Methods and devices for evaluating a person's need for support

Abstract

The invention relates to methods and devices for evaluating the need for support of a person assigned to a residential or accommodation unit (100). The proposed devices and methods are based on a non-interventionist approach according to which only usage data of the person in the residential or accommodation unit (100) are used. The usage data are examined for changes of the kind in which the person deviates from normal activities. Thus insidious behavioural changes typical of age-related diseases can be identified. The proposed methods and devices make it possible to identify whether a need for support is arising or is slowly developing for a person, specifically before the at least one person actually needs support. Thus proactive measures can be taken to counteract a need for support which is arising. This is of interest not only for family members but also for operators of care and assisted living facilities, who with the aid of the method can ensure higher safety for the residents in their facilities, and better care.

Description

[0001] Methods and devices for assessing a person's support needs

[0002] Description

[0003] Field of the invention

[0004] The invention relates to methods and devices for assessing the support needs of a person assigned to a residential or residential unit. The proposed methods are primarily based on an approach that is as non-invasive as possible, according to which consumption data of the person in the residential or residential unit is used to record and assess the person's behavior and in particular any support needs. The causes of the support need are often due to illnesses or diseases of the person, in particular those that are age-related or become more likely to occur with age (stroke, dementia, diabetes, etc.). However, they can also arise from other risks to the person. The need for support often manifests itself in the form of unexpected or conspicuous behavior on the part of the person.

[0005] Methods for collecting, transmitting, and storing consumption data are well known. This particularly includes consumption data describing the electricity, water, or gas consumption of a person in a residential unit. The analysis of temporally resolved consumption data using, for example, pattern recognition to derive various processes or activities, often referred to as "non-intrusive load monitoring," is also well known. These processes typically aim to assign consumption data to subordinate components and derive efficiency and savings measures from it.

[0006] The connection between changes in behavior and the onset of illness or the need for support in everyday life is also known. However, the recording and detection of these behaviors is often implemented with severe intrusions into privacy through the installation of cameras, additional sensors, or direct observation.

[0007] State of the art

[0008] The prior art contains a wide variety of methods and devices that use consumption data to assess whether a person requires assistance. For example, Swiss published patent application CH 719 116 A2 describes assistance methods and devices for determining an emergency situation within a residential unit. Chinese published patent application CN 114973596 A1 describes systems and methods for monitoring elderly people living alone using a combination of various consumption data sources (electricity, water, temperature, humidity, etc.). Similar devices and methods can also be found in published patent applications US 2022 / 207980 A1, WO 2015 / 124972 A1, and WO 2018 / 114035 A1, and patent applications EP 1 071 055 B1, EP 1 298 619 B1, and EP 2 159 770 B1.

[0009] What all of these methods and devices have in common is that a behavioral profile or behavioral reference is created based on a person's consumption data. The behavioral profile or behavioral reference is then used to assess the person's activities to determine whether they are normal or unexpected or abnormal. If an activity is identified as unexpected or abnormal, this is interpreted as meaning that the person needs support, e.g., because they are currently in an emergency situation or seriously ill (e.g., suffering from dementia). Using the known methods and devices, it is therefore possible to determine whether a person currently needs support, and to offer the necessary support to the person.

[0010] In many cases, however, this means that the person urgently needs medical care or is in a dangerous situation. Elderly people are often monitored using these procedures and devices, especially those who live alone or manage their daily lives independently. Once the aforementioned emergency situation occurs, or once it becomes clear that the person requires support and can no longer manage their daily life independently, placement in a residential care facility or the involvement of a nursing service is often unavoidable.

[0011] Task The task of the invention is to provide methods and devices that use consumption data and thus enable people, especially older people, to manage their everyday lives independently or at least as independently as possible for longer.

[0012] Solution

[0013] This problem is solved by the subject matter of the independent claims. Advantageous developments of the subject matter of the independent claims are characterized in the subclaims. The wording of all claims is hereby incorporated by reference into this description.

[0014] The use of the singular shall not exclude the plural, and this shall also apply in the reverse sense unless otherwise disclosed.

[0015] Individual method steps are described in more detail below. In a preferred variant of the invention, the steps are carried out in the order specified. However, the steps do not necessarily have to be carried out in the order specified, and the method to be described may also include additional, unmentioned steps.

[0016] To solve the problem, a method for assessing the support needs of at least one person assigned to a residential or accommodation unit is proposed, which comprises the following steps:

[0017] 1. Time-resolved consumption data are received by an evaluation device, wherein the time-resolved consumption data were recorded by at least one consumption meter and wherein the at least one consumption meter is assigned to the residential or accommodation unit.

[0018] 2. The time-resolved consumption data are analyzed by the evaluation device in which the evaluation device self-learningly identifies at least one typical pattern in the time-resolved consumption data with the help of pattern recognition.

