ALTER: Autonomous Land Transformation for Ecological Restoration

As climate change accelerates land degradation, vast areas are turning into semi-arid landscapes where rainwater rapidly runs off instead of nourishing the soil. This process drives erosion, lowers groundwater levels, and promotes desertification. A proven method to restore such environments is the construction of small water-retention bunds that capture rainfall and support vegetation recovery. However, deploying these structures across large and remote regions remains slow, costly, and labor-intensive.
Within the Autonomous Land Transformation for Ecological Restoration (ALTER) project, we investigate how advanced perception systems can enable future autonomous excavation for land restoration. By integrating state-of-the-art 3D sensing and computer vision technologies, ALTER focuses on giving excavation machines the ability to accurately perceive, interpret, and model complex, unstructured natural terrain. These perception capabilities will form the foundation for intelligent robotic systems in subsequent projects.
A key challenge lies in achieving reliable sensing in harsh outdoor conditions, including dust, strong sunlight, and highly variable soil textures, while operating under strict constraints on energy consumption and system size. Rather than focusing on optimizing individual technologies, in ALTER we take a holistic approach to optimize the complete perception systems for realistic excavation scenarios.
The project will deliver a systematic comparison of sensing technologies under challenging natural environments, a robust proof-of-concept perception system, and practical design guidelines for future autonomous excavation machines. By addressing the critical perception bottleneck, ALTER aims to accelerate the development of scalable technologies for sustainable landscape rehabilitation.