Goal-Conditioned Action Space Reduction for Deformable Object Manipulation

Authors: Shengyin Wang, Rafael Papallas, Matteo Leonetti and Mehmet Dogar
Conference: IEEE International Conference on Robotics and Automation (ICRA) 2023

Our work proposes a method to make it easier and faster for robots to manipulate deformable objects. By simplifying the geometric model and identifying key particles that are most relevant to achieve the goal, we can improve the efficiency and performance of motion planners. The ability to manipulate deformable objects is important in many industries, such as manufacturing, healthcare, and logistics. For example, in logistics, robots may need to handle packages of varying shapes and sizes. Our proposed method can improve the efficiency and accuracy of these tasks by reducing the computational cost of planning for deformable object manipulation.