Phil Knag
Papers
1
Total Citations
8
H-Index
1
About
Phil Knag is a leading researcher in energy-efficient hardware acceleration for 3D computer vision and edge computing. His work focuses on designing specialized CMOS architectures that enable real-time scene reconstruction for robotics and augmented reality, where power and latency constraints are critical. Knag’s most notable contribution is a ray-casting accelerator fabricated in 10nm CMOS, which simultaneously casts multiple spatially proximate rays to exploit voxel data-locality, dramatically reducing memory traffic. The design features a near-memory search for voxel address overlaps and an opportunistic approximate trilinear interpolation technique that saves energy with minimal accuracy loss. This work, published in 2020 and garnering 8 citations, demonstrates ray-casting of 3D scenes at high efficiency, laying the groundwork for future edge-AI vision systems. Knag’s research bridges the gap between algorithmic demands and silicon-level constraints, making him a key figure in the push toward autonomous, low-power perception in robotics and AR. His contributions are essential reading for students and engineers developing next-generation hardware for real-time 3D understanding.
Research Focus
Key Achievements
Top Papers
- 1