Automatic Subtask Segmentation Approach of the Timed Up and Go Test for Mobility Assessment System Using Wearable Sensors

Population aging is common phenomenon in the worldwide today. Maintaining and promoting the healthy mobility and mentality is crucial to enhance quality of life. The accuracy of mobility assessment in elderly people is an important issue of clinical practice. Many clinical tools are proposed for mobility assessment. The Timed Up and Go (TUG) test is one of the most widely accepted functional mobility test to measure basic mobility and balance capabilities. The TUG test consists of eight subtasks, including initial sitting, sit-to-stand, walking-out, turning, walking-in, turning around, stand-to-sit and end sitting. The detail information about subtask is essential to aid clinical professional and physiotherapist about making assessment decision. The main objective of this study is to develop an automatic subtask segmentation approach during TUG test execution. Activity-defined window technique and decision rules are designed and employed in the proposed subtask segmentation approach. To ensure feasibility of proposed segmentation approach, the experiment recruits ten volunteers, including five healthy people and five patients with severe knee osteoarthritis. Each volunteer performs three times 10m and 5m TUG and collects the motion data with wearable sensors. There are 60 instances, including 30 instances of 5m TUG and 10m TUG test, which are used to explore the performance of the proposed segmentation approach. The overall performances of the accuracy in the TUG test for healthy volunteers and patients with severe knee osteoarthritis are 95.47% and 95.28%, respectively. The results show that the proposed segmentation approach can fulfill the reliability of automatic subtasks segmentation during the TUG test.