Shijia Peng

Robotics Research (United States), Shenzhen Academy of Robotics, Shenzhen University

Papers

Total Citations

H-Index

About

Shijia Peng is a leading robotics researcher whose work centers on dual-arm coordination, generative digital twins, and 3D semantic understanding for autonomous manipulation. Peng’s most influential contribution is the **RoboTwin** benchmark, which tackles the critical shortage of diverse, high-quality demonstration data for dual-arm robots. By introducing generative digital twins that align with real-world conditions, this work has rapidly accumulated over 30 citations since its 2024–2025 release, establishing a new standard for evaluating complex object manipulation. In parallel, Peng developed **G3Flow**, a novel framework that fuses geometric precision with semantic understanding for diffusion-based imitation learning, enabling pose-aware and generalizable manipulation. This work bridges the gap between low-level control and high-level reasoning, a key step toward human-level dexterity. Peng’s research is distinguished by its practical impact: by creating scalable, simulation-to-real pipelines, it directly addresses the data scarcity bottleneck that has long hindered advanced robotics. For students and researchers, Peng’s work offers a blueprint for building robust, generalizable robotic systems—combining cutting-edge generative AI with rigorous benchmarking to push the boundaries of what autonomous robots can achieve.