GraphSLAM - Advanced Robotics - Lecture Slides, Slides of Robotics

Download GraphSLAM - Advanced Robotics - Lecture Slides and more Slides Robotics in PDF only on Docsity!

GraphSLAM

Graph-based Formulation

n Use a graph to represent the problem n Every node in the graph corresponds to a pose of the robot during mapping n Every edge between two nodes corresponds to the spatial constraints between them n Goal : Find a configuration of the nodes that minimize the error introduced by the constraints

Approaches

• n Lu and Milios, ‘ n 2D approaches:
• n Montemerlo et al., ‘
• n Howard et al., ‘
• n Dellaert et al., ‘
• n Frese and Duckett, ‘
• n Olson et al., ‘
• n Grisetti et al., ’
• n Tipaldi et al.,’
  • § Nuechter et al., ‘ § 3D approaches:
  • § Dellaert et al., ‘
  • § Triebel et al., ’
  • § Grisetti et al., ’08/’

Graph-Based SLAM in a Nutshell

n Problem described as a graph n Every node corresponds to a robot position and to a laser measurement n An edge between two nodes represents a data- dependent spatial constraint between the nodes [KUKA Hall 22, courtesy P. Pfaff & G. Grisetti]

Graph-Based SLAM in a Nutshell

n Once we have the graph, we determine the most likely map by “moving” the nodes [KUKA Hall 22, courtesy P. Pfaff & G. Grisetti]

Graph-Based SLAM in a Nutshell

n Once we have the graph, we determine the most likely map by “moving” the nodes n … like this. [KUKA Hall 22, courtesy P. Pfaff & G. Grisetti]

Graph-based Visual SLAM

Visual odometry Loop Closing [ courtesy B. Steder]

The KUKA Production Site

The KUKA Production Site

scans 59668 total acquisition time 4,699.71 seconds traveled distance 2,587.71 meters total rotations 262.07 radians size 180 x 110 meters processing time < 30 minutes