-leaving and return samples.

[0056] The path closure judgment module is used to compare the real-time generated nighttime bed-leaving return path node sequence with the individualized path closure benchmark to determine whether the nighttime bed-leaving return path node sequence forms a closed path from the bed-leaving node to the bed-re-bed-staying node.

[0057] The abnormal interruption location module is used to determine the abnormal interruption location and abnormality type when the nighttime bed-leaving return path node sequence does not form a closed path, based on the last valid path node, the next expected path node, the path node dwell time, the adjacent path node transfer time, and the node confidence.

[0058] The alarm output module is used to determine the recommended viewing area based on the abnormal interruption location and abnormality type, and output graded monitoring alarm information containing the recommended viewing area;

[0059] The path node sequence generation module includes a node mapping unit, a node confidence calculation unit, and a missing node compensation unit.

[0060] The present invention has the following beneficial effects:

[0061] Instead of relying solely on the total time spent out of bed, alarms are now triggered based on the sequence of return path nodes at night to determine whether the elderly person has completed a normal path closure, which can reduce false alarms caused by normal nighttime toilet visits.

[0062] Based on historical normal nighttime bed-leaving and return samples of the elderly, individualized transfer duration thresholds and node stay duration thresholds between adjacent path nodes are established to adapt to the differences in mobility among different elderly people.

[0063] This invention can determine the abnormal interruption location based on the last valid path node and the next expected path node when the path is not closed, and distinguish between bedside getting up, outbound path interruption, bathroom stay and return to bed interruption, thereby improving the positioning accuracy of alarm information;

[0064] It can output suggested areas to check based on the location of abnormal interruption, so that caregivers do not have to search through multiple rooms or areas one by one, thus improving the efficiency of on-site handling.

[0065] By using node confidence and missing node compensation mechanisms, even when a single path detection device misses a detection or goes offline, compensation judgments can still be made by combining adjacent path nodes, human movement direction, and changes in bed status, thereby improving the stability of path recognition in complex living environments. Attached Figure Description

[0066] Figure 1 This is a flowchart illustrating an IoT-based elderly care method proposed in this invention.

[0067] Figure 2 This is a schematic diagram of the spatial path node layout for the IoT-based elderly care method of the present invention;

[0068] Figure 3 This is a schematic diagram of the missing node compensation judgment process of the present invention;

[0069] Figure 4This is a structural block diagram of an elderly care IoT system proposed in this invention. Detailed Implementation

[0070] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0071] Example 1:

[0072] This embodiment is applied to the scenario of home-based elderly care. For example... Figure 2 As shown, a mattress sensor is installed at the bedside in the elderly person's bedroom, a bedside human body sensor is installed in the area next to the bed, a passage millimeter-wave radar and a passage infrared sensor are installed in the passage area between the bedroom and the bathroom, a bathroom door magnetic sensor is installed at the bathroom entrance, and a bathroom human body sensor and a bathroom millimeter-wave radar are installed inside the bathroom.

[0073] Reference Figure 1 The system is set to provide nighttime care from 22:00 to 6:30 the next day.

[0074] At 2:10 AM, the mattress sensor detected that the bed status changed from lying down to being out of bed. The bed leaving trigger module generated a nighttime bed leaving trigger event, and the monitoring window opening module opened the bed leaving return monitoring window.

[0075] At 2:10:12, the bedside human body sensor detected that the elderly person was in the bedside area, and the path node sequence generation module generated a bedside node.

[0076] At 2:10:45, the millimeter-wave radar detected the elderly person moving from the bed area to the bathroom area, and the system generated a destination channel node.

[0077] At 2:11:20, the bathroom door magnetic sensor is triggered, the bathroom human body sensor detects the presence of a human body, and the system generates a bathroom node.

[0078] At 2:16:50, the millimeter-wave radar detected the elderly person moving from the bathroom area to the bed area, and the system generated a return route node.

[0079] At 2:18:00, the mattress sensor detected that the elderly person had returned to a bedridden state, and the system generated a "re-bedridden" node.

