Jianming Zhang

Adobe Systems (United States), China Academy of Engineering Physics

Papers

Total Citations

H-Index

About

Jianming Zhang is a robotics researcher whose work bridges the gap between intelligent navigation and bio-inspired locomotion. His primary research areas include visual navigation, reinforcement learning for robotics, and quadrupedal locomotion, with a particular focus on how robots can perceive and interact with complex, unstructured environments. Zhang's most notable contribution is the development of **GAPLE (Generalizable Approaching Policy LEarning)**, a framework that enables robots to actively search for and approach objects in indoor environments using only visual input, achieving 26 citations. This work addresses a critical limitation in prior visual navigation methods by learning a generalizable action policy that adapts to new scenes without retraining. Additionally, Zhang has investigated the mechanics of quadruped running, specifically analyzing how spine motion influences contact time during locomotion—a study that, while less cited (6 citations), provides foundational insights for designing more agile and efficient legged robots. His research is characterized by a systems-level approach, combining perception, control, and biomechanics to create robots that move and explore with greater autonomy and adaptability. Zhang's work is particularly relevant for students and researchers interested in embodied AI, active perception, and the intersection of computer vision and robotics.