As the CARE project is progressing, the biologically inspired Stereo Vision Sensor Hardware and Software components were put to the acid test. The test phases 0 – 2 (module- and integration tests) of the CARE Test Plan have been performed successfully at AIT in Vienna from 16th to 17th June 2011.   These tests involved personnel of AIT and Sensocube as external tester. The collaboration between AIT and Sensocube as external tester was productive and prosperous.

CARE prototype in climate chamber

CARE prototype in climate chamber

The sensor had to prove its functionality in various conditions, in heat and cold in the climate chamber and under moist conditions in the bathroom. The tests ranged from power measurements, test recordings in different light intensities, to basic software tests.

Recordings of the field of view test

Recordings of the field of view test

In summary we can say that the SVS system meets all specifications that were defined during the project and the whole CARE project is headed in the right direction. The implemented tests and the results permit a release for test phase 3 – the recording of different fall scenarios with the help of a stuntman. This test starts at the 4th July 2011 in Bremen with the help of AIT, Sensocube and BME. BME will then start to train the artificial neuronal network for reliable and safe fall detection.

See the EURONEWS TV documentary on CARE, which was broadcasted in the first week of February 2013:

 

The video below shows the principle of fall detection with the neuromorphic vision sensors

CARE - Safe Private Homes for Elderly Persons

Printer-friendly versionSend to friendPDF version

The CARE project was an R&D activity running under the Ambient Assisted Living Joint Programme, which has been co-funded by the EC and by several participating countries. The project aimed to realise an intelligent monitoring and alarming system for the independent living of elderly persons. Specifically, this project targeted the automated recognition and alarming of critical situations (like falls) using a visual sensor and real-time processing while preserving the privacy and taking into account system dependability issues, especially ensuring reliability, availability, security, and safety from a holistic point of view.

A biologically inspired, neuromorphic vision sensor from AIT was integrated into the Everon alarm, security, and monitoring system for seamless analysis and tracking of elderly persons at home. This real-time information has been exploited for incident detection (e.g., fall detection, immobilised person; see images below), and instantaneous alarming of the concerned parties.

 

Fall detection with the neuromorphic vision sensors

Fall detection with the neuromorphic vision sensors: (1) picture from the scene (top left), (2) its corresponding data representation from a pair of neuromorphic vision sensors (bottom), (3) the resulting stereo representation (top right).

 

In relation to the Ambient Assisted Living Joint Programme, CARE proposed an innovative technology and system concept for real-time services to nursing homes and service homes. This initiative was an end-user driven R&D activity where the end-users represented the major market players in Ambient Assisted Living activities as they had direct relation and responsibility towards elderly persons ensuring their safety and independent living.

The R&D partners comprised research institutes (AIT, BME), SMEs (Everon, SensoCube) and end-users (Weser, Yrjö ja Hanna). The selected elderly homes in Germany and Finland were involved in the evaluation and demonstration of the CARE concept.

The project started on July 1, 2009 with a duration of 30 months. As the on-site tests with end-user involvement needed more time than originally planned, to project was extended by nine months, and thus ended on September 30, 2012.