Pokemon Go and Education

“Gee offered the observation that, though the USA had long suffered from a stubborn academic achievement gap between poor minority children and white middle-class kids, he knew of no “Pokémon gap.” – USAToday

Update! With new ‘native’ AR APIs for Android and iOS for single camera devices, we’ll be seeing these tools get better and better, and with more battery life. BTW, Unity3D has a plugin for the iOS ARKit and will surely have an Android version soon.

The devices are ubiquitous. And behind closed doors and walled off areas, educational opportunities abound. For developers such as myself who want to help educate and perhaps make the world a better place outside of the trouble riddled world of public education, producing an augmented reality approach to education seems to be an obvious route.

Pokemon Go smartly generates content around the globe algorithmically. I need to think about how such a thing could be done for content that is very particular to the location. Wouldn’t it be great if user generated content could come into play here?

Some ideas :

• Historical markers and other points of interest could be augmented to show speeches and moments along the timeline where important things occurred. How could that be user generated?
• Layers of rock in a cliffside could be augmented to show the layer where the KT event occurred and show the abundance of life before and the relatively quiet period after.
• Outside of an electronics station or power plant, students can learn about the insides of the plant and satisfy their interests about how power is produced there.
• Hooking up with a site like Coursera would allow course content to be developed around a single topic which would be augmented to show meaningful real-world things and how they work.

Problem points

1. The devices which are ubiquitous have capability for image tracking, imprecise gps, and an IMU. Pokemon GO likely has an accuracy of around 9 meters. This isn’t good enough for pinpointing interesting things. Notice – in GO, you can’t walk around a Pokemon. This is because the tracking is poor – likely something like a Vuforia user generated image target, which gives loose tracking to whatever is in front of the camera. These image targets are shareable but are likely very hard to reproduce. Lets say a person takes a snapshot of a rock formation. That image target uses points of contrast (or perhaps a SLAM approach, which is a combo of GPS + CV https://en.wikipedia.org/wiki/Simultaneous_localization_and_mapping). Changes to the environment such as lighting or plant growth would dramatically make such a thing hard to place. Pokemon did it right though.. a gps coord + a loose tracking may just work.  (edit – this is no longer the case, as SLAM tracking has gotten really good!)
2. Too much content. We are talking about a world worth of content. Server infrastructure alone could break this (this is likely workable..). However, developing content for this as a studio would be impossible.  So user generated content is likely the only way to go!

I want to think more on this. Please stay tuned.