Device-free activity recognition for assisted living

Abstract

Nowadays, many elderly people need to stay at home or care home due to their physical condition. To enhance their life pattern and provide more support for this assistive living, a newly emerging field, Device-Free Activity Recognition, has been introduced. This research has contributed to this field for developing Applications in two scenarios. A framework needs to be easily deployable and affordable to cope with real-world scenarios. We have used Received Signal Strength Indicator (RSSI). RSSI is used in most wireless devices, and it is easily collectible without any additional hardware. This research has contributed to the field of device-free activity recognition (DFAR) by developing applications for two scenarios: device-free estimated energy expenditure (DFEEE) and RSSI-based device-free walking direction detection for passers-by. The majority of wireless devices employ the received signal strength indicator (RSSI), which is simple to collect without any additional hardware. The DFEEE system was compared with a smart watch and was able to calculate expenditure with up to 95% accuracy. The device-free walking direction detection system was able to detect different passing directions with 72% accuracy. The suggested framework is ideal for use in real-world applications because it is inexpensive and simple to deploy. This study has the potential to improve elderly people’s quality of life and support assisted living.