[0080] At this time, the real-time path node sequence is:

[0081] Leaving the bed node—Bedside node—Outbound route node—Bathroom node—Return route node—Re-lying in bed node.

[0082] The path closure determination module compares the real-time path node sequence with the individualized path closure benchmark. It confirms that the transfer time between adjacent path nodes and the time spent in the bathroom do not exceed the corresponding thresholds, thus determining that a closed path has been formed. The system does not trigger any abnormal alarms, but only records the process of leaving and returning to bed during this night.

[0083] This embodiment illustrates that the present invention can identify the complete return path of a normal nighttime toilet visit, avoiding false alarms caused by relying solely on the total time spent away from the bed.

[0084] Example 2: Location of Abnormal Stay in the Toilet

[0085] At 3:05 a.m. one day, the mattress sensor detected that the elderly person had left the bed, and the system generated an exit node and opened the exit return monitoring window.

[0086] At 3:05:15, the system generates a bedside node.

[0087] At 3:05:50, the system generates the outbound channel node.

[0088] At 3:06:30, the bathroom door magnetic sensor and the bathroom human body sensor are triggered, and the system generates a bathroom node.

[0089] Based on the elderly person's historical normal nighttime bed-leaving and return samples, the individualized dwell time threshold for the bathroom node is 12 minutes. If, by 3:19:00, the bathroom human body sensor still detects a human presence and no return path node is generated, the path closure judgment module determines that the path is not closed.

[0090] The abnormal interruption location module determines that the last valid path node is the toilet node, the next expected path node is the return route node, the abnormal interruption location is the toilet area, and the abnormality type is toilet lingering abnormality. The alarm output module outputs a second-level care alarm, with alarm information including the abnormal interruption location "toilet area" and the suggested inspection area "toilet".

[0091] In this embodiment, after receiving an alarm, caregivers can go directly to the bathroom to check, without having to search the bedside, corridor, or living room one by one.

[0092] Example 3: Outbound channel interruption anomaly

[0093] At 1:30 a.m. one day, the mattress sensor detected that the elderly person had left the bed, and the system generated a bed-leaving node.

[0094] At 1:30:10, the bedside human body sensor detected that the elderly person was in the bedside area, and the system generated a bedside node.

[0095] At 1:30:40, the millimeter-wave radar detected the elderly person moving from the bed area to the bathroom area, and the system generated a destination channel node.

[0096] Based on the elderly person's historical normal data, the individualized transfer time threshold from the outbound channel node to the toilet node is 80 seconds. If, by 1:32:20, the system has still not generated a toilet node, and the channel millimeter-wave radar continues to detect low-speed human movement or lingering in the channel area, the system determines that the path is not closed.

[0097] The abnormal interruption location module determines that the last valid path node is the outbound channel node, the next expected path node is the toilet node, the abnormal interruption location is the area from the channel to the toilet entrance, and the abnormality type is outbound channel interruption abnormality.

[0098] The alarm output module outputs a first-level care alarm, triggers the lighting equipment in the passageway, and pushes an alarm message to the caregiver's terminal: "Suspected abnormal lingering in the passageway to the bathroom entrance area, please check."

[0099] This embodiment illustrates that the present invention can detect anomalies in advance based on path stage transfer timeouts before the overall bed-out time reaches the traditional bed-out timeout threshold.

[0100] Example 4: Abnormal Interruption of Return Trip to Bed

[0101] At 4:00 AM one day, after the elderly person got out of bed, the following nodes were generated in sequence: bedside node, outbound path node, and bathroom node.

[0102] At 4:07, the millimeter-wave radar detected the elderly person moving from the bathroom area to the bed area, and the system generated a return route node.

[0103] Based on the elderly person's historical normal data, the individualized transfer time threshold from the return route node to the re-bedding node is 100 seconds. If, by 4:09, the mattress sensor still has not detected the elderly person re-bedding, and the bedside human body sensor detects the presence of a human body in the bedside area, the system determines that the path is not closed.

[0104] The abnormal interruption location module 7 determines that the last valid path node is the return channel node, the next expected path node is the re-bed rest node, the abnormal interruption location is the area from the return channel to the bedside, and the abnormality type is the return return to bed interruption abnormality.

