Source-linked AI summary
Agricultural Robotics: The Future of Robotic Agriculture
Tom Duckett, Simon Pearson, Simon Blackmore, Bruce Grieve, Wen-Hua Chen, Grzegorz Cielniak, Jason Cleaversmith, Jian Dai, Steve Davis, Charles Fox, Pål From, Ioannis Georgilas, Richie Gill, Iain Gould, Marc Hanheide, Alan Hunter, Fumiya Iida, Lyudmila Mihalyova, Samia Nefti-Meziani, Gerhard Neumann, Paolo Paoletti, Tony Pridmore, Dave Ross, Melvyn Smith, Martin Stoelen, Mark Swainson, Sam Wane, Peter Wilson, Isobel Wright, Guang-Zhong Yang
TL;DR
Global agriculture faces pressures from population growth, climate change, migration-related political pressures, rural-to-urban population drift, and an aging population. This white paper reviews agricultural Robotics and Autonomous Systems and identifies research and innovation needs, concluding that robust platforms and integrated capabilities are required for wider deployment.
Problem
UK agricultural robotics research is fragmented, projects are too few and small-scale, and many varied or disadvantaged farms remain insufficiently served by existing projects.
Method
The white paper reviews the state of the art in agricultural Robotics and Autonomous Systems and explores research and innovation needs across sensing, planning, manipulation, learning, interoperability, and human-robot collaboration.
Results
The paper identifies robust agricultural platforms and improved integrated capabilities as research and innovation needs for delivering the future vision of robotic agriculture.
Takeaways & Limitations
Agricultural RAS development should address both discrete technologies and their large-scale integration, interoperability, safety, and user acceptance.
Takeaways & Limitations
Vertical production systems may not be economically or environmentally competitive with existing systems for the foreseeable future.
Abstract
from arXiv · showhide
Agri-Food is the largest manufacturing sector in the UK. It supports a food chain that generates over £108bn p.a., with 3.9m employees in a truly international industry and exports £20bn of UK manufactured goods. However, the global food chain is under pressure from population growth, climate change, political pressures affecting migration, population drift from rural to urban regions and the demographics of an aging global population. These challenges are recognised in the UK Industrial Strategy white paper and backed by significant investment via a Wave 2 Industrial Challenge Fund Investment ("Transforming Food Production: from Farm to Fork"). Robotics and Autonomous Systems (RAS) and associated digital technologies are now seen as enablers of this critical food chain transformation. To meet these challenges, this white paper reviews the state of the art in the application of RAS in Agri-Food production and explores research and innovation needs to ensure these technologies reach their full potential and deliver the necessary impacts in the Agri-Food sector.
UKRAS.ORG
The white paper frames RAS as an enabler of Agri-Food transformation, identifies UK capability and coordination barriers, and sets out a vision for interoperable robots working with people.
- UKRAS.ORG: Agri-Food faces pressures from population growth, climate change, migration, rural-to-urban drift, and an aging population, motivating RAS-enabled transformation.The UK sector generates over £108bn annually, employs 3.9m people, and exports £20bn of manufactured goods.
- UKRAS.ORG: The UK RAS community is small and dispersed, lacks dedicated training paths, and has insufficient low-TRL basic research to support industry innovation.The paper also reports that current projects are too few and small-scale for complex integration challenges.
- UKRAS.ORG: The proposed vision is for smart, flexible, interconnected robotic teams that self-organise, coordinate with humans, and support sustainable intensification and food security.Examples include interchangeable tools, low-tillage systems, soft robotic grasping, and distributed sensing.
- UKRAS.ORG: The paper calls for robust agricultural platforms and improved sensing, planning, manipulation, learning, interoperability, and safe human-robot collaboration.These capabilities are presented as research and innovation needs for delivering the proposed vision.
- UKRAS.ORG: Adoption is expected to proceed gradually, with humans and robots collaborating as robots take on increasingly complex work.The paper associates this transition with productivity gains and movement of human jobs up the value chain.
- UKRAS.ORG: Successful delivery requires close collaboration across RAS, academic, industry, and Agri-Food communities, while the UK has an opportunity to establish global leadership.The paper highlights cross-disciplinary activity and the need for coordinated networks.
