Patients with lower limbs problems are an increasing population in the US and many of them require surgery and its subsequent post-op Physical Therapy (PT). For all these patients, tracking their progress and evolution towards full recovery is very important. To assess the patients and track their progress, patients are usually required to perform very specific tests administered by a physical therapist. These tests either require very expensive equipment or rather require the subjective experience of the physical therapist who administer them. One of these tests is the Functional Gait Assessment (FGA) test, perhaps the most widely adopted one for gait assessment.
This thesis presents a system for Clinical Gait Assessment using exclusively the sensors embedded in today’s smartphones. The system processes the raw sensor data to perform the FGA test and calculate additional metrics, capable of identifying problems in the human gait. The system is therefore objective, as it is based on measurements; cheap, as it only requires a smartphone; mobile, as it can be used pretty much anywhere; and self-care, as it does not need the presence of a physical therapist.
The system was designed and tested on the Android OS with the phone attached to the back of the user using a belt or elastic band. It includes a new step detection algorithm with a mean absolute error of ±1 and algorithms to detect the deviation from a straight path with an accuracy of 90%, 80%, 35%, and 30% for each of the required deviation levels of the FGA test. Additionally, the system includes autocorrelation and DTW metrics, which provide additional information to detect different impediments of the user gait.
Identifer | oai:union.ndltd.org:USF/oai:scholarcommons.usf.edu:etd-7546 |
Date | 30 June 2016 |
Creators | Perez Leon, Andres Alfredo |
Publisher | Scholar Commons |
Source Sets | University of South Flordia |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | Graduate Theses and Dissertations |
Rights | default |
Page generated in 0.0018 seconds