Stephen Boyd
Papers
6
Total Citations
2,684
H-Index
6
About
Stephen Boyd is a leading figure in convex optimization and its transformative applications across engineering, computer vision, and robotics. His foundational work on second-order cone programming (SOCP) has been cited over 2,400 times, establishing a cornerstone for modern optimization theory and practice. Boyd’s research bridges rigorous mathematical frameworks with real-world impact, particularly in autonomous systems and perception. He pioneered the use of ellipsoid fitting for efficient collision detection and distance computation in robotics, enabling faster and more reliable obstacle avoidance. More recently, Boyd has tackled the critical challenge of making computer vision robust to real-world sensor imperfections. His influential "Dirty Pixels" series addresses the gap between idealized algorithms and noisy, blurred sensor data, proposing end-to-end architectures that integrate image processing with high-level perception—essential for applications like autonomous driving. He has also contributed to trajectory generation using sum-of-norms regularization, improving smoothness and control in tracking systems. With over 2,600 total citations, Boyd’s work continues to shape how optimization and perception systems are designed for reliability in the physical world.
Research Focus
Key Achievements
Top Papers
- 1Applications of second-order cone programming2,415 citations · 1998
- 2Obstacle Collision Detection Using Best Ellipsoid Fit103 citations · 1997
- 3
- 4Dirty Pixels: Towards End-to-end Image Processing and Perception59 citations · 2021
- 5Trajectory generation using sum-of-norms regularization22 citations · 2010
- 6Efficient distance computation using best ellipsoid fit17 citations · 2003