Yinbo Liu

Hisense (China), Tianjin University, Qingdao University

Papers

Total Citations

H-Index

About

Yinbo Liu is a researcher whose work spans robotics, computer vision, and autonomous navigation, with a particular focus on enabling intelligent agents to perceive and move through indoor environments. His research integrates deep learning with classical control, addressing challenges from high-level scene understanding to low-level motion control. A key contribution is his work on **indoor layout estimation (ILE)** , where he developed a method using HRNet to segment monocular RGB images into structural components like floors, walls, and ceilings—a task critical for scene understanding, reconstruction, and robot localization. Liu has also advanced **active exploration** by applying generative adversarial imitation learning to camera view planning, allowing robots to autonomously learn optimal viewing strategies for mapping unknown spaces (16 citations). Earlier in his career, he designed a laser navigation control system using EPOS motion controllers, demonstrating improved precision for mobile robot wall-following tasks. While his citation counts reflect a growing body of work, his integration of imitation learning with exploration and his systematic approach to indoor perception mark him as a researcher building foundational tools for autonomous systems in complex, real-world environments.