[0019] Identifying a typical pattern in the temporally resolved consumption data can be done using pattern recognition methods. A well-known pattern recognition method is the so-called clustering method (also known as time-series clustering). In this method, time series segments are clustered into patterns based on identified characteristics. Other methods are described in: o Jan-Peter Seevers, Kristina Jurczyk, Henning Meschede, Jens Hesselbach, John W. Sutherland: "Automatic Detection of Manufacturing Equipment Cycles Using Time Series", J. Comput. Inf. Sci. Eng. Jun 2020, 20(3), https: / / asmedigitalcollection.asme.org / computingengineering / article-abstract / 20 / 3 / 031005 / 1074102 / Automatic-Detection-of-Manufacturing-Equipment; htps: / / doi.orq / 10.1115 / 1.4046208, or o JP. Seevers, J. Johst, T. Weiß, H. Meschede, J.Hesselbach: "Automatic Time Series Segmentation as the Basis for Unsupervised, Non-Intrusive Load Monitoring of Machine Tools", Procedia CIRP, Volume 81, 2019, Pages 695-700; https: / / www.sci-encedirect.com / science / article / pii / S2212827119304834; https: / / doi.Org / 10.1016 / j.procir.2019.03.178.

[0020] Assigning a pattern to a possibly unknown activity or event can be done using unsupervised learning methods. The evaluation device can thus be designed as a self-learning system that independently identifies patterns as new patterns and associated unknown activities or events.

[0021] If the database was already populated with information at the start of the process, the recognized typical patterns can be assigned to known activities. Patterns that do not correspond to any known activity can be assigned to new, unknown activities, which may be created automatically.

[0022] 3. If the evaluation device has self-learningly identified at least one typical pattern in the time-resolved consumption data, the at least one typical pattern is assigned to at least one household appliance.

[0023] 4. The detected use of at least one household appliance is used to infer the person’s activity.

[0024] 5. From a multitude of recognized activities, the person’s activity profiles are identified through self-learning.

[0025] 6. A behavioral reference for the person is determined through self-learning from the person’s activity profiles.

[0026] One way to determine the behavioral reference is to use a Bayesian network with a probabilistic hybrid knowledge base. The advantages of such a Bayesian network are manifold:

[0027] • Formulation of knowledge in probability intervals (even in the case of very vague expert statements such as “I am quite sure that A is true.”);

[0028] • Possibility of integrating diverse sensor data inputs, possibility of integrating additional knowledge and formulated expert rules (also retrospectively);

[0029] • The result statement of the network is linked to a probability (it does not issue an alarm: yes or no, but an alarm with a certain probability).

[0030] This brings advantages for alarm data management.

[0031] The advantage of a Bayesian network over a pure neural network is that it can integrate probabilistic expert opinions from physicians and nurses, thus minimizing the number of false alarms. The overall process is significantly more robust than a pure neural network.If the evaluation device has identified an activity profile of the at least one person in the time-resolved consumption data, the evaluation device can assign a similarity value to the activity profile using the behavioral reference, wherein the similarity value quantitatively indicates whether and to what extent the activity profile deviates from a normal activity of the at least one person, and wherein the evaluation device and the behavioral reference are additionally designed to quantitatively record a change in the manner in which the at least one person deviates from a normal activity using the similarity value and thus to assess whether the at least one activity indicates that the at least one person is developing a need for support before the at least one person actually requires support.

[0032] The evaluation device generates a message if it detects that the at least one person develops a need for support before the at least one person actually requires support; and

[0033] The message is issued by the evaluation device.

[0034] The proposed method thus makes it possible to identify whether at least one person is developing or slowly developing a need for support, even before the at least one person actually requires it. This enables a third party, for example, to proactively approach the at least one person to counteract the developing need for support, reduce it if necessary, or at least slow the further development of the need for support. This is of interest not only to relatives of the at least one person, but also to operators of social housing and assisted living facilities, who can use the method to ensure increased safety for the residents of their properties and better care.It is also possible that the message from the evaluation device is received by the at least one person themselves and that the at least one person themselves initiates appropriate measures to counteract the increasing need for support, to reduce it if necessary, or at least to slow down its further development.

[0035] Assessing whether the person or persons develop a need for support before they actually require it may involve comparing it with typical behavioral change patterns associated with illnesses. This often enables early diagnosis of these illnesses and the initiation of preventative measures.

[0036] The evaluation device can be installed directly on-site, i.e., for example, in the residential or accommodation unit or in its surroundings (e.g., in the basement of a house in which the residential unit is located, or in the basement of the residential unit if the residential unit itself is a house, for example). It can be part of the at least one consumption meter. However, it can also be located elsewhere. It can be designed as a separate computing unit or part of a computing unit, such as a server. It can also be implemented as part of a cloud service or using a distributed computing structure.

[0037] The at least one consumption meter can be a smart meter, which not only records consumption data but is also equipped with means to transmit the recorded consumption data, for example, via a Wi-Fi network and / or the internet, to a computing unit or an evaluation device. In particular, existing consumption meters can be used for the proposed method, meaning that the installation of additional sensors can be omitted in many cases. If the existing consumption meters do not have means for transmitting consumption data, these means can often be retrofitted.

