Has it really already been five months since the last post? Ugh. Well, nothing unusual for my weblog.
I’ll start with a bit of a catchup on what’s been going on in the meantime…
The Way Way Waaaay Back Machine
After “Rogue Done?“, I decided that once I finished Artie Deco and/or BBL-BBQ, I would divert my energies toward building my own Astromech droid. As you may recall from The AD Project, this was really what I’d wanted all along, but couldn’t afford. My research after Dragon Con has led me to believe an Astromech was closer to being in reach than I’d thought. Still, given my financial limitations, I expect a proper Astromech project will take years.
WED-15-77 on Tatooine
So I also decided to do a shorter project while I researched and planned the astromech: a Treadwell droid. This has long been one of my favorite “other” droids from the original films (after the astromechs). I find something endearing in how ungainly they are.
Before I could get started on that, however, I had to shift my attention to my traditional Halloween prep. Got a late start, but it was going pretty well.
Unfortunately, my printer glitched about a week and change before the big day, in a way that made me think that either the control board or the thermistors had failed. (The latter issue could cause a thermal runaway, and thus a fire, if not caught quickly. Luckily, the system did catch the error and shut it down.) As I was too busy at the time to test for reliability or even basic functionality, I simply decided to take a break from printing and CAD until after Halloween.
Then — between a cold that I started coming down with just before the US election, the election itself, another cold I caught the week before Christmas, and all the nonsense leading up to and following the inauguration — my motivation promptly went into a total nosedive.
It’s only now — more than two-thirds of the way through February — that I’ve finally recovered enough gumption to turn my attention back toward my printer and robot projects again. The printer now appears to be operating normally (based on printing several objects over the last few days), so I’m chalking the incident in October up to a glitch. I’ve taken the opportunity to upgrade the hotend, giving me the ability to print more durable materials.
But I’ve lost months to illness and other distractions. I haven’t done any work on my Space Drones, other than establish that Artie’s head mechanism needs to be revised. I still don’t even have a battery (though one is now on order) or a solid drive train design.
So in order to get moving on Treadwell, I’ve got to put Artie and BBL-BBQ on the back burner for now.
Treadwell
Treadwells are seen a few times in Star Wars (WED-15-I662 is aboard the Jawa Sandcrawler as is the multi-armed Septoid, WED-15-1016 is working on the Falcon on Hoth while Luke and Han are talking, there are several in Clone Wars, etc.)
The epic battle between Assajj Ventress and Treadwell 327-T begins
A Treadwell has even been the star of a parody video.
I have some preliminary thoughts on this droid. Firstly, I’m not planning to do a screen-accurate replica of any particular Treadwell. It will clearly be a Treadwell, but not a specific Treadwell. It will be (a bit) more anthropomorphic than the ones seen in the films, with elements reminiscent of 327-T. Treadwells are fairly diverse, so that’s fine.
Unlike Artie, my current intent is to use brushless motors for the drivetrain. The head will be 3D printed. The post-like body will likely be PVC pipe. The base will probably be some combination of wood and styrene.
Treadwells have more trouble with door sills than Daleks do stairs
One thing I want to do with this droid is to make it expressive. In particular, I want it to not only be capable of head motion, but also “eye” motion (rotating up/down and left/right, and including internal mechanics reminiscent of a camera zoom.
Inspired by James Bruton’s work on Project Ultron, I considered adding a Pixy (camera with onboard object tracking) to the head to permit face tracking once that update became available. However, shortly after I mentioned the possibility to my friends, the page was updated to say the feature was no longer in active development.
As a result, it’s looking like I’ll need to use a Raspberry Pi with OpenCV instead. As an ordinary camera is potentially much cheaper than the pixy, and the Pixy would have needed a microcontroller or a small computer like the Pi as well, this could work out to be about the same price or less.
Like James, I am contemplating having the bot’s expressiveness be determined by an AI. For example, how diligently it focuses on someone can be determined by the AI … does it stare, get distracted, or simply look for something else to pay attention to? One benefit to using a RasPi — I can code this AI in Python.
For controls, I’m currently thinking of something akin to the SHADOW control scheme from the R2 builder’s club, which uses a PS3 move Nav controller (you may recall I mentioned encountering this in previous posts). But as I recall, SHADOW runs on Arduino, so I’ll either have to have the microcontroller AND a Pi, or roll my own code.
I’ve no idea how long this will take.
I’ll periodically update the status here, but if you want the super-detailed version, there’s a build log over on rebeldroids.net
dyne3d 