Papers
3
Total Citations
68
H-Index
2
About
Jishiyu Ding is a roboticist advancing the frontiers of intelligent manipulation and multi-agent coordination. Their research centers on three pivotal areas: safe and object-specific robotic grasping in cluttered environments, multirobot path optimization under congestion risk, and curriculum learning for multi-goal reinforcement learning. Ding’s most impactful contribution is the development of REGRAD, a large-scale relational grasp dataset that enables robots to perceive object relationships and execute sophisticated tasks like searching for and grasping specified targets in clutter—a leap beyond traditional grasping methods. This work has garnered 43 citations, underscoring its significance in the field. In multirobot systems, Ding proposed an improved PRM* algorithm for multiobjective coverage, eliminating zig-zag paths to enhance efficiency in obstacle-dense, congestion-prone settings (23 citations). Additionally, their density-based curriculum for multi-goal RL addresses the challenge of sparse rewards, reducing the need for laborious reward engineering while avoiding bias. Ding’s research bridges perception, planning, and learning, offering scalable solutions for real-world robotic applications. Their work is essential reading for researchers tackling clutter, congestion, and reward sparsity in autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3