Thesis Code: 23004

Thesis Type: M.Sc. thesis in Computer Science, Mechatronics, Electronics, Information Technology

Research Domain: Connected Systems & Cybersecurity


  • Strong knowledge in software development (C/C++)
  • Good knowledge of MATLAB
  • Good knowledge in mathematical derivation
  • Proactive mindset, problem-solving oriented
  • Data processing skills
  • Familiarity with embedded platforms (Raspberry PI) and Linux environment

Recently, autonomous navigation of Unmanned Aerial Vehicles (UAVs) has received increasing great attention from the research community. UAVs are going to be widely adopted also in indoor environments to support a variety of applications. These include logistic operations, automated inventory management in warehouses, inspection, and monitoring tasks in industrial plants, as well as precision agriculture in greenhouses. In order to enable autonomous navigation in such dynamic scenarios and challenging indoor environments, it is crucial to accurately estimate the UAV’s position and attitude in real-time. Continuous and high-rate updates of these estimates are essential for the navigation module of the UAV platform.
Typically, Ultra-Wideband (UWB) technology is used for position estimation, providing accurate Time of Arrival (ToA) measurements. For attitude estimation, Inertial Measurement Unit (IMU) sensors, which integrate accelerometer and gyroscope, are employed to estimate roll, pitch, and yaw angles of the UAV. In cases where UWB connectivity becomes temporarily unavailable due to UWB signal interference or attenuation, IMU sensor data can be combined with UWB ranging measurements. This fusion leads to a more robust and smoother estimation of both position and attitude. However, it is important to remark that the integration of IMU data over time is subject to drift effects. Therefore, having an accurate model of the IMU sensor is extremely useful to optimize performance.

The goal of this thesis is to firstly derive an accurate model of the selected IMU and subsequently design a hybrid localization algorithm that combines both UWB and IMU data enabling autonomous operations of UAVs in indoor environments. Specifically, the hybrid algorithm will combine ToA and IMU measurements to estimate both position and attitude of the UAV. A recursive method, such as the Extended Kalman Filter (EKF), will be employed for this purpose.
Initially, the designed algorithm will be tested through computer simulations and optimized iteratively. Subsequently, the optimized algorithm will be implemented in UWB devices, and its performance will be evaluated in the Robotic Laboratory made available by LINKS. In particular, the localization performance will be evaluated by using the VICON system as the ground truth. The VICON system is a precise localization solution based on infrared cameras, which offers an accuracy of 0.1 mm at high refresh rate (about 100 Hz).

Duration: 6-8 months.

Contact: please send a resume with attached the list of exams passed during the Bachelor of Science and Master of Sciences to or specifying the thesis title.