☰

Projects

Model Predictive Control

Model predictive control (MPC) is a versatile framework for feedback control. Recently, we have focused on combining MPC with Bayesian regression techniques to improve adaptability and safety, tailoring solvers to specific control problem to exploit structure and reduce the computational complexity, integrating planning and control modules to simplify the control pipeline, and designing predictive safety filters.

Multi-Agent Control

Coordinating multiple agents involves challenges such as scalability, communication limits, and uncertainty. We have developed distributed MPC and estimation approaches for cooperative exploration, coverage, and fleet rebalancing.

Hardware Demonstrators

Algorithmic developments are validated on experimental platforms, ranging from stratospheric balloon flights to miniature robotic vehicles. These demonstrators provide realistic conditions for testing control and estimation algorithms, while also serving as benchmarks for the robotics and control community.