1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS
The paper links agricultural RAS to economic, social, and environmental pressures, including resource constraints, soil degradation, water scarcity, pesticide dependence, and machinery impacts.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: The UK Agri-Food chain generates over £108bn annually, employs 3.7m people, and produced £20bn of exports in 2016.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: Global food production faces pressure from population growth, productivity needs, climate change, migration politics, rural-to-urban drift, and population aging.The paper notes that Brexit-related uncertainty is already affecting migrant worker confidence and availability in the UK.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: Advanced RAS technologies are presented as a potential route to transform Agri-Food, with food-manufacturing digital technologies estimated to add £58bn of UK GVA over 13 years.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: Food production uses 18% of UK energy consumption, while heavy machinery, pesticides, and expanding fields contribute to soil compaction and wildlife risks.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: Soil degradation losses in England and Wales are estimated at around £1.2bn annually, with compaction also increasing waterlogging, runoff, nitrous oxide emissions, and habitat loss.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: Agriculture uses 70% of global freshwater supplies even as 4bn people live in regions with water scarcity, increasing the need for water-use efficiency.
- 1. ECONOMIC, SOCIAL AND ENVIRONMENTAL DRIVERS: The paper describes a transition from large fossil-fuel platforms toward smaller electric robotic fleets as a possible route to lower emissions and safer, more efficient equipment.Electric drives can improve efficiency and reduce risks associated with mechanical linkages.
1.3 PRECISION AGRICULTURE
Precision agriculture, also known as smart farming, uses sensing and automation to manage spatial and temporal variability and target agricultural inputs more effectively.
- 1.3 PRECISION AGRICULTURE: Precision agriculture uses monitoring and intervention techniques, enabled by sensing technologies and automation, to improve efficiency.Its development addresses variability from farm scale down to field and sub-field scales.
- 1.3 PRECISION AGRICULTURE: Small, intelligent autonomous machines can reduce waste, improve economic viability, reduce environmental impact, and increase food sustainability.
- 1.3 PRECISION AGRICULTURE: Robotic systems can expand intervention opportunities by operating on wet soils and at night.
- 1.3 PRECISION AGRICULTURE: Field robots collect data on soils, seeds, livestock, crops, costs, equipment, water, and fertiliser, while IoT and analytics support planning and resource decisions.
1.4 LIVESTOCK AND AQUACULTURE
RAS is already used in livestock production and is expanding into monitoring, welfare, and farm operations, while aquaculture requires systems robust to hostile and difficult-to-access environments.
- 1.4 LIVESTOCK AND AQUACULTURE: Robotic milking is commercially deployed, and robots are also used or developed for waste removal, feed handling, livestock monitoring, and field-data collection.An EU foresight study predicts that around 50% of European herds will be milked by robots by 2025.
- 1.4 LIVESTOCK AND AQUACULTURE: Timely animal data can reduce waste and environmental pollution while improving welfare and farm productivity.Precision farming can support continuous, animal-centric assessment of individual condition and state.
- 1.4 LIVESTOCK AND AQUACULTURE: Maintaining animal health and welfare is constrained by monitoring demands, feed costs, environmental regulation, consumer concerns, and possible low-welfare imports after Brexit.
- 1.4 LIVESTOCK AND AQUACULTURE: Aquaculture robots and sensors must be robust because production occurs in hostile, remote, and difficult-to-access environments.
1.5 DISADVANTAGED FARMS
Previous agricultural robotics projects have largely targeted large, flat, monoculture or controlled indoor environments, leaving varied and hilly smaller farms comparatively underserved.
- Most previous agri-robotics projects have focused on large flat monoculture fields and controlled industrial-scale indoor growing.
- Northern and western Britain contain more varied, hilly terrains with smaller family farms, smaller vehicles and more manual work, especially dairy and sheep farming.
- These disadvantaged farms have largely been left behind by successive waves of automation, including agricultural robotics.
- Large machines cannot navigate some hill-farm terrains that human workers historically used, contributing to the disuse of certain farms and moorlands.
1.6 NON-CONVENTIONAL CLOSED (‘VERTICAL’) FARMING
Closed vertical farming combines controlled indoor environments with robotics, sensing and crop design to expand production possibilities, although its competitiveness remains uncertain.
- Vertical farming controls environmental factors including nutrients, temperature, humidity and lighting within indoor closed environments.
- Required vertical-farm components include environmental control, crop-nutrient chemistry, hydroponic or other growing media, semiconductor lighting, non-invasive sensing and robotics.
- Crops designed specifically for controlled systems can maximise desired outputs without breeding for conventional field constraints such as pests and weeds.
- Robotic systems could support novel crop traits, reduce downstream energy and waste, and produce plants compatible with autonomous care and harvesting.
- Vertical production systems may not be economically or environmentally competitive with existing systems for the foreseeable future.
- Their rationale may depend on urban and industrial integration, including use of plants as sinks for effluents, heat and excess energy.