[0038] The temporally resolved consumption data can be analyzed by the evaluation device in order to recognize at least one pattern in the temporally resolved consumption data. If at least one pattern in the temporally resolved consumption data has been recognized by the evaluation device, the at least one pattern is assigned by the evaluation device to at least one event that is independent of the at least one person or to at least one activity of the at least one person. The at least one event can be, for example, the operation of the compressor of a refrigerator or an automatically operated device in the residential unit, such as a WI_AN router that is only switched on at certain times and automatically goes into sleep mode from 1 a.m. to 6 a.m.In this way, patterns in the temporally resolved consumption data that play no or at least a minor role in the assessment of the support needs of at least one person can be filtered out and disregarded. This avoids errors in the assessment of support needs and thus improves its quality.

[0039] As an alternative to a corresponding configuration of the behavioral reference, a behavioral change reference can also be provided, wherein the evaluation device and the behavioral change reference are configured to use the similarity value to quantitatively record a change in the manner in which the at least one person deviates from a normal activity and thus to assess whether the at least one activity indicates that the at least one person develops a need for support before the at least one person actually requires support. In step 5, instead of using the behavioral reference, the similarity value of the at least one activity can be assessed by the evaluation device to determine whether the at least one person develops a need for support before the at least one person actually requires support.The behavior change reference can also be designed as part of the behavior reference. Changes in the way in which at least one person deviates from a normal activity can therefore, in principle, be quantitatively recorded and assessed using the same methods and procedures that can be used to quantitatively record and assess any deviation of at least one activity from at least one normal activity. While the variables examined and any variables derived from them, as well as the underlying research question, are different, they can be quantitatively examined, recorded, and assessed using appropriate adaptations to the evaluation device, the behavior reference, and / or the behavior change reference.

[0040] The at least one consumption meter can record time-resolved consumption data for electricity, water, and / or gas consumption, with the electricity, water, or gas consumption being assigned to the residential or residential unit. The time-resolved consumption data can therefore represent the electricity, water, and / or gas consumption assigned to the residential or residential unit. Such meters are usually already present in residential or residential units, so that existing sensors can be used. Furthermore, consumption data is often not considered personal data by at least one person or others. The sharing and storage of data for the purpose of evaluation by the evaluation device is therefore more easily accepted than the sharing and storage of data recorded by cameras or microphones.

[0041] In many cases, it is advantageous to consider not just electricity, water, or gas consumption, but a combination of these consumption data—that is, a combination of two or three of these consumption data types. Combining different consumption data such as electricity, water, and / or gas can often achieve a greater level of detail and differentiation in the detection of activities.

[0042] To improve the method, at least one sensor can be provided to monitor the presence of at least one person in the living or recreational unit in order to assign the consumption recorded by the at least one consumption meter to the at least one person. The at least one sensor for monitoring presence can then record presence data of the at least one person, which is suitable for assigning the time-resolved consumption data to the at least one person. This presence data can be transmitted from the at least one sensor to the evaluation device. The presence data can then be received by the evaluation device. The evaluation device can then use the presence data to detect the at least one activity. In this way, activities can be assigned more reliably to the at least one person. Furthermore, false alarms can be avoided.

[0043] The behavioral reference can be a generic or standard reference. In many cases, however, it is advantageous to adapt the reference to the at least one person or even to generate it using only data and information about the at least one person, i.e. to use and / or establish a personal reference. For this purpose, the behavioral reference can be updated, supplemented and / or created by the evaluation device using the time-resolved consumption data. This is done by converting the consumption data into patterns, household appliances, activities and ultimately activity profiles. In many cases, it is worthwhile to repeat this step as soon as new consumption data is available. Furthermore, it can be advantageous to store or save the recorded time-resolved consumption data in a database for later analysis and evaluation.

[0044] To update, supplement, and / or create the behavioral reference, the evaluation device can identify at least one typical pattern in the temporally resolved consumption data. Once the evaluation device has identified at least one typical pattern in the temporally resolved consumption data, it can assign the at least one typical pattern to at least one household appliance and thus to at least one typical or normal activity of the at least one person or to at least one event that is independent of the at least one person. The identification and assignment of typical patterns can be repeated, e.g., regularly and / or continuously.

[0045] The evaluation device can store the at least one typical pattern and / or the at least one typical or normal activity of the at least one person and / or the at least one event that is independent of the at least one person in a database. They can be stored in the database together with a timestamp to enable them to be ordered chronologically. The database can be empty at the beginning of the method and can be enriched step by step with typical patterns, activities and events by repeating the identification and assignment steps. The patterns, activities and events stored in the database can be used by the evaluation device to update, supplement or create the behavior reference. The behavior reference can thus be updated, supplemented and / or created using previous activities of the at least one person.

