Papers
5
Total Citations
12
H-Index
2
About
Koen Vos is pioneering the integration of semantic world models with multi-agent robotic control, fundamentally advancing how heterogeneous robots navigate and coordinate in complex environments. His research sits at the intersection of supervisory control theory, model predictive control, and semantic mapping, with a core focus on enabling scalable, intelligent multi-agent systems. Vos's most impactful work introduces a shared semantic map architecture that dynamically configures Model Predictive Controllers (MPC) for multiple robots, allowing them to solve navigation problems while sharing environmental resources—a contribution that has already garnered 5 citations since its 2024 publication. He further extended this paradigm through his work on semantic path planning from building digital twin data (2025, 3 citations), demonstrating practical pathways for real-world deployment. Notably, Vos developed RoboSC, a domain-specific language that streamlines the synthesis of supervisory controllers for ROS applications, bridging formal methods with practical robotics development. His recent contributions in 2025 include a hybrid decision-making framework for scalable multi-agent navigation and a novel approach to synthetic dataset generation using 3D Gaussian splatting, addressing critical bottlenecks in vision-based robotics training. Through this body of work, Vos is establishing himself as a leading voice in the next generation of autonomous robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5