23 June 2007

Robots for the future of farming

Major h/t to Instapundit linking to this Wired article on automated harvesting done by robots. This was also seen in an earlier Modern Marvels on the History Channel, which caused me to ask, at the time: If robotic 'beta versions' are being tested *now* then how long until production versions come around?

That centered on the multi-spectral sensors on gripping hands to judge fruit quality and send tactile feedback for picking of same. Since such sensors, which would include chemical sensors, do not rely on sunlight (looking more to IR and other non-visible spectra) and using chemical sensors to determine fruit quality and ripeness, one is soon in the position of picking delicate fruit (peaches, pears, apples, etc.) without the need for hand harvesting while keeping quality high. So to get an idea of jobs that humans will not need to do, by and large, lets start hitting that sector of the economy and find out just how much longer there will be jobs that *people* will need to do in the agricultural sector of the US.

To get started lets head over to the UK for a look at a back-breaking job if ever there was one: picking cauliflower. The Pera company points to its first pre-production model of an automated cauliflower harvester, that should be coming to market in the next few years.

Well, no one said these things would be a joy to look at! Still, this is a finicky vegetable that does require good handling and picking, and getting *that* into the automated realm is a huge plus to eliminating back breaking work in the world.

Not to be outdone, Giulio Reina's work at the University of Salento's Department of Innovative Engineering (on leave this year to Tokyo's Space Robotics Laboratory) points to a paper for robotic harvesting of raddicchio. While a pretty sturdy plant, it does need some specialized handling and the paper goes through the parameters of vision, examination and picking of it via an automated process. Who knows, the price of the stuff might come down a bit!

The University of Western Australia has been developing a sheep shearing robot, of all things!

And after seeing the Dirty Jobs episode of shaving alpaca, I can understand *why* someone would want to automate this for sheep... and if you can do this in a production line environment then the Shear Magic is the robot for your sheeply needs.

Yes, many jobs to go by the wayside in the sheep business with this, plus a reduced risk of getting things like anthrax. Apparently that is something that Australians would prefer that NO ONE gets from doing that job.

The TimesOnline (UK) reported in 2005 on the automated cow milking machine, so that one can have no fuss nor muss when having to get that task done day in and day out. There appear to be lots of jobs that folks in the UK and Australia just don't want to do!

Well, far be it from the US to lag behind, and Carnegie Mellon University has been at the forefront of automated harvesting of crops in an unattended fashion. Yes, set the harvester to go, day or night, and off it goes into the field to harvest the crops. To step just a second to another dirty job, how would you like something to help automate the fun job of paint removal?

M3500 Ultra Strip

Brought to you by the researchers at CMU! For those of you who don't think paint stripping is the best of all possible jobs on the planet.

CMU is famous for having been awarded a contract for the Unmanned Ground Combat Vehicle after the last DARPA Grand Challenge.

Apparently getting ammunition, supplies and such to soldiers on roads that are in or near hostile territory or frequented by IEDs is something that soldiers do, but would prefer to have robots do for them. And with this add-on package, called the UPI, the vehicle will perform even better against unknown parts of terrain or sudden problems.

But back to those *other* jobs Americans don't want to do!

How about lawn mowing for large areas? Yes the little home robot mowers are fine for the family yards, but, for things like golf courses... well... you need something a bit better. And that is what CMU is partnering with Toro on!

Ah, why should someone need to do *that* job, if a robot can do it for you? On the larger scale, for fields and orchards and such, there is the automated water sprayer system, also from CMU and partnering with John Deere.

Why should anyone have to just drive around and around hauling a spray system? Sounds very robotic so let a robot do it! One can get a full run-down of the CMU program here in their Projects vs Capabilities page.

The University of Kentucky with BAE systems has been working very hard at automating the tobacco harvest, as seen at their page on same. Their paper on mechanizing the harvest is here. Who would have guessed that Americans would prefer to have machines pick tobacco? The UK/BAE work is also looking into precision agriculture for forecasting, watering and soil analysis. Getting good and solid info on soil type, moisture content, plant nutrient needs and such is critical to doing robotics and better farm management, so that higher yields and lower utilization of pesticides and fertilizer allow for precision application *just* in the amounts needed and only where it is needed.

In 2001 Loyola College spun of its World Technology Evaluation Center into a corporation, WTEC, and it held a robotics conference in 2006 (full page here), looking across a broad array of uses of robotics across industries and environments. That conference looked at the wide array of types of robotics currently in use, in design and proposed, from such things as the Sony Aibo and Roomba to proposed environmental research vehicles to find the precursors to off-shore algal plumes that endanger fisheries. Yes, getting that data is, apparently, a job that is better suited to a robot than a scientist.

What all of this is leading to, of course, is RoboFarm. The first test of the automated farm is being done in Monterey County, to work with things like lettuce, strawberries and other farm produce that currently needs a lot of hard work to get picked. This is the first real integration of all the technology from soup to nuts, from GPS analyzed fields via automated systems to automated planting, watering, fertilizing, harvesting and packaging. The elimination of humans in the farm to make it more efficient. There will *always* be a place for human experience and the ability to integrate non-linear systems, of which farming is a prime one, but the removal of drudgery and labor intensive tasks that are limited to the fallible human for judgement and time on the job, means that better crops, more crops and cheaper crops will result due to lowered overhead and consumable expenditures via better crop management.

Within 10 years the first real hard changes will be towards eliminating humans completely from the mundane tasks of farming. And that will shift the areas of expertise to those areas that machines just don't do such a good job.

I am sure that we will become so lazy as to no longer wish to procreate.

Don't worry, there will be robots to help there, too.

No comments: