Dottorato in INGEGNERIA INFORMATICA

Titolo della tesi: DAMIAN: a Delay-Aware MultI-Aerial Navigation Environment for Cooperative DRL-based UAV Systems

A customizable environment based on Deep Reinforcement Learning (DRL) is here provided. The environment, called DAMIAN (Delay-Aware MultI Aerieal Navigation), allows to use different strategies for handling cooperative multi-UAV (Unmanned Aerial Vehicles) scenarios when delays are involved in the decision-making process for tasks such as spotting, tracking, coverage and many others. DAMIAN implements a realistic benchmark for delay-aware DRL solutions of cooperative multi-UAV tasks and is useful to develop and compare new solutions for this problem. Users can choose among various combinations of tasks and parameters and customize the scenarios by implementing new desired functionalities. Different approaches can be compared, taking into account either implicitly or explicitly the delays applied to actions and observations for both single-agent and multi-agent scenarios. The awareness of the delay, along with the possible usage of real-world-based external files, definitely increases the reality level of the environment by possibly easing the knowledge transferability process of the learned policy from the simulated environment to the real one. The extensible and modifiable environment developed along with the baselines use cases provided can generate new benchmarking tools for use cases where delays need to be considered.