Smart old-age care remote state monitoring method and device in combination with user portrait

Abstract

The invention relates to an intelligent old-age care remote state monitoring method and device combined with a user portrait, and relates to the technical field of data processing, and the method comprises the steps: communicating a positioning patch array of a first old-age care user, and receiving a positioning time sequence matrix of a first time zone; processing a first pension user portrait and the positioning time sequence matrix through a health sample N-level adjacency searcher to obtain an adjacency sample state; aggregating the adjacent sample states to generate state baseline information; the communication state monitor is used for collecting state monitoring information of the first time zone; and when the state monitoring information does not belong to the state baseline information, performing abnormal state identification on the state monitoring information. The technical problem that in the prior art, due to the fact that flexible change analysis of the health value interval cannot be achieved, the monitoring accuracy of abnormal states of old people is insufficient is solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and device for remote status monitoring of smart elderly care combined with user portraits. Background Art

[0002] As the proportion of the aging population gradually increases, the cost of traditional personal care continues to rise, and elderly care is gradually tending to be intelligent, remote, and unmanned, that is, smart elderly care. Abnormal monitoring of the elderly's state is an important part of smart elderly care. By real-time monitoring of the elderly's physiological state information and performing abnormal analysis, timely alarms for abnormal conditions of the elderly can be achieved, improving the reliability of smart elderly care.

[0003] In the research on abnormal physiological status monitoring of the elderly, the health value of each elderly person's physiological status is particularly important. There are currently two determination strategies. One is to customize the health value interval of the physiological status by professionals, and the other is to determine the health value interval of the elderly's physiological status by analyzing the static portrait of the user. The daily activity data of the elderly user will also affect the health value interval of the physiological status, which leads to greater flexibility in the actual health value interval. However, the existing strategy cannot realize the analysis of flexible changes in the health value interval, which leads to the technical problem of insufficient accuracy in abnormal monitoring of the elderly's status. Summary of the invention

[0004] In order to solve the technical problem that the prior art cannot realize the analysis of flexible changes in health value intervals, resulting in insufficient accuracy in abnormal monitoring of the elderly's status, the present invention provides a smart elderly care remote status monitoring method and device combined with user portraits.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In the first aspect, the present invention provides a remote status monitoring method for smart elderly care combined with user portraits, including: communicating with a positioning patch array of a first elderly care user, receiving a positioning timing matrix of a first time zone; processing the first elderly care user portrait and the positioning timing matrix through a healthy sample N-level adjacent searcher to obtain adjacent sample states, 5≥N≥2, N is an integer; aggregating the adjacent sample states to generate status baseline information; a communication status monitor to collect status monitoring information of the first time zone; when the status monitoring information does not belong to the status baseline information, marking the status monitoring information as abnormal.

[0007] In the second aspect, the present invention provides a smart elderly care remote status monitoring device combined with user portraits, including: a positioning timing matrix receiving module, used to communicate with the positioning patch array of the first elderly care user, and receive the positioning timing matrix of the first time zone; an adjacent sample state search module, used to process the first elderly care user portrait and the positioning timing matrix through the healthy sample N-level adjacent searcher, and obtain the adjacent sample state, 5≥N≥2, N is an integer; a state baseline information configuration module, used to aggregate the adjacent sample states and generate state baseline information; a state monitoring module, used for communicating with the state monitor, collecting the state monitoring information of the first time zone; an abnormal state identification module, used to identify the abnormal state of the state monitoring information when the state monitoring information does not belong to the state baseline information.

[0008] In a third aspect, the present invention provides a non-transitory computer-readable storage medium, in which a computer software program is stored. When the computer software program is executed by a processor, the method for remote status monitoring of smart elderly care combined with user portraits described in the first aspect is implemented.

[0009] The beneficial effects of the present invention are as follows: through the user's positioning patch array, the user's positioning timing matrix in the first time zone can be monitored in real time, which can characterize the motion state of the elderly user; then the motion state is used as a dynamic portrait, and the first elderly user portrait is used as a static portrait for adjacent sampling to obtain the adjacent sample state; further, the adjacent sample state is data fused to obtain the user's state baseline data under the portrait state constraint; if the state monitoring information does not belong to the state baseline information, the state monitoring information is marked as abnormal. By integrating dynamic portraits and static portraits, the state baseline of the elderly user is automatically evaluated by relying on the data fusion algorithm, the reliability of the state baseline is improved, and the technical effect of ensuring the accuracy of abnormal monitoring is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of the flow of the smart elderly care remote status monitoring method combined with user portraits provided by the present invention;

[0011] Figure 2 A schematic diagram of the adjacent sample state aggregation process of the smart elderly care remote state monitoring method combined with user portrait provided by the present invention;

[0012] Figure 3 A schematic diagram of the structure of a smart elderly care remote status monitoring device combined with user portraits provided by the present invention;

[0013] Figure 4 A schematic diagram of the structure of a computer-readable storage medium provided by the present invention.

[0014] In the drawings, the components represented by the reference numerals are listed as follows: computer readable storage medium 600 , computer program 611 . DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0016] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0017] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0018] Embodiment 1:

[0019] like Figure 1 As shown, the embodiment of the present invention provides a remote state monitoring method for smart elderly care combined with user portraits, including the steps of:

[0020] S10: communicating with the positioning patch array of the first elderly care user and receiving the positioning timing matrix of the first time zone;

[0021] Further, the positioning patch array of the first elderly user is communicated to receive the positioning timing matrix of the first time zone, and step S10 includes the following steps:

[0022] S11: The positioning patch array includes a waist positioning patch, a wrist positioning patch, an ankle positioning patch, a knee positioning patch, an elbow positioning patch, and a neck positioning patch;

[0023] S12: The first time zone includes the first time domain to the Kth time domain, where K is a preset number of time domains, K≥15;

[0024] S13: collecting first time-domain waist positioning information, first time-domain wrist positioning information, first time-domain ankle positioning information, first time-domain knee positioning information, first time-domain elbow positioning information, and first time-domain neck positioning information of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch in the first time domain to construct first time-domain positioning information;

[0025] S14: until the Kth time-domain waist positioning information, Kth time-domain wrist positioning information, Kth time-domain ankle positioning information, Kth time-domain knee positioning information, Kth time-domain elbow positioning information, and Kth time-domain neck positioning information of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch of the Kth time-domain are collected to construct the Kth time-domain positioning information;

[0026] S15: construct the positioning timing matrix according to the first time domain positioning information to the Kth time domain positioning information, and further, the time domain step is 1 second.

[0027] In a preferred embodiment, the first elderly care user refers to any elderly care user served by the smart elderly care remote status monitoring device combined with user portrait provided in the embodiment of the present application, and the positioning patch array refers to a flexible patch attached to the user's body, which can remotely communicate with the smart elderly care remote status monitoring device combined with user portrait and upload the positioning data of the patch.

[0028] The first time zone refers to the time zone step for abnormal status analysis of elderly care users preset by the management end, preferably, at least 15 seconds. The positioning timing matrix refers to the matrix constructed by the positioning data of the positioning patch array patch in the first time zone. Preferably, the matrix form is described as follows, each row represents a positioning patch at a position, and each column represents the positioning data of the same time domain. Whenever the first time zone is met, the positioning timing matrix is ​​collected once, and the first time zone is reset to 0, and the timing starts again.

[0029] Furthermore, preferably, the positioning patch array includes a waist positioning patch, a wrist positioning patch, an ankle positioning patch, a knee positioning patch, an elbow positioning patch, and a neck positioning patch. Through the time series data of the waist, two wrists, two ankles, two knees, two elbows and neck, the user's motion state can be identified. The first time zone includes the first time domain to the Kth time domain, K is the preset number of time domains, K≥15, and the time domain step is 1 second. Furthermore, the administrator can also customize the number of K and the value of the time domain step.

[0030] When the first time zone starts timing again, the real-time positioning data of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch in the first time domain are collected and stored as the first time domain waist positioning information, the first time domain wrist positioning information, the first time domain ankle positioning information, the first time domain knee positioning information, the first time domain elbow positioning information, and the first time domain neck positioning information to construct the first time domain positioning information; until the real-time positioning data of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch in the Kth time domain are collected and stored as the Kth time domain waist positioning information, the Kth time domain wrist positioning information, the Kth time domain ankle positioning information, the Kth time domain knee positioning information, the Kth time domain elbow positioning information, and the Kth time domain neck positioning information to construct the Kth time domain positioning information.

[0031] Furthermore, the preferred arrangement of the positioning timing matrix is ​​as follows:

[0032] The first time-domain waist positioning information to the K-th time-domain waist positioning information is used as the first row of data of the matrix; the first time-domain wrist positioning information of the left and right hands to the K-th time-domain wrist positioning information is used as the second and third row of data of the matrix respectively; the first time-domain ankle positioning information of the left and right feet to the K-th time-domain ankle positioning information is used as the fourth and fifth row of data of the matrix respectively; the first time-domain knee positioning information of the left and right knees to the K-th time-domain knee positioning information is used as the fifth and sixth row of data of the matrix; the first time-domain elbow positioning information of the left and right elbows to the K-th time-domain elbow positioning information is used as the seventh and eighth row of data of the matrix respectively; the first time-domain neck positioning information to the K-th time-domain neck positioning information is used as the ninth row of data of the matrix, and the data in the same time domain is used as the same column data, to construct a positioning timing matrix.

[0033] The positioning patch can be used to monitor the movement status of elderly users in real time to ensure the accuracy of dynamic portraits.

[0034] S20: Processing the first elderly care user portrait and the positioning time series matrix through a healthy sample N-level neighbor searcher to obtain a neighbor sample state, 5≥N≥2, where N is an integer;

[0035] In a preferred embodiment, the healthy sample N-level adjacent searcher is a virtual function module for searching adjacent users for the first elderly user portrait and the positioning timing matrix, that is, the abnormal state indicator characteristic value of similar users. The adjacent sample state is the abnormal state indicator characteristic value of the adjacent user. N represents the adjacent level to be searched. The meaning of 5≥ is to avoid too large an adjacent level. On the one hand, it avoids excessive computing power leading to reduced processing efficiency. On the other hand, too large an adjacent level causes the sample to be far away from the first elderly user portrait and the positioning timing matrix, which reduces the accuracy of the adjacent sample state. The meaning of ≥2 is to increase the sample size and avoid falling into local search, which leads to the one-sidedness of the adjacent sample state.

[0036] Preferably, the detailed workflow of the healthy sample N-level adjacency searcher of the embodiment of the present application is as follows:

[0037] Further, the first elderly user portrait and the positioning time series matrix are processed by the healthy sample N-level adjacency searcher to obtain the adjacency sample state. Step S20 includes the following steps:

[0038] S21: The first elderly care user portrait includes at least the first elderly care user's age, the first elderly care user's gender, and the first elderly care user's historical illness type;

[0039] S22: performing a first-level adjacency search of the health samples according to the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user, and the positioning time series matrix through the health sample N-level adjacency searcher to obtain a first-level adjacency sample state, wherein the first-level adjacency sample state has a first-level adjacency sample user portrait and a first-level adjacency sample record positioning time series matrix;

[0040] S23: Performing a healthy sample secondary adjacency search according to the primary adjacency sample user portrait and the primary adjacency sample record positioning time series matrix through the healthy sample N-level adjacency searcher to obtain the secondary adjacency sample status;

[0041] S24: performing a healthy sample adjacency search through the healthy sample N-level adjacency until an N-level adjacency sample state is obtained;

[0042] S25: Add the first-level adjacent sample state, the second-level adjacent sample state, and finally the N-level adjacent sample state into the adjacent sample state.

[0043] In a preferred embodiment, the first elderly care user portrait represents a static portrait of the user, including at least age, gender, and historical disease type, and is stored as the first elderly care user age, the first elderly care user gender, and the first elderly care user historical disease type.

[0044] The state of the first-level adjacent sample refers to the state data of the adjacent sample collected by the healthy sample N-level adjacent searcher with the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user and the positioning timing matrix as the screening benchmark. Therefore, the state of the first-level adjacent sample has a first-level adjacent sample user portrait that characterizes the static portrait of the adjacent sample, and a first-level adjacent sample record positioning timing matrix that characterizes the dynamic portrait of the adjacent sample. The state of the second-level adjacent sample refers to the state data of the adjacent sample collected by the healthy sample N-level adjacent searcher with the first-level adjacent sample user portrait and the first-level adjacent sample record positioning timing matrix as the screening benchmark. Iterate the search until the N-level adjacent sample state is obtained, then add the first-level adjacent sample state, the second-level adjacent sample state until the N-level adjacent sample state into the adjacent sample state, and output the adjacent sample state.

[0045] Statistically collecting N-level adjacent samples can increase the amount of sample data to ensure the comprehensiveness of the sample data and avoid falling into local search, which leads to insufficient accuracy of the sample status.

[0046] Further, by the healthy sample N-level adjacency searcher, a first-level adjacency search of healthy samples is performed according to the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user and the positioning timing matrix to obtain a first-level adjacency sample state. Step S22 includes the steps of:

[0047] S221: Obtain the status of healthy samples to be sorted, wherein the status of healthy samples to be sorted includes a user portrait of the samples to be sorted and a time series matrix of record positioning of the samples to be sorted;

[0048] S222: performing multi-valued processing on the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix through a piecewise function to obtain first multi-valued reference information;

[0049] Further, the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user and the positioning time series matrix are multi-valued by a piecewise function to obtain first multi-valued reference information. Step S222 includes the following steps:

[0050] S2221: The piecewise function includes an age piecewise function, a gender piecewise function, a disease type piecewise function, and a positioning piecewise function, wherein the piecewise function is used to convert the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix into a characteristic numerical representation;

[0051] S2222: Perform multi-value processing on the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix according to the age piecewise function, the gender piecewise function, the disease type piecewise function, and the positioning piecewise function to obtain a first elderly care user age feature value, a first elderly care user gender feature value, a first elderly care user historical disease type feature value, and a positioning feature value time series matrix;

[0052] S2223: Add the first multi-valued reference information according to the first elderly care user's age characteristic value, the first elderly care user's gender characteristic value, the first elderly care user's historical disease type characteristic value, and the positioning characteristic value time series matrix.

[0053] S223: Performing multi-valued processing on the user portrait of the sample to be sorted and the record positioning time series matrix of the sample to be sorted by using the piecewise function to obtain multi-valued comparison information to be sorted;

[0054] S224: performing similarity evaluation on the first multi-valued reference information and the multi-valued comparison information to be sorted to obtain a similarity coefficient;

[0055] Further, similarity evaluation is performed on the first multi-valued reference information and the multi-valued comparison information to be sorted to obtain a similarity coefficient. Step S224 includes the following steps:

[0056] S2241: The first multi-valued reference information includes the first elderly user's age feature value, the first elderly user's gender feature value, the first elderly user's historical disease type feature value, and the positioning feature value time series matrix;

[0057] S2242: The multi-valued comparison information to be sorted includes the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the time series matrix of the location feature value to be sorted;

[0058] S2243: Construct similarity evaluation function:

[0059]

[0060] Among them, SIM1 represents the user portrait similarity, SIM2 represents the positioning similarity, and w i Characterizes the weight of the i-th dimension portrait index, x i1 Characterizes the feature value to be sorted of the i-th dimension image index, x i0 represents the benchmark characteristic value of the i-th dimension image index, a and b represent small constants, y jz1 Characterize the feature value of the jth positioning patch in the zth time domain to be sorted, y jz0Characterize the z-th time domain benchmark feature value of the j-th positioning patch, where K is the preset time domain number;

[0061] S2244: According to the similarity evaluation function, the age feature value of the first elderly care user, the gender feature value of the first elderly care user, the historical disease type feature value of the first elderly care user, and the positioning feature value time series matrix are processed with the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the positioning feature value time series matrix to be sorted to obtain the user portrait similarity coefficient and the positioning similarity coefficient, and add them into the similarity coefficient.

[0062] S225: When the similarity coefficient is greater than or equal to the similarity coefficient threshold, the state of the healthy sample to be sorted is added to the state of the first-level adjacent sample;

[0063] S226: When the first-level adjacent sample state includes 5 groups, output the first-level adjacent sample state.

[0064] When searching for adjacent samples, the main thing that needs to be clarified is the sorting rules of adjacent samples. The traditional way of selecting adjacent portraits is mainly through the comparison of different indicators. If the deviation is less than the corresponding threshold, it is considered adjacent. However, the application scenario of the embodiment of the present application is timeliness monitoring, and the refined comparison calculation of each indicator involves the comparison of type indicators, the comparison of numerical indicators, and the calculation of fine numerical values, resulting in high computational complexity and fine calculation granularity, making the comparison process more redundant, so the traditional way of selecting adjacent portraits cannot be effectively applied to the adjacent sample sorting of the embodiment of the present application.

[0065] Preferably, the rule flow of adjacent sample sorting in the embodiment of the present application is described by taking the sorting rule of the first-level adjacent sample state as an example:

[0066] In a preferred embodiment, the status of the healthy sample to be sorted is the status data of a healthy sample randomly extracted by the N-level adjacency searcher of the healthy sample from a website or database preset by the administrator. The healthy sample includes a user portrait of the sample to be sorted representing the static portrait of the healthy sample, and a record positioning timing matrix of the sample to be sorted representing the dynamic portrait of the healthy sample.

[0067] The piecewise function can convert the data of the first elderly care user's age, the first elderly care user's gender, the first elderly care user's historical disease type, and the positioning time series matrix into a certain value. Preferably, for example, the age between 50 and 55 years old is a value 8, the age between 56 and 60 years old is a value 9, the male gender is a value 1, and the female gender is a value 2. Different disease types also have corresponding numerical representations, and the positioning coordinates within a certain area can also be represented by corresponding numerical values. In this way, fine-grained calculation comparisons and multi-modal comparisons can be avoided. Only numerical comparisons need to be performed, which improves calculation efficiency.

[0068] The first multi-valued benchmark information refers to the result obtained by multi-valued processing of the first elderly care user's age, the first elderly care user's gender, the first elderly care user's historical disease type and the positioning time series matrix through a piecewise function, including the first elderly care user's age feature value, the first elderly care user's gender feature value, the first elderly care user's historical disease type feature value, and the positioning feature value time series matrix.

[0069] The detailed determination process is as follows:

[0070] The segmentation function includes an age segmentation function, a gender segmentation function, a disease type segmentation function, and a positioning segmentation function. The age segmentation function, the gender segmentation function, the disease type segmentation function, and the positioning segmentation function can be used to blur the type or specific values ​​distributed in a certain interval into feature values ​​to simplify the subsequent calculation process. According to the age segmentation function, the gender segmentation function, the disease type segmentation function, and the positioning segmentation function, the first elderly care user age, the first elderly care user gender, the first elderly care user historical disease type, and the positioning time series matrix are multi-valued to obtain the first elderly care user age feature value, the first elderly care user gender feature value, the first elderly care user historical disease type feature value, and the positioning feature value time series matrix. According to the first elderly care user age feature value, the first elderly care user gender feature value, the first elderly care user historical disease type feature value, and the positioning feature value time series matrix, the first multi-valued reference information is added.

[0071] The multi-valued comparison information to be sorted refers to the result obtained by multi-valued processing of the user portrait of the sample to be sorted and the record positioning time series matrix of the sample to be sorted through the piecewise function, including the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the positioning feature value time series matrix to be sorted. The acquisition process is the same as the first multi-valued benchmark information generation process.

[0072] Then follow the preset similarity evaluation function: Calculate the similarity coefficients between the multi-valued comparison information to be sorted and the first multi-valued reference information, including the user portrait similarity coefficient and the positioning similarity coefficient.

[0073] Furthermore, the similarity coefficient threshold includes a user portrait similarity coefficient threshold and a positioning similarity coefficient threshold. When the user portrait similarity coefficient is greater than or equal to the user portrait similarity coefficient threshold, and the positioning similarity coefficient is greater than or equal to the positioning similarity coefficient threshold, the healthy sample state to be sorted is added to the first-level adjacent sample state. The first-level adjacent sample state that meets the requirements is searched with the same rule to include at least 5 groups, and the first-level adjacent sample state is output for the next step of searching.

[0074] The adjacent sample sorting rules provided in the embodiments of the present application can avoid fine-grained calculation comparisons and multi-modal comparisons, and only numerical comparisons need to be performed, thereby improving calculation efficiency.

[0075] S30: Aggregate the adjacent sample states to generate state baseline information;

[0076] Further, such as Figure 2 As shown, the adjacent sample states are aggregated to generate state baseline information, and step S30 includes the following steps:

[0077] S31: Obtaining a t-th dimension state feature value set according to the adjacent sample states;

[0078] S32: Perform LOF algorithm outlier detection on the t-th dimension state feature value set to obtain a t-th dimension state non-outlier feature value set;

[0079] S33: Counting the distribution interval of the set of non-outlier characteristic values ​​of the t-th dimension state to construct the t-th dimension state baseline information;

[0080] S34: Add the t-th dimension state baseline information into the state baseline information.

[0081] In a preferred embodiment, the process of adjoining sample states is as follows:

[0082] The LOF algorithm can be used to detect outliers. Preferably, in the embodiment of the present application, the outlier factor threshold is usually set to 3. If the outlier factor score of a data point is greater than 3, the data point is considered to be an outlier. If it is less than or equal to 3, the data point is considered to be a non-outlier. The t-dimensional state feature value set refers to the adjacent sample state feature value set of the physiological state indicator of any dimension. The t-dimensional state feature value set is subjected to LOF algorithm outlier detection to obtain the t-dimensional state non-outlier feature value set; the distribution interval of the t-dimensional state non-outlier feature value set is counted to construct the t-dimensional state baseline information. The state baseline information of all dimensions is analyzed, and the state baseline information is output as a judgment criterion for subsequent evaluation of abnormal states.

[0083] S40: a communication status monitor collects status monitoring information of the first time zone;

[0084] S50: When the state monitoring information does not belong to the state baseline information, an abnormal state mark is performed on the state monitoring information.

[0085] In a preferred embodiment, the status monitor may be various medical devices worn on the user to monitor physiological indicators, and collects and stores the status monitoring information in the first time zone.

[0086] Furthermore, the status monitoring information includes the first time domain monitoring information to the Kth time domain monitoring information. If the status indicator of any dimension of each time domain does not belong to the status baseline information, the corresponding time domain is deemed to be an abnormal time domain. The first time domain monitoring information is traversed until the Kth time domain monitoring information is compared with the status baseline information, and the proportion of abnormal time domains is counted. Preferably, if the proportion of abnormal time domains is greater than the threshold of the proportion of abnormal time domains, 15 of 15%, the status monitoring information is deemed not to belong to the status baseline information, and the status monitoring information is further marked as abnormal. Preferably, the threshold of the proportion of abnormal time domains can be customized and corrected by the management personnel.

[0087] The smart elderly care remote status monitoring method combined with user portrait provided by the embodiment of the present invention has at least the following technical effects:

[0088] 1. Through the user's positioning patch array, the user's positioning timing matrix in the first time zone can be monitored in real time to characterize the movement state of the elderly user; then the movement state is used as a dynamic portrait, and the first elderly user portrait is used as a static portrait for adjacent sampling to obtain the adjacent sample state; further, the adjacent sample state is data fused to obtain the user's state baseline data under the portrait state constraint; if the state monitoring information does not belong to the state baseline information, the state monitoring information is marked as abnormal. By integrating dynamic portraits and static portraits, the state baseline of the elderly user is automatically evaluated with the help of data fusion algorithms, the reliability of the state baseline is improved, and the technical effect of ensuring the accuracy of abnormal monitoring is achieved.

[0089] 2. The adjacent sample sorting rules provided in the embodiments of the present application can avoid fine-grained calculation comparisons and multi-modal comparisons, and only numerical comparisons need to be performed, thereby improving calculation efficiency.

[0090] Embodiment 2:

[0091] like Figure 3 As shown, based on the same inventive concept of the smart elderly care remote status monitoring method combined with user portrait provided in Example 1, the embodiment of the present invention also provides a smart elderly care remote status monitoring device combined with user portrait, including:

[0092] A positioning timing matrix receiving module, used for communicating with the positioning patch array of the first elderly user and receiving the positioning timing matrix of the first time zone;

[0093] An adjacent sample state search module, used to process the first elderly care user portrait and the positioning timing matrix through a healthy sample N-level adjacent searcher to obtain an adjacent sample state, 5≥N≥2, where N is an integer;

[0094] A status baseline information configuration module, used to aggregate the adjacent sample states to generate status baseline information;

[0095] A status monitoring module, used for a communication status monitor, to collect status monitoring information of the first time zone;

[0096] The abnormal state identification module is used to identify the abnormal state of the state monitoring information when the state monitoring information does not belong to the state baseline information.

[0097] Furthermore, the positioning timing matrix receiving module executes the steps including:

[0098] The positioning patch array includes a waist positioning patch, a wrist positioning patch, an ankle positioning patch, a knee positioning patch, an elbow positioning patch, and a neck positioning patch;

[0099] The first time zone includes the first time domain to the Kth time domain, K is a preset number of time domains, K≥15;

[0100] Collecting first time-domain waist positioning information, first time-domain wrist positioning information, first time-domain ankle positioning information, first time-domain knee positioning information, first time-domain elbow positioning information, and first time-domain neck positioning information of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch in the first time domain to construct first time-domain positioning information;

[0101] Until the Kth time-domain waist positioning information, the Kth time-domain wrist positioning information, the Kth time-domain ankle positioning information, the Kth time-domain knee positioning information, the Kth time-domain elbow positioning information, and the Kth time-domain neck positioning information of the waist positioning patch, the wrist positioning patch, the ankle positioning patch, the knee positioning patch, the elbow positioning patch, and the neck positioning patch of the Kth time-domain are collected to construct the Kth time-domain positioning information;

[0102] The positioning timing matrix is ​​constructed according to the first time domain positioning information to the Kth time domain positioning information, and further, the time domain step length is 1 second.

[0103] Furthermore, the adjacent sample state search module executes the steps including:

[0104] The first elderly care user portrait at least includes the first elderly care user's age, the first elderly care user's gender, and the first elderly care user's historical illness type;

[0105] By means of the healthy sample N-level adjacency searcher, a first-level adjacency search of healthy samples is performed according to the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user, and the positioning time series matrix, to obtain a first-level adjacency sample state, wherein the first-level adjacency sample state has a first-level adjacency sample user portrait and a first-level adjacency sample record positioning time series matrix;

[0106] Through the healthy sample N-level adjacency searcher, a healthy sample secondary adjacency search is performed according to the primary adjacency sample user portrait and the primary adjacency sample record positioning timing matrix to obtain the secondary adjacency sample status;

[0107] Performing a healthy sample adjacency search through the healthy sample N-level adjacency until an N-level adjacency sample state is obtained;

[0108] The first-level adjacent sample state, the second-level adjacent sample state, and finally the N-level adjacent sample state are added into the adjacent sample state.

[0109] Furthermore, the adjacent sample state search module executes the steps including:

[0110] Obtaining the status of healthy samples to be sorted, wherein the status of healthy samples to be sorted includes a user portrait of the samples to be sorted and a time series matrix of record positioning of the samples to be sorted;

[0111] Performing multi-valued processing on the age of the first elderly care user, the gender of the first elderly care user, the historical illness type of the first elderly care user, and the positioning time series matrix through a piecewise function to obtain first multi-valued reference information;

[0112] Performing multi-valued processing on the user portrait of the sample to be sorted and the record positioning time series matrix of the sample to be sorted by the piecewise function to obtain multi-valued comparison information to be sorted;

[0113] Performing similarity evaluation on the first multi-valued reference information and the multi-valued comparison information to be sorted to obtain a similarity coefficient;

[0114] When the similarity coefficient is greater than or equal to the similarity coefficient threshold, the state of the healthy sample to be sorted is added to the state of the first-level adjacent sample;

[0115] When the first-level adjacent sample state includes 5 groups, the first-level adjacent sample state is output.

[0116] Furthermore, the adjacent sample state search module executes the steps including:

[0117] The piecewise function includes an age piecewise function, a gender piecewise function, a disease type piecewise function, and a positioning piecewise function, wherein the piecewise function is used to convert the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix into a characteristic numerical representation;

[0118] According to the age segmentation function, the gender segmentation function, the disease type segmentation function, and the positioning segmentation function, the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix are multi-valued to obtain a first elderly care user age feature value, a first elderly care user gender feature value, a first elderly care user historical disease type feature value, and a positioning feature value time series matrix;

[0119] The first multi-valued reference information is added according to the first elderly care user's age characteristic value, the first elderly care user's gender characteristic value, the first elderly care user's historical disease type characteristic value, and the positioning characteristic value time series matrix.

