[{"_path":"/tools/weeding","_draft":false,"_partial":false,"_locale":"en","_empty":false,"title":"The Rover for Weeding","description":"We designed the ROMI Rover as a farming tool to assist vegetable farmers in maintaining vegetable beds free of weeds. It does this by regularly hoeing the surface of the soil and thus preventing small weeds from taking root. It can do this task mostly autonomously and requires only minor changes to the organization of the farm.","excerpt":{"type":"root","children":[{"type":"element","tag":"section-cover","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:title":""},"children":[{"type":"element","tag":"h1","props":{"id":"the-romi-rover"},"children":[{"type":"text","value":"The ROMI Rover"}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"The ROMI Rover is a farming tool that assists vegetable farmers in maintaining vegetable beds free of weeds. It does this by regularly hoeing the surface of the soil and thus preventing small weeds from taking root. A weekly passage of the robot should be sufficient to keep the population of weeds under control."}]},{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"base-embed","props":{"id":"0-Focjjzc7k"},"children":[]}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/tools/weeding/1-rover.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"The ROMI Rover can perform the weeding task mostly autonomously and requires only minor changes to the organization of the farm. It is designed for vegetable beds between 70 cm and 120 cm wide (not including the passage ways) and for crops up to 50 cm high."}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/tools/weeding/2-lettuce.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"It currently handles two types of crops, lettuce and carrots. The lettuce can be planted out in any layout, most likely in a quincunx pattern. In this configuration the rover uses a precision rotary hoe to clean the soil both between the rows and the plants. This process is slower than classical mechanical weeding. The rover can cover up to 600 m²/day."}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/tools/weeding/3-carrots.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"For carrots, the rover uses classical mechanical tools, such as stirrup hoe, to regularly clean the soil in between the rows. In this configuration, the carrots should be sown in line. In this classic configuration, the rover can cover a surface of 7200 m²/day."}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/tools/weeding/4-vegetable-beds.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"In addition to weeding, the embedded camera can be used to collect images of the vegetable beds. This images can be used by "},{"type":"element","tag":"a","props":{"href":"/tools/crop-monitoring"},"children":[{"type":"text","value":"the Farmer's Dashboard"}]},{"type":"text","value":"."}]}]}]},"img":"/assets/tools/weeding/0-cover.png","docs":"https://docs.romi-project.eu/Rover/","_type":"markdown","_id":"content:2.tools:3.weeding.md","_source":"content","_file":"2.tools/3.weeding.md","_extension":"md"},{"_path":"/research/adaptive-systems","_draft":false,"_partial":false,"_locale":"en","_empty":false,"title":"Adaptive systems","description":"Farming robots must be able to work in complex and variable environments. For example, plants are complex, time-varying objects. Outdoor fields are very uncontrolled environments, too.","excerpt":{"type":"root","children":[{"type":"element","tag":"section-cover","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:title":""},"children":[{"type":"element","tag":"h1","props":{"id":""},"children":[{"type":"element","tag":"binding","props":{"value":"$doc.title"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"We investigate advanced, open-ended learning techniques to gain insight in how farming robots can adapt their image processing capacities when facing plants on which they have not been trained, and insights in how they can learn to optimise the collection of visual information when facing complex plant scenes."}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/research/adaptive-systems/1-curiosity.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"We work on curiosity-driven goal-directed exploration behaviours to move an image sensor around a plant. The artificial curiosity system assigns interest values to pre-defined goals, and drives the exploration towards those that are expected to maximise the learning progress."}]}]},{"type":"element","tag":"section-split","props":{},"children":[{"type":"element","tag":"template","props":{"v-slot:media":""},"children":[{"type":"element","tag":"p","props":{},"children":[{"type":"element","tag":"img","props":{"alt":"dashboard screenshot","src":"/assets/research/adaptive-systems/2-rl.jpg"},"children":[]}]}]},{"type":"element","tag":"p","props":{},"children":[{"type":"text","value":"We also train agents to move the camera using Reinforcement Learning. In this technique the agent has to learn how to map situations to actions so as to maximize a numerical reward. The learner is not told which actions to take, but instead has to discover which actions yield the most reward by trying them. In our case, the reward is derived from building an accurate 3D representation of a plant using a small number of images."}]}]}]},"img":"/assets/research/adaptive-systems/0-cover.png","_type":"markdown","_id":"content:3.research:2.adaptive-systems.md","_source":"content","_file":"3.research/2.adaptive-systems.md","_extension":"md"}]