On beaches, in parks, and in [BDM]’s back yard, there’s a lot of liter everywhere. The normal solution to this problem is to hire someone or find some volunteers to pick up all this trash. We’re living in the future, though, and that means robots. For his Hackaday Prize entry, [BDM] is building a robot that picks up trash.
A robot that picks up litter is a very, very interesting problem. It can’t be controlled by a person, or else it would be more efficient to just get out there and kill your back picking up bottles. This means it must work autonomously, and that means identifying litter, picking it up, and disposing of it.
For the identification part of the problem, [BDM] is using computer vision that captures an RGB image and discriminates against natural objects. Right now the computer vision is far from perfect, but it does a very good job, all things considering.
The next biggest problem is picking the trash up and disposing of it. For this, [BDM] has repurposed a Power Wheels and attached a DIY robot arm. It’s not a very powerful arm, and a children’s toy probably isn’t the best platform, but it is the start of something very, very cool.
You can check out [BDM]’s video for the project below.
Very cool idea, would be very useful in the windiest places to clean up the fences at trash dumps.
Afterwards you have to sort the trash, we dont want big unknown waste dumps anymore. http://zenrobotics.com
Reminds me of this:
https://youtu.be/te47vQJ8J4s?t=1m35s
HA, that is about what it feels like building this project. It is my first robotic attempt and what a learning experience.
That is a funny video and about what it feels like trying to build a robot. There is always something to tweak and improve.
You’re crazy Benchoff- a PowerWheel is probably the best platform he could be working from. I certainly can’t think of anything better that has more working for it than what he’s using.
(size) is about perfect for adding gear like more batteries. The pickup model is great because you have a built-in place for the collected (payload). Human reproductive and growth rates assures that a supply that has little more wrong than having been outgrown is (readily available). It’s (rugged) enough to handle most of the terrain that ot could be expected to experience. It’s already electric which helps with interfacing with the (control) system… travel is pretty much done, just need to tackle steering.
Add more batteries for weight and runtime, maybe some cheap harbor freight tube tires for steering and handling… if it’s going to run on grass with hills I’d probably stud the tires with short screws to really take advantage of the added battery weight… definately need a more robust arm eventually, but maybe make it side mount to make it easier to reach the bed.
If you wanted to get serious about it add an automatic recharging system and a conveyor belt in the bottom of the bed and a hopper that feeds into a sorting system with autonomous bins that take themselves to the curb on trash morning and back to the sort area once they detect being picked up and emptied.
Anyone know where I can find a PowerWheel for cheap?
Craigslist, ask neighbors to keep an eye out, or roaming neighborhoods on heavy trash day are easy ways to find one. For me it was the easiest robotic platform w/o spending $$$$.
Addition of trash pickup robot resulted in neighborhood kids throwing more trash in backyard, to watch cool robot work, and see what it would or would not pick up. Today trash bin contained a bobcat. Would not buy again.
Pick-up Roadkill – put in my mouth. Yum.