Master thesis presentation : Anna Ragnerius and Frida Widelund presents their master thesis “Extraction of foot strike patterns using a sock with piezoelectric fibres”
On Wednesday 8 June, Frida Widelund and Anna Ragnerius will be presenting their master theses with the title “Extraction of foot strike patterns using a sock with piezoelectric fibres”.
When? 10:00 am, Wednesday 8 June 2016
Where? Lunnerummet (room 3311), Hörsalsvägen 11, 3rd floor, Chalmers
Examiner: Stefan Candefjord
Information about a runner’s foot strike pattern is interesting as the foot strike is not only believed to impact the runner’s performance but also the risk of getting running-related injuries. In this thesis a software system for recognition of foot strike patterns has been developed. The system makes use of signals from a sock instrumented with textile piezoelectric sensors in heel and toe. The purpose of the thesis was to develop software for automatic classification of runners as either heel-, mid- or toe-strike and provide information about foot strike patterns based on signals from the instrumented socks.
Data was collected on a treadmill while following a protocol. The protocol included walking and running in different speed with the three strike types; heel-, mid- and toe-strike, resulting in a database with tagged sequences. A pattern recognition method with two decision layers and a segmentation algorithm was developed. The first layer classifies data as activity or no activity based on periodicity, the segmentation algorithm isolates each foot step while the second decision layer is a neural network that classifies foot-strike patterns.
The resulting system succeeds to classify foot strike patterns correctly up to 97.9\%. This shows that it is possible to use the piezoelectric textile sensor to classify a runner’s foot strike pattern, though the system has some limitations. The changing properties of the instrumented sock during use disable logging sessions throughout a full protocol. Therefore, it would be necessary to improve the sock, and possibly the hardware, before conducting software test on a larger test group and continuing the research on other areas of use than running.