Put wild animals to work for a fee, follow my project, hardware, software AND, how I trained the bird.
Things used in this project
The entire project depends on mechanical design, electronics, software, and the fantastic opportunity to build a machine using parts from my 3D printer. I have trained wild magpies in my garden to exchange litter for food. I have been working on this project for a number of years, but it has had an exciting development recently, as I mention in the title. Now the magpies work as garbage collectors, payed with food!.
I work everyday with industrial applications of artificial intelligence (AI). so the whole rig for the magpie project must be:
Fully remote controlled, autonomous, dependable, adaptable, and capable of recording and logging both data and video
I spent remarkably much time creating the actual food dispenser.
After testing with several Thingiverse designs of pet feeders, often based on a rotating feeder screw, I left these techniques. The dispenser must never get stuck and I want to be able to feed out individual peanuts. I looked at industrial solutions and got hooked on vibrating feeders. The base of the feeder is a fantastic vibration feeder from thingiverse. https://www.thingiverse.com/thing:2118961
For the four dog-bone pieces, select ABS or PETG; PLA does not last very long.
To this design I have added a number of extensions and funnels, I upload these as STL files. But, these are quick and dirty designs, you have to attach the parts on your own with small sheet-metal sheets, or in another way. Sorry for this, but when I build I am often experimenting and usually dont know exactly what the solution will look like… My parts to extend the vibro feeder can be found here:https://www.thingiverse.com/thing:4601125
One of these funnel sections has a 5mm IR-LED and photodiode socket. With this setup, I can turn up the intensity of the dispenser until I see a peanut, at which point I can turn off the vibrating motor.
The vibrating motor is a simple 4.5 v DC motor. I made an unbalanced flywheel to this motor, included here https://www.thingiverse.com/thing:4601125 as STL The BirdBox is controlled by a Raspberry Pi 4 with an add-on board. The add-on board includes a DC motor controller.
Theres also two Arduinos in the design. A Python application for the raspberry is in charge of the feeder’s logic. I use AnyDesk and the VNC feature to get complete remote access to the BirdBox. A straightforward GUI is included in the Python code so that I can monitor and record the BirdBox’s progress and status.
An Arduino code-based 3D-printed metal detector is used to identify and accept the bottle caps.
I’ve developed a basic graphical user interface (GUI) for the Raspberry Pi, which I’ve combined with Anydesk and a wide-angle Picam to monitor the experiments. This has been really helpful because I’ve had to conduct the experiments entirely on my own, which involves setting up the experiment, leaving for work, and then checking the results later that day.
Later on this page, I will go over the electronics and software. History of the project, and how I trained the magpies. The project has been running for several years. The video clips that are displayed on this page are the outcome of numerous steps and tenacious work. The first step is to arouse the birds’ curiosity and familiarity with the feeder. I advise anyone considering conducting comparable research to begin with that For your site to be included in the birds’ patrol schedule, you must feed them on a regular basis. For this reason, I begin to publish the designs that are required for this first stage. The feeder and an experiment site similar to the standard food bowl found in the STL files
https://youtu. be/zvmhl-E4QQ0 I would love to tell everyone who is interested about this project and my experiences. What might be discovered through comparable studies in which a sizable community participates?
Now I’m working really hard to get my project files and other stuff ready to share. more to come.
The next scheduled action is to try to encourage the birds to gather fallen fruit from my garden. Are apples to heavy ?.
I believe that training for practically anything would be feasible in the long run. Bottle-caps was relatively easy to classify with a metal-detector. Other objects, like slugs, candy paper, and cigarette butts, will need a more universal classification system. Fortunately I spend my days with applied AI and machine-learning. In my line of work, we create intelligent sensors using tiny Running such networks on tiny platforms like the RPI 4 is totally doable. This could serve as the basis for a general classification technique for a device such as the BirdBox.
A common question is. In this video, I attempt to explain how you trained the birds:
Here are a few videos that go over the system in more detail:
Bottlecaps? I can categorize and identify bottle caps at this time. This is done with a dedicated, arduino-powered, metaldetector. The description of that system is as follows: In parallel with this, I’m developing an AI-powered general-purpose detector that uses cameras. I want to use this to experiment with things like slugs, dandelions, and cigarette butts. etc.
Part two of of metal-detector
What’s your next step? is a question I get asked a lot. To put it briefly, I’m currently attempting to see if I can teach the birds to gather soda cans. I just need to make a new version of the metal detector and an implementable training program in order to accomplish this. I describe my attempts so far in these videos.
A little about myself: My parents gave me an electric building kit for my sixth birthday. This ignited a passion for electronics, robotics, and artificial intelligence (AI) that has lasted a lifetime.
I have fifteen years of experience researching and developing robotic lawn mowers, so I am very knowledgeable about positioning and autonomous systems. We are using AI with Deep Learning (DL) in an ongoing project to enhance an autonomous product’s current customer value. Deep Learning offers incredible opportunities in the most diverse industries! Performance in this Neural Network surpasses everything I’ve seen before! Opportunities I want to develop further!.
My lectures and speeches are appreciated. I live with my family in the outskirts of Gothenburg. I like to spend my free time with friends, music, tennis, downhill skiing, and classic Italian sports cars.
In nature, corvids collect and store food for later use. Additionally, they see to it that the food is consumed before it goes bad. They are used to defending their caches against thieves. But according to a journal Science article, they are not above committing a little theft themselves and are able to anticipate what other ravens will do.
After an hour, the birds had the opportunity to choose a token from three unimportant, distracting items that were placed on a tray. These trays could hold up to three tokens because they were presented three times in a row. A different researcher who had no prior experience exchanging tokens for food made these available.
Fifteen minutes later the apparatus was brought back in. The ravens selected the appropriate tool to open it most of the time. One resourceful woman figured out how to open the box without using a tool. After her removal from the test, 86% of the birds that remained chose the correct tool.
The cockatoo is another intelligent feathered friend that possesses the self-control to refuse food that is right in front of it if waiting is rewarded with an even better morsel. This kind of parrot is also capable of picking the proper tool to open a box; if the incorrect tool is provided, it won’t be duped.
FAQ
Can you train birds to find money?
Can you barter with Ravens?
Do birds know how do you trade?
What birds trade with humans?