Skip to content

Latest commit

 

History

History
23 lines (18 loc) · 1.2 KB

roadmap.md

File metadata and controls

23 lines (18 loc) · 1.2 KB
Raw

KinDynVIO

KinDynVIO is a factor graph-based, tightly-coupled multimodal sensor fusion framework that aims at fusing inertial, kinematic, visual and centroidal factors of a humanoid robot in order to achieve a reliable whole body estimation.

Roadmap

  • Add points tracker class for image features
  • Add lines tracker class for image features
  • Add features manager class for tracking the features across images
  • Add an example of BLF filter with GTSAM iSAM2 based fixed lag smoother with odometry factors
  • Add IMU preintegration class
  • BLF filter as odometry factor + IMU preintegration for kinematic-inertial smoothing
  • Add an example of Visual-inertial smoothing based on image features
  • Test on a realsense recorded dataset with Aruco marker tracking as ground truth
  • Connect Kinematic-inertial smoother with visual-inertial smoother
  • Add centroidal dynamics factor
  • Add centroidal kinematics factor
  • Add legged odometry based velocity-preintegration factor
  • Test kinematic-inertial-centroidal smoother with iCub 2.5 dataset
  • Add Gazebo simulated experiment of kinematic-inertial-visual-centroidal smoother with iCub3 robot walking experiment
  • ...