With a title as absurd as AI Powered Chicken Targeting Water Turret, one might suspect clickbait, but I assure you that you'll find none of that here. What you are about to read should be read with caution. Everything I'm sharing about this project and the process is for educational purposes only. I can't make such a project without recognizing what technology like this could be used for. That being said, I share this because it's fun, it's exciting, and my article is hardly the first to outline the process of making such a device.
Figure 1: My husband with our chickens
Now that the important topics have been discussed, let's dive into what I'm making here: an automated turret. Why? Because my chickens keep pecking my house, and I have been yelling at them and shooing them away for two years now. I like fresh eggs, I like the company of chickens, but I don't want to be paying for new sidewall every other year on my house. There was a time where I tried using a pump based water sprayer to scare them away, and while effective, it was a very physically intensive and slow process, and it wasn't always the offending chicken that got sprayed.
What I want to create is true farmyard justice. There should be a camera scanning the perimeter, and when a chicken gets within a certain distance of the house, just spray them. We don't cluck negotiations with chickens who are out to destroy my home.
Silly, yet serious, introductions out of the way, I started to research how something like this would be done. You can't just buy turrets on the internet, or if you can I certainly don't want to be put on the list of people looking for that. I also don't feel entirely comfortable just googling how to make an automatic turret, so a lot of this research has been trial and error.
I had a basic list of requirements that I could think of: The system needed to have two axis of rotation, up and down as well as left and right; To spray chickens with water, I needed some sort of spraying system; There would need to be some sort of targeting system, and since I'm familiar with AI object recognition I decided to go with a camera; Finally, I would need a computer system of sorts to control everything.
Now, this article isn't a record of the first prototype for this project. Officially version two, the first prototype of this system was made with cardboard, household DC motors, and controls from an old washing machine months ago. If you're interested in that version, you can check it out on TikTok. That being said, version two has been a much more intensive process, involving a ton of research, development, mistakes, and learning. I would also say that this latest version is one well worthy of proper documenting.
Figure 2: Prototype Version 1, small scale model of dual rotation design made from trash
Version two of the prototype was mostly designed in FreeCAD, a free and open source 3D modeling software. I had been practicing with FreeCAD for a few months when I started this project, and it seemed like a perfectly usable platform to mock everything up in before building it. As you can see in the image below, this involved the creation of a large number of complex and moving parts.
Figure 3: 3D model of two-dimensional rotation system
There are two main structures conjoined together in this part of the design. There is a lazy-susan acting as the horizontal rotational support, and a platform connected to two gears for vertical rotation. Surrounding these components are a number of brackets, various styles of bearings, some axles for support, and motors to power it all. You'll also note that there are no feet in this design to support the structure. This is because I used some old wooden couch feed I had lying around to hold everything up. For the feet, I got some rubber pads to screw into the bottom for extra stability and vibration reduction, and then made some adapter plates that attaches the top of the feet to the bottom of the horizontal motor mount.
Figure 4: The device so far
After around a month of work, this is the general structure that I've come up with. You'll also notice that there's an extra compartment on the bottom that's bolted into place between the bottom motor mount and the feet. This is to hold some of the electronics, but you'll have to wait until the next post to learn more about that! For now, I have a system that can rotate in two directions, but I still need to mount the camera and water jet to the rotating plate, as well as a decent bit more work with the electronics. All that being said, I feel great about what I've gotten done so far!
This is a personal project, so it gets worked on when I'm able to between client work, but it's something I've really enjoyed and want to continue. That being said, I have gotten this far with just my spare time between client projects. If you're interested in seeing what kind of work I can get done when I'm actually being paid, reach out via the contact form below!