Description
Among elderly persons especially older adults, physical injuries sustained by an unintentional or an unpredictable fall on a hard surface is one of the important reasons to cause injuries related to fall and sometimes loss of life. Each year close to 30% of adults above the age of 65 fall at least once. In the year 2015 close to 2.9 million falls were reported, resulting in 33,000 deaths. Elderly people staying at nursing care homes tend to fall 61% of the time during their initial days of stay. This may aggravate the situation leading to bone fracture, concussion, internal bleeding, or traumatic brain injury when immediate medical attention is not offered as soon as the fall has occurred. Delay in course of the event may sometimes lead to death as well. These days many studies have come up with wearable devices that are small, compact, wireless, battery operated with low power consumption and mobile for detection falls. These devices are now commercially available in the market. This thesis discusses the findings that the optimal location for a Fall Detection Sensor on the human body is in front of the Shin bone. This is with regards to the features recorded by the Inertial Measurement Unit sensors strapped at 16 different locations on the human body while the obtained data is trained-tested using the Convolutional Neural Networks machine learning model. The ultimate goal is to develop a wireless, mobile, wearable, low-power device that uses a tiny Lattice iCE40 FPGA which detects whether the device wearer has fallen or not. This FPGA is capable of realizing the Neural Network model. This In-Situ or Edge inferencing wearable device can provide real-time classifications without any Transmitting features and Receiving class over a wireless communication channel.
