Robots on the farm

Citation

(2003), "Robots on the farm", Industrial Robot, Vol. 30 No. 5. https://doi.org/10.1108/ir.2003.04930eaa.002

Robots on the farm

Robots on the farm

Nick Tillett

Nick Tillett is a Research Leader at Silsoe Research Institute, Wrest park, Silsoe, Bedfordshire, MK45 4HS, UK. Tel: +44 (0) 1525 864033; E-mail: nick.tillet@bbsrc.ac.uk

The Emerald Research Register for this journal is available at http://www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0143-991X.htm

Keywords: Farming, Automation, Robotics

Farming and primary food productions are evolving to reconcile the need for international competitiveness with pressure to produce safe food with a reduced environmental impact. The public's nostalgia for traditional approaches to farming is not generally realizable due to a shortage of labour, but robotic technologies might offer solutions. In this paper, low input farming does not mean low-tech farming. On the contrary reducing inputs is about conducting operations with precision, a task well matched to robotics. For example, hand weeding is probably not a long-term alternative to chemical herbicides, but robotic weeding might be.

Robotics in farming is not new. Researchers have been at work since the late 1970s tackling those tasks that have proved unsuitable for conventional approaches to mechanization. Harvesting apples for the fresh market was an early research topic, as the process is labour intensive and the product is susceptible to damage. The technical challenges typify many in agriculture. The product is not uniform, it grows on branches that obstruct visually and physically, ambient lighting cannot be controlled and utilization is poor as the task is seasonal. Sadly, despite considerable effort and fine technical advances through the 1980s and early 1990s both in Europe and the US, to date there has been no commercial robotic apple harvesting. The bottom line has been that seasonal migrant manual labour has proved more cost-effective. Broadly similar stories can be told about harvesting oranges, cauliflower and mushrooms, automating the process of micro-propagation and shearing sheep.

Happily, there have also been some notable successes. The major milking machine equipment manufacturers now offer robotic milking. Some of the most advanced glasshouses use robotic warehouse technology to move trays of plants between standing out areas and work stations. Vision and GPS based guidance systems are now steering tractors and implements, albeit with a man on board to supervise the operation and conduct headland maneuvers. The most advanced pack houses, whilst not yet using robots to handle raw product, are at least using robotic palletizing technology in their final packing areas.

So what lessons can we learn about past successes and failures? Generally, successful applications offer more than just a replacement for cheap labour. Often they involved collaboration between robotic experts and specialists from the application area, including those from biological sciences. Such collaborations result in better more integrated systems. For example, the milking robot manipulator that places teat cups onto the cow's udder is only one part of a sophisticated system that can increase yield and improve animal welfare due to more frequent milking. The labour replaced is relatively skilled and is required at unsocial hours. Automation in glasshouses improves packing efficiency and can be linked to stock control. The glasshouse is also an unpleasant environment in which to work with health and safety implications due to use of crop protection chemicals. GPS based automatic steering systems are most widely used in Australia and North America where the scale of operation is so large that drivers find it difficult to match adjacent drill bouts. GPS reduces overlap, minimizing inputs and making better use of expensive capital equipment. Vision based guidance is used primarily to guide inter-row cultivation equipment more accurately and at higher speeds than that achieved manually. This makes mechanical weeding more effective and economic leading to reduced herbicide use.

Prospects for the future seem bright with scope for many of the successful applications mentioned above and others to be developed further and used more widely. Farm rationalization to form a smaller number of larger farming enterprises, whilst painful to some, will provide better opportunities for capital investment. Perhaps the biggest uncertainty lies in political decisions that affect the availability of cheap migrant labour, which whilst helpful to growers in the short term, acts as a disincentive to robotic development.

Sensors will form an important role in future developments. Computer vision is an important tool due to its flexibility and low cost, but other techniques will be required. The robotic approach to precision targeting does not always lend itself to modern high work rate platforms. It might be more appropriate to run autonomously from small vehicles. Similarly, data gathering might best be done from autonomous field-walking robots. Restricting vehicle size and speed might satisfy safety, one of the principle barriers to autonomy in the open access farm environment. Whilst large high-powered autonomous machines are unlikely to be acceptable in Europe, the same technology could also have a wider role on tractors in reducing driver workload.

Robotic assistance to the farmer need not be restricted to primary food production alone. In his role as a custodian of the countryside the farmer will be called upon to manage conservation areas, hedgerows etc. Robotics may also have a role in managing and monitoring these assets.

In conclusion, farming is on the change due to economic, food safety and environmental pressures. Robotic technology can help the next generation of farmers to deliver against their new goals.