Shenlong Wang

Papers

Total Citations

H-Index

About

Shenlong Wang is a researcher working at the intersection of robotics control systems and 3D computer vision, with particular focus on autonomous systems and model predictive control. His most influential work centers on the DiffTune framework, a gradient-based automatic tuning methodology that leverages autodifferentiation to optimize controller parameters without requiring manual hand-tuning — a notoriously difficult problem given the nonlinear nature of robotic dynamics. This foundational contribution was extended through DiffTune-MPC, which applies closed-loop learning specifically to Model Predictive Control pipelines, enabling more adaptive and performant autonomous systems that can respect real-world constraints while continuously improving. Together, these papers have garnered nearly 30 citations since 2024, signaling rapid community uptake in the robotics and controls community. Beyond control systems, Wang has also contributed to 3D perception research, notably developing methods for building rearticulable models of arbitrary everyday objects from 4D point cloud data — work that advances how machines understand and reason about articulated objects in dynamic environments. His research portfolio reflects a strong commitment to bridging principled mathematical frameworks with practical robotics applications, making him a researcher of growing significance for those working on intelligent autonomous systems.