Carlo Rizzardo
Papers
2
Total Citations
11
H-Index
2
About
Carlo Rizzardo is a robotics researcher whose work centers on bridging the critical gap between simulation and real-world deployment, particularly in visual non-prehensile manipulation and precision agriculture. His most cited paper, "Sim-to-real via latent prediction: Transferring visual non-prehensile manipulation policies" (2023, 7 citations), introduces a novel approach that leverages latent prediction to enable reinforcement learning policies trained in simulation to transfer effectively to physical robots. This work addresses a fundamental challenge in robotics—the sim-to-real gap—by allowing end-to-end visual policies to operate without custom perception systems. In his earlier study, "The Importance and the Limitations of Sim2Real for Robotic Manipulation in Precision Agriculture" (2020, 4 citations), Rizzardo critically examines where simulation accuracy remains essential despite advances in domain randomization and model-based learning. This research is particularly significant for agricultural robotics, where environmental variability demands robust transfer methods. Rizzardo’s contributions are shaping how roboticists think about deploying learned manipulation policies in the real world, offering practical insights for both industrial and agricultural applications. His work continues to influence the development of more reliable, simulation-to-reality transfer techniques in robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2