The AgriRobot project responds to a challenge that has been set by the Jiangsu-UK Industrial Challenge Programme. The challenge addressed by the project was to build a prototype proof of concept semi-autonomous agricultural robot capable of intelligently spraying Orchard crops using automatic navigation in a field, rows between plants, trees, lane following.
The innovation provided is the creation of a turnkey solution offering simple autonomous agricultural robot straight line navigation and intelligent spraying with remote monitoring capabilities with sensing and control systems enabling full autonomy of existing crop sprayers.
The need that the project addresses is for intelligent solutions to reduce labour and material costs. Future farming utilising intelligent crop spraying is seen as the answer to improving farming productivity and cost reduction overall in China.
The technical challenge lied in producing a fully automated system for crop spraying in rows in a field. Market for these solutions has yet to be fully developed and is still in its infancy. Some elements of intelligent spraying are evident, but not full driverless autonomy with automatic sense spraying. We plan to exploit this gap in the China market place.
The contribution of the project is a prototype semi-autonomous fully automated intelligent crop spraying farming machine using existing light weight independently driven crop sprayers. The improved nearest state is the addition of simple low costs novel navigational control combined with fully automated intelligent spraying control moves us to improved state.
The industrial research outcomes are the provisioning of a fully working prototype device with remote sensing and monitoring capabilities registered within the UK and ideally used under licence in China.
The project provides solutions to two challenges that form a single solution for the autonomous and optimised efficient and effective delivery of pesticides in orchard environments.
- Navigation System: The developed autonomous agricultural vehicle has the capability to automatically travel along a row with an orchard or farm without collisions and can also intelligently manoeuvre to change from one row to another. To achieve this the system applies a computer vision and AI-based approach to detect trees and other obstacles such as humans from on-board RGB-Depth cameras, and to navigate the vehicle as required in order to operate the onboard pesticide delivery system.
- Spraying System: The developed sensor-based variable-rate air-assisted sprayer can analyse the presence, size, shape, and density of target trees/plants, providing a unique and precision spraying operation applying only the necessary amount of pesticide when needed, leading to more efficient spraying while reducing the demand of human operators.
The project is split into a UK- and a Jiangsu-based project, which together encompass a number of key aspects:
- Identification of real-world requirements, use cases and the overall system architecture require to address the project topics
- Design and implementation of a sensing subsystem for autonomous vehicle navigation through an orchard environment (see also further information provided by Loughborough University here)
- Design of Navigation and collision avoidance subsystems that plan and carry out navigation actions in a safe manner
- Development of an automatic spray control subsystem that ensures efficient and effective dispersal of pesticides as required in the application setting
- Integration of the system components and of the system with prototype vehicle controls for evaluation, testing and demonstration purposes
- Dissemination, exploitation and business model development during and after completion of the co-funded research projects
The AgriRobot research project was active from April 2018 to March 2020, and has delivered the subsystems mentioned above as well as a prototype vehicle implementation for development and testing.
The AgriRobot project has been supported by Innovate UK (Project No. 104016) and Jiangsu Province.