1.7 FOOD MANUFACTURING AND PROCESSING
Robotics research extends beyond the farm gate into food manufacturing and processing, where robotisation, collaboration and tracking may improve productivity, safety and supply-chain information.
- Post-harvest activities present additional research needs, including meat cutting tasks facing shortages of suitably skilled workers.
- Cobots, which work together with humans, offer an alternative for increasing productivity, improving health and safety, and attracting skilled workers and graduates.
- Agricultural robotics could enable earlier labelling and tracking throughout manufacturing, improving origin information and speeding responses to food-safety issues.
- Food metadata could be fed back to field operations to further improve primary production.
- Potential synergies between agricultural robotics and downstream processing could unlock whole-supply-chain efficiencies through future RAS applications.
1.8 ETHICAL ISSUES
Ethical issues concern employment, labour transitions and control over agricultural data as robotics and autonomous systems become more prominent in Agri-Food.
- Agricultural robotics raises concerns about AI’s impact on employment amid reliance on migrant labour and an ageing agricultural workforce.The average age of a UK farmer is 58 years, while the sector relies on approximately 65,000 migrant labourers.
- Robotic automation is presented as both performing undesirable work and creating rewarding employment that can move human jobs up the value chain.
- Data ownership raises ethical concerns because a small number of companies may control most information and infrastructure.
- Smart Farming could develop as closed proprietary systems or open systems that let stakeholders choose technologies and business partners.
- Open-source data, publicly funded collection networks and secure, objective measurements are identified as considerations for agricultural data governance.
- Autonomous agricultural robotics also require attention to liability frameworks and the reuse of robot-collected data.
2. TECHNOLOGICAL FOCUS
The paper focuses on applying robotics and digital technologies to repetitive agricultural tasks, while enabling a gradual transition from existing farming systems toward coordinated human-robot production.
- Agricultural robotics research targets repetitive tasks where automation can outperform traditional human or large-machine approaches.Priority areas include crop-proximate platforms, advanced manipulation, tactile interaction, and soft-fruit picking.
- Integrating automation, remote sensing, large datasets, IoT, Big Data and artificial intelligence could improve crop and livestock production.These technologies can fuse and interpret data, assess crop status, and plan interventions for weather, disease and pest changes.
- The technology vision combines smart, robust, interconnected robots with interchangeable tools, soft grasping, sensors and machine learning across farms and food factories.The intended benefits include sustainable intensification, environmental protection, food quality and manufacturing productivity.
- Only a gradual transition is realistic because few large companies can afford full-automation disruption and some technologies remain insufficiently robust or cost-effective.Soft-fruit picking still requires fundamental research in sensing, manipulation and soft robotics.
- Human-robot collaboration is fundamental in the short term, with mixed systems combining existing farm practices and increasingly autonomous implements.Human-driven tractors could tow robotic implements for selective harvesting or weeding before autonomous vehicles replace legacy vehicles over time.
3. ENABLING TECHNOLOGIES FOR FUTURE ROBOTIC AGRICULTURE SYSTEMS
Future agricultural robotics depends on adaptable platforms, robust sensing and perception, integrated heterogeneous fleets, manipulation suited to variable food products, and safe human-robot collaboration.
- Robotic platforms: Agricultural robots may be domain- and task-specific or generic, but both types must address farm-specific infrastructure and operating conditions.Early systems may function only on individual farms and have limited transferability across sites.
- Robotic platforms: Commercial platforms must improve robustness, reliability, power management, locomotion, real-time control and operation in changing weather and soil conditions.Locomotion should be co-designed with sensing and collection capabilities because tasks such as fruit collection affect movement requirements.
- Robotic platforms: Platform weight and locomotion affect soil and crops, motivating alternatives such as tracked, wheeled and potentially legged robots.Legged systems could reduce footprint and move flexibly through narrow spaces while carrying specialised sensors.
- Sensing and perception: Integrated sensor systems, satellite and drone remote sensing, and machine vision can support field mapping, crop monitoring, animal assessment and targeted interventions.Machine vision is already used for animal weight estimation, body-condition monitoring, illness detection and identification.
- Sensing and perception: Robotic vision needs open-ended learning and continual adaptation to seasonal changes, diseases, pests and new crop varieties.Most existing work focuses on pre-deployment training rather than long-term model adaptation, creating a need for ground-truthing and semi-supervised interfaces.
- Planning and coordination: The potential of agricultural robotics depends on heterogeneous fleets that coordinate ground and airborne vehicles centrally or distributively.UAVs support monitoring but their payload and durability constrain larger-scale intervention and treatment delivery.
