Robotic Following Cart to Replace Light Duty Tractor Tasks?

[candersen10]

Would the cart have the ability to follow a user through a field, create a map of where it follow and them autonomously repeat that mapped trial through the field? If it can autonomously repeat a mapped route of field then I could see potential application for sprayer application for an aqueous insecticide or fertilizer. Heck it may be stretch but it could be used to deliver metered water to crops in a dry/drought situation to extend the crops life...I couldn't see it engineering enough water to keep crops gong during a drought but it it could extend their life another week or two for better weather to hit there may be a market.

mcj115 - our idea is that Burro would follow a user to record a path, and could then simply re-run that path with basic obstacle avoidance (i.e. if something was in front of Burro, he would stop and wait for that thing that moved into the path to move out of the way again). Burro is relatively small, but certainly I'd imagine he could run a small sprayer on the back for that type of thing (spraying or watering on a relatively human scale).

We believe there are infinite applications for Burro, but our target msrp of $4.5-5K is as low as we can build Burro with what is available today hence the quest to maximize utility and handiness to a customer base with a real pain point for this type of repetitive small-payload shuttling labor.

 

[mcj115]

I would also recommend there be the option for an automated switch control also be added to mapped route to operate as solenoid or actuator.

 

[candersen10]

2WD with rear caster versus 4WD? What are thoughts here on the merits of each?

2WD takes us to around $5K MSRP. 4WD takes us to around $6K. That's with following, and likely with path retrace.

Scale is 24 inches wide, 13 inch drive wheels in both cases, 2WD we would likely run a 24 x 36 inch bed, 4WD we would aim for a 24x48 inch bed.

 

[daugen]

if this is going to be used outdoors, 4wd would eliminate many of your traction concerns, because 2wd is really only one tire drive. I doubt your little rig
will have limited slip but i would be interested in seeing the steering gear, which assuming birectional operation would be similar on both ends.
Plus 4wd allows you to offer a larger bed surface.
Indoors on hard floors no reason for 4wd usually so it would add cost unnecessarily to the product.

I'd worry about that caster getting hung with something, like a hose.

 

[BigToeTN]

I like where you are going with this, however I have a few thoughts:
1) It appears the following and route assignment technology is the value and can be applied to just about any cart design.
2) The price seems high because most of us hobby farmers just use our multi purpose tractors that can't be replace with this cart.
3) Route memorization has potential to be significantly valuable to the right kind of orchard or plant nursery farmers, maybe even certain soft vegetable crop farmers who manually harvest.
4) I would spend time with the specialty farmers perhaps even work with them for a few weekends to see if you can lock down a specific need for repetitive trips to and from say a barn or during harvest to haul baskets of produce in and out of rows to a central truck loading station. I would then design the cart size and capacity around their needs as they perform these task day in and day out so $5000 for this cart wouldn't seem too much if it can pay for itself in productivity over say 4-8 years.
5) You may need to self fund and trial a prototype with a grower so you have proof of concept, a reference, and proven labor savings to other growers/potential buyers.
6) $10,000 is a relatively low amount to crowd fund, if you connect with several of these growers and get their interest, they may each contribute if they think it will save them money down the road

Hope this helps,

 

[candersen10]

if this is going to be used outdoors, 4wd would eliminate many of your traction concerns, because 2wd is really only one tire drive. I doubt your little rig
will have limited slip but i would be interested in seeing the steering gear, which assuming birectional operation would be similar on both ends.
Plus 4wd allows you to offer a larger bed surface.
Indoors on hard floors no reason for 4wd usually so it would add cost unnecessarily to the product.

I'd worry about that caster getting hung with something, like a hose.

It is 100% electric drive with drive motors for each wheel, so there is not an issue with just one wheel getting power or any sort of differential.

 

[daugen]

It is 100% electric drive with drive motors for each wheel, so there is not an issue with just one wheel getting power or any sort of differential.

3 or 4 tire drive is darn good. Just ask audi and subaru...electric motors with appropriate controller software actually makes this simpler. Now controller software much less radar guided software with lots of protective devices built in is not simple at all to most of us, but that seems to be what you know well.
I agree you need to build one and have someone demo it, give them one free someday if need be but references and pictures of machines in action vs glossy marketing sheets
will get you phone calls. From then on you have a true educational sale.

route memorization and gps guidance may not be plug and play out there but it's sure been around awhile in large ag, so perhaps trickle down will bring
advanced features at not too high incorporation cost. You can always get the client to buy an advanced "tech package", maybe a mandatory option for many
but keeps the basic model cost down

 

[candersen10]

4WD chassis. Still working on the coding...

 

[quicksandfarmer]

View attachment 501897

4WD chassis. Still working on the coding...

If you want to quickly get something out in front of potential users I suggest you "mechanical Turk" it. That means have a human pretending to be a computer operate it. (Named in honor of an 18th century fraudulent chess playing machine -- The Turk - Wikipedia ). I don't mean that you would deceive anyone, but you would write up a list of its capabilties, and then have a human operator who is told he can only do the things on the list. Then have him work with a potential user for a few days and ask him if he feels it's useful. You'll learn a lot more in a few days than you would building and debugging prototypes and field-testing them.

It looks like you could easily fit it with a controller for an RC car. Depending on what type of computer you are using you could probably put in control over bluetooth or WIFI. As you figure out what sensors it needs you can add the sensors and tell the human operator he can only react to the sensors, not his own eyes and ears.

 

[daugen]

http://www.mobileye.com/our-technology/
I bet these folks have some coding for you...likely for an indigestible cost

I like the 4wd prototype, keep at it.
I'd like one programmed to go up and down my orchard rows of blooming fruit trees and spray water on them each night in the freezes ahead.