The home environment is equipped with vital data acquisition devices which are available on the market. These are pretty ordinary home sensors like smoke detectors and sensors at the doors, windows, light switches, refrigerators, stove, water faucets and pressure sensors that can be used to detect the presence in the bed. But there are also more complex devices like person trackers based on so called ToF scanners which recognize when a resident has fallen and is lying on the ground. There is also the alternative of infrared detectors or a smart membrane for retrofitting the domestic environment. These sensor systems can not only be used as an alarm sensor but also as a motion detector build up a heat map which shows where the elderly moves.
The software analytics, which is based on Big Data methods and machine learning techniques au-tomatically, detects deviations from the normal all-day course of the resident based on long term observations. It reacts without the resident needs to be active. For instance, it could be stated that the resident falls more often, has not switched on the stove, or has not opened the apartment door and since he uses a new medication. A possible danger situations can be derived and the es-calation chain will be started; E.g. the home care service is informed of the changes in the apart-ment. The nurses can view detailed information on the situation and then initiate a reaction. In addition, long-term evaluations can be used to determine if the resident cannot live longer alone.
Access to Smart Service Power is primarily provided to the residents via tablets and natural lan-guage assistant systems via speakers. It is likely that the Microsoft NXP System will be used. It has the big advantage that it offers an open system interface. Assistants like Amazon Alexa or Google home are not suitable because data autonomy is not possible in closed systems. With the help of the tablets and the natural language assistants, information can be retrieved or on-off functions can be activated or deactivated. The devices can for example remind the resident to take his or her medication. In addition to the use of tablets or speakers in the home, it is also possible to display information, for example, via the TV or a display integrated in the mirror.
A goal of the project is the development of a differentiated data usage concept while respecting data protection and data security. The resident can specify how often his or her data can be used (one-time, constantly), who can receive it (all or only a certain group, maybe as a basis for a service offer from the concierge service). The person can also determine in which granularity or anonymity the data can be used. The resident can also decide in which context the information should be available (for example in an emergency).
The data is provided with permanent static or temporary access rights, depending on the usage and data protection class.
The data of the sensors are stored first in an encapsulated virtual homeserver (in the cloud) the so-called HomeCollector. There the information is bundled and automatically pre-evaluated and transmitted according to data usage concept in real time to the cloud. The system uses a hybrid cloud approach.
After a long and careful review process the standard IoT Platform Microsoft Azure IoT-Hub was selected because it offered the best fit for the requirements. The implementation task of this project is to integrate the horizontal level of the AAL, the care and emergency systems, from E-Health to Smart Home. On the vertical level, it integrates various gate-ways with the corresponding sensor technology, device management and integration, communica-tion, data analysis, Big Data and Machine Learning. It also involves payment solutions, license man-agement and connectivity solutions. Part of the integrated platform is a user interface with different views for the individual target groups, which differ primarily in the granularity of the data. Doctors, relatives, emergency services, housing associations and the residents themselves - everyone needs a different view and a user interface that is oriented towards their needs. The elderly, for example need an age-appropriate graphical user interface, which can be operated simply and intuitively. The relatives, on the other hand, want information about what he or she has done all day. The emer-gency service is only interested in the fact that the person is lying on the ground and cannot get up. The surface is equipped with a traffic light system so that everyone can immediately recognize whether there have been any conspicuities. The IoT-platform already includes several different security mechanisms which have to be config-ured carefully. This is a big relief in the implementation process.