Simone Rotondi
Robotics researcher at Inria and co-founder of Localyte. Previously GNSS R&D at Thales Alenia Space.
Research: in the Rainbow team at Inria Rennes, I study how robots can control the shape of deformable objects when their models are imperfect.
Localyte: I build software that reduces IMU drift, so drones, robots, vehicles and satellites keep navigating when GNSS or other external references fail.
Research
Shape control under model uncertainty
Deformable objects change shape as a robot manipulates them. My PhD investigates how visual feedback can guide this process while accounting for uncertainty in the model of the object.
The aim is to make shape control more robust when the real object behaves differently from its model.
Closed-loop shape control
Supervised by Paolo Robuffo Giordano, Alexandre Krupa, and Tommaso Belvedere. PhD started in November 2025.
Publications
-
S. Rotondi, V. Becchetti, S. Carletta, A. Giuseppi. Safe Reinforcement Learning for Autonomous Rendezvous and Proximity Operations: Balancing Safety and Performance in CubeSat Docking Missions. IAF Astrodynamics Symposium, 76th International Astronautical Congress (IAC 2025), Sydney, Australia, 2025. PDF
Soft Actor-Critic extended with a safety critic, so a CubeSat learns to dock autonomously while respecting operational safety constraints.
All publications on Semantic Scholar →
Research profile at Inria →
Localyte
IMUs drift. When GNSS, vision or odometry drop out, that drift is what breaks navigation on drones, robots, vehicles and small satellites. Localyte is a software layer between the IMU and the existing navigation stack: it corrects drift, bias and non-linear errors in real time, with physics-aware AI and no hardware redesign.
Applications: UAV navigation in GNSS-challenged missions, gyro drift on small satellites, and motion tracking for industrial testing (NDT).
Background
I work at the intersection of control theory, robotics and navigation, with a focus on making physical systems reliable despite imperfect models and measurements. Based in Rennes, France.
-
PhD researcher · Inria Rennes, Rainbow team
– present
Visual shape control of deformable objects under model uncertainty.
-
Co-founder · Localyte
– present
AI-enhanced inertial sensing for navigation when GNSS and other external references fail.
-
System Engineer · Thales Alenia Space, Advanced Projects
– · Rome
Precise localization on Earth and Moon: GNSS R&D including the Galileo High Accuracy Service (HAS) and Moonlight, plus GNSS radio-occultation work for atmospheric estimation and weather forecasting.
-
Control Engineer · CRAT, with Thales Alenia Space
– · Rome
High Accuracy Service for the Galileo satellite system, with the Thales Alenia Space R&D team.
Selected work
-
Nonlinear optimal control (iLQR / Gauss-Newton shooting)
Iterative Gauss-Newton shooting methods for underactuated systems, from cart-pole swing-up to a two-wheeled inverted pendulum. MATLAB solvers with Julia visualizations.
-
Colored Ball Balancing
A physical PID demonstrator that makes feedback control visible, used with students and at university Open Days.
Contact
Research and academic collaboration:
simone.rotondi@inria.fr
Localyte, industry and partnerships:
simone.rotondi@localyte.tech
INRIA: Research Profile
Publications: Semantic Scholar
GitHub: @spbisc97
LinkedIn: simone-rotondi