Dandan Peng

Northwestern Polytechnical University, Huazhong University of Science and Technology

Papers

Total Citations

H-Index

About

Dandan Peng is a researcher advancing intelligent fault diagnosis and robotic vision for heavy industrial machinery. Their work centers on two key areas: multi-joint robot health monitoring and high-precision visual positioning for shield tunneling machines. Peng’s most impactful contribution is the development of SMNet, a novel compositional generalization model for compound fault diagnosis in industrial robots. This work addresses a critical gap in the Industrial Internet of Things, where simultaneous degradation of multiple joints poses severe reliability challenges—a problem conventional single-fault methods cannot solve. With 15 citations since 2026, SMNet is Peng’s most recognized paper. In parallel, Peng has tackled the harsh underground environment of shield machines, proposing a UNet-based high-precision segmentation method for disc cutter holder positioning, and an improved monocular-vision pose measurement technique. These methods overcome low illumination, dust, and sand deposition to enable robotic disc cutter replacement. Though early in their career, Peng’s work demonstrates a clear trajectory toward solving real-world industrial challenges at the intersection of deep learning and robotics.