The use of machines to automate previously manual tasks has revolutionised many aspects of agriculture, such as seed planting, weeding and harvesting of some fruits and vegetables. However, many jobs in agriculture are still manual, despite their apparent simplicity, largely because of a combination of the programming/set-up costs for robots to do them are prohibitively expensive, or these deceptively simple tasks are surprisingly difficult for robotics. Robots are good at executing repeated tasks with identical objects. Many agricultural tasks are varied, however, involving irregular objects in irregular and/or non-uniform manipulations that require feedback from the environment and a basic "understanding" of the task to be completed. While straightforward for humans, such tasks remain largely out of reach for robots. Lack of automation limits productivity for the agriculture sector, which is compounded by chronic labour shortages for low-skilled repetitive manual work, and contributes to food waste. In this project, we will build on recent developments in Artificial Intelligence (AI) and Machine learning (ML) to develop low-cost robotics solutions for a number of currently hard-to-automate agricultural tasks. We aim to drastically reduce the time and cost to set-up robotics installations, as well as improve accuracy and reliability. As a demonstration of this new technology, two processes will be taught to the robot - processing celeriac (removal of leafy top, unwanted roots and imperfections) and replanting of germinated pak choi (picking small pak choi growing in ellepots and moving it to be planted out in the glass house). These tasks were identified by our project partner, M&W Mack, as being ones where automation is urgently needed. Our robotics solution will be adaptable to a wide range of currently manual agricultural processing jobs, leading to savings for food producers and processors, reduced waste and lower prices for consumers. Our existing relationships with large farming and food-companies will allow us to bring our innovation into the market quickly, creating significant revenue and jobs for our company, and helping the agricultural sector to improve the efficiency and productivity of many vital processing tasks.

Our project will demonstrate the feasibility of a new robot motion planning algorithm that will increase the productivity of industrial robots. We will demonstrate the algorithm on a robot packing fruit and vegetables. The technology can be applied much more broadly to help manufacturers reduce costs and the CO2 footprint of their industrial robots.

University of Lincoln and Xihelm are collaborating to deliver _Qualicrop_ - aiming to develop crop sorting for produce, lowering costs to consumers. It will build sophisticated sorting systems to disintermediate the supply chain - unlocking technology only available to other sectors, improving the quality of sold fruit & vegetables and lowering prices to consumers and to make automation technology affordable for smaller farmers. Extensive research will be made into using modern imaging technology and machine learning to detect issues with crops in a just-in-time manner. Both partners are dedicated to advancing Equality, Diversity and Inclusion as part of this project. The system once commercialised will allow tight-margin participants in the value chain to increase their margins, lowering chargebacks and wastage, and lower CO2 impact by reducing food miles. Furthermore, it can cost effectively open new markets for different crops, and support farmers in England to accelerate their technology adoption.