Classification Different Types of Fall For Reducing False Alarm Using Single Accelerometer

摘要: Fall is one of the major causes serious injury, which include fractures, traumatic brain and death, to elderly. True fall detection in time will improve chances survival increases likelihood normal behavior recovery by up 80%. Many researchers use accelerometers detect as its convenience, low power portable. However, simple threshold method can lead false alarm several ADLs (Activities daily living) such lying or sitting types fall. This paper presents a algorithm reduce using predetermined multi – thresholds three phase events. The performance this techniques evaluated signals generated during lab experiments that record user’s movement activities (walking, up/down stairs, standing up, down down) variety cases. It was found our able classify 6 different type with accuracy 92%, comparable other methods.