Yuxiang Peng

University of Delaware, Sichuan University

Papers

Total Citations

H-Index

About

Yuxiang Peng is a rising researcher at the forefront of robotics, computer vision, and human-robot interaction, whose work bridges the gap between perception, navigation, and soft materials. His most impactful contribution is **NeRF-VINS**, a real-time visual-inertial navigation system that leverages Neural Radiance Fields (NeRF) as a prior map to overcome the limitations of conventional keyframe-based localization. By addressing challenges like sub-optimal viewpoints and constrained motion, this work (19 citations) offers a robust solution for consistent localization in complex environments—critical for autonomous systems like AR/VR devices and drones. Peng also pioneers in soft robotics with his 2025 paper on **intrinsically soft, fully recyclable robotic sensors** (16 citations), which integrate quadruple sensing functions for reliable human-robot interactions, emphasizing sustainability and safety. His critical analysis in "Is Iteration Worth It?" (2025) further refines sliding-window VIO, questioning computational trade-offs in edge-device deployment. With a focus on real-time efficiency, material innovation, and algorithmic rigor, Peng’s work is shaping next-generation autonomous systems that are both intelligent and environmentally conscious.