Periodic Reporting for period 2 - I-Wood (Forest Intelligence: robotic networks inspired by the Wood Wide Web)
Período documentado: 2022-11-01 hasta 2024-04-30
Plants are connected by an underground fungi network that provides them with nutrients, helps share resources, and extends their perception abilities. This mycorrhizal network, known as the Wood Wide Web, is crucial in maintaining healthy natural ecosystems and limiting global warming. Thus, it must be preserved to mitigate the speeding up of the carbon cycle and its effects on climate change. Robotics and Artificial Intelligence can offer concrete solutions for a deeper analysis of natural processes at the basis of this global change and for developing sustainable technologies. Based on that, I-Wood proposes a new virtual and physical robotic network paradigm inspired by the belowground fungus-mediated inter-plant communication and the associated collective behaviours. Specifically, I-Wood will study, extract and formalize the rules of plant-fungus interaction mechanisms to develop a plant-inspired decision-making model and a new generation of root-like robots to explore the soil with plant-inspired behaviours. By imitating plants, these distributed intelligent systems will co-develop morphology and behaviour in a dynamic environment. The impact and feasibility of the proposed approach will be tested in a mixed social network, scale-down in a confined environment, where robots will interact with natural plants to facilitate the development of mycorrhizal networks. Grounded on a solid multi-disciplinary approach, I-Wood will pave the way for new paradigms in robotics and embodied AI based on solutions that overcome the current animal-based or brain-based model, novel approaches for the use of robotics in biology, and for new scientific knowledge on plant community with a major significance for biodiversity and climate protection.
Major results achieved so far include the production of experimental protocols and methodologies, hardware and software, to investigate, observe, and analyze the plant-fungus symbiotic relationship and the mycorrhizal network formation in complex three-dimensional (3D) experimental settings. A computational model has been proposed for analyzing plant behaviors and the mycorrhizal role in the underground ecosystem. It will be adopted to control artificial agents and exploited for the navigation of 3D environments. A 3D Discrete Element Method-based model has been developed to simulate soil intrusion by mimicking plant root morphology and growth strategies. This model is instrumental in investigating conditions where imitating plant root features brings advantages with respect to classic approaches to soil penetration. The future development of artificial robotic root-like intruders will be guided by analysis performed with this tool to implement efficient movements in soil. Finally, new artificial growth and sensing strategies are under investigation for embodiment in artificial autonomous systems acting in soil.
Future developments will involve the creation of robotic agents capable of penetrating soil through apical growth, embodying root-like features that facilitate movement in such a complex environment. These agents will navigate using a bioinspired control architecture, incorporating interactions similar to root-fungus and root-environment relationships. Incorporating plant-like behaviors will require energy-efficient actuation, promoting non-destructive intrusion for the safe coexistence of robots and the environment. These results represent significant milestones in realizing innovative approaches to soil exploration, enabling the monitoring of soil conditions, and aiding in developing management strategies for compromised or endangered ecosystems.