Title

Simulator for Planning Collision-Free Trajectories in Manipulation of Objects Applications

DOI

https://doi.org/10.1007/978-3-030-45096-0_58

Document Type

Conference Proceeding

Publication Date

1-1-2020

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Abstract

This article presents the functional description of a simulator that plans trajectories in an automatic fashion in order for a Motoman HP20D industrial robot to move box-like objects from an initial to a final position avoiding obstacles in its workspace. The simulator sets up robot positioning based on the data entered by the user as well as the position of obstacles (boxes) by using an offline workspace reconstruction based on a computer vision tool, the main component of which is an RGB-D camera. The vision system estimates the position, orientation and size of a set of boxes to create a virtual representation in the simulator. In addition to that, a collision detection module, cost-efficient in computational terms, is described as the structural block of the trajectory planning process. Finally, the work space road map is obtained through a lazy PRM algorithm, as a global planner, and a type-A* search algorithm, as a local planner. This way, a collision-free trajectory is obtained and animated in the simulator as the final result of the general planning process.

Volume

12014 LNCS

First Page

477

Last Page

484

ISSN

03029743

ISBN

9783030450953

Identifier

SCOPUS_ID:85083998552

Documento original

Compartir

COinS