Software design for UAV applications in GNSS-denied environments

Nowadays the unmanned aerial vehicles (UAVs) are getting more and more interest, their velocity and precision make them a powerful tool for a wide range of applications, from aerial photography to search and rescue operations. These systems rely on the global navigation satellite system (GNSS) as primary source of position information for navigation, without reliable signal, a UAV's ability to maintain accurate positioning and autonomy is compromised, for this reason the UAVs are not suitable to work in environments where GNSS signals are either weak or unavailable, so called GNSS-denied environments. However, there are many operations, such as indoor and urban inspections, that take place in GNSS-denied environments that would benefit from the use of a drone. The design and development of UAV software capable of enabling reliable navigation in the absence of GNSS signals has become an area of intense research. Currently, most of the solution to this problem leverage alternative sensor modalities such as: visual odometry, LiDAR, inertial measurement units (IMUs) and barometric pressure sensors. These modalities, in conjunction with advanced estimation algorithms, ensure autonomous flight. Another possible solution is to triangulate the UAV position, but the estimate could be inaccurate because of the obstacles in the GNSS-denied environments. The ultra-wideband (UWB) technology is an innovative technology with low power consumption that implements a communication defined by a wide frequency spectrum, this characteristic offers high precision and robustness against interference, making it ideal for indoor and GNSS-denied environments. This thesis focuses on the design and development of software for UAVs operating in a GNSS-denied environment. The software uses ROS2 humble, PX4 (v 1.15.2 dev) and the DDS-XRCE protocol to create the necessary architecture to enable and control the flight of the drone. It was developed with the help of LINKS foundation for a specific hardware setup, for this reason the implementation of some parts of the software are strictly hardware dependent. After the development, the software has been tested in two different environments: the first one leverage a mocap system to obtain position information; the second one the UWB technology. The first scenario was mainly used to improve the software correctness and the UAV performances while the second to check the system’s performance in real-world scenarios.