I am a Project Assistant in the Vehicle Engineering and Autonomous Systems division of the Department of Mechanical Engineering at Chalmers University of Technology in Gothenburg.
In October 2026 I completed my PhD in Automotive Engineering for Intelligent Mobility at the University of Bologna, supervised by Prof. Stefano Mattoccia at the Computer Vision Lab. My research is about vision-based perception for autonomous systems: end-to-end driving, depth estimation with event cameras, and metric visual odometry in natural environments, where GNSS is unreliable. During the PhD I spent six months as a visiting researcher at Chalmers, in the same group I have now joined.
From 2020 to 2026 I was part of the UniBo Motorsport Driverless 1:10 team, which I led from 2024. We won the RoboRacer World Championship at ICRA 2025 and took 2nd and 3rd place at ICRA 2026.
Before the PhD I earned a BSc in Physics and an MSc in Artificial Intelligence at the University of Bologna, spending the second year of the master's at NTNU in Trondheim.
Under submission
Recovers metric trajectories from monocular headcam video using the camera height above the ground as a cheap metric anchor, for forests where GNSS degrades under the canopy.
IET Image Processing, 2026
MADNet-ε: a lightweight recurrent stereo network that predicts disparity from 10 ms event windows at about 145 FPS, plus ε-CARLA, a synthetic event-stereo dataset.
O-Conference 2026, Gothenburg (extended abstract)
Preliminary version of the forest trajectory work: estimating an orienteer's route from an RGB headcam video, the starting position and heading.
ICCV 2025, Honolulu
Distills dense proxy depth labels from a vision foundation model on RGB frames to train monocular depth networks on event cameras.
ICRA 2025, Atlanta
FlowFuser: adds optical flow as an auxiliary task to the perception module of an end-to-end driving model, improving collision avoidance in dense traffic in CARLA.
Proc. SPIE 12188, Advances in Optical and Mechanical Technologies for Telescopes and Instrumentation V, 2022
University of Bologna, Automotive Engineering for Intelligent Mobility, 2023–2026. Supervisor: Prof. Stefano Mattoccia.
University of Bologna, Artificial Intelligence, 2023 (110/110 cum laude). Supervisor: Prof. Andrea Asperti. A deep Q-network that plans sailing routes from satellite wind data, compared against the GPS tracks of the Vendée Globe 2020/21 skippers.
Study of the spectroscopic mode of the SuperFOSC instrument for the Cassini telescope
University of Bologna, Physics, 2021 (110/110). Supervisor: Prof. Matteo Lombini. Compares spectroscopic configurations for the instrument used for space debris surveillance at the Loiano observatory.
I joined the UniBo Motorsport Driverless 1:10 team in 2020 and led it from 2024 to 2026, coordinating about twenty bachelor's and master's students. We build and program 1:10 scale autonomous cars for the RoboRacer competitions held at ICRA: open practice to map the track, time trials, and a head-to-head tournament where detecting and overtaking the opponent is essential.
For the 2026 season I replaced our pure-pursuit and Frenet overtaking stack with a single nonlinear Model Predictive Controller (CasADi + IPOPT): it follows the raceline and avoids walls and opponents in one optimisation, so overtaking emerges from the same cost function.
I also ported our algorithms to Rust in AuRoRus, an open-source autonomous racing stack for RoboRacer / F1TENTH cars with simulator, SLAM, planning and a web GUI.
Autonomous RoboRacer Rust: an autonomous racing stack in Rust for RoboRacer (F1TENTH) cars, with simulator, SLAM, race-line planning, opponent detection, Frenet overtaking, MPC and a web GUI. Code on GitHub.
Map utilities in Rust for orienteering: renders OpenOrienteering Mapper maps to images from the command line, with tools to benchmark the output against a reference renderer.
Code of my master's thesis: deep reinforcement learning for sailing boat route planning on real wind data.