Papers
11
Total Citations
523
H-Index
8
About
Thomas Lew is a robotics and autonomy researcher whose work sits at the intersection of trajectory optimization, safe learning, and autonomous systems. He is best known for his highly influential tutorial on convex optimization for trajectory generation (2022, 233 citations), which has become a foundational reference for engineers and researchers designing reliable motion planning algorithms for autonomous vehicles, spacecraft, and robots. His contributions to the DARPA Subterranean Challenge through the NeBula framework—developed with TEAM CoSTAR at JPL—demonstrate his ability to translate theoretical advances into award-winning real-world systems, with the team's Phase II victory standing as a landmark achievement in field robotics. Beyond trajectory optimization, Lew has made significant contributions to safe learning under uncertainty, proposing theoretically grounded frameworks for exploration-exploitation in unknown environments and risk-averse trajectory optimization under epistemic and aleatoric uncertainty. His work on learning-based warm-starting for sequential convex programming bridges machine learning and classical optimization, accelerating solve times for real-time applications. More recently, he has addressed challenges in out-of-distribution robustness and whole-body robot manipulation. Collectively, his research portfolio—spanning over 500 cumulative citations—reflects a rare combination of mathematical rigor and practical impact in modern autonomous robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6Risk-Averse Trajectory Optimization via Sample Average Approximation19 citations · 2023
- 7Contact Inertial Odometry: Collisions are your Friends12 citations · 2022
- 8
- 9A System-Level View on Out-of-Distribution Data in Robotics7 citations · 2022
- 10