XSCAVE is a Horizon Europe-funded project that aims to revolutionize the autonomy of heavy mobile machines—such as excavators, forestry forwarders, and urban logistics vehicles—through AI-driven terrain-adaptive control systems. By focusing on real-world tasks in Earth-Moving, Forestry, and Urban Logistics sectors, XSCAVE seeks to improve operational efficiency, safety, and adaptability in challenging environments, ultimately addressing labor shortages and environmental concerns.
The project is motivated by the pressing economic, societal, and environmental challenges in sectors dependent on heavy machinery. Increased autonomy in such machines can address labor shortages and reduce environmental impact, such as minimizing soil damage and fuel consumption. The complex interactions between terrain and machine, which vary across tasks and environmental conditions, require a new level of intelligent adaptation that current systems lack.
XSCAVE represents a critical step toward safe, sustainable, and efficient automation in some of Europe’s most vital industries.
XSCAVE will develop high-fidelity simulators for terrain-machine interaction and distill these into differentiable neural models. These simulators will support perception-driven planning and control, incorporating real-world sensor inputs and enabling scalable training and testing environments.
The project will create robust AI planning and control software, embedding safety features such as collision avoidance and vehicle stability. This includes structured priors from control theory and optimization, allowing for high-performance, adaptable control across diverse terrain and tasks.
To bridge the gap between simulation and real-world deployment, XSCAVE will use uncertainty estimation and online re-training techniques. These tools will ensure that AI models remain reliable under real-world dynamics and enable rapid adaptation to new environments and tasks.
XSCAVE focuses on trust, explainability, and human-AI collaboration. Using explainable AI and reward learning based on human input, the project will empower users to understand and customize machine behavior. Furthermore, XSCAVE will support industrial adoption through demonstration campaigns, stakeholder engagement, and exploitation planning.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the HADEA. Neither the European Union nor the granting authority can be held responsible for them.