Periodic Reporting for period 2 - GALIRUMI (Galileo-assisted robot to tackle the weed Rumex obtusifolius and increase the profitability and sustainability of dairy farming)
Okres sprawozdawczy: 2021-06-01 do 2023-09-30
The main objective of GALIRUMI is to harness the possibilities that GALILEO brings to deliver robot weeding for herbicide-free weed control in dairy farming. Robotic weeding will reduce the environmental impact of dairy farming by eliminating herbicide use and reducing exposure of farm workers to herbicides. It will also remove an important obstacle for dairy farmers to switch to organic production, thereby contributing to an increase in production of organically produced milk and higher incomes for farmers.
Why robotic weeding now? There have been great advances in relevant technology over the last years:
* New weed control methods: electrocution and frequent defoliation
* Galileo differentiators (multiple frequencies with modern modulations) enable robust navigation for autonomous agricultural robots even in locations where other GNSS do not work sufficiently well.
* Mobile robot platforms are available commercially and need no longer be painstakingly constructed in projects.
* Visual detection has become much more robust with the use of machine learning (e.g. deep learning) techniques.
GALIRUMI will significantly reduce:
• Manual labour for weeding
• Weed management cost
• Damage to grassland from the application of herbicide
• Impact of dairy farming on the environment
• Dairy cow discomfort
GALIRUMI specific objectives are:
• Objective 1: to integrate an appropriate Galileo capable receiver for robot positioning in order to provide the required robustness and accuracy (in the order of cm).
• Objective 2: to improve the existing camera-based detection methods for broadleaved dock.
• Objective 3: to construct a prototype robot that damages the weed by applying a high-voltage current to it.
• Objective 4: to construct a prototype robot that degrades a weed over time by removing its leaves several times in a growing season using laser.
• Objective 5: to develop software for robot navigation and mission planning & control.
• Objective 6: to test the robots (O3, O4) in an operational environment.
• Objective 7: The final and most important objective is to evaluate the complementary business models of offering a medium-sized, electrocuting robot as-a-service and offering a small, laser-based defoliating robot for sale.
Why robotic weeding now? There have been great advances in relevant technology over the last years:
* New weed control methods: electrocution and frequent defoliation
* Galileo differentiators (multiple frequencies with modern modulations) enable robust navigation for autonomous agricultural robots even in locations where other GNSS do not work sufficiently well.
* Mobile robot platforms are available commercially and need no longer be painstakingly constructed in projects.
* Visual detection has become much more robust with the use of machine learning (e.g. deep learning) techniques.
GALIRUMI will significantly reduce:
• Manual labour for weeding
• Weed management cost
• Damage to grassland from the application of herbicide
• Impact of dairy farming on the environment
• Dairy cow discomfort
GALIRUMI specific objectives are:
• Objective 1: to integrate an appropriate Galileo capable receiver for robot positioning in order to provide the required robustness and accuracy (in the order of cm).
• Objective 2: to improve the existing camera-based detection methods for broadleaved dock.
• Objective 3: to construct a prototype robot that damages the weed by applying a high-voltage current to it.
• Objective 4: to construct a prototype robot that degrades a weed over time by removing its leaves several times in a growing season using laser.
• Objective 5: to develop software for robot navigation and mission planning & control.
• Objective 6: to test the robots (O3, O4) in an operational environment.
• Objective 7: The final and most important objective is to evaluate the complementary business models of offering a medium-sized, electrocuting robot as-a-service and offering a small, laser-based defoliating robot for sale.
Different activities have been made towards fulfilling GALIRUMI objectives. Specifically, the first integration of the different modules in a weeding robot has been completed. The Galileo capable receiver has been integrated for robot positioning and is currently under optimization with the rest of the modules (towards objective O1). This GNSS module has been individually characterized and its performance has been reported. Moreover, the weed detection module has also been integrated, and is being optimized in order to improve the existing camera-based detection methods (towards objective O2). All these modules are currently integrated into the first version of the laser weeding robot, that already implements all the functionalities. This first version is being optimized towards the construction of the final prototype robot (Objective O5). The same modules except the weeder itself are to be integrated into the electrocution weeding robot. Investigation activities related to the specific characteristics of the weeder are currently ongoing, to select the most appropriate parameters (towards objective O4). To design, develop and manufacture the different modules, a huge work has been performed, specially under WP3.
Moreover, even if O6 and O7 are expected to be achieved by the end of the project, several activities have been undertaken. The first version of the robot has been tested on PEK and KOO premises, in order to get feedback from its use under real field conditions. Several test campaigns have been made, and some others are already planned (towards objective O6). In addition, consortium is working on the first version of the business plan (towards objective O7).
