Ugo Rosolia
Papers
13
Total Citations
311
H-Index
7
About
Ugo Rosolia is a robotics and control systems researcher whose work sits at the intersection of safe reinforcement learning, model predictive control (MPC), and autonomous systems. His research addresses some of the most pressing challenges in deploying intelligent robots in uncertain, real-world environments — particularly ensuring safety and efficiency under dynamic constraints. Rosolia is perhaps best known for developing SAVED (Safety Augmented Value Estimation from Demonstrations), a deep model-based reinforcement learning framework that enables safe robotic learning from sparse cost signals, accumulating over 90 citations. His Branch MPC framework for interactive multi-modal motion planning (72 citations) has advanced how autonomous vehicles and robots reason about the unpredictable behavior of surrounding agents. His contributions extend to multi-robot coordination, where his decentralized task and path planning framework (59 citations) offers scalable solutions for collaborative robot teams. Beyond these highlights, Rosolia has tackled quadrupedal locomotion control, safety-critical nonlinear systems via differentially-flat architectures, and probabilistic safety guarantees for motion planning. Collectively, his portfolio demonstrates a rigorous, mathematically grounded approach to making autonomous systems both capable and reliably safe — a contribution of growing importance as robotics moves from controlled labs into complex, human-shared environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2Interactive Multi-Modal Motion Planning With Branch Model Predictive Control72 citations · 2022
- 3Decentralized Task and Path Planning for Multi-Robot Systems59 citations · 2021
- 4
- 5
- 6
- 7
- 8Reactive motion planning with probabilistic safety guarantees7 citations · 2020
- 9
- 10Decentralized Task and Path Planning for Multi-Robot Systems4 citations · 2020