It's probably not a big deal, but there are a few things to sign yet. Most of my analysis work is related to developing peptide-drug conjugates, so this will be a departure. The general idea is to improve on a specific QC process for acquired field data, adjusting correction factors on the basis of observed phenomena. The results should boil down to improved efficiency in some regard, but I'm not sure about the specifics yet.
The problem is that a farm is a dirty, corrosive, and stochastic place.
Equipment needs to be tough as nails and able to deal with each unique farmers challenges.
Things I see on the horizon :
Robotic weeding is going to be a HUGE deal. Improved robotic harvesting also. These both are almost entirely reliant on robust computer vision that can survive days in the field.
Drones are going to also be a huge deal. Imagine a quad copter that can go out and sample your 1000 hectares of fields, spotting disease or nutrient deficiency from images taken from above.
The promise is big, but integrating tech seamlessly is going to be damn hard.
"Drones are going to also be a huge deal. Imagine a quad copter that can go out and sample your 1000 hectares of fields, spotting disease or nutrient deficiency from images taken from above."
I don't know if this is meant to be pithy or serious, but I think I disagree. Agriculture is big business, farmers are smart people, and there is a history of lots of mechanization and interesting technology being applied. The productivity gains that we've already seen in the last hundred years (and before) are tremendous.
I'm sure that there are areas to be improved upon, but I expect a lot of the low hanging fruit has been picked, as it were.
I also tend to think of farmers as the original hackers, they often have problems, and they cobble together a solution that might not be pretty, but works.
He's right. I don't know about openings where absolutely nothing is being done, but I see opportunities for simplifying and bringing costs down. A lot of farm machinery is geared towards large producers with prices to match, leaving smaller farms to either DIY or do without.
I have/had a few ideas I was researching and at least one I plan on testing on my own acreage this summer. Figuring out how to market to a population that doesn't seem to have a big online presence is the hurdle I haven't crossed yet.
Unfortunately, some of these productivity gains have been in the wrong direction from a sustainability and health standpoint.
One of the things that I'm looking at (as a hobby research topic) are automating managed intensive rotational grazing for cattle and remote herd management.
Unless that was an excuse for a pun (and I have no problem with that), you are underestimating how much work has gone into farm automation. All the industrialized countries have worked on this for decades, with some successes (GPS-guided tractors with pesticide spay patterns guided by aerial photography, good old milking machines, etc). This is a multi-billion dollar business in most countries and also a strategic issue of food security.
Israel in particular has spent a fortune on research into robotizing fruit production, since they have money but have complex relations with neighbours that make cheap labour immigration unreliable. They are working very hard on all the literal hanging fruit.
One of the first stories I did as a freelancer, back in the early 80s, was about radio tracking collars being used on cows so that farmers could manage the herd better.
So this is something in the robotics field that has been long-coming. If you'd like to get an idea of what the future holds for people, take a look at cows, a creature that has a lot of financial value. The idea is that the cow is tagged to be uniquely identified, when the cow approaches a service center, for say salt or food, the system reads the ID and delivers an appropriate form of nourishment. When the cow is uncomfortable due to lack of milking, the system milks the cow. Next up, I'm sure, will be automated detection of sick cows along with some basic veterinary interdiction. Finally it would be very easy to add a little bit of behavioral control to the system, either a mild shock, a noise, or some other stimulus to gently train the animal over time. I could even imagine a farm without fences, managed completely by robots and conditioned cows. The cows are happy, the people are happy.
Pretty cool. Also a bit unnerving.
I don't think there needs to be a robot revolt. I think over a few centuries they'll just slowly manage us into extinction, where we will most likely belong.
A tractor can make a lot more assumptions about its environment than a car can, and things like field topography/row direction can be programmed once and then just replayed. Automating them should be relatively simple (and it looks like it is being done [0]). I think the necessary algorithmic/sensor complexity is probably closer to a Roomba than a Google car.
Farmers (here in the UK) have been pretty good jumping on technology, they have sophisticated navigation software, probes they use to assess crops before and during harvesting (down to which part to harvest with first).
Where I live they grow a lot of pea's and the period between "ready to harvest" and "pig food" is pretty short, it's a spectacular sight at night watching them go up and down the hills near here in the dark (always reminds me of War of the Worlds as they have rotating spotlights) harvesting.
"There's already a prototype robot that brings cows in to be milked"
Long-term, that is a dead end. The thing is, you don't have to bring in your cows, and certainly not all at the same time, if you have a milking robot. They will come in by themselves if they want to be milked.
