Balakumar Sundaralingam
University of Utah, Nvidia (United States), Georgia Institute of Technology
Papers
32
Total Citations
1,013
H-Index
15
About
Balakumar Sundaralingam is a leading robotics researcher whose work sits at the intersection of robot manipulation, deep learning, and motion planning. His research has made transformative contributions to how robots perceive, grasp, and manipulate objects — both rigid and deformable — with particular emphasis on bridging the gap between simulation and real-world deployment. His most cited work, "Deep Object Pose Estimation for Semantic Robotic Grasping" (2018, 283 citations), demonstrated how synthetic training data could be leveraged to achieve robust real-world grasping performance. This sim-to-real theme continues through "DeXtreme" (2023), which showcases agile dexterous in-hand manipulation transferred from simulation using deep reinforcement learning. His 2023 paper "CuRobo" introduced parallelized GPU-accelerated motion planning, offering a highly practical tool for the robotics community. Sundaralingam has also advanced multi-fingered grasp planning through probabilistic and differentiable deep network approaches, and contributed foundational benchmarking methodology for in-hand manipulation evaluation. His body of work — spanning trajectory optimization, natural language-guided planning, and deformable object grasping — reflects a researcher consistently pushing the boundaries of dexterous robotic manipulation toward real-world applicability.
Research Focus
Key Achievements
Top Papers
- 1Deep Object Pose Estimation for Semantic Robotic Grasping of Household\n Objects283 citations · 2018
- 2DeXtreme: Transfer of Agile In-hand Manipulation from Simulation to Reality88 citations · 2023
- 3CuRobo: Parallelized Collision-Free Robot Motion Generation82 citations · 2023
- 4Correcting Robot Plans with Natural Language Feedback67 citations · 2022
- 5
- 6
- 7
- 8Benchmarking In-Hand Manipulation48 citations · 2020
- 9
- 10