Digitally Augmented Skateboard

In this the Selfie Age (perhaps not so much ‘The “Anthropocene’, but ‘The Egocene’?), it has become normal to record and broadcast your own life, including your recreation activities. It didn’t happen if its not on video (and on the Internet). We can quibble about the merits of this practice (personally, I think that turning everything into a superhero/action flic rather detracts from actually getting good and enjoying the sport), but it is at least extending the self-expression inherent in play.

But the increasing use of sensors and data processing to monitor (and, inexplicably, broadcast) exercise and play is a bit more problematic.

Last month Benjamin H. Groh and colleagues at Friedrich-Alexander-Universität Erlangen-Nürnberg published anarticle about “Wearable Real-Time Skateboard Trick Visualization” <<link>>. That’s right, the quintessential laid-back, let-it-flow avocation, “augmented” with sensors to track, in sub-millimeter precision, the skateboard. Sigh.

https://www.computer.org/cga/2016/12/06/wearable-real-time-skateboard-trick-visualization-and-its-community-perception/

The research employs sophisticated wireless motion detectors that can measure and report in realtime the 3D position and movements of the skateboard as the skater performs complex tricks.

Their approach is to detect the (correct) execution of various standard moves including various flips. To do this, the rich data streams were fed to a Naïve Bayes classifier, which learned to distinguish each identified stunt. These classifications are used to record the successful completion of particular tricks, which would allow an automatic record of a sequence of stunts, even if human observers cannot see.

The project also created a visualization which illustrates the path and behavior of the board, based on the sensor readings. The visualization elucidates the path of the board, which provides visual feedback for the skater.

Finally, the researchers surveyed a small sample of skaters, to explore the reception of the system. In general, the opportunity for feedback was welcomed, though more information from the sensors is desired. Not surprisingly, it is less clear whether the classification system should be used in competition.

My own reaction was, first, “why?” Reading the article, I can now see that, with work, the visualization system might be quite useful as an aide to practice, especially id it can be delivered instantly to a mobile device or even eyewear. (Maybe it would be marketed for a helmet + visor system, sneakily enticing skaters to wear a helmet, for goodness sakes.)

I am not a skateboarder, but I see little point in competitive skating, or, for that matter, in some kind of rigorous definition of skateboard tricks for any reason. So the classification scheme looks pointless to me.

I would note that there is some question in my mind just how robust their classifier really is. From experience I know that you need to have a good training sample, with enough diversity, or else you can easily overlearn the specific training set.

In regular English: if they only had one or a few example, the computer may have learned to recognize how these guys do the flips. That may ore may not be accurate for other people, it’s hard to know without experimenting.

As far as the community survey, I chuckled at the phrasing of the item, “I do not fear that technology might strongly invade in sports.” Fear? Skateboarders “afraid”? Skateboarders saying they are afraid? Never!

Regardless of “fear”, my point above indicates a very important caveat: it is important to validate that the classifier is actually measuring what you want it to measure. Are you trying to measure something that is clearly defined in terms of rotation of the board, etc.? Or something that looks close enough, and is delivered with panache? Or something that has more than one “right way” to do it? I don’t think these questions have been asked or answered, so it’s not possible to say the software is working or not.


Of course, this technology could be combined with other technologies.  For example, the sensor traces could automatically trigger capture and upload to social media, zinging a short video of every good flip out to your (soon to block you) friends.

The sensor data also might feed into a music synthesis or mixing program, so you can “direct” a sound track or even “play” the synthesizer with your skateboard moves.  Combined with your friends, you could have a live concert by an orchestra of skateboarders.


Overall, this is an interesting project, and it shows how much sophisticated sensing and data processing is in the reach of ordinary users. The particular application will be useful for skill development, but probably not for judging.  And it might be combined with other technology to create some very interesting “immersive” experiences.


  1. Benjamin H, Groh, Benjamin H., Jasmin Flaschka, Markus Wirth, Thomas Kautz, Martin Fleckenstein, and Bjoern M. Eskofier, Wearable Real-Time Skateboard Trick Visualization and Its Community Perception. IEEE Computer Graphics and Applications, 36 (5):12-18, 2016.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s