Also, I find it curious that this is news. Milking robots have been used for decades (since the early 1990's)
Economically, we have about reached the point where the robot is economically feasible (a robot increases milk production, but costs money, both in investments and in electricity. It's not a given that the increase in milk production pays for the extra costs)
>I don't think there needs to be a robot revolt. I think over a few centuries they'll just slowly manage us into extinction, where we will most likely belong.
Its amazing how difficult these problems are. This type of work relies on things that humans are optimized to do - forage and produce food. We are arguably the best robot for some of these tasks.
We are currently much better than lime jello at these tasks, too. "Currently" implies an expectation that things will change, which is a claim that needs some kind of argument in support. Rhetoric alone doesn't advance the discussion.
We are in the midst of a massive trend towards automation and robotics. Every year, we are seeing more tasks and more jobs successfully completed by robots, and more and more robots completing "work" tasks more efficiently than humans.
I am not qualified to statistically prove the automation trend, but I believe you are being obtuse in rejecting the premise that said trend exists at all.
"Its amazing how difficult these problems are. This type of work relies on things that humans are optimized to do - forage and produce food. We are arguably the best robot for some of these tasks."
It really is amazing how difficult some of these problems are. That was the point of GP's post. By saying "currently" you wave away that point without addressing it.
While robots are getting better, there are lots of things we take for granted about human and animal competence and that we have made very slow progress on replicating in robots.
It's much more fun to think and talk about those tricky bits than to simply expect them to be solved.
I don't need statistical proof that there is an automation trend. I'd like an argument about why the interesting hard parts are going to turn out to be doable.
You might infer from my username that I am moderately optimistic.
That's old news, such technology is available for several years. I have seen such systems in the early 1990s in Europe.
The downside is that you need a lot of antibiotics for the cows. In the end, people will eat the meat and drink the milk of such animals plus a dose of antibiotics. More and more get an immunity.
So worse is better. (aka low tech agriculture with less animals in one farm building)
I agree that this is old news, however, the comment about the antibiotics is not accurate. I visited an organic farmer last summer who was using a robotic milker for his 40 or so dairy cows. He was very happy with the system.
How is my text about the antibiotics not accurate?
I was writing about big farming companies with many cows. To keep a lot of animals healthy in a narrow space you have to inject each cow antibiotics in advance.
Antibiotic in livestock is a real problem already - at least in the Europe. The media and the doctors speak open about the issues, yet some big farming companies don't care.
Everyone is fine if farmer inject antibiotics to ill cows. But giving all cows antibiotics all the time (in advance), just in case, is the problem. People will drink & eat it and if they get ill or are in hospital, the same antibiotics has no good effect anymore. And people already die because the aftermath (hospital bug).
> How is my text about the antibiotics not accurate?
It was inaccurate because the robotic milkers do not necessitate antibiotics in any way. There's no increased need for antibiotics because a cow is milked by a robot instead of a human.
> Antibiotic in livestock is a real problem already
In the dairy industry, if a cow is on antibiotics her milk is dumped at the order of the company that buys the milk for bottling and distribution to consumers. At least that's how it works on my neighbor's farm, who sells milk to a regional dairy distributor.
Due to having heard this topic in detail at university, I think the link between high antibiotic use and robotic milking is not the case.
One robot can only deal with a relatively little amount of cows in a day. I remember something around a hundred. And industrial scale human-involved systems are at a much bigger scale. If anything, these large-scale systems would require more antibiotics.
Over time the antibiotic management should even improve through the availability of per-quarter, per-cow data on the performance of the cow. Some antibiosis is often necessary, and probably can't be avoided if you don't want cows suffering unduly, or getting culled because without treatment the cow won't be profitable anymore. But that is a whole different kind of animal... err.... discussion
To qualify this a little, there is a way the robots can increase antibiotic use, and that would be if the hygiene of the robot is worse than that of other milking systems. But this does not seem to be the case. The robot would probably serve less "customers" with more thorough cleaning in between...
It has nothing todo with robots per se. Robots just make it a lot easier to feed a lot more animals in a narrow space with little human interaction.
In the United States, the use of antibiotics in livestock
is still prevalent. The FDA reports that 80 percent of
all antibiotics sold in 2009 were administered to
livestock animals, and that many of these antibiotics are
identical or closely related to drugs used for treating
illnesses in humans. Consequently, many of these drugs
are losing their effectiveness on humans, and the total
healthcare costs associated with drug-resistant bacterial
infections in the United States are between $16.6 billion
and $26 billion annually
No, traditional milking systems allow more intensive cattle stables than robotic systems. Which doesn't matter in Europe, because the space requirements per cow are a lot higher than the minimum possible. If anything, the robot allows for a less stressful milking and probably "happier" cows.
