Lorenzo Pasqualetto Cassinis

About

Lorenzo Pasqualetto Cassinis is a leading researcher in spacecraft relative navigation, specializing in vision-based pose estimation for uncooperative targets. His work directly addresses critical challenges in In-Orbit Servicing and Active Debris Removal, where servicer spacecraft must autonomously interact with inactive satellites. His most cited paper (2022, 43 citations) pioneers the use of convolutional neural networks (CNNs) for monocular pose estimation, tackling the domain shift problem between synthetic training data and real space imagery—a fundamental barrier to deploying deep learning in orbit. Building on this, his 2023 work (14 citations) introduces an adaptive Unscented Kalman Filter that leverages neural network uncertainty to robustly fuse CNN predictions with dynamics, validated at Stanford’s robotic Testbed for Rendezvous and Optical Navigation. This integration of uncertainty-aware learning with classical filtering represents a significant advance in reliable, real-time spacecraft navigation. Cassinis’s research bridges the gap between computer vision and astrodynamics, offering practical solutions for autonomous rendezvous with uncooperative objects. His work is essential reading for engineers developing next-generation guidance systems for space sustainability and debris mitigation.

Research Focus

Key Achievements

Top Papers

1. 1
   On-ground validation of a CNN-based monocular pose estimation system for uncooperative spacecraft: Bridging domain shift in rendezvous scenarios

   43 citations · 2022
2. 2
   Leveraging neural network uncertainty in adaptive unscented Kalman Filter for spacecraft pose estimation

   14 citations · 2023

Key Collaborators

Contact & Links