Since the early 1990s dairy farmers have had the option of milking their herds robotically. Far more than just a milking machine, an Automated Milking System (AMS) fully automates the process of getting cows in and out of the milking shed. The most modern systems have no need for recourse to human labour at all. About 35,000 such systems are thought to be in operation around the world today.
Traditionally whole dairy herds are corralled twice daily through a milking parlour. Each time, they must queue to have machines manually fitted to their udders before being milked and driven back to pasture.
Animals in automated systems, by contrast, approach the milking machines at times of their own choosing (hence the alternative term Voluntary Milking System). The cows are motivated by the pressure of a full udder together with the presence of feeding stations in the machines. Thanks to electronic tagging the system can tell one cow from another and by means of smart gates only allows in those animals it knows to be ready for milking.
After automatic brushes clean the udder, pumps are sensitively attached by a robotic arm with the aid of positioning sensors. Other sensors monitor the milk flow from each quarter and ensure that the animal is neither over- nor under-milked. While the cow is cleaned again the milk is routed to a bulk tank where it is filtered and cooled.
Other aspects of the environment, such as automated manure removal, may be under the system’s control.
At its most basic level, mechanisation in dairy farming existed as long ago as the Industrial Revolution itself. Early examples of milking machines were being patented in the middle of the nineteenth century. Over the decades countless innovators refined the design of teat-cups and the supporting claw, the use of vacuum as an extraction force and the deployment of pulsation technology to optimise flow and reduce trauma to the teat.
It is the intelligent aspect of the Automated Milking System that is new. The system remembers individual animals, using its knowledge of the shape of each udder, for example, in its application of the teat-cups. Biosensors have the potential to detect abnormalities in milk by responding to visible or non-visible light transmission in the liquid. And the system’s software builds a detailed ongoing data profile of the herd’s physical and behavioural health.
One milking robot services about 60 cows and costs around €120,000. To date, they have been most popular in small northern European countries where firms such as Lely (Netherlands) and SAC (Denmark) have pioneered much of the research. The British company Fullwood has been designing milking robots for over twenty years.
The farming old guard – and, no less vocally, the environmentalist lobby – has registered some anxiety about the trend towards total automation. The subtle complex of skills and instincts that make up the quality known as ‘stockmanship’ finds no equivalent, they say, in a world governed by robotics and artificial intelligence.
An experienced cowman can put a nervous animal at ease. He or she can spot the earliest signs of common diseases such as mastitis. Milk is traditionally inspected for quality by eye, and this practice has been reflected in industry standards.
Among those who have made the transition to automated milking, however, the word on the farm is positive. Happier herds, it seems, are one of the main results. These are cows, after all, who get milked when they choose rather than according to an imposed regime. Not only this, but the animals’ welfare is looked after by the system’s data which can show abnormalities or inconsistencies too subtle for even an experienced herdsman’s eye.
Milk yield is generally about 10% higher in an automated system, a consequence of more frequent milking but possibly a reflection, too, of the cows’ greater contentment.
It is the freeing up of time that modern dairy farmers cite most often as the key dividend of AMS. The owner of one 100-strong herd in Ireland claims that robotic milking has reduced her daily workload by at least three hours. These time-gains enable better running of the farm, allowing the farmer to respond to other day-to-day issues, or indeed to issues flagged for attention by the system software.
But what may ultimately render these systems inevitable rather than simply desirable is the growing problem with available farm labour. The EU agricultural outlook report for 2017-2030 predicts a decline in the agricultural workforce for this period by 28%. No less significant is the traditionally high proportion of immigrants in this sector – about half, according to a recent American survey. Few farmers in the UK or America will have failed to register the threat to their labour pool posed by the Brexit phenomenon and Trump presidency respectively.
The human implications of robotic milking are not just that fewer people will work on farms, but that the nature of farming work itself will become more managerial. Herein lies the real challenge of the Automated Milking System.
Today’s dairy farmers must be as comfortable with computers as they are with cows. The amount of data generated by the system can be daunting: a 24/7 stream of information about cows’ body weight, milk yield, milking time, preparation time, rumination time, milk protein-fat ratios and so on.
The farmers of the future will require training in system management and in the use of sophisticated data analysis software. Without access to the appropriate software that can quickly analyse huge datasets, future farmers may risk simply trading hours in the milking parlour for hours in front of the monitor. With those skills in place, however, and with appropriate support from manufacturers, the potential of robotic milking to deliver a new industry norm of healthy, efficient dairy farms may soon be realised.