Arduino animatronics- make your awesome costumes more awesome!

Step 15 FAQs

Frequently asked questions-

Do you offer a servo board kit?
Nope- at least not yet. I might be coaxed into making a fully assembled board though...just message me if you want one.

Will you do custom work?
On occasion- message me to discuss your project.

I need help with my own code/project- is there a forum for help?
The Arduino forum is the place to go-
http://arduino.cc/cgi-bin/yabb2/YaBB.pl


The Predator setup isn't really a true head tracking system- why?
When I first started looking at this I had just envisioned a simple system that was controlled using bend sensors on your fingers and it wouldn't be able to look up and down. That really was a bit cumbersome to operate and wasn't very natural.

Next I looked at keeping the bend sensor that would raise the cannon, but I figured out a system that had a cable mounted under the mask that turned a pulley/spring loaded potentiometer (mounted under the cannon mechanism) that would allow the cannon to follow your head rotation. It still wasn't able to look up/down. Now I got a natural (and smooth) rotation but fixing/routing the cable was difficult and cumbersome. Mask removal could also be a real problem.

Then I started looking at different kinds of sensors. Gyroscopic, compass and accelerometers. Combo boards, IMUs (inertial measurement units), etc. I have a friend that is a data acquisition engineer as well as a friend that is a programmer and I spoke to several electrical engineers for help. Compasses are affected by tilt, so you have to compensate for that and then they don't work when you rotate your whole body vs. just your head, so dual compasses would be necessary and you would have to write code to differentiate between the two.

The problem with gyros is that you need two of them, preferably with an accelerometer to account for drift as well as front/rear tilt (pitch). One gyro really won't work well because it will respond to whole body rotation as well as just head rotation, so you would need one IMU at the head and another at the cannon base and then you have to write some code to differentiate between the two. Gyros really only sense change in angular velocity so its output only changes with respect to movement and it will reset itself once you stop rotating your body- the cannon will quickly become out of sync with your movements. That's why to really do it right you need a multiple DOF (degree of freedom) IMU and you need two of them for true head tracking. Most of the systems I've seen that people post online as a head tracking system for R/C or video control that use these types of sensors have pretty poor yaw control and I think this is why. And they don't even have to deal with the body vs. head rotation issue as they can sit in a chair- not an option for costuming.

There are IMUs and digital compasses available now that have on board processing and are programmed to account for drift and tilt so some of the hard work is done for you but you would still have to factor in the difference output of two of them and then generate your necessary servo movement values from that. It can be done but it's pretty darn expensive.

Most of the solutions I found were pretty complex in terms of programming requiring sophisticated algorithms and/or extremely expensive on the hardware side. There are also pretty severe space restrictions for fitting sensors inside the Predator Bio helmet as some of the IMUs available are pretty large physically. 

Then I found that I could modify the sensor board out of a Wii nunchuck controller and interface it with a microcontroller over its I²C bus. This provided me with an inexpensive and reliable multi axis accelerometer and two finger pushbuttons with an easy to use interface. I ended up writing code that would allow me to access all of the nunchuck's functions- the accelerometer outputs, the joystick and the two pushbuttons.

When it was all said and done the rotation was still a bit of a problem as the accelerometer only really gives you a stable output with respect to gravity so you have to tilt it to get the rotation. What I found was that if I mounted it as level as possible in Predator helmet it really didn't need much tilt at all to get a stable rotation. The beauty of this system is that there are only only two finger buttons to control everything- it's also pretty easy for me to modify the code. I haven't yet taken apart a Wii Motion Plus to combine it with a nunchuck to create an inexpensive IMU but I'll post the results when I do.

What I did was to basically fake it by strictly using the accelerometer inputs- when looking sideways you just need to tilt your head to the side like a bird for yaw (rotation) control. The accelerometer reads the tilt and moves everything accordingly- the pitch function is normal. It does take a bit of getting used to but after a while it becomes more natural and it's very convincing.

For costuming all you need to do is provide the illusion. When I started the project with a friend three years ago (he's sculpting the Pedator backpack) we wanted something that anyone could build from readily available building blocks to add some cool animatronics to their Predator costume without spending a ton of cash.

Whew...

I want to power my project and I only want to use one battery to supply different voltages- what do I do?
I would use a switching step down voltage regulator like this one-
http://www.pololu.com/catalog/product/2110
It's far more efficient than using a traditional voltage regulator, especially if you need to use something like a 12V battery and drop it down to 5V for servo power.  If you need to drop 5V to 3.3V then a traditional voltage regulator is fine. Just don't plug it into the JST connector on the controller board and then plug in a USB cable into the USB mini-B port for the charging circuit- that would be bad.