Pedro Porto Buarque de Gusmão
Papers
6
Total Citations
253
H-Index
4
About
Pedro Porto Buarque de Gusmão is a leading researcher in robust state estimation for autonomous systems, specializing in multi-modal sensor fusion for challenging, visually-denied environments. His work addresses the critical failure of conventional visual odometry in conditions like heavy smoke, darkness, or fog by pioneering the use of alternative sensing modalities. His most impactful contribution, **milliEgo** (130 citations), introduces a single-chip mmWave radar and deep sensor fusion approach for accurate egomotion estimation, offering a low-cost, robust alternative to optical methods. Complementing this, his **DeepTIO** (82 citations) system innovatively combines thermal-inertial odometry with a "visual hallucination" network, enabling reliable navigation where standard cameras fail. Gusmão has further advanced the field by developing **RadarLoc** (26 citations), the first learning-based relocalization method for FMCW radar, and a graph-based thermal-inertial SLAM system using probabilistic neural networks. His work is foundational for deploying autonomous robots and augmented reality in industrial inspection, search-and-rescue, and underground or night-time operations, consistently pushing the boundaries of perception in the most adverse conditions.
Research Focus
Key Achievements
Top Papers
- 1milliEgo130 citations · 2020
- 2DeepTIO: A Deep Thermal-Inertial Odometry With Visual Hallucination82 citations · 2020
- 3RadarLoc: Learning to Relocalize in FMCW Radar26 citations · 2021
- 4
- 5Graph-based Thermal-Inertial SLAM with Probabilistic Neural Networks3 citations · 2021
- 6DeepTIO: A Deep Thermal-Inertial Odometry with Visual Hallucination2 citations · 2019