[0105] The alarm output module 8 outputs a third-level care alarm and pushes emergency alarm information to the family terminal, duty terminal, or nursing station terminal.

[0106] This embodiment illustrates that the present invention can identify abnormal situations where an elderly person has left the bathroom but has not been able to go back to bed, thus avoiding the situation where the abnormal location is unclear due to only outputting "time out of bed".

[0107] Example 5: Missing Node Compensation Judgment

[0108] Reference Figure 4 In practical applications, the bathroom door magnetic sensor may miss detections. During a nighttime bed-leaving process, the system had already generated bed-leaving, bedside, and outbound passage nodes, but the bathroom door magnetic sensor did not trigger. Simultaneously, the bathroom human body sensor and the bathroom millimeter-wave radar detected the presence of a person in the bathroom area.

[0109] At this point, the missing node compensation unit generates a toilet node based on the previous valid path node being the outbound channel node, the presence of a human body in the toilet area, and the human body's movement direction pointing towards the toilet area, and marks this node as a compensation node.

[0110] If the system subsequently generates a return route node and a re-bedization node, and the overall path sequence is normal, the system determines that the path is closed and does not trigger an abnormal alarm.

[0111] If a return route node is not generated for an extended period after the toilet node is generated as compensation, the system can still determine that the toilet is experiencing an abnormal stay based on the toilet node's dwell time exceeding the time limit.

[0112] This embodiment illustrates that the node compensation mechanism does not simply ignore device missed detections, but rather repairs the path sequence while preserving node confidence, enabling the system to reduce false alarms while maintaining its anomaly identification capabilities.

[0113] like Figure 3 As shown, the present invention also provides an elderly care IoT system, including a bed-leaving trigger module, a monitoring window opening module, a path data acquisition module, a path node sequence generation module, a path closure benchmark module, a path closure judgment module, an abnormal interruption location module, and an alarm output module.

[0114] The bed-leaving trigger module is used to generate nighttime bed-leaving trigger events based on bed status data generated by the bed detection equipment during nighttime care periods.

[0115] The monitoring window opening module is used to open the bed exit return monitoring window in response to a nighttime bed exit trigger event.

[0116] The path data acquisition module is used to collect path status data generated by path detection devices in the bedside area, outbound path area, toilet area, return path area, and bed area within the bed departure and return monitoring window.

[0117] The path node sequence generation module is used to map path status data into a nighttime bed exit and return path node sequence.

[0118] The path node sequence generation module includes a node mapping unit, a node confidence calculation unit, and a missing node compensation unit.

[0119] The node mapping unit is used to generate the bedside node, bedside node, outbound route node, toilet node, return route node, and re-bedding node based on the path status data.

[0120] The node confidence calculation unit is used to determine the node confidence based on the number of source devices, the status of source devices, the duration of data triggering, the matching relationship between adjacent path nodes, and the consistency of human movement direction.

[0121] The missing node compensation unit is used to generate a missing path node or determine an untrusted path interruption when a path node is missing, based on the previous valid path node, the next valid path node, the direction of human movement, changes in bed status, and the presence status of human beings in the area.

[0122] The path closure baseline module is used to establish individualized transfer duration thresholds between adjacent path nodes and path node dwell time thresholds based on the elderly’s historical normal nighttime bed-leaving and return samples.

[0123] The path closure judgment module is used to compare the real-time generated nighttime bed-out return path node sequence with the individualized path closure benchmark to determine whether the nighttime bed-out return path node sequence forms a closed path from the bed-out node to the bed-back node.

[0124] The abnormal interruption location module 7 is used to determine the abnormal interruption location and abnormality type when the sequence of return path nodes at night has not formed a closed path, based on the last valid path node, the next expected path node, the dwell time of the path node, the transfer time of the adjacent path node, and the node confidence.

[0125] The alarm output module is used to determine the recommended viewing area based on the location and type of abnormal interruption, and output graded monitoring alarm information containing the recommended viewing area.

