Small Robot Company - An opportunity for farmers to own shares in us

[Devil's advocate]

Hi - yes this is a fair challenge and one that we get regularly - why are we focusing on boring old wheat when we could be selling these robots into higher value markets like lettuce, celery, onions etc. who all have big weed control issues and much more cash than your average combinable crops farmer?

A few key reasons - firstly, as has been mentioned, is the size of the market opportunity.

What we are trying to do is to develop a system which has the maximum possible impact on the way we use technology to produce food. We were worried that if we designed something for any of the above mentioned crops, we would end up with something too bespoke which then only works for a tiny number of customers. On the flip side of that, if we are able to deliver something that works for wheat then the leap to barley and oats is small (some re-training of the AI required but hardware remains the same). There would then need to be a bit more development done to get to pulses and oilseeds and maize but we would quite quickly be able to get to a solution where farmers do away with existing equipment, leasing robots instead, and re-invest some of that capital in other projects.

The second reason has to do with what we are ultimately trying to achieve with our business, which is the digitisation of the farm. We define digitisation as a per plant view of the field and a per square metre view of the soil. We do not see robots as simply a tool to automate labour. Some automation will happen, but that is not the ultimate objective. The objective is a truly digital view of the field, because once we have that all sorts of new and exciting ways of managing the farm become possible, some of which have been mentioned above in this thread.

The third reason is a more practical one which is that we see wheat as a much better test case for developing a weeding robot than almost anything else. We will be able to do 80% plus of our weeding work between October and March when the canopy is open and the weeds are easier to target. A robot working on a field through this time period will be able to hugely reduce the weed burden in field, but crucially the timeliness of operation is not critical. When the first robots start working in fields, they will be slow. That's why it is a more difficult sell to a lettuce farmer who needs their weeding done TODAY!! The technology has to be more developed to be really useful to those guys, which is not to say that we don't want to serve those customers with aspects of our service in time, but not immediately.



Finally, there's very little competition in this area. There is some, but most other companies see robotics as an automation tool pure and simple. By focusing on wheat, we are operating in a relatively open field (at least for now!)

I'm out, he is turning his back on the only market that will make this tech pay.

 

[Daniel Larn]

I see you're overfunding at the moment, congratulations!

 

[SmallRobotCo]

I like what your doing, when do you visage making a profit?

Next couple of years are understandably about R&D spend, but should be making an operating profit in about 4 years.

 

[SmallRobotCo]

I see you're overfunding at the moment, congratulations!

Thanks! We are indeed overfunding - we went live today and we hit our £500k target within minutes. Lots of farmers are already on board - thank you to those who have already invested.

If you haven't been on yet, there is still plenty of time!

https://www.crowdcube.com/companies/small-robot-company/pitches/b3AQ7l

 

[SmallRobotCo]

 

[Exfarmer]

I think the robot armyis a potentially a hugestep forward in farming today.
Rather like the tractor, its going to take a long time to move the first steam engines to todays tracked monsters.
The first issue i would have qualms about is the insistence on being battery pwered.this ensures that you need a source of such power in the field and also means the robot will spend a lot of its life at the charging point andtravelling too and from. There are simple answers to this but the easiest for a start would be a small petrolengine, possibly as fitted to model aircraft, to supply a cheap and reliable power supply possibly meaning refuelling only once a day or less, and then only taking seconds not tens of minutes.
Next I think although electrocution of plants has potential surely micro jet nozzeles as used in printers using high strength round up will surely be a much quicker development. I know the process of electrocution of weeds has been tried for many years but with extremely variable results.
Of course I would not rule out hoeing for many weeds, but i also know there are many prople with quite advanced working models of these.
For broadacre crops perhaps the style of seed placement could also be looked at, precision placing of seeds will mean the unit has a far easier task.
One thing for certain I do know , there should be a very easily tackled market for such a unit across the country killing broadleaved weeds such as docks and particularly ragwort in established grassland. If it could be just set to work and left to navigate across fields on there own. Look at the thousands of horse paddocks!

 

[e3120]

One thing for certain I do know , there should be a very easily tackled market for such a unit across the country killing broadleaved weeds such as docks and particularly ragwort in established grassland. If it could be just set to work and left to navigate across fields on there own. Look at the thousands of horse paddocks!

I'd be a bit worried about Tom's ability to cope with a sticky dung pat, rooting holes or the tongue of an inquisitive bullock. I suppose the first 2 could be detected and avoided, but the 3rd is a challenge for both parties.

 

[Y Fan Wen]

I'd be a bit worried about Tom's ability to cope with a sticky dung pat, rooting holes or the tongue of an inquisitive bullock. I suppose the first 2 could be detected and avoided, but the 3rd is a challenge for both parties.

As you say AI could easily avoid the first 2 and an electric fence type reaction would solve the second.

 

[Devil's advocate]

While I wish the robots well.

I see a robotic 50 hp tractor, that could be driven a more versatile piece of equipment, with a well proven diseal power unit or bio fuel or electric battery.

Fleets on large farm but still affordable for one on smaller operations.

Implements already available although if the demand was there they could be much better, carbon fibre booms on small sprayers, small kit with farm qualty bearing/chains etc.

1.8 ton on wide tyres is light enougth.

Can be driven to sites or on small lorry or trailer.

A decent cab would protect the high tech equipment & nice enviroment for any human operator, perhaps not like a big Fendt.

