Atsutake Kosuge
Papers
8
Total Citations
89
H-Index
6
About
Atsutake Kosuge is a leading researcher at the intersection of robotics, hardware acceleration, and embedded systems, whose work is pivotal in enabling faster, more intelligent industrial robots. His primary contributions lie in developing specialized FPGA-based accelerators that dramatically speed up computationally intensive algorithms for robotic manipulation. Notably, Kosuge has created an accelerator for the Iterative Closest Point (ICP) algorithm, achieving up to 4.8× faster object-pose estimation for picking robots—a breakthrough that directly addresses the low throughput of conventional systems. His innovations extend to motion planning, where he introduced an SoC-FPGA accelerator capable of handling graphs with over 10 million edges for dual-arm robots, and to non-contact data connectors, achieving 6 Gb/s data transfer for robot arms. With his most-cited work garnering 28 citations, Kosuge’s research has been validated in real-world scenarios like the Amazon Picking Challenge. His fusion of RGB and infrared cameras for human pose estimation in low light further showcases his versatility, making him a key figure in advancing autonomous, collaborative robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8