/uwb-localization

Accurate 3D Localization for MAV Swarms by UWB and IMU Fusion.

Primary LanguageC++Apache License 2.0Apache-2.0

UWB Localization

Accurate 3D Localization for MAV Swarms by UWB and IMU Fusion. ICCA (International Conference on Control and Automation) 2018, Anchorage, USA. Jiaxin Li, Yingcai Bi, Kun Li, Kangli Wang, Feng Lin, Ben M. Chen

The paper is available here.

@article{li2018accurate,
      title={Accurate 3D Localization for MAV Swarms by UWB and IMU Fusion},
      author={Li, Jiaxin and Bi, Yingcai and Li, Kun and Wang, Kangli and Lin, Feng, and Chen, Ben M.},
      journal={arXiv preprint arXiv:1807.10913},
      year={2018}
}

More codes are comming...

Datasets

Indoor UAV fly test with VICON ground truth

Link to rosbag. Provides IMU, UWB, VICON measurements

  • /ref_generation/current_reference: reference sent to UAV. ~50Hz.
  • /time_domain/full_range_info: UWB raw measurements. ~80Hz.
  • /mavros/imu/rpy_acc_short: IMU raw measurements, provides roll, pitch, yaw & acceleration. ~50Hz.
  • /mavros/vicon/position: VICON ground truth, precision of 0.1cm, ~20Hz.

UWB anchor positions are acquired by Least Square Optimization over the measurements between each pair of UWB anchors, in NWU coordinate.
anchor_list: [101, 102, 103, 104, 105, 106]
anchor_101: [0, 0, 0]
anchor_102: [6.09394, 8.20272e-05, 0.00105595]
anchor_103: [0.161215, 6.21084, -0.0104368]
anchor_104: [5.65533, 5.99776, 2.55325]
anchor_105: [5.9612, 0.178178, 2.54669]
anchor_106: [-0.257302, 3.58951, 2.5628]

vicon-xy

Introduction

A ROS based library to perform localization for robot swarms using Ultra Wide Band (UWB) and Inertial Measurement Unit (UWB).

The algorithm has been deployed to a multiple drone light show performace in Changi Exhibition Center of Singapore, during the opening ceremony of Unmanned System Asia 2017, Rotorcraft Asia 2017. Video link can be found here.

Features:

  • EKF/UKF localization algorithms that integrate UWB and IMU. Up to 90Hz with the accuracy of 5cm.
  • C++/ROS interface for TimeDomain UWB sensors
  • Least square optimization to calibrate the positions of UWB anchors

Usage

time_domain

C++/ROS based driver for TimeDomain UWB sensors.

slam_pp

Localization algorithms to perform real-time UWB and IMU fusion.

uwb_calibration

Least square optimization to determine the 3D position of the anchors. Depends on the Ceres Solver.