[0046] Unknown patterns (patterns not stored in the database) can be created as new patterns in the pattern database after a suitable period of time (e.g. 24 hours) and used as additional reference patterns for further evaluations.

[0047] The behavioral reference can also be updated, supplemented, and / or created using similarity values ​​from previous activity profiles of the at least one person. In this way, slowly developing deviations or gradual changes in the way the at least one person deviates from normal activity can be detected using personal data, and the behavioral reference can be adapted to the at least one person accordingly.

[0048] The at least one activity can be linked to at least one attribute, whereby the at least one attribute is used to calculate the similarity value and / or the at least one attribute is used to assess whether the at least one person is slowly developing a need for support. Using the at least one attribute can often achieve a higher level of detail and differentiation in the recognition of activities.

[0049] The at least one attribute can also be stored in the aforementioned database as a reference. Furthermore, all or a majority of the previously detected activities in the database can also be assigned such attributes. The attributes of previously detected activities can be compared with the at least one attribute to detect deviations. Deviations from the expected values ​​are automatically interpreted.

[0050] If a deviation in the attribute values ​​is considered critical, a warning or alarm is issued in this message. The attribute limits for an alarm or warning can be absolute or relative. An example of a critical attribute value deviation would be a shower lasting more than 60 minutes. If only a minor attribute value deviation has occurred, the attribute values ​​are compared with historical values. This check determines whether the deviation is a one-off, gradual, or permanent. A gradual or permanent deviation indicates a change in behavior. This change in behavior is compared with typical changes in behavior that could indicate certain clinical patterns. A gradual or permanent deviation does not trigger a direct alarm, but rather a message that at least describes the changed behavior if it could not be linked to any clinical patterns.If one or more symptoms are detected, they are also included in the message. The message or alarm is then transferred via suitable interfaces, for example, to a third party, and thus leaves the evaluation device. The system can then continue reading further measurement data. The third party is then responsible for interpreting the issued alarms, warnings, messages, and diagnoses.

[0051] In addition, at least one attribute can be used to update the database. The behavior reference can be updated and / or supplemented in this way.

[0052] The at least one attribute can describe a device assignment, a duration, a temporal occurrence, a frequency, and / or an intensity. The intensity can, for example, be a temporal integral of the measured consumption data.

[0053] The steps of one of the proposed methods can be performed repeatedly and / or in real time. This allows at least one person to be monitored as non-invasively as possible.

[0054] The similarity value of the at least one activity can be evaluated by the evaluation device using the behavioral reference to determine whether the at least one person requires support. The evaluation device generates a message if it determines that the at least one person requires support and outputs the message. Thus, the method can also be used to identify a need for support that does not require delay, but rather immediate action or intervention, for example, by a third party.

[0055] The temporally resolved consumption data can have a temporal resolution in the range from 0.2 s to 15 min, preferably in the range from 0.5 s to 10 min, particularly preferably in the range from 1 s to 1 min. The at least one consumption meter can be designed such that it measures consumption data with a temporal resolution in the ranges mentioned. It can also be designed such that it transmits consumption data with a temporal resolution in the ranges mentioned, even though it measures consumption data with a higher resolution. Furthermore, it is possible for the evaluation device to process consumption data transmitted by the at least one consumption meter in such a way that its temporal resolution lies in the ranges mentioned, even though consumption data is transmitted with a higher resolution.With the aforementioned limitation of the temporal resolution of the consumption data, the privacy of at least one person can be better protected and thus, among other things, the acceptance of the procedure can be further improved.

[0056] The at least one activity profile can include multiple activities. In many cases, the support need can be assessed better and more reliably based on multiple activities than with just one activity.

[0057] The at least one activity profile can also correspond to the activities of the at least one person on a day, in a week, or in a month. Using the plurality of activities, for example, an activity profile of the at least one person over the day, week, or month, i.e., a daily, weekly, or monthly activity profile, can be created and analyzed for behavioral changes. In particular, the behavioral reference can include a plurality of such daily, weekly, or monthly activity profiles or have been created with them in order to detect behavioral changes based on a daily, weekly, or monthly activity profile.

[0058] The object is also achieved by an evaluation device for assessing the support needs of at least one person, wherein the evaluation device is equipped with means suitable for carrying out the steps of one of the proposed methods.

[0059] In addition, a system with an evaluation device as just proposed and at least one consumption meter equipped with means to record time-resolved consumption data and transmit them to the evaluation device solves the problem.

[0060] A solution to the problem is also provided by a computer program comprising instructions that cause one of the evaluation devices just described or one of the systems just described to carry out one of the proposed methods.

[0061] A computer-readable medium on which such a computer program is stored also solves the problem. Further details and features emerge from the following description of preferred embodiments in conjunction with the figures. The respective features can be implemented individually or in combination with one another. The possibilities for solving the problem are not limited to the embodiments.

[0062] The exemplary embodiments are illustrated schematically in the figures. Identical reference numerals in the individual figures denote identical or functionally equivalent elements, or elements that correspond to one another in terms of their functions. In detail:

[0063] Fig. 1 shows a section of a residential unit;

[0064] Fig. 2 shows a system with an evaluation device;

[0065] Fig. 3 Time-resolved consumption data; and

[0066] Fig. 4 shows a flow chart of a method according to the invention

[0067] Fig. 1 shows a section of a residential unit 100 that is inhabited by a person and is thus assigned to the person. The section of the residential unit 100 comprises a kitchen 110 with a dining area 120, a bathroom 130, and a bedroom 140. In the kitchen 110 with the dining area 120, there is a stove 115. The stove 115 has a specific power consumption that indicates an activity 150 of the person, namely the activity of cooking 150. The bathroom 130 has a bathtub 135. Filling the bathtub 135 is reflected, among other things, in the water consumption and indicates another activity 160 of the person, namely the activity of bathing 160. The bedroom 140 is used by the person for rest breaks and for sleeping 170. The activity Sleeping 170 is characterized mainly by low consumption of electricity and water, i.e. mainly by inactivity of the person in the housing unit 100.The activities cooking 150, bathing 160 and sleeping 170 are therefore each characterized by characteristic values ​​in the electricity and water consumption of the person in the living unit 100, ie by characteristic courses of the corresponding time-resolved consumption data.

[0068] In addition to the appliances and consumption points mentioned above, residential unit 100 also has other appliances and consumption points. These include various lighting fixtures, a refrigerator, a kettle, a television, etc.

[0069] Fig. 2 shows the residential unit 100 and a system 200. The system 200 includes, in particular, an electricity consumption meter 210 and a water consumption meter 220. The electricity consumption meter 210 and the water consumption meter 220 are located in a basement 230. The basement 230 belongs to the residential unit 100. The electricity consumption meter 210 and the water consumption meter 220 are assigned to the residential unit 100.

[0070] The electricity and water consumption in residential unit 100 is measured using electricity consumption meter 210 and water consumption meter 220. Consumption meters 210 and 220 thus serve to record the electricity and water consumption of the person in residential unit 100. The recorded electricity and water consumption is billed to the person by an electricity and water supplier at regular intervals.

[0071] The consumption meters 210 and 220 are designed as smart meters and transmit time-resolved consumption data to the electricity and water supplier via an at least partially wireless network.

[0072] In addition, the consumption meters 210 and 220 are connected to an evaluation device 240 of the system 200. The connection is also established via an at least partially wireless network comprising connections 250, 251, and 252. Connections 250 and 251 are combined in connection 252.

[0073] Using connections 250 and 252, the electricity consumption meter 210 can transmit temporally resolved electricity consumption data to the evaluation device 240. The water consumption meter 220 transmits temporally resolved water consumption data to the evaluation device 240 via connections 251 and 252.

[0074] Fig. 3 shows temporally resolved electricity consumption data 300 of the person in the residential unit 100, which was transmitted from the electricity consumption meter 210 to the evaluation device 240. The temporally resolved electricity consumption data 200 shown in Fig. 3 extend over a period of two hours, i.e., over 7200 s. During this period, electricity consumption varies significantly. The variations can be attributed to various events that are independent of the person, as well as to activities of the person themselves. The temporally resolved electricity consumption data 300 have a temporal resolution of 5 seconds, i.e., the course of the temporally resolved electricity consumption data has one data point every 5 seconds, or one data point per 5-second interval.

[0075] At the beginning of the curve shown, no or at least almost no electricity is consumed. After approximately 800 s, electricity consumption increases significantly. After approximately 1600 s, it drops back to the previous level. The electricity consumption from approximately 800 s to approximately 2400 s can be attributed to the operation of the refrigerator. This is indicated by the corresponding refrigerator symbol in Fig. 3. The operation of the refrigerator is automatic and is therefore an event that is independent of the person, more precisely independent of the presence and / or activity of the person in the living unit 100. From approximately 3650 s to 3900 s, electricity consumption increases significantly. At this time, the person in the living unit 100 operates the kettle. The operation of the kettle requires an action by the person and can therefore be attributed to a person's activity. This is indicated in Fig. 3 by a kettle symbol.The activity of boiling the kettle can be assigned an attribute, for example, the duration attribute. In this case, boiling the kettle took approximately 350 seconds. The duration attribute is therefore assigned the value 350, which represents a duration of 350 seconds.

[0076] The refrigerator is then operated twice more. The person also turns on the television at approximately 4100 seconds. The television is turned off again after approximately 3000 seconds. During this time, the person was apparently watching television, which is indicated by a television symbol in Fig. 3.

[0077] The temporally resolved power consumption shown in Fig. 3 can therefore be broken down into individual events and activities. Activities, in particular, enable monitoring of the person. While not continuous, they nevertheless indicate a possible need for support, particularly dangerous situations and conspicuous activities. The latter are evident, for example, in the continuous operation of the stove 115 or the television.

