PreFalls

The main objective of this project is to develop an innovative software

system that can dynamically perform spatial-temporal gait analysis and

physical activity classification in order to assess the mobility and falls of

elderly people remotely in real-time. The system is designed to be utilized

as a tool that can precisely identify changes in gait parameters (e.g. stride

length) and recognize types of physical activities of the elderly based on the patterns of acceleration data. In our system, a mobile phone equipped with

motion sensors (an accelerometer and a gyroscope) has to be worn on the

waist of elderly people when they perform physical activities. It then

continuously submits sensor measurements (e.g. acceleration force angular

velocity and calculated stride length) to our remote monitoring system. The

system then processes the sensor measurements dynamically as they

arrive, performs gait analysis and activity classification with low latency,

and finally graphically displays certain gait parameters and types of

mobility-related activities of the elderly in real-time to support medical

staffs in diagnosing diseases and planning the treatment. In the case of fall

risk is identified or a fall event is detected, an instant notification will be

sent to caregivers who provide care and support for the elderly via a

dedicated mobile application. Furthermore, in case that immobility is

detected (elderly people sit or lie in bed for long periods of time), the

system will notify caregivers immediately as well.