1. Systems dealing with walking and possible fall of the elderly

1.1 Development of a system to predict the risk of falling into the home of elderly people

International models for the prevention and treatment of falls start either from the voluntary preventive appearance of the elderly person in a special centre in order to check his individual risk of falling or from his appearance in an emergency unit after a fall. Therefore, the burden of prevention falls on the elderly person himself, whose negligence to operate according to the predictions of the algorithm of each model leads to the failure of the entire system of prevention/treatment of falls. This makes it imperative to develop a fall prediction system, without an active role of the elderly person himself, which operates continuously but discreetly, without interfering with the person's usual lifestyle and daily activities (Rajagopalan et al, 2017).

A fall prediction system should recognise all scenarios of a possible fall and provide a framework for preventing it through targeted and personalised interventions. The system should collect data on a 24-hour basis on the most important fall risk factors, using methods that are more reliable, less subjective and more effective than the widely used questionnaires, fall diaries and telephone interviews (Rajagopalan et al, 2017). 

The fall of the elderly person in his or her home is in any case the development of a balance disorder that has not been successfully addressed. Therefore, the basic approach to the development of a fall prediction system should start from the assessment of the factors affecting the stability of the elderly person, i.e. his ability to successfully deal with small or larger imbalances.

Achieving a stable posture and gait has three basic conditions (Bruijn et al, 2013):

  1. the person must be able to regain his balance from or limit minor disturbances during posture or at every step during walking (e.g. due to sloping or uneven posture and movement surfaces) 
  2. the person must be able to regain balance from his major disorders, which require changes in his motor behaviour, and
  3. the maximum disturbance from which the person can marginally regain his balance must be greater than the disturbances that will occur

The objectives of this solution are:

  1. the determination of the parameters that quantify the extent to which the above conditions appear in basic daily activities of the elderly person in his or her home.
  2. the documentation of the ability of the above parameters to assess changes in the behaviour of the elderly person in the direction of reducing the risk of falling.

1.2 Development of an Elderly Fall Detection System

1.2.1 Development of System Softwares for Defined Elderly Fall Detection Radar at Home

The purpose of this action is to develop a Doppler-based radar system (Radar FM-CW) to detect the accidental fall of elderly people when they do not carry a sensor and either voluntarily or inadvertently do not cooperate. The system is based on software defined radio (SDR, software defined radio) of microwave and/or mileometric frequencies. This applies a linear FM technique and determines the speed of the target (Older), its rate of change (acceleration) and its relative position. The installation of multiple Radar FM-CW is considered to avoid false alarms.

1.2.2 Detection of the elderly's fall with optical-acoustic sensors

In this mode, sound and visual sensors will be able to detect an elderly person's fall and will automatically be able to call for help. 

1.3 Gait Analysis of the elderly

In this mode, we can use a camera in a hallway of the house to record the movement of the elderly. From this pictorial series, we can draw conclusions about the kinesiological performance and behavior of the elderly and inform the attending physicians in cases of change.  

2. Development of a robotic system to assist the lives of older people

2.1 Development of a total support system for the elderly

The main objective is to develop a system integrated into a real AAL robotic platform (ambient assisting living). The overall system will understand both spatially and semantically its environment and will have the ability to function as an auxiliary in the everyday life of the elderly person. At the same time, its intelligent navigation will make its presence discreet in the environment of the elderly, while helping to identify and predict any risks that will occur (e.g. fall).

2.2 Development of a robotic fall support system

The system aims at the timely mobilisation of the robotic system at the point of decline to verify the situation and provide first aid as well as – if necessary by the system – to notify the competent health authorities of support. In the event that the elderly person falls, the first aid-equipped domestic robotic vehicle will rush directly to the exact location by extracting photographic material (critical photos, video and sound after the fall) which will be sent wirelessly to the cooperating medical authority: At the sound urging and if the elderly person responds to the basic stimuli of the robotic assistant the system will export additional photographic and audio material which will be sent to the competent authority where it will be verified whether there is a serious injury and intervention by professionals or hospitalization is necessary.

