Source code

The attached zip file contains all the source codes, which can also be found in a GitHub repository.

List of source codes:

• leader-follower

  ROS node that implements the leader-follower algorithm.

• Object-tracker-ROS

  ROS node for UWB and UVDAR data fusion. Could be used for other types of position/distance sensors as well.

• UWB_MRS_ranging_ros_node

  ROS node for communication with DWM1001-DEV.

• UWB-MRS-ranging

  Zephyr RTOS application for DWM1001. The nRF Connect SDK is needed for compilation and flashing.

• UWB-workspace

  Complete ROS Noetic workspace for UWB. Includes all necessary packages with real-world tmux scripts and gazebo simulations.

The source code for the thesis is at https://github.com/vitpetrik/UWB-thesis and can be viewed as a web page at https://vitpetrik.github.io/UWB-thesis/.

All the source codes are tested on Ubuntu 20.04 and should compile right away. In case of any trouble, feel free to contact the author of this thesis at petrivi2@fel.cvut.cz, vit.petrik@gmail.com, or GitHub.

ISO-OSI network model

Figure 1: OSI-ISO network model.

Object tracker and UWB evaluation tests

Multiple tests were run to evaluate the function of the Object tracker. As a test environment, the Gazebo simulator was used. The first UAV, an observer, remained at the coordinates \(\begin{bmatrix} 0 & 0 & 5 \end{bmatrix}^T\) for the entire duration of the tests. The second UAV, a performer, followed a pre-planned trajectory.

Range error is the difference between the ground truth distance between the UAVs and the distance calculated via unscented transform from Object tracker pose estimation.

Screenshot from Gazebo simulator showing the two UAVs.

Circle path with UWB fusion.

Circle path with standalone UVDAR.

Square path with UWB fusion.

Square path with UWB fusion.

Flower path with UWB fusion.

Flower path with standalone UVDAR.

Figure 2: Test of not moving UWBs.