Hibiki Kawai
Papers
2
Total Citations
7
H-Index
2
About
Hibiki Kawai is a robotics researcher focused on advancing visual perception and state estimation for mobile robots, particularly those with articulated or tilting upper bodies. His primary research areas include deep learning–based camera attitude estimation, sensor fusion, and robust pose inference for terrestrial robots operating in challenging environments. Kawai’s major contribution lies in rethinking camera pose estimation as a classification problem rather than conventional numerical regression—an approach inspired by optical character recognition (OCR) that improves robustness in dynamic conditions. His 2022 paper on this method has garnered 4 citations, while his subsequent 2023 work, which fuses aggregated information from both camera and depth images to enhance attitude estimation, has received 3 citations. These studies address a critical limitation of traditional IMU- or gyro-based systems, which are susceptible to ground-induced noise. By leveraging neural networks to integrate visual and depth data, Kawai’s research enables more reliable and agile robot control, particularly for platforms that require free upper-body tilting. His work represents a meaningful step toward making autonomous robots more resilient in real-world, unstructured settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2