The ‘Multipurpose Orchard Robotics’ project is a four year collaboration between Robotics Plus Ltd, University of Auckland, University or Waikato and Plant and Food Research aiming to automate the harvesting and pollination of kiwifruit and apples. The team successfully secured over $10M in funding, with a mix of public and private investment, including $7.6M grant from the Ministry of Business Innovation and Employment (MBIE). The aim is to produce world leading export technology, based on RPL’s current kiwifruit harvesting robot, to meet global labour shortages, rising orchard productivity costs and to address issues like bee colony collapse disorder, which affects fruit pollination, reducing fruit size and yields.
The shortage and struggle with attracting and retaining suitable labour into horticulture has driven many of RPL’s projects. These labour issues are of great concern to the sustainable and competitive future of the horticultural sector. There is a large demand to mechanise many on-orchard tasks to help reduce the reliance on a seasonal and often migrant labour force.
In addition, “People can only afford to pay so much for food but the rising cost of labour and increasing cost of orchard inputs such as fertiliser means it is becoming more expensive for orchardists to produce their fruit and remain competitive on the international market.” Steve Saunders. Robotics can help maintain a static, or even declining cost to produce crops. Over time robotics will become cheaper to produce and their role will become more advanced. The technology will also allow resources, like fertiliser or sprays, to be used more efficiently providing additional environmental and cost benefits.
Developing a foundation of core technology that is transferable across many horticultural tasks and crops for horticulture globally. The centre of this approach is a sensing system that is capable of determining what is happening in the machines environment. For instance, being able to distinguish fruit for harvesting, flowers for pollination, or objects around the machine for navigation. The biggest challenge of making these systems reliable is developing them to automatically cope with the vast amount of variability within both crop and environment. Unlike most automation within factories that deal with man-made parts, everything from item appearance, size, shape, position, etc. are inconsistent. With a reliable sensing system, different mechanical systems, like arms, can be commanded to perform the necessary tasks like driving around the orchard, picking fruit or pollinating flowers.
The central system is developing an Autonomous Mobile Modular Platform (AMMP) that task performing systems attach to. The AMMP will be capable of driving around an orchard by itself, stopping at the appropriate spots for tasks to be performed. The modularity aspect of the platform allows for other systems to be attached to it. These other systems, like sensing systems, custom arms or spraying systems, perform the necessary tasks without needing to build vehicles for every seasonal task.
Robotics Plus are both the commercial partner and Dr Scarfe is the technical leader on this project. RPL is driving a commercial focus to the research, much of which will be based at our facility in Tauranga. The research builds Dr Alistair Scarfe’s PhD work, where he successfully demonstrated harvesting of kiwifruit. The prototype kiwifruit harvester was assessed as being a world leader in discrete harvesting of fruit or nuts by an international expert, it also featured in the NZ Story and NZ Innovation Showcase 2011 movies.
“Our academic partners will bring a huge amount of knowledge and expertise to this project. It’s also important we retain a commercial focus to make sure the technology being developed is suitable for the environment it will be working in.” Dr Alistair Scarfe.
University of Auckland, led by Dr Bruce MacDonald, are focusing on developing the sensing systems. These systems will allow the robotics to interact with and navigate around their operational environment.