Dynamic Cloth Manipulation Considering Variable Stiffness and Material Change Using Deep Predictive Model with Parametric Bias

TL;DR

This research shows that Musashi-W, a musculoskeletal humanoid with a variable stiffness mechanism, can dynamically manipulate cloth while detecting changes in the physical properties of the manipulated object.

Abstract

Dynamic manipulation of flexible objects such as fabric, which is difficult to modelize, is one of the major challenges in robotics. With the development of deep learning, we are beginning to see results in simulations and in some actual robots, but there are still many problems that have not yet been tackled. Humans can move their arms at high speed using their flexible bodies skillfully, and even when the material to be manipulated changes, they can manipulate the material after moving it several times and understanding its characteristics. Therefore, in this research, we focus on the following two points: (1) body control using a variable stiffness mechanism for more dynamic manipulation, and (2) response to changes in the material of the manipulated object using parametric bias. By incorporating these two approaches into a deep predictive model, we show through simulation and actual robot experiments that Musashi-W, a musculoskeletal humanoid with variable stiffness mechanism, can dynamically manipulate cloth while detecting changes in the physical properties of the manipulated object.

Figures (15)

• Figure 1: Dynamic cloth manipulation by the musculoskeletal humanoid Musashi-W considering body control with variable stiffness and adaptation to material change.
• Figure 2: The overview of our system: deep predictive model with parametric bias (DPMPB), controller using DPMPB for dynamic cloth manipulation, variable stiffness controller for musculoskeletal humanoids, data collector for DPMPB, and online updater of parametric bias (PB).
• Figure 3: Operational stiffness ellipsoid when adding force of 1 N while changing $f^{const}=\{10, 30, 50, 70\}$ [N].
• Figure 4: Experimental setup: the simulated simple robot and the musculoskeletal humanoid Musashi-W used in this study and cloth materials of soft and hard type polyethylene foam.
• Figure 5: Simulation experiment: the trained parametric bias when setting $C_{damp}=\{0.03, 0.05, 0.07\}$ and $C_{mass}=\{0.05, 0.10, 0.15\}$, the trajectory of online updated parametric bias when setting $(C_{damp}, C_{mass})=\{(0.05, 0.05), (0.07, 0.10), (0.03, 0.15)\}$, and the trajectory of online updated parametric bias when setting $(C_{damp}, C_{mass})=(0.03, 0.10)$ for the integrated experiment.