Pasit TangkongchitrMingkamon BuathangThanavit UnsuwanKonlakorn WongpatikasereeMahidol University2019-08-232019-08-232018-07-022018 International Joint Symposium on Artificial Intelligence and Natural Language Processing, iSAI-NLP 2018 - Proceedings. (2018)2-s2.0-85065097273https://repository.li.mahidol.ac.th/handle/20.500.14594/45612© 2018 IEEE. The main reason for hospital admission of elderly people worldwide is falls. The longer one has to stay on the floor due to inability to get up after falling, the more severe the injury can become. Existing fall detection devices that need to be worn on the body is not very useful because the elderly simply forget or do not want to wear them. From this, we have created a fall detection system named 'SaFloor' - a soft mat with force sensors embedded inside. SaFloor can be placed in fall-prone areas, such as by the bedside, bathroom, at the bottom of stairs, etc. It can distinguish between a real fall and other impacts such as walking and dropped objects, and send out a notification to a family member or a care giver via Line message when a fall is detected. The experiment consisting of 14 participants with different weights and heights shows that SaFloor has a successful fall detection rate of 88%.Mahidol UniversityComputer ScienceMedicineConference PaperSCOPUS10.1109/iSAI-NLP.2018.8692857