Figure 10
From: Manipulation of free-floating objects using Faraday flows and deep reinforcement learning
The difference in object trajectories, \(r_{DTW}\), during repetitions of each task for different object diameters. A lower \(r_{DTW}\) value indicates a highly repeatable trajectory, representing the minimized local cost measure during dynamic time warping. Values are normalized using the container radius \(r_c\).