[0078] Fig. 4 shows a flowchart of an embodiment of a method 400 according to the invention for assessing the support needs of at least one person assigned to a residential or accommodation unit. The method 400 is used in particular to monitor the person in the residential unit 100, using the system 200 and time-resolved consumption data from the consumption meters 210 and 220, as shown, for example, in Fig. 2.

[0079] The method 400 begins with step 410. In step 410, the system 200 is initialized. In particular, a database is created that includes patterns of typical activities and corresponding assignments or links of the patterns to the typical activities, as well as a behavior reference.

[0080] After initialization of the system 200 in step 410, the consumption meters 210 and 220 record time-resolved consumption data of the electricity and water consumption in the residential unit 100 in step 420.

[0081] The time-resolved consumption data are transmitted in step 430 via the connections 250, 251 and 252 to the evaluation device 240 and stored there in the database created in step 410. In addition, the transmitted time-resolved consumption data are examined in step 430 for patterns, household appliances and activities. If an unknown pattern or an unknown activity is discovered that is not yet stored in the database, the unknown pattern or the unknown activity is added to the database after a period of 24 hours as a new pattern or as a new activity, together with the corresponding link or assignment of the new pattern to the new activity or to a new household appliance. The unknown pattern or the unknown activity is then used as an additional reference pattern or additional reference activity for further evaluations. The recognition of a known pattern orA known activity in step 430 also causes the database to be updated, i.e., the known pattern or activity is also added to the database in order to adapt the pattern or activity recognition and the database step by step to the individual. Patterns in the time series of temporally resolved consumption data are recognized using conventional time series analyses according to the state of the art. Known patterns are assigned to known household appliances and activities by comparing the patterns stored in the database with the assignments of patterns to household appliances and activities stored in the database.

[0082] In step 440, a decision is made as to how method 400 continues. If no activity was detected in the consumption data transmitted in step 430, the process continues with reading in additional consumption data, ie, method 400 continues with step 420. If activity was detected, method 400 continues with step 450.

[0083] In step 450, attributes and corresponding attribute values ​​are assigned to the activity detected in step 430, which characterize the detected activity as accurately as possible. The attributes describe the activity based on, for example, its duration, its temporal occurrence, or its temporal integral of the measured consumption data. Furthermore, the activity attributes are added with a timestamp to the database, in which all previously detected patterns, activities, attributes, and attribute values ​​are also stored.

[0084] In step 460, the behavior reference is used to determine whether the attribute values ​​of the activity detected in step 430 were expected or not. To do this, expected values ​​are formed using the detected pattern or activity, and the expected values ​​are compared with the attribute values ​​determined in step 450.

[0085] In step 470, any existing deviations of the attribute values ​​determined in step 450 from the expected attribute values ​​determined in step 460 are evaluated using the behavior reference. If a deviation in the attribute values ​​is considered critical, a warning or alarm is generated. The attribute thresholds for generating the alarm or warning can be absolute or relative. An example of a critical attribute value deviation would be a shower duration of more than 60 minutes.

[0086] If the assessment of any existing deviations between the attribute values ​​determined in step 450 and the expected attribute values ​​determined in step 460 results in the attribute values ​​not deviating from one another or deviating only slightly, the attribute values ​​determined in step 450 are further examined to determine whether any existing deviations are only a one-off, gradual, or permanent. Gradual or permanent deviations indicate a change in behavior and / or whether the at least one person develops a need for support before the at least one person actually requires support. This change in behavior or this developing or increasing need for support is compared with typical changes in behavior that may indicate certain clinical pictures. A gradual or permanent deviation does not trigger a direct alarm or warning in this case.However, it does result in a report being generated that at least describes the altered behavior if it could not be linked to clinical symptoms. If one or more clinical symptoms could be identified, they will also be included in the report.

[0087] In addition, in step 470, the behavior reference is supplemented or updated with any deviations detected.

[0088] In step 480, a decision is made as to how method 400 continues. If no warning, alarm, or message was generated in step 470, the process continues with reading in additional consumption data, ie, method 400 continues with step 420. However, if a warning, alarm, or message was generated, method 400 continues with step 490.

[0089] In step 490, the message, warning, or alarm is output via suitable interfaces, for example, to third parties. Method 400 can then be aborted. Alternatively, method 400 can also continue with step 420. The interpretation of the output alarm, warning, or message is then the responsibility of the third parties, for example.

[0090] glossary

[0091] activity

[0092] An activity is a specific action performed by a person, such as cooking, showering, watching television, or resting or sleeping. Activities particularly include activities of daily living, or ADL for short. In many cases, an activity refers to an entire class of activities, since describing an activity as cooking, showering, or watching television does not yet contain any information about the time, duration, or other details of the activity. The term activity therefore encompasses a class or series of activities that can generally be very different from one another, but still have one or more common characteristics. The common characteristics can be described using attributes, and corresponding values ​​can be assigned to the attributes. These include, for example, the duration, frequency, or intensity of an activity.