[0126] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for elderly care using IoT, characterized in that, Includes the following steps: S1. During the nighttime care period, acquire bed status data generated by the bed detection device. When the bed status data changes from bed-lying state to bed-away state, generate a nighttime bed-away trigger event. S2. In response to the nighttime bed leaving trigger event, open the bed leaving return monitoring window, and collect path status data generated by path detection devices in the bedside area, outbound passage area, bathroom area, return passage area and bed area within the bed leaving return monitoring window; S3. Map the path status data into a nighttime bed-leaving return path node sequence. The nighttime bed-leaving return path node sequence includes multiple path nodes arranged according to the occurrence time. Each path node includes at least the node type, node trigger time, node region, node direction identifier, source device identifier, and node confidence level. S4. Based on the historical normal nighttime bed-leaving and return samples of the elderly, establish individualized transfer duration thresholds between adjacent path nodes and path node dwell time thresholds to form individualized path closure benchmarks. S5. Compare the real-time generated nighttime bed-leaving return path node sequence with the individualized path closure benchmark to determine whether the nighttime bed-leaving return path node sequence forms a closed path from the bed-leaving node to the bed-re-bed node. S6. When the nighttime bed-leaving return path node sequence does not form a closed path, the abnormal interruption location and abnormality type are determined based on the last valid path node, the next expected path node, the path node dwell time, the adjacent path node transfer time, and the node confidence. S7. Based on the location and type of the abnormal interruption, determine the recommended viewing area and output the hierarchical monitoring alarm information containing the recommended viewing area.

2. The elderly care IoT monitoring method according to claim 1, characterized in that, The path nodes include the bed exit node, bedside node, outbound path node, toilet node, return path node, and re-bed node; The outbound passage node is a passage node generated when the elderly person moves from the bed area to the bathroom area; The return route node is the route node generated when the elderly person moves from the bathroom area to the bed area.

3. The elderly care IoT monitoring method according to claim 2, characterized in that, When the bed detection device detects that an elderly person has left the bed, a bed departure node is generated; When the bedside human body sensor or bedside millimeter-wave radar detects that an elderly person is in the bedside area, a bedside node is generated. When the channel detection device detects that an elderly person is moving from the bed area to the bathroom area, a destination channel node is generated. A bathroom node is generated when the bathroom door magnetic sensor, bathroom human body sensor, or bathroom millimeter-wave radar detects that an elderly person has entered or stayed in the bathroom area. When the access detection device detects that an elderly person is moving from the toilet area to the bed area, a return access node is generated; When the bed detection device detects that an elderly person has returned to a bedridden state, a "re-bedridden" node is generated.

4. The elderly care IoT monitoring method according to claim 1, characterized in that, The node confidence level is determined based on the number of source devices, the status of source devices, the duration of data triggering, the matching relationship between adjacent path nodes, and the consistency of human movement direction. When the same path node is triggered by at least two path detection devices within the same node confirmation time window, the confidence level of the node is increased. When the source device is offline, has low battery, has delayed upload, or does not match with adjacent path nodes, the confidence level of the node is reduced.

5. The elderly care IoT monitoring method according to claim 1, characterized in that, The individualized path closure benchmark is established in the following manner: Obtain samples of nighttime bed-getting and returning to bed that were marked as normally completed within a preset historical period for elderly individuals; Extract the transfer times of adjacent nodes between the bed-leaving node and the bedside node, the bedside node and the outbound path node, the outbound path node and the bathroom node, the bathroom node and the return path node, and the return path node and the re-bed-down node from the nighttime bed-leaving return sample; Extract the node dwell time of bedside node, outbound route node, toilet node and return route node from the nighttime bed-leaving and return samples; Individualized transfer duration thresholds and path node dwell duration thresholds are determined based on the adjacent node transfer duration and node dwell duration, respectively.

6. The elderly care IoT monitoring method according to claim 1, characterized in that, The individualized transfer duration threshold is determined by the following formula: ； in, Indicates the first Individualized transfer time threshold for adjacent path node stages This indicates that the elderly person's historically normal nighttime bed-getting return sample is the first Average transition time of adjacent path node stages This represents the standard deviation of the corresponding transfer duration. This represents the threshold relaxation factor.

