Sherdil Niyaz

About

Sherdil Niyaz is a leading researcher in robotic manipulation and motion planning, with a focus on bridging simulation and real-world performance. His most influential work, *Dex-Net 2.0* (2017), has accumulated over 1,400 citations and revolutionized deep learning for robotic grasping by training models on a massive synthetic dataset of 6.7 million point clouds and grasps. This approach dramatically reduced the need for time-consuming physical data collection, enabling robust grasp planning directly from simulated data. Niyaz has also made significant contributions to agricultural robotics, as demonstrated in his work on *Robotic Lime Picking* (2021), where he innovatively treated leaves as permeable obstacles to improve fruit harvesting in dense foliage. In surgical robotics, he advanced motion planning for concentric tube robots by using nearest-neighbor graphs to follow complex surgical trajectories (2020), and developed a bounded evaluation method to optimize motion-planning problem setups efficiently (2019). His research consistently tackles real-world constraints—from cluttered orchards to delicate surgical paths—making him a key figure in practical, deployable robotics.

Research Focus

Key Achievements

Top Papers

1. 1
   Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

   1,162 citations · 2017
2. 2
   Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

   277 citations · 2017
3. 3
   Robotic Lime Picking by Considering Leaves as Permeable Obstacles

   13 citations · 2021
4. 4
   optimizing Motion-Planning Problem Setup via Bounded Evaluation with Application to Following Surgical Trajectories

   11 citations · 2019
5. 5
   Following Surgical Trajectories with Concentric Tube Robots via Nearest-Neighbor Graphs

   10 citations · 2020

Key Collaborators

Contact & Links