Fruit picking robots at work in the fields and automated machinery removing the weeds growing among crops until very recently seemed like fantasy from the pages of a science fiction novel – but it has fast become reality, with robot technology being trialled extensively in field conditions. Trials are underway on numerous projects in the UK and Israel, for instance, to analyse the viability of various automated systems designed to undertake tasks of this kind. The science behind the technology exists and is fairly well proven; all that remains to be perfected is the engineering aspect, as Chris Roberts, head of Industrial Robotics at Cambridge Consultants, tells Produce Business UK.
“It is no longer science fiction to think of vegetables in fields being harvested by robotic technology. It is now possible,” says Roberts. “All that has to be worked out is the engineering. The result might not look like people’s idea of a robot moving around the fields, it is more likely to be like a plough or a harrow attached to a tractor to pick and harvest fruit and vegetables. One of the projects we are looking at, and expect to be in use by 2018, involves robots moving and collecting items in fields and packaging fruit.”
Harvesting in Israel
Over in Israel, a robotic picker underwent field tests in an apple orchard last year. The harvester is the brainchild of Fresh Fruits Robotics. The automated robot is able to navigate apple orchards and pick fruit from the trees. Avi Khanani is the CEO of Fresh Fruits Robotics. He says: “The harvester is in its prototype stage. The team developed its technology hand in hand with growers, academic experts and horticulture experts, defining the specific requirements for each type of fruit. This resulted in a special software package, including image processing and analysis, designed to identify the location, colour and size of each type of fruit, in order to bring the best fruit quality to the market. Fruit is picked without any bruising and a module for each type of fruit has been developed, distinguishing between the various varieties that are to be picked. Once fully developed, Khanani says the harvester will be able to pick 10,000 fruits an hour and it could work 24/7.”
The company anticipates that commercial harvesting using the equipment will begin next year. It will be made available for purchase first in Israel and the EU, followed by the US and eventually in the Far East. Khanani says that initial contacts with strategic partners in the EU are already established.
The creation of such robots has become possible due to the development of multi-disciplinary technology. Traditionally robots have been used for repetitive tasks within a controlled environment, but until now, this technology has struggled with tasks that involve a changing environment or the performing of a variety of tasks – which is generally a requirement within a warehouse or field, where robots need to be able to adjust to fast-evolving needs, cope with irregular items and adapt to changing environments.
Recognising this, researchers at Cambridge Consultants created a robot ‘hand’ which is capable of tasks like choosing the type of grip, lifting and placing items, calculating in which order to pick fruit and recognising different objects. “We use multi-disciplinary technology which includes machine vision, electronics and mechanical engineering expertise,” Roberts says. “Our industrial sensing and control team has combined high-powered image-processing algorithms with low-cost sensors and commodity hardware to allow ‘soft’ control of robots when the task is not rigidly defined.
“The system is capable of handling objects for which no detailed computer-aided design (CAD) exists – a necessary step to using a robot with natural objects which, although they share some characteristics, are not identical.”
Cambridge Consultants' robot hand can pick up and place different fruits
Hand of robot
The result was a robot ‘hand’ that could move fruit from one place to another. Possessing soft-suction silicon grips, the hand could distribute weight across several grippers thus allowing it to pick up soft oranges and bananas as well as hard apples without causing any damage. It could also sort fruit by colour, dividing red apples from green apples. This demonstrator hand highlighted the increasing versatility of robot technology.
As a result of this innovation, Cambridge Consultants has been asked to design a robot that could deal with plants in a field, identifying spacing between plants, and removing the weeds between them.
“There are many different applications for this technology. These techniques can be combined with harvesting, sorting and classifying fruit and vegetables, filling packaging in a field, even washing and disinfecting,” Roberts explains.
Employing robot technology for these purposes may well have implications for the way produce is actually grown, he adds. “Using this technology in harvesting fruit such as apples could lead to changes in the shape of the trees, with producers growing them thinner and more slender to suit the equipment. Or take lettuce growing - instead of growing them close to the ground for the human picker to cut, they might be developed to have longer stems that could be cut by machinery. [As I said] it is now a problem of engineering rather than science. All it needs is for the engineer to put the systems in place to answer questions such as how the equipment could be fastened onto a tractor rather than plugged into a wall, and how the robot could deal with dust and dirt.”
Anyone who has seen a robot lawnmower in use around a garden can immediately appreciate how easy it would be to set up necessary lines to allow a machine to move around along pre-determined routes.
“Strawberries grown in raised beds or hanging down from staging would be ideal for this type of robot system. The fruit is exposed and hangs away from the leaves allowing it to be harvested by robots moving down the sides. This is already done in the large scale growing of tomatoes and peppers,” Roberts says.
Vege-bot and broccoli harvesting
Trials of alternative styles of robot technology are underway at the University of Cambridge and the University of Lincoln. The University of Cambridge is trialling a Vege-bot which can handle and cut iceberg lettuces.
Harvesting broccoli by robot could well be possible within two years, researchers believe. The University of Lincoln is also involved in a BBSRC/Innovate UK project designed to create an automated broccoli harvesting machine.
Dr Simon Pearson of the University of Lincoln’s National Centre for Food Manufacturing (NCFM) says: “We have been focusing on the creation of systems to enable robots to see and identify products, followed by cutting, harvesting and packing. 3D cameras give robots the co-ordinates to identify where a product is. The problem is that most 3D cameras have been used indoors in a controlled environment. It is a different matter outside with varying weather conditions and light levels. It can be very light and sunny, then cloudy and this affects the light levels. We had to investigate different cameras to explore what could be used outdoors so as to develop the algorithms necessary. “
He continues: “We think we have the camera technology and are now working on the next stage of the project combining camera technology and harvesting techniques. We are two years away from making a working machine that can be deployed in the industry.”
Alongside this robotic broccoli-harvesting concept, the NCFM is also researching technology for precision robotic weeding and the creation of 3D mapping techniques to improve the precision of agricultural sprayers enabling only those plants which need spraying to receive the spray. Pearson points out that there are some limitations involved since some crops are more suitable for robotic weeding and harvesting than others. For example, the presence of leaves around cauliflowers makes it harder for robots to identify the cauliflower and to weed around it, and harvest it.
“We are at the first stage of identifying broccoli crops,” he says. “Then it moves onto crops with cluster fruits such as tomatoes while the third level with plants that cannot be seen immediately such as cauliflowers. This is a subject that will run and run. There will be a lot of technological development over the next five to 10 years.”
Costs are a key issue when it comes to the creation of robot technology as Pearson indicates. “It is important to make this technology affordable and low cost. Fundamentally the systems are not expensive, it is the research costs, which are the big cost.
“This is where the government can help by funding the research needed to make these systems work. There are a lot of long-term benefits we can open up with the technology, not just for agriculture but for the robotics industry generally, since they will be able to sell the technology worldwide.
“Farmers are asking for these types of products, because they are worried about the cost of labour and the availability of labour in the long term. The industry is being very supportive and proactive in this research.”