Skip to content

Tutorial on pose graph optimization using g2o

Notifications You must be signed in to change notification settings

Quantum-Entropy/g2o_tutorial

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 

Repository files navigation

g2o_tutorial

  1. This repository is the tutorial on the general edge types(EDGE_SE2 and EDGE_SE3) and vertex types(VERTEX_SE2 & VERTEX_SE3) found in g2o, using python. The accompanied documentation on results can be found in the notion page.
  2. We have covered two common scenarios found in SLAM: Pose Graph SLAM and Landmark based SLAM using synthetic dataset. The main goal is to cover how to formulate a graph based optimization problem in proper g2o format, using g2o's vertices, edges and information matrix.
  3. Problem statement:
    1. Pose Graph SLAM: A robot is travelling in a oval trajectory. It is equipped with wheel odometry for odometry information and RGBD sensors for loop closure and ICP transformation. Due to noise in wheel odometry it generated a noisy estimate of the trajectory. Our task is to use loop closure pairs to correct the drift.
      Alt Text
    2. Landmark SLAM: A robot observes a cube from five different locations. The robot is equipped with RGBD sensors and would be using those to calculate odometry and the map of the cube. Due to noise in the sensors, it obtained a erroneous estimate of its poses and vertices of the cube. Our task is to couple odometry measurements and cube's vertices(landmarks) measurements to generate a consistent and better estimate of those values.
      Alt Text
  4. Installation:
    1. You can either install using g2o binaries using command: sudo apt install ros-kinetic-libg2o or build g2o from source.
    2. Other common dependencies are:
      open3d, scipy, numpy, math, matplotlib
  5. Directory structure:
    1. pgSLAM:
      1. pgSlam.py: Python script creating dataset, generating g2o file and visualizing pose graph at each stage.
      2. pgSlam.ipynb: Modular and detailed implementation of each function in the pgSlam.py.
    2. landmarkSlam:
      1. landmarkSlam.py: Python script creating cube and robot positions, generating g2o file and visualizing all the intermediate steps.
      2. landmarkSlam.ipynb: Modular and detailed implementation of each function in the landmarkSlam.py.

About

Tutorial on pose graph optimization using g2o

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 98.9%
  • Python 1.1%