Asim Munawar
Papers
7
Total Citations
400
H-Index
6
About
Asim Munawar is a robotics and machine learning researcher whose work sits at the intersection of deep reinforcement learning, robot manipulation, and intelligent automation. He is best known for his highly cited 2017 work on deep reinforcement learning for high precision assembly tasks, which demonstrated that robots could autonomously learn tight-clearance peg-in-hole insertion without the tedious manual parameter tuning traditionally required in manufacturing — a contribution that has garnered over 300 citations and significantly influenced the field of learned robot manipulation. Munawar has also made notable contributions to unsupervised anomaly detection for industrial robots using spatio-temporal deep learning, enabling surveillance and fault detection through monocular cameras without labeled training data. His research extends into recurrent neural networks for task planning and robot teaching, human-like motion prediction using LSTMs, and safe exploration in reinforcement learning using spatio-temporal Gaussian processes. To support reproducible research, he released a dedicated force-torque dataset for multi-shape robotic insertion tasks. Across his body of work, Munawar consistently addresses practical challenges in deploying intelligent robots in real-world manufacturing and exploration settings, making his research highly relevant to both academic roboticists and industry practitioners.
Research Focus
Key Achievements
Top Papers
- 1Deep reinforcement learning for high precision assembly tasks300 citations · 2017
- 2
- 3Deep Reinforcement Learning for High Precision Assembly Tasks29 citations · 2017
- 4Human-Like Hand Reaching by Motion Prediction Using Long Short-Term Memory10 citations · 2017
- 5Teaching a Robot Pick and Place Task using Recurrent Neural Network6 citations · 2016
- 6
- 7