[0093] household appliance

[0094] The term "household appliances" encompasses all devices that generate consumption (electricity, water, etc.) that can be recorded by a utility meter. Typical household appliances in this sense include refrigerators, kettles, televisions, stoves, hotplates, bathtubs, showers, coffee makers, lighting, toilet flushes, faucets, etc.

[0095] Pattern

[0096] A pattern is a structure characterized by repeated or parallel, uniform occurrence—that is, by uniform repetition (reproduction). Patterns can reflect design or behavior patterns or corresponding sequences of actions. In temporally resolved consumption data, patterns can indicate activities, actions, or events that repeat or occur repeatedly.

[0097] Need for support

[0098] A person's need for support is defined as the extent to which they require assistance. Individuals may have little or no need for support. In these cases, individuals can generally manage their daily lives independently without assistance from third parties or outside sources. A high need for support is primarily characterized by a person no longer being able to manage their daily lives independently or being dependent on help or support, i.e., requiring support. Consumption data, time-resolved

[0099] Consumption data is usually recorded by consumption meters to record the consumption of, for example, electricity, water or gas in a residential unit. In many cases, the actual consumption is recorded in real time and added up to form a total consumption value. Smart meters or consumption meters equipped with corresponding tools can also record the consumption of, for example, electricity, water or gas with a time resolution, i.e. in addition to the total consumption value, the actual consumption is also recorded at a number of points in time or in a time series. The temporal resolution of the time series or the time-resolved consumption data is determined by the number of data points that are recorded in a specific time interval. For example, a time series that represents consumption over one minute contains 12 data points with a resolution of 5 s (or 13 data points if the data point at the beginning and / or end of the time series is also counted).

[0100] Consumption meter

[0101] A consumption meter measures the consumption of, for example, electricity, water, or gas in a residential unit. Consumption meters are used, for example, by municipal utilities or basic suppliers to record a household's electricity, water, or gas consumption, primarily for billing purposes. A consumption meter can be a smart meter, which not only records consumption data but is also equipped to transmit the recorded consumption data, for example, via a Wi-Fi network and / or the internet, to a computing unit or evaluation device. If a consumption meter does not have any means for transmitting consumption data, these means can often be retrofitted. Smart meters or consumption meters equipped with the aforementioned means can in many cases be used to record time-resolved consumption data and transmit it to corresponding evaluation or computing units.

[0102] Behavioral reference

[0103] A behavioral reference makes it possible to quantitatively record the extent to which an activity is normal. A behavioral reference can be used to assess whether a person's activity indicates that at least one person needs support. In addition, a behavioral reference can be designed to identify changes in the way a person deviates from a normal activity that can still be considered normal or that indicate a change in behavior and related clinical pictures. In such designs, a behavioral reference makes it possible to assess whether a person is developing a need for support even before the person actually needs support. Behavioral references can, for example, be constructed using statistical methods, whereby expected values ​​for attributes and / or characteristics of certain activities are compared with measured values ​​or values ​​determined using consumption data.

[0104] Reference symbol

[0105] residential unit

[0106] Kitchen

[0107] stove

[0108] Dining area

[0109] bathroom

[0110] Bathtub

[0111] bedroom

[0112] Cook

[0113] Bathe

[0114] Sleep

[0115] system

[0116] Electricity consumption meter

[0117] Water consumption meter

[0118] cellar

[0119] Evaluation device

[0120] Connection

[0121] Connection

[0122] Connection of time-resolved consumption data

[0123] Proceedings

[0124] initialization

[0125] Recording time-resolved electricity and water consumption data

[0126] Transfer and evaluation of time-resolved electricity and water consumption data

[0127] Decide how to continue the procedure 400

[0128] Assigning attributes and attribute values

[0129] Comparison with expected attribute values

[0130] Assessment of any deviations

[0131] Decide how the procedure 400 continues 490 Issue a message, warning or alarm

[0132] cited literature cited patent literature

[0133] CH 719 116 A2

[0134] CN 114973596 A1

[0135] EP 1 071 055 B1

[0136] EP 1 298 619 B1

[0137] EP 2 159 770 B1

[0138] US 2022 / 207980 A1

[0139] WO 2015 / 124972 A1

[0140] WO 2018 / 114035 A1 cited non-patent literature

[0141] Jan-Peter Seevers, Kristina Jurczyk, Henning Meschede, Jens Hesselbach, John W. Sutherland: "Automatic Detection of Manufacturing Equipment Cycles Using Time Series", J. Com- put. Inf. Sei. Eng. Jun 2020, 20(3), https: / / asmedigitalcollection.asme.org / computin- gengineering / article-abstract / 20 / 3 / 031005 / 1074102 / Automatic-Detection-of-Manu- facturing-Equipment; https: / / d0i.0rg / l 0.1115 / 1.4046208

