Video presentation

We present a method for fusing visual-inertial odometry (VIO) with single-antenna GNSS to provide accurate, long-term robot localization. This approach overcomes the inherent drift in position and heading associated with VIO while maintaining reliability during GNSS signal outages.

Specifically, we developed an Extended Kalman Filter (EKF) that utilizes GNSS measurements to correct drift in VIO-estimated poses. Embedded real-time experiments demonstrate that the algorithm effectively bounds position and heading errors, providing drift-free 6-DoF estimates that achieve decimeter-level position and degree-level heading accuracy.

Publications

  • Nguyen Hoang Khoi Tran and Vinh-Hao Nguyen, “An EKF-Based Fusion of Visual-Inertial Odometry and GPS for Global Robot Pose Estimation,” in IEEE International Conference on Recent Advances and Innovations in Engineering (ICRAIE), Kedah, Malaysia, 2021. Available here.

Citing

If you find this work useful please cite

@inproceedings{tran2021viogps,
  author    = {Tran, Nguyen Hoang Khoi and Nguyen, Vinh-Hao},
  title     = {An {EKF}-Based Fusion of Visual-Inertial Odometry and {GPS} for Global Robot Pose Estimation},
  booktitle = {IEEE International Conference on Recent Advances and Innovations in Engineering (ICRAIE)},
  year      = {2021},
  address   = {Kedah, Malaysia}
}

Updated: