Safety at home

The CARE project, whose aim is to automatically detect and signal falling, a constant hazard to older people, has arrived to the beginning of its last third. As a result of the developments, the safety of older people spending long hours or days alone in their homes can be significantly improved.

 We are living in an ageing society. Caring for older people is a more and more difficult task for families. “Many of whom live alone in an apartment or do not have anybody with them most of the day do not feel safe. They are afraid because if they fall, which is a particularly common accident in bathrooms or bedrooms, they can be seriously hurt and it is also possible that they get no help for hours or even for days. The CARE (Safe Private Homes for Elderly Persons) project seeks solutions to this problem,” says Péter Hanák, director of Biomedical Engineering Knowledge Centre, Budapest University of Technology and Economics (BME).

 Based on an idea from Austria

Austrian Institute of Technology (AIT), the leading partner of CARE, has been studying how information and communication technologies could effectively help living, since 2008. The institution has presented impressive results. It was also in 2008 when, after being introduced to the AAL Joint Programme, Ahmed Nabil Belbachir, a neuroinformatics professional of the Safety and Security Department of AIT came up with the idea of the CARE project.

“As soon as I read the AAL’s first call for proposals I recognized that we had much to contribute to the efforts aiming at solving the issues of ageing. Arriving back to Vienna we started to work out a concept. After this I contacted BME, Everon and Yrjö & Hanna Foundation from Finland, and SensoCube from Germany. We formulated the project proposal defining specific steps and the final results we aim for together with the members of the consortium. Senioren Wohnpark Weser joined the consortium after accepting the project,” says Ahmed Nabil Belbachir.

 Fall detection with a visual sensor

We could not begin the development process immediately after having been granted subsidy. First we had to find answers to some important questions. For example: what are the most common critical situations; to what extent help is needed; can older people accept having a visual sensory device reminding of a camera in their homes; what amount of one-time or regular expenses can these people afford etc.

After reviewing human aspects we had to clarify what exactly can be counted as a fall. This was necessary to avoid false alarms and making sure of a secure recognition of real falls. (It cannot be considered a fall if for example someone intentionally sits down on the floor, abruptly tumbles down to a bed, drops to a chair etc. It is of course a possibility in these cases that later the person cannot get up, which also has to be recognized.) After this we had to decide how the system should detect a fall. The members of the consortium had already voted for fall detection based on visual sensing before handing in the project proposal. The system, which was set up by developing further a special visual sensor already available at AIT and working out an algorithm processing the data provided by this sensor, can recognize if the person at home falls, and can send the appropriate signal or alarm to the telemonitoring centre.

“If we perform continuous monitoring with a traditional camera we get a huge amount of data most of which is redundant. In addition, using a camera you really feel as if somebody was constantly watching. To tackle this issue, in CARE we have chosen a special visual sensor that transmits data of changes only. The system “unravels” the information indicating a fall from the characteristics of the change (e.g. some of the pixels change from a light shade to a dark one), the position of the change (i.e. if there is a change in the highest point, the centre of mass, the ratios of the limiter box or cylinder), the time and speed of the change etc. The real image, the actual view cannot even be restored from the information. The device provides only the result indicating if the monitored person fell or not, is incapable of moving or not. With this method we can avoid intruding on the private sphere of people,” points out Ferenc Vajda, associate professor of the Department of Control Engineering and Information Technology, BME.

 Silicon retina sensors

“The embedded sensor system we use to detect falls includes two silicon retina sensors, which apply the working principle of human eyes, in a stereo configuration. This enables three dimensional optical sensing,” explains Martin Litzenberger (sensor technologies, Safety and Security Department, AIT) the technical details of the system.

In contrast to traditional image sensors, silicon retina sensors contain 304x240 independent optical movement sensors in a CMOS chip of approximately 1 cm square size. These sensors were chosen because of their big time resolution (>1/1000 seconds), wide dynamic range (120 dB) and “anonymous” optical sensory capabilities. The above parameters far outstrip the features of traditional image sensors. This makes it possible that silicon retina sensors can detect the monitored person’s fall by precisely following a human head in a 3D space, even in poor lighting conditions (for example in a dark room or in bright sunshine next to a window).

“I am more than satisfied with the results we have achieved in the project so far: we are proceeding according to plan. System development is under process and we will soon evaluate results to be obtained in two testing environments in Germany and in Finland. During the interviews we made with end users many further questions have arisen regarding human dignity throughout ageing. These can be extremely useful for the future, and present new challenges for us,” says Ahmed Nabil Belbachir.

AAL Joint Programme

The Ambient Assisted Living (AAL) Joint Programme of the 23 member countries of the European Union and the European Commission from 2008 to 2013 aims at developing information and communication solutions that help older people in continuing with their independent life in their usual environment. The programme´s planned total budget is 700 million Euros. There have been four calls for proposals so far and one more can be expected each year. The basic condition of application is that researchers, developers and users of the applying consortium have to come from at least three different countries. The CARE project is one of the winners of the first call of the AAL Joint Programme in 2008.

Most important data of CARE

• A consortium of Austrian, Finnish, German and Hungarian members
• Hungarian member: Budapest University of Technology and Economics
• Duration: July 2009 to December 2011.
• Budget: 2.36 million Euros out of which 1.75 million Euros are granted by the European Union

Ethical issues

In the area of ambient assisted living ethical behaviour and operation means that every collaborator of a given service acts considering the interests and feelings of the people receiving assistance and of other people involved (care personnel, service providers).

The primary target group of home monitoring are the older people. Professor Lajos Izsó, head of Department of Ergonomics and Psychology at BME points out the main ethical issues of home monitoring. These are the followings: protection of personal data of the monitored people and their rights regarding their own bodies; acceptable amount of control; the truth of the promise of immediate, professional assistance; possibilities of false alarms.