[0120] Furthermore, the adjacent sample state search module executes the steps including:

[0121] The first multi-valued reference information includes the first elderly user's age feature value, the first elderly user's gender feature value, the first elderly user's historical disease type feature value, and the positioning feature value time series matrix;

[0122] The multi-valued comparison information to be sorted includes the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the time series matrix of the location feature value to be sorted;

[0123] Construct a similarity evaluation function:

[0124]

[0125] Among them, SIM1 represents the user portrait similarity, SIM2 represents the positioning similarity, and w i Characterizes the weight of the i-th dimension portrait index, x i1 Characterizes the feature value to be sorted of the i-th dimension image index, x i0 represents the benchmark characteristic value of the i-th dimension image index, a and b represent small constants, y jz1 Characterize the feature value of the jth positioning patch in the zth time domain to be sorted, y jz0 Characterize the z-th time domain benchmark feature value of the j-th positioning patch, where K is the preset time domain number;

[0126] According to the similarity evaluation function, the age feature value of the first elderly care user, the gender feature value of the first elderly care user, the historical disease type feature value of the first elderly care user, and the positioning feature value time series matrix are processed, and the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the positioning feature value time series matrix to be sorted are obtained to obtain the user portrait similarity coefficient and the positioning similarity coefficient, which are added to the similarity coefficient.

[0127] Furthermore, the state baseline information configuration module executes the steps including:

[0128] According to the adjacent sample states, obtaining a t-th dimension state feature value set;

[0129] Performing LOF algorithm outlier detection on the t-th dimension state feature value set to obtain a t-th dimension state non-outlier feature value set;

[0130] Counting the distribution interval of the set of non-outlier characteristic values ​​of the t-th dimension state to construct the t-th dimension state baseline information;

[0131] The t-th dimension state baseline information is added to the state baseline information.

[0132] Embodiment three:

[0133] See also Figure 4 , Figure 4 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4 As shown, this embodiment provides a computer-readable storage medium 600, on which a computer program 611 is stored. When the computer program 611 is executed by a processor, the following steps are implemented:

[0134] The positioning patch array of the first elderly care user in communication receives the positioning timing matrix of the first time zone;

[0135] Process the first elderly care user portrait and the positioning time series matrix through a healthy sample N-level adjacency searcher to obtain an adjacency sample state, 5≥N≥2, where N is an integer;

[0136] Aggregating the adjacent sample states to generate state baseline information;

[0137] A communication status monitor for collecting status monitoring information of the first time zone;

[0138] When the state monitoring information does not belong to the state baseline information, an abnormal state mark is performed on the state monitoring information.

[0139] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0140] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0141] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0142] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0143] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0144] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0145] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A remote status monitoring method for smart elderly care combined with user portraits, characterized in that: include: The positioning patch array of the first elderly care user in communication receives the positioning timing matrix of the first time zone; Process the first elderly care user portrait and the positioning time series matrix through a healthy sample N-level adjacency searcher to obtain an adjacency sample state, 5≥N≥2, where N is an integer; Aggregating the adjacent sample states to generate state baseline information; A communication status monitor for collecting status monitoring information of the first time zone; When the state monitoring information does not belong to the state baseline information, an abnormal state mark is performed on the state monitoring information.

2. The method according to claim 1, characterized in that The positioning patch array of the first elderly care user in communication receives the positioning timing matrix of the first time zone, including: The positioning patch array includes a waist positioning patch, a wrist positioning patch, an ankle positioning patch, a knee positioning patch, an elbow positioning patch, and a neck positioning patch; The first time zone includes the first time domain to the Kth time domain, K is a preset number of time domains, K≥15; Collecting first time-domain waist positioning information, first time-domain wrist positioning information, first time-domain ankle positioning information, first time-domain knee positioning information, first time-domain elbow positioning information, and first time-domain neck positioning information of the waist positioning patch, wrist positioning patch, ankle positioning patch, knee positioning patch, elbow positioning patch, and neck positioning patch in the first time domain to construct first time-domain positioning information; Until the Kth time-domain waist positioning information, the Kth time-domain wrist positioning information, the Kth time-domain ankle positioning information, the Kth time-domain knee positioning information, the Kth time-domain elbow positioning information, and the Kth time-domain neck positioning information of the waist positioning patch, the wrist positioning patch, the ankle positioning patch, the knee positioning patch, the elbow positioning patch, and the neck positioning patch of the Kth time-domain are collected to construct the Kth time-domain positioning information; The positioning timing matrix is ​​constructed according to the first time domain positioning information up to the Kth time domain positioning information.

3. The method according to claim 2, characterized in that The time domain step size is 1 second.

4. The method according to claim 1, characterized in that The first elderly care user portrait and the positioning time series matrix are processed by the healthy sample N-level adjacency searcher to obtain the adjacency sample state, including: The first elderly care user portrait at least includes the first elderly care user's age, the first elderly care user's gender, and the first elderly care user's historical illness type; By means of the healthy sample N-level adjacency searcher, a first-level adjacency search of healthy samples is performed according to the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user, and the positioning time series matrix, to obtain a first-level adjacency sample state, wherein the first-level adjacency sample state has a first-level adjacency sample user portrait and a first-level adjacency sample record positioning time series matrix; Through the healthy sample N-level adjacency searcher, a healthy sample secondary adjacency search is performed according to the primary adjacency sample user portrait and the primary adjacency sample record positioning timing matrix to obtain the secondary adjacency sample status; Performing a healthy sample adjacency search through the healthy sample N-level adjacency until an N-level adjacency sample state is obtained; The first-level adjacent sample state, the second-level adjacent sample state, and finally the N-level adjacent sample state are added into the adjacent sample state.

