Lorenzo Bellone
Papers
1
Total Citations
24
H-Index
1
About
Lorenzo Bellone is pioneering the frontier of autonomous nano-drones, focusing on making tiny flying robots both intelligent and efficient. His key research areas span embedded deep learning, ultra-low-power computer vision, and multi-task inference for resource-constrained robotic platforms. Bellone’s major contribution is demonstrating that complex neural networks can be aggressively compressed—through quantization, pruning, and architecture search—to run in real-time on milliwatt-scale processors like the PULP platform. His most cited work, “Tiny-PULP-Dronets” (2022, 24 citations), shows how to squeeze multi-task neural networks into pocket-sized drones weighing tens of grams, enabling simultaneous object detection, tracking, and collision avoidance without cloud connectivity. This breakthrough directly addresses the challenge of deploying AI on nano-drones for safe human-robot interaction and visual inspection in confined spaces. Bellone’s research is notable for bridging the gap between theoretical model compression and practical, real-world deployment, achieving faster and lighter inference without sacrificing accuracy. His work is shaping the next generation of autonomous micro-robots, where intelligence must fit within milliwatts and cubic centimeters.
Research Focus
Key Achievements
Top Papers
- 1