At the same time, the system will also be supported by a built-in smart device which during the recovery of the elderly will provide detailed instructions adapted to the specificities and location of the elderly person's fall (visual and audio material – video) for his repair and the use of first aid. The smart device, tablet type, can also be used to communicate directly with the medical center.

3. Development of a system of daily observation and taking of vital measurements of the elderly

3.1 Continuous Tele-Monitoring of Basic Biological Parameters of Elderly At Home

The aim is to develop a system of wearable Sensors for continuous monitoring of vital parameters (Temperature, pressure, heart rate, blood oxygenation, geographical stigma and acceleration) of elderly people living autonomously in their home. The sensors will be tested and provide data to a processor embedded in a properly configured wristband. It is investigated that this wristband includes a mobile phone terminal or that it communicates with a short-range system with a corresponding access point located within the home. This system will provide uninterrupted monitoring of elderly health, process data locally and operate at different alert levels. The information is transmitted at a self-adjusting rate to the alert level at a partner health centre. Depending on the severity of the alarm, an on-site crew will be on standby. In consultation with the responsible doctor, if necessary, it will be done by the same system and electrocardiogram. This will be transmitted to the health center and the Doctor will decide on the further.

3.2 Exploitation of thermometry for the development of a disease prediction model in the third age 

The system will incorporate thermal observation/thermometry for the early detection of musculoskeletal problems and other diseases in order to provide early treatment and diagnosis in elderly people. Thanks to thermal IR observation and sending of images and videos to the responsible diagnostic centre, the competent authority will be able to carry out telediagnosis and further verification of any health problems or dangerous situations for the elderly. At the same time, a computationally efficient software tool for applications for early treatment of diseases and adverse conditions based on an adaptive & plug-n-play methodology of optimal closed-loop control will be developed that will work with a detailed model of the elderly's thermal behavior to predict the development of health problems. The system will have internal machine learning mechanisms for automated decision-making based on the real-time data collected from the automation ecosystem that has the elderly's thermal profile. The logic of implementation with the help of a prediction model will follow the standards of cognitive optimal control function, assisted by simulations (simulation-assisted optimized control), similar to Model-Predictive-Control (MPC) methods with the ultimate goal of preventing potentially unpleasant situations and early response to cases of local physical overheating or body temperature change. It should be noted that the special software tools will be designed so that they can be hosted on platforms for high-data communication supporting the system.

3.3 Detection of basic activities of the daily life of the elderly 

In this mode, sound and optical sensors will be combined to classify the activity of the elderly into 6 basic activities (Food Preparation, Bathroom, Cooking, Television, Sleep, Unspecified Activity). Thus, we can have an automated recording of the activities of the elderly, which is important for the medical staff who monitor that person. Here, in cooperation with medical staff, we can propose increasing the basic activities that we want to record. 

3.4 Contactless Detection of heart rate or other vital measurements of the elderly

Of course, in many phases of the day, the elderly are familiar with taking some basic heart measurements (pulse, pressure, etc.). However, it would be very useful if we could monitor the heartbeat intact at times when an elderly person is unable to make measurements, e.g. when sleeping or watching TV. So, we could have an automatic call system to help, in case the pulse stops. Here, the drive aid technique will be used and it will also be extended if we can extract some other vital measurements from this reinforced signal. 

3.5 Assessment of indoor hazards

Assessing the risk of indoor areas where elderly people move (houses, institutions, etc.) in accordance with technical legislation and current technical guidelines is a necessary procedure. The comfort of movement in corridors and scales and the reduction of comfort due to the presence of objects such as furniture, game openings, etc. Are assessed. The evaluation methodology may include the following:

  • Division of space into elementary surfaces (cells) and performance in each cell of comfort and danger parameters.
  • Development of an algorithm based on Cellular Automatics and Random Walks to control and predict the movement of the elderly in the areas under control.
  • Combination of the above algorithms with learning techniques to adapt them to the personality and routine of the elderly.
  • Implementation of algorithms in FPGA integrated circuit and integration of the system into a portable wearable device. The device alerts the user with an optical and auditory signal whenever it predicts that its next movements may end in a fall.
  • The device will also be able with different signals to prompt the user to stop, move right or left in order to avoid dangerous areas of space.
  • Algorithms implemented on an integrated circuit will be added to the device to assess the physical condition of the user at the time and change the risk parameters of the cells in the space. For example, on a scale with a low degree of risk, a greater degree can be attributed if the physical condition of the user at the moment is not good (e.g. limb tremor, etc.) 
  • Also, a re-evaluation of the risk of the cells in the area will be carried out every time the user has transported medical items, e.g. hand in plaster, wheeled serum, etc. 
  • Also, the system will be able to include cases where while the user was moving freely he will now move with walking aids such as crutches or "pi".

Care will also be taken to ensure the safe entertainment of the elderly with vision or hearing aids for watching TV, radio and internet use. 

4. Development of infrastructure to support assisted living

The goal here is the development of innovative technologies, which will support the above activities as well as the smooth electrical and network operation of a smart home for the elderly.

4.1 Development of innovative sensors for patient monitoring

The goal is to build an economical system with sensors and electronic sensors on a flexible substrate with water resistance (waterproof). In order to achieve this objective, it is necessary to achieve the following objectives:

  • Find sensors such as accelerometer and gyroscope and build a skin conductivity measurement sensor with the necessary specifications for the application. 
  • Inkjet printing using conductive silver and graphene nanoparticle inks on the flexible substrate offering biocompatibility, elasticity factors similar to those of the skin, very good adhesion to the skin, durability, etc. for the interconnection of sensors (accelerometer, skin conductivity measurement, etc.)
  • Design, manufacture and adaptation of an energy storage device (rechargeable lithium-ion battery or super condenser depending on the energy requirements of the other components of the system). The goal is to integrate battery with solid electrolyte for greater safety for the user. 
  • Design and construction of an antenna for the sending of data (in collaboration with the telecommunications interconnections team) on the flexible substrate using inkjet printing.
  • Design circuits to collect and send data in such a way as to ensure low power consumption. 

4.2 Development of electrical network infrastructure for the uninterrupted supply of electricity and lighting

The aim is to develop a comprehensive proposal that incorporates smart exploitation ideas and the exploitation of existing and potential electrical energy infrastructures in older people's homes for their safe and quality of life. To this end, the proposal includes the design and development of individual electrical consumption management and control systems, production units and storage systems, with the following sub-systems for the elderly:

  • Ensuring high-level lighting conditions and climate of home spaces 
  • Ensure safe, quality and economic living under every possible operating condition of the electrical installation of the house, even after a power outage with the contracted provider.

4.3 Development of a sensor network for the interconnection and management of IoT devices in health

The network includes a gateway-type IOT node that periodically detects space to discover devices of various technologies. It then interfaces with low-power wireless transmission technology (such as Bluetooth Low Energy) and continuously collects real-time data, which it first collects locally and then sends to an appropriate repository. Adding or removing an IoT device from the network will be dynamically supported with an easy process. The network supports the interconnection of devices of many different network technologies by compiling their data from one package format to another through wrappers that will be developed in the project that can contribute to the homogenization and synchronization of data. The aim of the research is to allow the network to manage the devices adaptively, allowing, for example, an increase in the frequency of sampling when a need arises. Finally, the network can also include an appropriate hierarchy of more than one node to cover a wide area.

A platform to display their network nodes will also be developed in an appropriate topology e.g. on a map as well as a user interface for device management e.g. to send updates.

4.4 Development of a smart power system for electronic devices and appliances

An innovative, intelligent power supply system (iPowerKit) will be developed to cover portable electronic devices, robotic and telecommunications systems. Its main characteristics will be the very high reliability and at the same time the quality of the power supplied to these loads – thus protecting them from premature aging and/or malfunctions