AI Fleet

AI-FLEET develops an AI-driven fleet management framework for coordinating heterogeneous robotic systems on dynamic construction sites, including mobile robots, autonomous construction machinery, and other robotic agents involved in assembly, disassembly, logistics, and as-built monitoring. The project integrates AI planning and reasoning with site representations to move construction workflows beyond rule-based automation toward adaptive, data-driven coordination. At its core, the system combines a capability-aware task allocation mechanism with a semantic knowledge graph that formally represents robots, tasks, materials, spatial zones, and constraints. This semantic layer is coupled with voxel-based spatio-temporal site models that capture the evolving context and accessibility of the construction environment. Building on this shared representation, the framework supports task allocation, scheduling, navigation, and coordinated manipulation across multiple autonomous robots. Given the current state of the site, the planner searches for a near-optimal solution to the coordination problem, balancing task dependencies, robot capabilities, and resource availability. As conditions change, the system continuously reasons over the updated knowledge graph and replans, allowing robots to adapt without halting the broader workflow. The developed system will be verified and validated at the CARE*Site.

Further information can be found on the homepage: www.exc-care.com

Duration of the Project

January 2026 until December 2027

Researchers

Pranav Shevkar (shevkar@ip.rwth-aachen.de), Chu Han Wu (wu@ip.rwth-aachen.de) and Dr.-Ing. Davide Picchi

Funding

The project is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 3115 – 533767731

Project Partners