Deploying the Robot Operating System (ROS) for Complex Path Planning on Planar Surfaces
Loading...
Date
2021-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Industrial robotics are characterized by intrinsically closed systems and present a unique difficulty to the factory integration efforts of the next industrial revolution. Homogenizing the controls of a factory, specifically a robotic operation, allows engineers to design for modularity. To meet this need, collaborative academic and industry efforts have developed the Robot Operating System (ROS) as a package-based programming framework. Developers assembling motion control systems have a diversity of options for key packages such as the path planner, the kinematic motion planner, the collision checker, and the robot-specific support package.
The goal of this study is to configure and deploy a motion control system for an industrial robotic cell using existing ROS packages. Further, a path planning package will be developed to navigate complex planar paths on a flat surface. Existing path planners for curved, planar-like surfaces are designed to output a raster (parallel rows of linear lines) type motion. A method for creating complex, non-linear paths across a surface is needed to support applications such as non-destructive inspection and surface finishing. To provide context to this study, a demonstrative application of navigating a robotic arm through a grid-based maze presented to the robot work cell was employed. Initial work configured the software environment including the robot support, motion planning, and sensor packages necessary to form the motion planning pipeline and testing tools.Machine vision was enabled through the use of an RGB/stereo depth camera and static frames were filtered with the DREAM.3D data processing software. Filtered data was used to solve and navigate the presented maze. A tool for generating and visualizing these three-dimensional pathways was created. This study successfully stood up a motion control system based on the MoveIt planning framework and deployed a novel planar path planning technique. Analysis determined the MoveIt framework to be insufficient for a semi-constrained, short throw paths and suggested future work to explore alternative motion planners. With many existing tools focusing on parallel paths, this tool offers greater flexibility to system designers in achieving unique, complex motion.
Description
Keywords
Robot Operating System (ROS), Path Planning, Industrial Robotics, Robotics, MoveIt, System Integration, Applied Machine Vision