Manuel Belke

Papers

Total Citations

H-Index

About

Manuel Belke is a robotics researcher focused on advancing autonomous manipulation and perception in industrial environments. His work centers on three key areas: deep reinforcement learning for robotic grasping, 6D object pose estimation, and autonomous 3D reconstruction. Belke’s major contributions include developing a Q-learning-based optimization framework for industrial robot grasping, which addresses the challenge of adapting to unstructured and dynamic environments where traditional planning algorithms fall short. He has also pioneered a synthetic data generation pipeline for object pose estimation, enabling robust, machine-learning-based 6D pose estimation with faster processing times—a critical capability for production automation. In his most recent work, Belke applies deep reinforcement learning to Next Best View planning, allowing a robot-guided sensor to autonomously select optimal viewpoints for 3D reconstruction. While his citation counts are still growing (3, 2, and 1 citations respectively), these early publications demonstrate significant potential for impact in industrial robotics. His research bridges the gap between simulation and real-world application, offering practical solutions for automating complex, perception-driven tasks in manufacturing and beyond.