MindWave Cat Ears: Interview With Josh DiMauro
Earlier this year, videos circulated revealing prototypes of Japanese “nekomimi”1 — robotic, wearable cat costume ears that bend, twist and re-orient based on a wearer’s mood and brain activity. This impressive, if whimsical, engineering feat makes a great deal more sense within the context of the endless repetition of the motif of catgirls in Japanese manga, anime and cosplay.2 But even beyond these cultural associations, anyone who has spent time with a housecat has probably noticed the expressive qualities of cat ears.
Ever since the high-end tech demos surfaced, DIY Makers world-over have been experimenting with how to accomplish this type of project using components at least a few orders of magnitude cheaper than research-grade brain scanning equipment. In fact, MakerBot R&D staffer and anime-fan Benjamin Rockhold has a folder full of mechanical and arduino sketches to address this very chibi-awesome design challenge.3
Well, MakerBot Operator Josh DiMauro not only beat everyone to the punch, he has brought the whole project in at a price that has kickstarted an entire branch of DIY, affordable brain-mappable appendages with his MindWave Cat Ears project on Thingiverse. And he was able to accomplish his mission quickly over the course of scores of iterations thanks to his MakerBot Cupcake!
Follow below the fold for a quick interview with Josh in the wake the posting of his own tech demo video (at the head of this article).
Interview with Josh DiMauro
MakerBot: Can you tell us a bit about how you tackled this project?
Josh DiMauro: The approach is the hard part of anything like this. It’s very easy to look at it and go, “Oh hell, I’m never going to be able to do this.” But any big project is made of lots of little bits, and those little bits are a lot more manageable. So to begin with, I asked myself, “what are the pieces of this thing that can be worked on as their own problem?”
For example, the headset returns a variable from 1-100, called “attention,” so I could work on the control software on its own, without needing the Arduino to talk to the headset. And the rest of the project had a few similarly loosely-coupled components, so I broke it down like this:
- Designing the physical mounting and movement of the ears: how they would sit on the headset and move around when the black box of the control circuit was complete.
- Designing the control circuit itself: how the black box of a set of 2-4 servo motors would be wired to a power supply and Arduino, and how the whole thing might sit on the head and human body.
- Programming the arduino to receive a value from 1-100 (attention) and control a set of servos.
- Connecting the MindWave headset to an arduino and have it provide the control variables.
These pieces all affected each other as I worked, but each problem domain seemed a lot easier to manage than the whole thing by itself. As I got more involved in them, I ended up breaking the project down into sub-components. I manage projects in OmniFocus, and the project tree for this thing ended up having something like sixteen or twenty major branches. It was pretty crazy.
Another bonus to this approach is, I’m lazy and easily distractible. So when I got frustrated or bored with one problem, I could move on to the next one, and when I hit a wall with that, moved on, and so on.
MakerBot: And you used a MakerBot Cupcake to prototype this project?
The beauty of using a MakerBot for this project was that I could custom-fit the mounts to the headset, and fine-tune the positioning and angle of the ears as I continued to design. I used my Cupcake to make the mounting hardware for the servo motors first, and then hot-glued cardboard to the servo horns to fake out the ears, so that I could play with different movement styles and patterns.
Fussing with those mounts was really my main area of innovation in terms of physical design. I’ve got well over a dozen iterations under my belt, and I’m still not happy with them. I couldn’t possibly have done that without my MakerBot.
MakerBot: What was your biggest surprise when interacting with your completed cat-ears prototype?
The biggest surprise was how different it is to have the ears on your head, where you can’t see them, versus watching them on a dummy head while you wear a headset and control them. It’s very easy to control the ears when you’re watching them. When they’re just on your head, there’s no feedback about their position. I’m thinking of ways to add feedback.
MakerBot: Where are you taking this project next?
For features, I’d like to add a pair of small microphones, and have the ear closest to a loud noise track towards it, like a cat’s ear does.
But the main thing I’m working on is making the control circuit smaller and simpler. Right now, the headset communicates wirelessly via a USB dongle, which needs 3 volts (rather than the arduino’s 5 volts), and that adds a lot of complexity and pain. It’s easy when you’re using a desktop arduino, like you can see here and here.
[flickrslideshow acct_name=”jazzmasterson” id=”72157627137879911″ width=”700″ height=”395″]
But it’s a REAL pain when you’re trying to run this thing off batteries. So my next design change is, I’ve cracked open the MindWave headset and soldered a terminal wire to the “ThinkGear” chip, and have successfully controlled the ears with it. That cuts like 70% of the electronics hardware and complexity out.
Of course, it voids your warranty, too. So there’s a trade-off.
My goal with this was mainly to see if it was possible, and now I’m less interested in the cat ears themselves, than I am in making this into more of a generalized toolkit.
Thanks again, Josh, for talking with us about this project. We look forward to seeing further iterations and the whole swath of Yoda-ears, bunny-ears, elf-ears, etc. that will no doubt follow in the wake of your great project!
Error - could not find Thing 13390.