Selected work

Representative Papers

Learning-augmented robotic automation for real-world manufacturing

Apr 24, 2026

This work addresses the limitations of traditional industrial robots in dynamic environments and the reliability and safety challenges of purely learning-based control in real-world production lines. The authors propose a hybrid architecture that integrates a learned task controller with a neural 3D safety monitoring module, seamlessly embedded into existing industrial workflows. For the first time, this approach enables fully automated deformable cable insertion and welding on a physical motor assembly line without protective fencing. Requiring only minimal real-world data, the system operated continuously for 5 hours and 10 minutes, successfully assembling 108 motors with a 99.4% quality inspection pass rate. Cycle times approached manual operation levels, while variability in weld quality and cycle duration was significantly reduced, demonstrating the method’s effectiveness in ensuring safety, consistency, and human-robot collaboration.

0 citationsRead paper

Learning Fast, Tool aware Collision Avoidance for Collaborative Robots

Aug 28, 2025

Collaborative robots face safety-critical obstacle avoidance challenges in dynamic environments due to tool changes and partial observability. Method: This paper proposes a tool-aware, real-time adaptive collision avoidance framework integrating a learning-based point cloud perception model—featuring tool-aware filtering and occlusion-aware collision prediction—with a constrained reinforcement learning controller, enabling millisecond-level response and online decision-making under dynamic task switching. Contribution/Results: It is the first approach to explicitly embed tool geometry and interaction patterns into the perception–control closed loop, achieving sub-millimeter-accurate safe navigation under partial observability. Experiments demonstrate a 60% reduction in computational overhead compared to Artificial Potential Field (APF) and Model Predictive Path Integral (MPPI) methods. The framework has been successfully deployed on a real collaborative robot system, demonstrating strong safety guarantees, motion smoothness, and modular extensibility.

0 citationsRead paper

Recent publications

Latest Papers

Learning-augmented robotic automation for real-world manufacturing

Apr 24, 2026

This work addresses the limitations of traditional industrial robots in dynamic environments and the reliability and safety challenges of purely learning-based control in real-world production lines. The authors propose a hybrid architecture that integrates a learned task controller with a neural 3D safety monitoring module, seamlessly embedded into existing industrial workflows. For the first time, this approach enables fully automated deformable cable insertion and welding on a physical motor assembly line without protective fencing. Requiring only minimal real-world data, the system operated continuously for 5 hours and 10 minutes, successfully assembling 108 motors with a 99.4% quality inspection pass rate. Cycle times approached manual operation levels, while variability in weld quality and cycle duration was significantly reduced, demonstrating the method’s effectiveness in ensuring safety, consistency, and human-robot collaboration.

0 citationsRead paper

Learning Fast, Tool aware Collision Avoidance for Collaborative Robots

Aug 28, 2025

Collaborative robots face safety-critical obstacle avoidance challenges in dynamic environments due to tool changes and partial observability. Method: This paper proposes a tool-aware, real-time adaptive collision avoidance framework integrating a learning-based point cloud perception model—featuring tool-aware filtering and occlusion-aware collision prediction—with a constrained reinforcement learning controller, enabling millisecond-level response and online decision-making under dynamic task switching. Contribution/Results: It is the first approach to explicitly embed tool geometry and interaction patterns into the perception–control closed loop, achieving sub-millimeter-accurate safe navigation under partial observability. Experiments demonstrate a 60% reduction in computational overhead compared to Artificial Potential Field (APF) and Model Predictive Path Integral (MPPI) methods. The framework has been successfully deployed on a real collaborative robot system, demonstrating strong safety guarantees, motion smoothness, and modular extensibility.

0 citationsRead paper