[0142] J-P. Seevers, J. Johst, T. Weiß, H. Meschede, J. Hesselbach: "Automatic Time Series Segmentation as the Basis for Unsupervised, Non-Intrusive Load Monitoring of Machine Tools", Procedia CIRP, Volume 81 , 2019, Pages 695-700; https: / / www.sciencedirect.com / sci- ence / article / pii / S22128271 1 9304834; https: / / doi.Org / 10.1016 / j.procir.2019.03.178

Claims

Patent claims 1. A method (400) for assessing the support needs of at least one person assigned to a residential or accommodation unit (100), comprising the following steps: 1.1 time-resolved consumption data (300) are received by an evaluation device (240); 1.1.1 wherein the time-resolved consumption data (300) were recorded by at least one consumption meter (210, 220); 1.1.2 wherein the at least one consumption meter (210, 220) is assigned to the residential or accommodation unit (100); 1.2 the time-resolved consumption data (300) are analyzed (430) by the evaluation device (240) in which the evaluation device (240) self-learningly identifies at least one typical pattern in the time-resolved consumption data with the aid of pattern recognition; 1.3 the evaluation device has at least one typical pattern in the time-resolved Consumption data is identified in a self-learning manner, and at least one typical pattern is assigned to at least one household appliance; 1.4 the detected use of at least one household appliance is used to infer an activity of the person; 1.5 From a number of detected activities, the person’s activity profiles are identified in a self-learning manner; 1 .6 a behavioral reference of the person is determined from the activity profiles of the person in a self-learning manner (410); 1.7 If a behavioral reference is available, the person’s daily activity profiles are measured; 1 .8 these daily activity profiles of the person are compared with the person’s behavioral reference; 1.9 At least one of the person’s daily activity profiles deviates beyond a given Tolerance deviates from the behavioral reference, the evaluation device (240) generates a message (470) and outputs it (490) to signal the person's possible need for support.

2. Method (400) according to the preceding claim, characterized in that 2.1 that the time-resolved consumption data (300) represent electricity, water and / or gas consumption; 2.2 whereby the electricity, water or gas consumption is allocated to the residential or accommodation unit (100).

3. Method (400) according to one of the preceding claims, characterized in that 3.1 that presence data of the at least one person are received by the evaluation device (240); 3.1.1 where the presence data was recorded by at least one sensor; 3.1.2 wherein the evaluation device (240) is designed to assign the time-resolved consumption data (300) to the at least one person using the presence data.

4. Method (400) according to one of the preceding claims, characterized in that the behavior reference is updated, supplemented and / or created (430) by the evaluation device (240) with the aid of the time-resolved consumption data (300).

5. Method (400) according to one of the preceding claims, characterized in that the behavioral reference is updated, supplemented and / or created (450) using previous activities (150, 160, 170) of the at least one person.

6. Method (400) according to one of the preceding claims, characterized in that the behavioral reference is updated, supplemented and / or created (450) using similarity values ​​of previous activity profiles (150, 160, 170) of the at least one person.

7. Method (400) according to one of the preceding claims, characterized in that 7.1 that the at least one activity (150, 160, 170) is linked to at least one attribute (450) will be; and 7.2 that the at least one attribute is used to form the similarity value (460); and / or 7.3 that the at least one attribute is used to assess (470) whether the at least one person develops a need for support before the at least one person actually needs support.

8. Method (400) according to the preceding claim, characterized in that the at least one attribute describes a time duration, a temporal occurrence, a frequency and / or an intensity.

9. Method (400) according to one of the preceding claims, characterized in that the steps of the method (400) are carried out repeatedly and / or in real time.

10. Method (400) according to one of the preceding claims, characterized in that 10.1 that the similarity value of the at least one activity profile (150, 160, 170) is evaluated (470) by the evaluation device (240) with the aid of the behavioral reference to determine whether the at least one person has a need for support; 10.2 that the evaluation device (240) generates a message (470) if it determines that the at least one person has a need for support; and 10.3 that the message is output (490) by the evaluation device (240).

11. Method (400) according to one of the preceding claims, characterized in that the time-resolved consumption data (300) have a time resolution which is in the range from 0.2 s to 15 min, preferably in the range from 0.5 s to 10 min, particularly preferably in the range from 1 s to 1 min.

12. Method (400) according to one of the preceding claims, characterized in that the at least one activity profile comprises a plurality of activities.

13. Method (400) according to one of the preceding claims, characterized in that the at least one activity profile corresponds to the activities of the at least one person in one day, in one week or in one month.

14. Evaluation device (240) for evaluating the support needs of at least one Person, wherein the evaluation device (240) is equipped with means suitable for carrying out the steps of a method (400) according to one of the preceding method claims.

15. System with an evaluation device (240) according to the preceding claim and at least one consumption meter (210, 220) which is equipped with means to record time-resolved consumption data (300) and to transmit them to the evaluation device (240).

16. Computer program comprising instructions that cause an evaluation device (240) according to claim 14 or a system according to claim 15 carries out a method (400) according to any one of the preceding method claims 1 to 13.

17. A computer-readable medium on which the computer program according to the immediately preceding claim is stored.

Images