Qiliang Du

South China University of Technology, Ministry of Natural Resources, Guangdong Institute of Intelligent Manufacturing

Papers

Total Citations

H-Index

About

Qiliang Du is a researcher whose work bridges robotics, computer vision, and 3D representation learning. His early contributions focus on agricultural automation, most notably the design of an automatic weeding robot guided by visual navigation—a system that addresses the challenge of minimizing turning radius in paddy fields through a compact mechanical structure and vision-based control (10 citations). He has also explored miniature robotics, developing a fuzzy logic-based path-following controller for a piezo-driven robot intended for desktop micro-factory applications (2 citations). More recently, Du has advanced into 3D deep learning with MD-Mamba, a feature extractor that leverages multi-view depth for robust 3D representation (4 citations), and is investigating graph neural networks that incorporate historical node state increments to enhance point cloud feature learning (1 citation). This trajectory from practical agricultural robots to cutting-edge 3D perception demonstrates a versatile engineering mindset. Du’s work is particularly notable for its progression from hardware-constrained control systems to data-driven representation learning, reflecting a growing emphasis on AI in robotics.