Justin Romberg
Papers
5
Total Citations
29
H-Index
3
About
Justin Romberg is a leading researcher at the intersection of compressed sensing, low-power embedded vision, and multi-robot coordination. His work is defined by a drive to extract maximal functionality from minimal resources—whether that means energy, data, or computational power. In the domain of smart cameras, Romberg pioneered a paradigm shift: instead of first reconstructing a full image and then analyzing it, his systems perform gesture detection directly on compressed measurements. This approach, demonstrated in his highly-cited work on light-powered, "always-on" smart cameras, slashes energy consumption by orders of magnitude, enabling perpetual, battery-free computer vision. His contributions have been recognized for their potential to unlock new classes of IoT and edge-AI devices. Beyond vision, Romberg has made significant strides in multi-robot systems, developing globally optimal assignment algorithms for air-ground teams and reinforcement learning frameworks for sequencing complex multi-robot behaviors. His research elegantly bridges theoretical foundations with practical, energy-aware system design, making him a pivotal figure in creating efficient, autonomous, and collaborative robotic systems.
Research Focus
Key Achievements
Top Papers
- 1A Light-Powered Smart Camera With Compressed Domain Gesture Detection11 citations · 2017
- 2
- 3
- 4A Reinforcement Learning Framework for Sequencing Multi-Robot Behaviors3 citations · 2019
- 5Sequencing of multi-robot behaviors using reinforcement learning2 citations · 2021