Objectives of the service
This ESA activity involved running commercial trials of our scouting and weeding system for organic brassicas (initially broccoli) between 2019 and 2022.
Our service is immediately addressing the sector of organic vegetable farming, which has the following characteristics: (1) it is fragmented by crop and geography (2) labour cost and availability has made manual weeding and harvesting of all organic vegetables very difficult (3) most farms belong to large owner-operator farming companies while smaller farms are organised under large farming collectives.
The data collected by the farm rover is uploaded on the Farm Intelligence System. There, the scouting data can inform crop yield measurement and forecasting models, and it is also critical to launching higher-value services like selective harvesting.
Users and their needs
How do you walk 2,000 hectares of crop every week? You can’t – it’s too big and expensive to cover with people. This means farmers don’t get good data about what’s in their fields. Instead, they rely on estimates.
Not having accurate data means:
-
Overplanting crops – to meet stringent retailer contracts
-
Poor input choices – lack of precision (e.g. fertilisers reduce margins)
-
Poor harvesting choices – farmers harvest blind, out of spec crop is wasted
Accurate data is also the essential starting point for autonomous farming.
Service/ system concept
We use computer vision, AI and edge processing to track every single plant from planting to harvest. We know where each plant is to 1cm accuracy.
We then give farmers actionable information on crop size, health, and forecast. This scouting data is integrated with other data the farmers may already have for an information solution.
Our scouting solution is rover-based and fully autonomous. Our rover system will also be able to pull a weeding implement which uses the scouting information “on the fly” to eliminate weeds one by one.
We are commercialising 3 core technologies:
-
A fully autonomous scouting and concentrated-light weeding system
-
A cloud-based Farm Intelligence System
-
Operating system for PFaaS (precision farming as a service)
Our technology stack includes ROS and the business information system Tableau.
Space Added Value
The project utilised RTK GPS to use as localisation data for the plant level data from the rover-mounted system. The localisation data is also used for fleet management.
The project uses technology derived from the ExoMars project to build the fourth generation of the rover system.
Current Status
The final phase of development entailed the ruggedisation of our core technology stack, specifically the weeding module and the CLAWS rover. Both tasks have been successfully completed.
In the next phase, the weeding module underwent extensive lab testing, yielding positive results. Based on these results, the company proceeded to initiate production of three additional modules. Subsequently, the three modules were integrated onto the CLAWS rover and the company is currently conducting incremental field testing and development of the CLAWS.
In the commercial version of the CLAWS rover, solar panels have been mounted to reduce or eliminate the need to charge rover.
