So I was wondering if anybody was familiar with robots that can sweep, avoid objects, and place dirt/junk into a single location.
This would be for industrial use.
Any ideas?
This would be for industrial use.
Any ideas?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Robots
- Thread starter eric/
- Start date
- Sort by reaction score
Yeah, but they appear to be vacuum cleaners more than anything. I need large pieces of things pushed into a shoot on the floor. Very light, , but ranging in size from 4 feet to dust size.
Lego Mindstorms could do that depending on the weight. They have IR sensors and can detect range via bounceback. So in the sense they can avoid obstacles. That's all the Roomba does anyway. There is a more advanced C based language you can use for them too in order to make them do more rigorous stuff or potentially map out the locale for it. As long as the motor can push the load, they could do the job.
Well, the thing is that it's not for personal use. It's for industrial use in an actual factory. So although I love Legos, I don't think they will work in this case.
Automatic.
----------
We'd have to do an evaluation. but really you could just put like a 3 foot orange flag on the top of it that collapses or something when it goes under things and it'll be just fine.
Yeah I make the best threads around
All I know of offhand for mobile robots is the Kiva system, which Amazon purchased last year. The company, I mean - Amazon purchased the whole Kiva company. But they still support other clients. (Gilt, Office Depot, Staples, Crate & Barrel, Toys R Us, and Saks Fifth Avenue.) I suppose it's possible to remove the pallet lifters, install pushers, and repurpose the control software. I can't find any information on cost, beyond "millions." Sounds like most Kiva customers use hundreds or thousands of the units, rather than one or two.
You would likely have to design it yourself (I would probably go AVR rather than mess with lego stuff on a personal project) along with an IR base so it could find its way home. Unless you are thinking of diving into the world of robotics yourself it would probably be cheaper and easier to buy a broom.
Oh yeah the guys on the floor have brooms and everything. I was just looking to see if there were any options. I might try and get a team together at my university to see if there is any interest in building one. I think it has some potential, at least for use here, and I don't think it needs to be expensive. We'll see.
Microcontrollers are cheap (30-40). Sensors will be cheap as well, motors, battery, and fabricating the actual design will probably be the most expensive.
I think actually the software necessary to get it to do what I want would be the most expensive, or most difficult.
In theory if you had GPS on it and some IR sensors, it could automatically map out the room fairly easily. You'd want to position the IR sensors just above the height of the debris. Probably the most difficult part would be ensuring that the debris was pushed towards a central location instead of just up against whatever the robot was travelling towards. But that shouldn't require too much effort. So long as it knew to push towards the central area. You could also have IR sensors lower down that could see if there was any debris in front of it. Then it wouldn't need to sweep areas with no debris. Of course it always sounds easier on paper I suppose.
If you have connections at the university it would make a good EE\CS project.
I've never had a Roomba but I think people think its doing more than it actually is. Unless they have updated it recently it doesn't even map the rooms, it just darts back and forth using collision detection until it has decided its covered enough ground and continues to move until it senses the ir from the base station.
If you have a set area where these scraps fall I would think you could fence the bot in with infrared walls and let it bounce around until complete and then look for the ir signal from a base station.
----------
Truth. Lol
Those are good points. But even the software to make it detect and not run into IR sensors has to be pretty complicated for university level stuff. Idk, we'll see. First problem is finding people who are interested in building something like this. 
Those are good points. But even the software to make it detect and not run into IR sensors has to be pretty complicated for university level stuff. Idk, we'll see. First problem is finding people who are interested in building something like this.
Going back to Lego Mindstorms, all they do is have an IR emitter and receiver. It shoots out an IR beam and determines how long it takes it to get back to figure out distance to the object in front of it. You program it to say "If you're in 5 inches, stop, turn around" or whatever. Mindstorms have everything you need at a small scale. You should buy a set to try it out yourself. It's stupid easy. Plus you get to buy some fancy Lego!
Going back to Lego Mindstorms, all they do is have an IR emitter and receiver. It shoots out an IR beam and determines how long it takes it to get back to figure out distance to the object in front of it. You program it to say "If you're in 5 inches, stop, turn around" or whatever. Mindstorms have everything you need at a small scale. You should buy a set to try it out yourself. It's stupid easy. Plus you get to buy some fancy Lego!
While I think that's really cool, the problem is this is for large-scale, industrial use. hell, if I could get something good made than it would perhaps be something I'd have patented.
Oh I know it's for large scale. But you can make a proof of concept with Lego and scale down your debris/warehouse to match.
Heck! http://www.robotgames.com/wp-content/uploads/2012/01/WCRG-2012-LEGO-Mindstorms-Landslide-Rev11.pdf
That is true. I'll have to see. Idk how much I'm really going to pursue this. I just wanted to bounce some ideas off and/or get info, which you guys have fortunately provided.