Harper Adams gets my vote/investment if they need help.

Robots are almost available, would think idea for really valuble sites like inside horticultural tunnels/glasshouses.

 

[oil barron]

Well done for trying

I think however you need to make some fundamental adjustments to your concept to stand any chance of it working. Wheat farming is a broad acre occupation and the weed burden is high and interspersed with the growing crop. Take a look at a picture of black grass in wheat. The idea that you can have a robot take high resolution imagery across thousands of acres, process that data using image recognition technology and then send out another robot to zap the weeds electronically is nonsensical.

1. The precision required would be in the mm range way beyond current DGPS
2. A growing crop is dynamic. By the time you get the imagery, process it and come back to do the zapping the image will have changed
3. It actually takes a lot of electric to kill a plant. Take a look at where an electric fence touches grass. It goes yellow but then recovers. Your battery might last 1m2 with the power required.
4. The data storage requirements and processing capacity is of the chart.

I think if you changed your mindset to a broad acre approach that recognized weeds or disease in generality and applied varying rates of herbicide and fungicide real time without a second pass you could be more successfull.

Also the amount of vegetables grown may be lower than cereals but the demand to reduce Labour is exponentially higher. The majority of cereal farmers will not part with their tractors. The majority of vegetable growers would send their Eastern Europeans home tommorow even if the technology came at a premium.

PS. If you could develop a robot that recognized stones and zapped them into thin air you would be an absolute legend.

 

[Devil's advocate]

Almost ready for market, just need the right tyres, compact direct drill, set of rolls, sprayer & fertliser spreader.

Use it for little jobs rest of the year, sorted.

Edit To really complete the job it needs to ferry grain in a small grain chaser to the nearest farm road & collect the straw, perhaps bulk pickup wagon (for home use) or small bale collector for sale.

Still not sure how to protect even this from vandals & theives.

Really clever spot hoeing/spraying kit is available from Garford & others

 

[beefandsleep]

I don’t think travelling through a crop post stem extension or canopy closure will be much of an issue. If the crop has been kept clean up to that point then little else will germinate or become overly competitive with the crop if it does. We would maybe need to adjust our mindset to accept a nearly clean crop rather than perfection as there would be no danger of weed issues getting out of control. It could even be a boon for rare arable plants if a tolerance of x plants per hectare were to be programmed into the robots. Fungicides are still going to require full field treatment if high input systems are used but I’m sure there could be a big swing toward organic systems if weed control could be manually automated, lower plant populations and more open canopies could reduce effects of fungal diseases also.

 

[Clive]

Other than the law what’s actualy stooping fully autonomous operation right now ?

Next spring we will have an x35 equiped tractor with auto turn enabled so in theory the tractor could drill a fiejd without any driver input

 

[Devil's advocate]

Other than the law what’s actualy stooping fully autonomous operation right now ?

Next spring we will have an x35 equiped tractor with auto turn enabled so in theory the tractor could drill a fiejd without any driver input

Be nice to have it, even with a driver in the cab

High qualty work, 24 hours a day (with shift working) no problems with sun in drivers eyes, operator fatique, working in the dark, meal breaks driver nodding off in warm afternoon sun but still someone sat on the seat to keep an eye on blockages, slipping out of adjustment & health/safety. Also of course to make decisions about changing weather conditions, perhaps update records on ipad.

In good conditions operator can nip back to farm for more seed/chem etc or take a meal break.

 

[Y Fan Wen]

Be nice to have it, even with a driver in the cab

High qualty work, 24 hours a day (with shift working) no problems with sun in drivers eyes, operator fatique, working in the dark, meal breaks driver nodding off in warm afternoon sun but still someone sat on the seat to keep an eye on blockages, slipping out of adjustment & health/safety. Also of course to make decisions about changing weather conditions, perhaps update records on ipad.

In good conditions operator can nip back to farm for more seed/chem etc or take a meal break.

I'm a pg farmer so have no knowledge of how these things work; how about 2 or 3 tractors in the field doing different tasks and supervised by one driver?

 

[Clive]

I'm a pg farmer so have no knowledge of how these things work; how about 2 or 3 tractors in the field doing different tasks and supervised by one driver?

I think Fendt did that (and it was commercially available even ?) several years ago now ?

 

[Clive]

yes 2012 it seems - wonder why it didn't take off ?

https://www.fwi.co.uk/machinery/fendt-demos-driverless-tractor

 

[curlietailz]

yes 2012 it seems - wonder why it didn't take off ?

https://www.fwi.co.uk/machinery/fendt-demos-driverless-tractor

Ahead of its time perhaps

GPS, self steering, etc are quite mainstream now
Controlling a tractor from the combine etc

 

[Devil's advocate]

yes 2012 it seems - wonder why it didn't take off ?

https://www.fwi.co.uk/machinery/fendt-demos-driverless-tractor

You would think it would work well on the steppes or praries.

Not ideal for small fields on the edge of the pennines.

Was thinking today some of the steeper fields to the west of where I live. Used to be limed by some pretty crazy guys, but with larger tractors & lime spreaders some is getting too scary.

Assuming as a society we want these marginal fields to be productive (the top soil is often better than you might think). A small cabless 4x4 perhaps with spiked duals with a one ton spreader with grounddrive could follow a route planned by quad bike or mapped on a smart phone..

 

[Chae1]

Quad bikes are surprisingly unstable in my experience.