- Planning and coordination: Fleet management must integrate goal allocation, motion planning, coordination and control while accounting for human presence, predicted actions and legible robot motion.These sub-problems have largely been studied in isolation, so integration and scaling to real-world scenarios remain research needs.
- Manipulation: Agricultural manipulation must handle variable size and shape, heterogeneous positioning and fragile products, while vision-guided grasping can fail under occlusion.Compliant grippers and tactile feedback could reduce sensing complexity and help adjust grasp actions during picking.
4. THE CHALLENGES
The paper identifies challenges across phenotyping, cultivation and crop care, especially the difficulty of transferring robotic assessment from controlled environments to diverse real-world fields.
- RAS contributions span crops, livestock and aquaculture, from phenotyping through primary production, with potential economic and ecological benefits.The paper separates technology challenges into breeding or phenotyping and farming or primary production.
- Laboratory: Laboratory breeding robotics can reduce manual intervention, but implementation cost, complexity, questionable reliability and limited technology readiness restrict uptake.The systems support identifying traits such as drought tolerance, disease resistance, shelf-life and nutritional quality.
- Field: Robotics enables mass direct in-field phenotyping under true farm conditions, but uncontrolled environments make it difficult to identify which specific trait produced a beneficial response.Repeated, detailed assessment of individual plants could shift agri-genetics toward field-based evaluation.
- Establishment and Seeding: Small smart electric robots could reduce soil compaction through micro-tillage and automate seed placement, mapping and targeted nutrient, soil and water management.Traditional cultivation is estimated to use 80%-90% of its energy repairing damage caused by large tractors.
- Crop Care: Autonomous aerial and ground robots could collect timely crop-health data, but fusing and interpreting measurements across devices and spatial-temporal resolutions remains challenging.
THE OPPORTUNITIES FOR ROBOTICS AND AUTONOMOUS SYSTEMS IN AGRICULTURE
Agricultural RAS opportunities include targeted crop operations, selective harvesting, in-field transport and parallel monitoring or protection activities, but selective harvesting remains technically demanding.
- In-field Transportation: Autonomous robots could support in-field transportation and conduct additional monitoring, equipment-security and wildlife-protection activities alongside other field operations.
- Robotic crop-care opportunities include vision-guided mechanical weeding, selective micro-spraying, laser weeding, targeted irrigation and pre-harvest yield forecasting.
- Selective Harvesting: Selective harvesting requires sensing quality before harvest and removing the desired product without damaging the remaining crop.Its central challenge is autonomous sensorimotor control under variable crop and growing-environment conditions.
- Selective Harvesting: Selective harvesting involves trade-offs between adapting robotic systems to a crop and adapting the growing environment to support robotic harvesting.The balance may differ across crops and environments, while compliant implements can shift precision demands toward software and sensing.
5. BARRIERS, CONCLUSIONS AND RECOMMENDATIONS
The paper identifies fragmented capacity, insufficient foundational research, limited training, and small-scale projects as barriers to realizing Agri-Food RAS potential. It recommends coordinated networks, skills development, large-scale integration projects, and broader research and international collaboration.
- Barriers and capacity: The UK Agri-Food RAS community is small and dispersed, lacks dedicated training pathways, and requires both defragmentation and expanded human-resource capacity.Recommendations include Network+ grants, larger multi-centred hubs, Centres for Doctoral Training, and skills development through apprentice level.
- Barriers and capacity: Insufficient low-TRL basic research threatens to underpin the onward delivery of heavily funded high-TRL Agri-Food RAS innovation.The paper highlights a potential mismatch between significant government investment in translational research and ongoing foundational research needs.
- Barriers and capacity: Current Agri-Food RAS projects are too few and too small-scale to resolve the integration and interoperability of technologies such as navigation, safety, manipulation, and perception.The paper calls for large-scale industrial applications alongside continued development of discrete technologies.
- Recommendations: Successful RAS delivery requires close collaboration between RAS specialists and academic and industry practitioners across Agri-Food domains.The paper gives crop breeding for robot-visible and robot-pickable phenotypes as an example of domain integration supporting RAS application.
- Recommendations: The paper recommends coordinated foresight across UK research councils to integrate RAS into wider Agri-Food research programmes.Such a review is intended to recognise cross-disciplinary challenges and encourage responsive-mode applications aligned with Agri-Food RAS.
- Recommendations: UKRI should commission a small number of large-scale integration or “moon shot” projects, while international collaboration can support responses to global challenges such as sustainable food security.These recommendations aim to demonstrate routes through interoperability problems and accelerate RAS technology development.