Deke Guo

National University of Defense Technology

Papers

Total Citations

H-Index

About

Deke Guo is a leading researcher in data-driven inertial navigation, mobile computing, and sensor fusion, with a focus on enabling robust positioning for augmented reality, robotics, and autonomous systems. His most influential work introduces VANE-IN, a velocity auto-encoder for inertial navigation that addresses the fundamental challenge of estimating accurate velocities from noisy IMU data. By developing a velocity regression network (VRN) that learns to extract motion patterns from raw inertial measurements, Guo has advanced the state of the art in dead-reckoning and localization without reliance on GPS or external infrastructure. This approach has significant implications for seamless indoor navigation and real-time AR experiences. With over 3 citations on his recent 2024 publication, his contributions are gaining traction in the mobile computing and robotics communities. Guo’s work bridges the gap between classical inertial navigation theory and modern deep learning, offering practical solutions for continuous, drift-reduced positioning in GPS-denied environments. His research continues to shape how devices understand and respond to human motion in complex, dynamic settings.