AF7
MAHMOOOD MERICAN AWARD FINALIST
Recognition Of Common Gait Patterns Using I-Pod Touch Tri-Axial Accelerometer Waveforms Data
Khoo CCH
Introduction :Current methods for gait analysis are laboratory-based and their use in clinical setting is difficult. We want to explore the potential of using a portable device based in an i-Pod Touch as an ambulatory gait detection and analysis system.Objective :1.To determine the accuracy and sensitivity of the tri-axial accelerometer embedded within the 4th generation i-Pod Touch as a gait detection device. 2. To describe a new algorithm for analysis of the tri-axial accelerometer data for gait and subsequently explore a new method for visual representation for 4 commonly seen gait patterns.Design & Methodology :Part 1 of the study involved a single 4th generation i-Pod Touch which was subjected to 30 free falls tests (10 tests each week for 3 consecutive weeks) and the resultant vectorial sum accelerations captured by the Accel4Pros application were validated against the gravitational force at Earth’s surface which is 1 G Force.Part 2 of the study involved 36 participants (9 each from normal gait, antalgic gait, short limb gait and bilateral Trendelenburg gait) walking normally on a 10 metres walkway with the 4th generation i-Pod Touch attached to the body centre of mass at the low back region whilst gait data was being captured using the Accel4Pros application. All 36 gait data were analysed and summarized into circular patterns which were then presented to 6 trained orthopaedic observers for Interobserver Agreement and Intraobserver Reliability Tests.Results :The 4th generation i-Pod Touch demonstrated a high accuracy and sensitivity abilities under dynamic conditions shown by a standard deviation of 1.58% from a mean acceleration of 0.999G with 95% confidence interval of 0.993G to 1.005G, in comparison to the standard unit of free fall on surface of earth, 1.000 G Force. Interobserver Agreement Test and Intraobserver Reliability Tests revealed a successful 100% for gait patterns identification. Conclusion :Our preliminary study shows great promise in terms of utilizing the accelerometer data from a portable device to detect characteristic features for waveforms of the 4 types of gait with trained observers. The tri-axial accelerometer embedded within the 4th generation i-Pod Touch represents a huge potential as a gait analysis tool based on its accurate and sensitive acceleration data output. More future researches involving multiple devices subjected to more complex motions over a longer duration are required.