5. The method according to claim 4, characterized in that By means of the health sample N-level adjacency searcher, a health sample first-level adjacency search is performed according to the age of the first elderly user, the gender of the first elderly user, the historical disease type of the first elderly user, and the positioning time series matrix to obtain a first-level adjacency sample state, including: Obtaining the status of healthy samples to be sorted, wherein the status of healthy samples to be sorted includes a user portrait of the samples to be sorted and a time series matrix of record positioning of the samples to be sorted; Performing multi-valued processing on the age of the first elderly care user, the gender of the first elderly care user, the historical illness type of the first elderly care user, and the positioning time series matrix through a piecewise function to obtain first multi-valued reference information; Performing multi-valued processing on the user portrait of the sample to be sorted and the record positioning time series matrix of the sample to be sorted by the piecewise function to obtain multi-valued comparison information to be sorted; Performing similarity evaluation on the first multi-valued reference information and the multi-valued comparison information to be sorted to obtain a similarity coefficient; When the similarity coefficient is greater than or equal to the similarity coefficient threshold, the state of the healthy sample to be sorted is added to the state of the first-level adjacent sample; When the first-level adjacent sample state includes 5 groups, the first-level adjacent sample state is output.

6. The method according to claim 5, characterized in that The age of the first elderly care user, the gender of the first elderly care user, the historical illness type of the first elderly care user, and the positioning time series matrix are multi-valued by using a piecewise function to obtain first multi-valued reference information, including: The piecewise function includes an age piecewise function, a gender piecewise function, a disease type piecewise function, and a positioning piecewise function, wherein the piecewise function is used to convert the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix into a characteristic numerical representation; According to the age segmentation function, the gender segmentation function, the disease type segmentation function, and the positioning segmentation function, the age of the first elderly care user, the gender of the first elderly care user, the historical disease type of the first elderly care user, and the positioning time series matrix are multi-valued to obtain a first elderly care user age feature value, a first elderly care user gender feature value, a first elderly care user historical disease type feature value, and a positioning feature value time series matrix; The first multi-valued reference information is added according to the first elderly care user's age characteristic value, the first elderly care user's gender characteristic value, the first elderly care user's historical disease type characteristic value, and the positioning characteristic value time series matrix.

7. The method according to claim 6, characterized in that Performing a similarity evaluation on the first multi-valued reference information and the multi-valued comparison information to be sorted to obtain a similarity coefficient includes: The first multi-valued reference information includes the first elderly user's age feature value, the first elderly user's gender feature value, the first elderly user's historical disease type feature value, and the positioning feature value time series matrix; The multi-valued comparison information to be sorted includes the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the time series matrix of the location feature value to be sorted; Construct a similarity evaluation function: Among them, SIM1 represents the user portrait similarity, SIM2 represents the positioning similarity, and w i Characterizes the weight of the i-th dimension portrait index, x i1 Characterizes the feature value to be sorted of the i-th dimension image index, x i0 represents the benchmark characteristic value of the i-th dimension image index, a and b represent small constants, y jz1 Characterize the feature value of the jth positioning patch in the zth time domain to be sorted, y jz0 Characterize the z-th time domain benchmark feature value of the j-th positioning patch, where K is the preset time domain number; According to the similarity evaluation function, the age feature value of the first elderly care user, the gender feature value of the first elderly care user, the historical disease type feature value of the first elderly care user, and the positioning feature value time series matrix are processed, and the age feature value to be sorted, the gender feature value to be sorted, the historical disease type feature value to be sorted, and the positioning feature value time series matrix to be sorted are obtained to obtain the user portrait similarity coefficient and the positioning similarity coefficient, which are added to the similarity coefficient.

8. The method according to claim 1, characterized in that Aggregating the adjacent sample states to generate state baseline information includes: According to the adjacent sample states, obtaining a t-th dimension state feature value set; Performing LOF algorithm outlier detection on the t-th dimension state feature value set to obtain a t-th dimension state non-outlier feature value set; Counting the distribution interval of the set of non-outlier characteristic values ​​of the t-th dimension state to construct the t-th dimension state baseline information; The t-th dimension state baseline information is added to the state baseline information.

9. A smart elderly care remote status monitoring device combined with user portraits is characterized by: include: A positioning timing matrix receiving module, used for communicating with the positioning patch array of the first elderly user and receiving the positioning timing matrix of the first time zone; An adjacent sample state search module, used to process the first elderly care user portrait and the positioning timing matrix through a healthy sample N-level adjacent searcher to obtain an adjacent sample state, 5≥N≥2, where N is an integer; A status baseline information configuration module, used to aggregate the adjacent sample states to generate status baseline information; A status monitoring module, used for a communication status monitor, to collect status monitoring information of the first time zone; The abnormal state identification module is used to identify the abnormal state of the state monitoring information when the state monitoring information does not belong to the state baseline information.

10. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, which, when executed by a processor, implements the smart elderly care remote status monitoring method combined with user portraits as described in any one of claims 1 to 8.

Images