Friday, 11 January 2013

The future of farming.

Jethro Tull
Sometimes as technology advances some areas get left behind. To a certain extent this happened to farming. It was always at the front of technological development and for long periods had been one of the driving forces behind development.
There are several candidates for what is man's 3rd oldest technology, as the first two are language and tool making, but agriculture is what put us on the road to modern society. With agriculture human society started growing. It's been a double edged sword ever since, it's been blamed for increased disease due to proximity, war, sexism and bad diet. It has been the spur and the limiting factor to human growth at various times.
In the first instance everything would have been done with simple hand tools such as a hoe. From there the technology moved on through ploughing from about 1000 BCE, with gradual improvements with the introduction of metal implements. It wasn't until Jethro Tull in the early 18th century that anything like modern agriculture started and the complete mechanization was not really complete until the 1940s.
In the days of horse and oxen power the farm was both the source of these animals and made extensive use of them in haulage. Steam power made an early appearance on the farm with the traction engine, the price of farm products was affected by the arrival of the train with the ability to move fresh goods over long distances.
By the 1930s chemistry was being used on an industrial scale to produce synthetic nitrogenous fertilizer. The 1940s saw the introduction of Herbicides and Insecticides came into common use.
In the processing of produce in buildings electricity was a common but made very little impact in field usage despite its ability to be used in battery powered vehicles.
Recently, along with every other vehicle driven by internal combustion the tractor and other farm vehicles have become covered in various digital devices from engine management units, GPS. Satellite and UAV system have been used to gather data for precision agriculture. Devices have been fitted to vehicles to sense the state of the crop to assess health and nutritional requirements.
These are all big machines with big costs that require constants human supervision, they are committing the computerization sin of producing a computerizing version of a manual system rather than designing a digital system from scratch. This usually leads to an inefficient system as allowances were made in the design of the original system for human frailties and strengths and will only partial take advantage of the computers frailties and strengths. So we  end up with a system that is unnecessarily constrained by some redundant requirements and at the same time fails to take full advantage of what the new tools available.
One of the major constraining factors with current farming machines is the need for constant human supervision. With the advent of technologies such as auto driving cars, it should be a simple process to adapt the technology to drive a tractor. In simply doing this we would be replacing the human being and the constraints which apply to them, these include the need to be paid, to eat, sleep, have somewhere comfortable to sit, be protected from the weather. On the other hand a human is a very flexible item capable of overcoming unexpected events and in the case of farmers changing implements, any system that does without people must either cope with the same possibilities or eliminate them or greatly mitigate the consequences of them occurring.
By eliminating people from the direct supervisor and guidance task we remove the need for speed. If we have one week to do something that is no longer 5*8 hour shifts or 7 *Daylight hours it is the full 24/7. If all the time was spent in the field then this would equate to driving at somewhere between about 1/2 or 1/3 of the speed. As going faster increases energy usage then this will inevitably use less energy, going fast also reduces the stresses and strains on the vehicle which would allow a reduction in strength and therefore mass, which further reduces stress and strain and also further reduces energy costs. In eliminating the driver we  remove their comfort requirements such as suspension, heating, lighting, seating and seeing. This will further reduce energy and mass requirements.
There are some jobs that require large amounts of power such as ploughing an whilst it is possible to envisage a system where ploughing is not needed it is unlikely to find acceptance with current farming techniques there is some indication that the current fashion for deep ploughing may cause some problems. This however is a something which is done infrequently and could like other specialisms be economically be outsourced to a contractor.
Several discrete and distinct processes need to occur, the first is the  collection of information. Once collected the observations have to analysed, a diagnosis made and a treatment if any selected. Finally the action if any needs to be taken.
The collection of information and the action can only be done in-situ whereas the data processing can be done at a base station This saves the need to carry around high power processing equipment and the power supplies necessary to run it, it also avoids the need for more expensive low power processing. The like amount of processing required means that a reasonable speced 2013 PC will be able to provide enough processing power.
There are various options for the radio network from 2G/3G/4G networking through to0 pieces of more local equipment such as WiMAX, would be good with a range of 31 miles, and covering several farms. Though shorter range systems may be better suited due to the lower power costs.
 Though it is possible that in the future with the improvement of battery technology and further reduction in power consumption could make it possible to carry enough processing power, but the test is a relatively simple one based purely on the weight and power consumption of each option, the data transmission will use energy and the Wireless Modem will have mass. Depending on the balance energy balance the processing load on the mobile unit could be minimal and handled by little more than a basic PIC, support by the required communications and positioning equipment with the remote host downloading a script of actions to be taken, by the various on-board devices, no intelligence at all would be needed. In fact I think it could all be run using an off the shelf M2M system.
For collecting the information we need sensors, preferable as off the shelf as possible. We will need two different kinds of sensor e one to collect location information and the other to collect the diagnostic data.
The location information can be done using a set of stereoscopic cameras, these have been successfully used by NASAs current Mars rovers.
To gather the analysis data we can look to Hyperspectral Imaging, this will allow collection of data from the far infra-red to the far ultraviolet using a single mechanism. If we couple that with a suitable illumination source to extract florescences data would provide powerful tools for diagnosis and identification. Amongst the many things that such a system could identify would be the degree if water stress the plant is under, fungal parasitic infections. In addition to classifying plant to allow decisions on whether to keep or not.
Once we have all the information has been collected  and analysed then action may need to be taken or scheduled to be taken. That action may be carried out by the same vehicle that collected the data or by any other vehicle at a later date. At the simplest level this action will be to kill an undesirable plant which will involve nothing more that physical manipulation. However the the best action may require reagents or possible some specialist action over a large area.
As the machine is working with a single individual plant the range of options is quite large from applying a specific treatment or depositing beneficial insects. If the number of individuals that require a treatment is large it will be difficult for a small light device to carry sufficient reagent, however the basis of this system is regular inspection weekly or more often of each plant. This reduces the chance of wide area needing to be treated as hopefully inspections will detect problems before they have become extensive.
The entire system may be of such low power that it can be run purely on environmental energy collected by each unit. This however is unlikely a system with a central energy collection point which converts the energy into something like methanol which is distributed to each unit as it needs it may be best allowing for storage of fuel generated during energy rich periods.
The essence of the entire system is to reduce things down to minimal simple tasks done as regularly as possible, a move away from the process of whole field treatment but individually tailoring treatment to each individual. The action does not have to be taken immediately as for a machine there is no tedium in coming back to the current spot in 6hrs to take some action, no loss of face in calling in another entity to take action.
We are swapping from a system of fast heavy machines for a fleet of slow light and hopefully cheap machines. Which seek to minimize the chemical and energy inputs to farming. Depending on the actions and agents used the system could work as well on an organic farm as in the mainstream.
Other methods of data acquisition including UAVs and satellites are possible, as are other modes of action. The essence of this system is that almost continuous monitoring of plant crops is possible and that by correct application of computer technology arable farming can use the work of chemists, biologist and farmers in a far more subtle way.

No comments:

Post a Comment