Developing an R&D field robot
A R&D platform based on the Twisted Fields Acorn rover. Using the existing electronics and software stack which features:
- CM4 CANBUS compute module carrier
- BLDC CANBUS steering/ motion drivers
- RTK integration
- Python software stack
- Optional ROS2 extension
Proposed BOM for R&D field robot
Project v1.0 – Mechanical weeding between crop rows
The project is to test and validate the Visual servoing code we have ported to ROS2 and integrate it into Twisted fields code. Using visual servoing as the local planner & Twisted fields as the global planner / GPS waypoint follower.
The robot will pull simple mechanical hoe between the crop rows
Project v2.0 – Automated charging
Automatic charging. Development of Navigation and mechanical components to enable the robot to return to a charging dock
Project v3.0 – Delta weeding
This will enhance the weeding capability using a AI driven Delta robot to enable more accurate weeding
Project v4.0 – Harvesting
Lease open hardware robots to farmers.
1) I am a working horticultural business and can provide a site near Bristol for robotic experiments in weeding, navigation etc. Agroecology lab will trade as https://nibleyleaves.co.uk/ this year. We sell mixed salad to Bristol restaurants.
2) I have some expertise in LoRa mesh networking and could provide a long-range comms system for Telemetry, RTK data based on the established https://github.com/meshtastic/ project
3) I have considerable experience in wireless mesh networking and could provide field-scale multi-radio mesh connectivity, https://libremesh.org/howitworks.html
4) I am collaborating on https://github.com/Twisted-Fields on the Acorn rover and have a particular interest in developing a more accessible scaled-down version of the rover. This could be used as an open-hardware experimental platform for RTK sensor fusion/ Visual servoing.
5) I have got this ported this project to ROS2 https://github.com/Agroecology-Lab/visual-multi-crop-row-navigation/tree/ROS2 and would be interested in developing it further as an in-crop visual-servoing navigation strategy.