I can't see any connection between increased use of antibiotics and autonomous milking....
At least in Europe? One might argue that Europe has its eyes open. That Wikipedia page states: "On 1 January 2006 the European Union banned the non-medicinal use of antibiotics in livestock production."
That takes care (formally) of a substantial part of Europe. Regulations such as these also tend to spread because they limit the ability to export produce that does not follow them (that's how lead-free soldering conquered the world based on a EU directive, and how Californian pollution limits for cars spread over the USA, etc)
If you think those rules get broken so often that this is more of a problem in Europe than in, say, the USA, Canada, Australia, or Asia, can you give a link?
"Confinement at high stocking density is one part of a
systematic effort to produce the highest output at the
lowest cost by relying on economies of scale, modern
machinery, biotechnology, and global trade. Confinement
at high stocking density requires the use of antibiotics
and pesticides to mitigate the spread of disease and
pestilence exacerbated by these crowded living
conditions. In addition, antibiotics are used to
stimulate livestock growth by killing intestinal
bacteria."
It's actually not that easy... There are other factors at work. Often antibiotics only avoid growth depression, and in many cases the space requirements are higher anyway. The robot doesn't come into this discussion at all...
I was actually most impressed by this little rotating robot. I actually had to this by hand some years ago, shoveling the feed so that the cows can get at it, and this rotating solution looks ingenious!
53 comments
[ 2.9 ms ] story [ 114 ms ] threadThe problem is that a farm is a dirty, corrosive, and stochastic place.
Equipment needs to be tough as nails and able to deal with each unique farmers challenges.
Things I see on the horizon :
Robotic weeding is going to be a HUGE deal. Improved robotic harvesting also. These both are almost entirely reliant on robust computer vision that can survive days in the field.
Drones are going to also be a huge deal. Imagine a quad copter that can go out and sample your 1000 hectares of fields, spotting disease or nutrient deficiency from images taken from above.
The promise is big, but integrating tech seamlessly is going to be damn hard.
I believe that one is already a reality. Some quick Googling turns up: http://www.precisiondrone.com/drones-for-agriculture.html I'm pretty sure I saw a TV special too.
Surely, these projects sow the seeds of eventual progress.
I'm sure that there are areas to be improved upon, but I expect a lot of the low hanging fruit has been picked, as it were.
I also tend to think of farmers as the original hackers, they often have problems, and they cobble together a solution that might not be pretty, but works.
Once I finished some other projects I'll looking into Arduino or something similar to start hacking at some ideas I have.
By the way, you are indeed right that farmers have already done most of the hacking themselves.
I have/had a few ideas I was researching and at least one I plan on testing on my own acreage this summer. Figuring out how to market to a population that doesn't seem to have a big online presence is the hurdle I haven't crossed yet.
One of the things that I'm looking at (as a hobby research topic) are automating managed intensive rotational grazing for cattle and remote herd management.
Israel in particular has spent a fortune on research into robotizing fruit production, since they have money but have complex relations with neighbours that make cheap labour immigration unreliable. They are working very hard on all the literal hanging fruit.
So this is something in the robotics field that has been long-coming. If you'd like to get an idea of what the future holds for people, take a look at cows, a creature that has a lot of financial value. The idea is that the cow is tagged to be uniquely identified, when the cow approaches a service center, for say salt or food, the system reads the ID and delivers an appropriate form of nourishment. When the cow is uncomfortable due to lack of milking, the system milks the cow. Next up, I'm sure, will be automated detection of sick cows along with some basic veterinary interdiction. Finally it would be very easy to add a little bit of behavioral control to the system, either a mild shock, a noise, or some other stimulus to gently train the animal over time. I could even imagine a farm without fences, managed completely by robots and conditioned cows. The cows are happy, the people are happy.
Pretty cool. Also a bit unnerving.
I don't think there needs to be a robot revolt. I think over a few centuries they'll just slowly manage us into extinction, where we will most likely belong.
Once driverless cars are perfected, driverless tractors will be right there with them too, they will be hugely popular.
If you think about it, driverless tractors are a much easier problem than driving a car.
[0] http://en.wikipedia.org/wiki/Driverless_tractor
Where I live they grow a lot of pea's and the period between "ready to harvest" and "pig food" is pretty short, it's a spectacular sight at night watching them go up and down the hills near here in the dark (always reminds me of War of the Worlds as they have rotating spotlights) harvesting.