7. The elderly care IoT monitoring method according to claim 1, characterized in that, The closed path refers to a sequence of nighttime bed-out and return path nodes that includes at least the bed-out node, outbound path node, toilet node, return path node, and re-bed-out node in sequence, and the actual transfer time between adjacent path nodes does not exceed the corresponding individualized transfer time threshold in the individualized path closure benchmark, and the actual stay time of the toilet node does not exceed the corresponding path node stay time threshold.

8. The elderly care IoT monitoring method according to claim 1, characterized in that, When there are missing path nodes in the nighttime bed leave return path node sequence, a missing node compensation judgment is performed. The missing node compensation judgment includes generating a missing path node or determining it as an untrusted path interruption based on the previous valid path node, the next valid path node, the direction of human movement, changes in bed status, and the presence status of human in the area. Among them, when the bathroom door magnetic sensor is not triggered, but the bathroom human body sensor or bathroom millimeter-wave radar detects the presence of a human body, and the previous valid path node is the outbound channel node, a bathroom node is generated as compensation. When the infrared sensor in the passage is not triggered, but the millimeter-wave radar in the passage detects a human body moving along the direction from the bed area to the bathroom area, a compensation is generated to create an outbound passage node. When the return route node is missing, but the bed detection device detects a return to bed status and the toilet node has already been generated, a return route node will be generated as compensation.

9. The elderly care IoT monitoring method according to claim 1, characterized in that, The location and type of the abnormal interruption are determined in the following way: When the last valid path node is a bedside node or a bedside node, and the next expected path node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the bedside area, and the abnormality type is determined to be the bedside getting up and staying abnormality. When the last valid path node is the outbound channel node, and the next expected path node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the outbound channel area or the channel to the toilet entrance area, and the abnormality type is determined to be the outbound channel interruption abnormality. When the last valid path node is the toilet node, and the actual dwell time of the toilet node exceeds the dwell time threshold of the corresponding path node, the abnormal interruption location is determined to be the toilet area, and the abnormal type is determined to be toilet stay abnormal. When the last valid path node is the return route node, and the re-bed rest node is not generated within the corresponding individualized transfer time threshold, the abnormal interruption location is determined to be the return route area or the bedside area, and the abnormality type is determined to be the return return to bed interruption abnormality. The suggested viewing area is determined based on the spatial connection relationship between the last valid path node and the next expected path node.

10. An IoT-based elderly care system, used to implement the IoT-based elderly care method according to any one of claims 1-9, characterized in that, include: The bed-leaving trigger module is used to generate nighttime bed-leaving trigger events based on bed status data generated by the bed detection device during nighttime care periods. The monitoring window opening module is used to open the bed exit return monitoring window in response to the nighttime bed exit trigger event; The path data acquisition module is used to collect path status data generated by path detection devices in the bedside area, outbound path area, toilet area, return path area and bed area within the bed departure and return monitoring window; The path node sequence generation module is used to map the path status data into a nighttime bed leave and return path node sequence; The path closure baseline module is used to establish individualized transfer duration thresholds between adjacent path nodes and path node dwell time thresholds based on the elderly’s historical normal nighttime bed-leaving and return samples. The path closure judgment module is used to compare the real-time generated nighttime bed-leaving return path node sequence with the individualized path closure benchmark to determine whether the nighttime bed-leaving return path node sequence forms a closed path from the bed-leaving node to the bed-re-bed-staying node. The abnormal interruption location module is used to determine the abnormal interruption location and abnormality type when the nighttime bed-leaving return path node sequence does not form a closed path, based on the last valid path node, the next expected path node, the path node dwell time, the adjacent path node transfer time, and the node confidence. The alarm output module is used to determine the recommended viewing area based on the abnormal interruption location and abnormality type, and output graded monitoring alarm information containing the recommended viewing area; The path node sequence generation module includes a node mapping unit, a node confidence calculation unit, and a missing node compensation unit.

Images