Adam Dai
Papers
3
Total Citations
9
H-Index
2
About
Adam Dai is a researcher at the forefront of safe autonomy, specializing in the intersection of neural network verification, robot motion planning, and simultaneous localization and mapping (SLAM). His work addresses a critical challenge in robotics: ensuring that learning-based systems can operate reliably in safety-critical, real-world environments. Dai’s most influential contribution is his pioneering use of **reachability analysis** to enforce safety constraints during neural network training, a method that has garnered significant attention (5 citations) for its potential to bridge the gap between deep learning and safety-critical applications like human-robot interaction. He further extended this framework to safeguard learning-based planners against motion and sensing uncertainties, demonstrating a robust approach to deploying planners trained in simulation onto physical robots. In a complementary line of work, Dai developed **PlaneSLAM**, a LiDAR-based SLAM system that leverages planar features for efficient 3D mapping, directly enabling downstream motion planning in structured environments. By tackling both the perception and planning pillars of autonomy, Adam Dai’s research provides a cohesive pathway toward trustworthy, real-world robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3