I am recently took a summer studentship as a part of the Bristol Interface and Graphics group. This group is part of the Computer Science department of the University of Bristol, where I did my undergraduate degree. The studentship is aimed at extending my earlier Master’s thesis into foot gestures into a full paper which has been accepted by the Computer Human Interaction (CHI) conference for 2012. Here’s a section of the draft of the abstract of ‘Putting Your Best Foot Forward: Investigating Real-World Mappings for Foot-based Gestures’:

The over-arching goal of this work is to understand user perceptions of the mapping of foot gestures to useful, real-world functions in a mobile-device context. This is achieved via a guessability study that probes users on suitable commands for common functions and an in-depth study of one continuous interaction area: map navigation.

So what does that actually mean? Essentially, I’m looking at tracking the motion of the feet and turning these movements into an input method. So the idea is that you stamp your foot once, and your phone recognises that and does something (like answer an incoming call). Or you want to skip to the next track on your MP3 player, and so you kick to the left. To skip back, you kick to the right. You aren’t limited to kicks and stomps either; you can also perform more complicated gestures such as tracing symbols.

Actually making up the pairings between gesture (what you do) and action (what actually happens on your phone or computer) is quite difficult to get right, so I performed a study to determine what optimal mappings might be. I took 19 participants, and asked them to think up a fitting gesture for 30 common mobile phone tasks (like answering or denying a call, or pausing music). Using the most popular responses, I identified 23 gesture-task mappings that seemed ideal.

To actually prove these mappings were both sensible and sensable  (e.g. can actually be picked up by sensors and used it the real world), I developed a classifier, which is essentially a program which must recognise which gesture is being performed. The classifier worked well, correctly identifying each gesture 87% of the time, and therefore validated the choice of the gesture-task mappings.

Performing this research was a lot of fun, and I hope to return to it soon. I’ve applied to look further into foot gestures (and whole body gestures) with a PhD in Computer Science or Communications, so we’ll see how that goes!