Long-term, that is a dead end. The thing is, you don't have to bring in your cows, and certainly not all at the same time, if you have a milking robot. They will come in by themselves if they want to be milked.
Also, I find it curious that this is news. Milking robots have been used for decades (since the early 1990's)
Economically, we have about reached the point where the robot is economically feasible (a robot increases milk production, but costs money, both in investments and in electricity. It's not a given that the increase in milk production pays for the extra costs)
Obligatory link to Manna (sci-fi story about robots and bots impact on humans) - http://marshallbrain.com/manna1.htm
Its amazing how difficult these problems are. This type of work relies on things that humans are optimized to do - forage and produce food. We are arguably the best robot for some of these tasks.
I am not qualified to statistically prove the automation trend, but I believe you are being obtuse in rejecting the premise that said trend exists at all.
"Its amazing how difficult these problems are. This type of work relies on things that humans are optimized to do - forage and produce food. We are arguably the best robot for some of these tasks."
It really is amazing how difficult some of these problems are. That was the point of GP's post. By saying "currently" you wave away that point without addressing it.
While robots are getting better, there are lots of things we take for granted about human and animal competence and that we have made very slow progress on replicating in robots.
It's much more fun to think and talk about those tricky bits than to simply expect them to be solved.
I don't need statistical proof that there is an automation trend. I'd like an argument about why the interesting hard parts are going to turn out to be doable.
You might infer from my username that I am moderately optimistic.
I'm far more optimistic on these problems, as I've seen many of them solved.
Lettuce picking (and other ground crops shortly) thinned and picked by machine? Check.
Strawberry sorting for plant quality? Check.
Strawberry harvesting? Check.
Harvesting of tree crops? Check.
Orange grove navigation by tractors? Check.
Ex: http://www.rec.ri.cmu.edu/usda/ http://farmofthefuture.net/#/slideshow/autonomous-tractors-t... http://farmindustrynews.com/precision-guidance/new-driverles... http://www.agprofessional.com/news/JD-autonomous-tractors-ar... https://www.asme.org/engineering-topics/articles/robotics/sm...
Sorry for that mistake, and thanks for the links.
Automatic berry picking for "tree type" bushes is a reality as well.
The downside is that you need a lot of antibiotics for the cows. In the end, people will eat the meat and drink the milk of such animals plus a dose of antibiotics. More and more get an immunity.
So worse is better. (aka low tech agriculture with less animals in one farm building)
I was writing about big farming companies with many cows. To keep a lot of animals healthy in a narrow space you have to inject each cow antibiotics in advance.
Antibiotic in livestock is a real problem already - at least in the Europe. The media and the doctors speak open about the issues, yet some big farming companies don't care.
Everyone is fine if farmer inject antibiotics to ill cows. But giving all cows antibiotics all the time (in advance), just in case, is the problem. People will drink & eat it and if they get ill or are in hospital, the same antibiotics has no good effect anymore. And people already die because the aftermath (hospital bug).
http://en.wikipedia.org/wiki/Antibiotic_use_in_livestock#Con...
It was inaccurate because the robotic milkers do not necessitate antibiotics in any way. There's no increased need for antibiotics because a cow is milked by a robot instead of a human.
> Antibiotic in livestock is a real problem already
In the dairy industry, if a cow is on antibiotics her milk is dumped at the order of the company that buys the milk for bottling and distribution to consumers. At least that's how it works on my neighbor's farm, who sells milk to a regional dairy distributor.
One robot can only deal with a relatively little amount of cows in a day. I remember something around a hundred. And industrial scale human-involved systems are at a much bigger scale. If anything, these large-scale systems would require more antibiotics.
Over time the antibiotic management should even improve through the availability of per-quarter, per-cow data on the performance of the cow. Some antibiosis is often necessary, and probably can't be avoided if you don't want cows suffering unduly, or getting culled because without treatment the cow won't be profitable anymore. But that is a whole different kind of animal... err.... discussion
The same goes for some countries in Europe. And as companies are importing goods from cheap farming companies such food can be found everywhere.
I can't see any connection between increased use of antibiotics and autonomous milking....
That takes care (formally) of a substantial part of Europe. Regulations such as these also tend to spread because they limit the ability to export produce that does not follow them (that's how lead-free soldering conquered the world based on a EU directive, and how Californian pollution limits for cars spread over the USA, etc)
If you think those rules get broken so often that this is more of a problem in Europe than in, say, the USA, Canada, Australia, or Asia, can you give a link?
An even better article is on the German Wikipedia: http://de.wikipedia.org/wiki/Intensivtierhaltung#Antibiotika...