IEEE International Conference on Robotics and Automation (ICRA) 2023

Learning-based Initialization of Trajectory Optimization for Path-following Problems of Redundant Manipulators

Learning-based Initialization of Trajectory Optimization for Path-following Problems of Redundant Manipulators

by Minsung Yoon, Mincheul Kang, Daehyung Park, and Sung-Eui Yoon

Korea Advanced Institute of Science and Technology (KAIST)

Received an outstanding planning paper award at ICRA 2023.
(As award papers, only 15 papers were selected among a competitive pool of 1,341 accepted submissions.)

Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7



Abstract: Trajectory optimization (TO) is an efficient tool to generate a redundant manipulator's joint trajectory following a 6-dimensional Cartesian path. The optimization performance largely depends on the quality of initial trajectories. However, the selection of a high-quality initial trajectory is non-trivial and requires a considerable time budget due to the extremely large space of the solution trajectories and the lack of prior knowledge about task constraints in configuration space. To alleviate the issue, we present a learning-based initial trajectory generation method that generates high-quality initial trajectories in a short time budget by adopting example-guided reinforcement learning. In addition, we suggest a null-space projected imitation reward to consider null-space constraints by efficiently learning kinematically feasible motion captured in expert demonstrations. Our statistical evaluation in simulation shows the improved optimality, efficiency, and applicability of TO when we plug in our method's output, compared with three other baselines. We also show the performance improvement and feasibility via real-world experiments with a seven-degree-of-freedom manipulator.

[ Graphical abstract ]

Video (6 min):




Supplementary video:


Contents:

* Paper: paper.pdf
* Slide: slides.pptx
* Poster: poster.pptx
* Source Code: Github Link