Online Trajectory Replanner for Dynamically Grasping Irregular Objects

TL;DR

A new trajectory replanner for grasping irregular objects that aims to achieve a “dynamic grasp” of the irregular objects, which requires continuous adjustment during the grasping process and a trajectory optimization framework that comprises two phases.

Abstract

This paper presents a new trajectory replanner for grasping irregular objects. Unlike conventional grasping tasks where the object's geometry is assumed simple, we aim to achieve a "dynamic grasp" of the irregular objects, which requires continuous adjustment during the grasping process. To effectively handle irregular objects, we propose a trajectory optimization framework that comprises two phases. Firstly, in a specified time limit of 10s, initial offline trajectories are computed for a seamless motion from an initial configuration of the robot to grasp the object and deliver it to a pre-defined target location. Secondly, fast online trajectory optimization is implemented to update robot trajectories in real-time within 100 ms. This helps to mitigate pose estimation errors from the vision system. To account for model inaccuracies, disturbances, and other non-modeled effects, trajectory tracking controllers for both the robot and the gripper are implemented to execute the optimal trajectories from the proposed framework. The intensive experimental results effectively demonstrate the performance of our trajectory planning framework in both simulation and real-world scenarios.

Figures (9)

• Figure 1: An example setup of grasping irregular objects.
• Figure 2: Schematic drawing of (a) the KUKA LBR iiwa R820, (b) the SDH 2 hand, (c) the camera Basler AVA 1000-100GC, and (d) the object. The $x$-, $y$-, and $z$-axis of each coordinate frame are depicted by red, green, and blue arrows, respectively. In the tuple of the rotation angle and its corresponding frame $(\mathcal{O}_{\mathrm{h},i},q_{\mathrm{h},i})$, the illustrating color of the joint angle $q_{\mathrm{h},i}$ matches the corresponding rotate axis.
• Figure 3: (a) Side view of the SDH2 without the finger 2 with $q_{\mathrm{h},1}=-q_{\mathrm{h},6}=-\pi/2$. (b) Side view of the gripping hand with the SDH2 state variables $\mathbf{q}_\mathrm{h}$ and the grasping state $\mathbf{q}_G$.
• Figure 4: Timeline
• Figure 5: Side view of the grasping object and fingertips with the potential function $p\left(0,y\right)$.