Papers
6
Total Citations
36
H-Index
4
About
Theja Tulabandhula is a researcher at the forefront of uncertainty-aware machine learning and edge autonomy, with a particular focus on making autonomous robotic systems more reliable, efficient, and trustworthy in real-world deployments. His work addresses a critical gap in modern AI-driven robotics: while deep learning models have achieved impressive accuracy, they often fail to quantify their own uncertainty — a dangerous shortcoming for safety-critical systems like surgical robots, insect-scale drones, and autonomous vehicles. Tulabandhula's most cited contributions span conformal prediction for visual odometry, multimodal 3D object detection, and sensor anomaly recognition. His STARNet framework, garnering 7–9 citations across iterations, tackles sensor trustworthiness under complex failure modes in LiDAR, RADAR, and event cameras. His work on mutual information-calibrated conformal feature fusion advances uncertainty quantification in 3D perception at the edge. He has also pioneered frugal sensing approaches, proposing that LiDAR systems can generate rather than sense predictable environmental data, dramatically reducing computational overhead. More recently, he has extended uncertainty-aware principles into deep reinforcement learning for mission-critical robotics. Collectively, his research is shaping next-generation autonomous systems that are not only accurate but genuinely dependable.
Research Focus
Key Achievements
Top Papers
- 1Lightweight, Uncertainty-Aware Conformalized Visual Odometry13 citations · 2023
- 2
- 3
- 4
- 5
- 6