Consortium: University of Lincoln, TU Delft, University of Bonn, ETH Zurich.
Duration: 2024 – 2025
Multi-robot mapping is increasingly used in applications like site inspection, search and rescue, and orchard monitoring. The advantage of multiple robots working together lies in their ability to communicate, share data, and create a unified view of the environment. Map integration is typically handled through map merging (aligning occupancy grids) or multi-robot SLAM (connecting pose graphs). Despite advancements in robot localization, integrating maps efficiently remains a challenge, particularly for heterogeneous fleets with different movement and sensing capabilities.
The GAIA project addresses this challenge by using semantic information to enhance scene understanding. By recognizing objects in the environment, GAIA integrates multi-perspective robotic maps more effectively. The project focuses on agriculture, where robots can improve precision farming by assisting in monitoring and harvesting.
Ground robots provide detailed close-up inspections, while UAVs cover large areas quickly and update incomplete maps on demand. UAVs also enhance UGV path planning by mapping obstacles like workers, tractors, and trolleys, ensuring ground robots navigate efficiently. Through real-time multi-robot collaboration, GAIA aims to advance sustainable and autonomous farming solutions.
Contact: Riccardo Polvara