The Kick off meeting was performed in Dalfsen (The Netherlands), quite close to contractors facilities for dairy farmers. System design was completed for the first iteration for all the robot subsystems (positioning, detection, weeding, navigation), having all the subsystems ready for a first integration in the robot platform, that was succesfully performed. A first prototype of the funcional laser weeding robot is available, and has been tested in outdoor environments. GALIRUMI project was presented in several conferences and events where key stakeholders in the areas of agriculture, robotic, communication and AI were involved. GALIRUMI webpage "galirumi-project.eu” was published, which summarized the DoA objectives and provide a overall overview of the project activities and partners involved.
Moreover, even if O6 and O7 are expected to be achieved by the end of the project, several activities have been undertaken. The first version of the robot has been tested on PEK and KOO premises, in order to get feedback from its use under real field conditions. Several test campaigns have been made, and some others are already planned (towards objective O6). In addition, consortium is working on the first version of the business plan (towards objective O7).
The Kick off meeting was performed in Dalfsen (The Netherlands), quite close to contractors facilities for dairy farmers. System design was completed for the first iteration for all the robot subsystems (positioning, detection, weeding, navigation), having all the subsystems ready for a first integration in the robot platform, that was succesfully performed. A first prototype of the funcional laser weeding robot is available, and has been tested in outdoor environments. GALIRUMI project was presented in several conferences and events where key stakeholders in the areas of agriculture, robotic, communication and AI were involved. GALIRUMI webpage "galirumi-project.eu” was published, which summarized the DoA objectives and provide a overall overview of the project activities and partners involved.
GALIRUMI is on the way of advancing beyond the state-of-the-art not only in the separate areas of weed detection, weed degradation, mobile plaform for degrading weeds, but also in the integration of these components into a truly autonomous precision agriculture robotic weeding system as a whole. Further, GALIRUMI will fully explore two options for commercial operation of robots on dairy farms: robot-as-a-service, offered by a farm contractor, a farmer-owned robots.
The result of GALIRUMI will consists of a robot which explores the pasture autonomously using Galileo positioning services. Mission planning and control software is used to instruct the robot where to search and to monitor progress remotely. Multimodal sensing detects the presence and position of weeds. Whenever the robot encounters a weed, a control action is immediately taken. This new weeding robot will be designed for helping dairy farmers to be less dependent on manual labour and herbicides for weed control. This will help minimize the impact on natural resources (by reducing chemical contamination of natural resources and minimizing soil compaction in agricultural land), mitigate climate change (through a reduction of fuel use/ CO2 emissions), and helps contribute to public health (by minimizing herbicide residues in the environment and by promoting safer working conditions for farmers). The breakthrough impact of GALIRUMI is mostly related to its compliance with the factors that further incentivize the dairy sector (organic and conventional) to invest in robotics and automation.
The GALIRUMI project is designed for maximum impact through its experimentation at local scale, and is expected to continue beyond it's execution timespan due to: firstly, the project’s partners who are farming businesses are committed to exploit and maintain the robotic innovations and associated services in their business. Secondly, the data acquired by the sensors used to recognize and identify weeds, represent a precious source of information and knowledge about the location and abundance of weeds which may in time be explained by farm management. Thirdly, the project supports the creation of businesses based on a share economy to lower the costs of expensive equipment, boosting organic agriculture in Europe and leading to a food production free of herbicides.
Other potential impacts are related to the areas of agronomic and environmental (weeding for organic production), animal welfare (support for organic cow outdoor pastures), social (more opportunities for younger agriculture with IT profiles), wellbeing (reduce of time consuming manual weeding labors).
The result of GALIRUMI will consists of a robot which explores the pasture autonomously using Galileo positioning services. Mission planning and control software is used to instruct the robot where to search and to monitor progress remotely. Multimodal sensing detects the presence and position of weeds. Whenever the robot encounters a weed, a control action is immediately taken. This new weeding robot will be designed for helping dairy farmers to be less dependent on manual labour and herbicides for weed control. This will help minimize the impact on natural resources (by reducing chemical contamination of natural resources and minimizing soil compaction in agricultural land), mitigate climate change (through a reduction of fuel use/ CO2 emissions), and helps contribute to public health (by minimizing herbicide residues in the environment and by promoting safer working conditions for farmers). The breakthrough impact of GALIRUMI is mostly related to its compliance with the factors that further incentivize the dairy sector (organic and conventional) to invest in robotics and automation.
The GALIRUMI project is designed for maximum impact through its experimentation at local scale, and is expected to continue beyond it's execution timespan due to: firstly, the project’s partners who are farming businesses are committed to exploit and maintain the robotic innovations and associated services in their business. Secondly, the data acquired by the sensors used to recognize and identify weeds, represent a precious source of information and knowledge about the location and abundance of weeds which may in time be explained by farm management. Thirdly, the project supports the creation of businesses based on a share economy to lower the costs of expensive equipment, boosting organic agriculture in Europe and leading to a food production free of herbicides.
Other potential impacts are related to the areas of agronomic and environmental (weeding for organic production), animal welfare (support for organic cow outdoor pastures), social (more opportunities for younger agriculture with IT profiles), wellbeing (reduce of time consuming manual weeding labors).