A. Biblias, E. Rigas, N. Bassiliades, “Monitoring Ships' Emissions Using Unmanned Aerial Vehicles”, Proc. of the 13th Hellenic Conference on Artificial Intelligence (SETN '24), Sep 11-13, 2024, Athens, Greece, Association for Computing Machinery, New York, NY, USA, Article 21, 1-6.
The use of Unmanned Aerial Vehicles (UAVs) to monitor airborne emissions from ships has become quite widespread, especially in maritime areas where stricter restrictions have been imposed to minimize air pollution caused by ships. This paper focuses on this application of drones and specifically aims to schedule the trips that will be made by a set of drones in order to monitor as many ships as possible given a set of spatial and temporal constraints. The general drone scheduling problem is broken down into two sub-problems: In the first sub-problem, the number of ships and the set of requests for ship surveillance, (i.e., the positions and times that require surveillance) are known in advance. This problem is modeled as an Integer Linear Programming (ILP) one and is solved offline and optimally. In the second sub-problem, similarly to the first, the number of ships is known in advance; however, the set of requests is not, thereby requiring that, during the period the drones are either on their trips or idle, new monitoring requests may be generated. To solve this sub-problem, an online heuristic algorithm was developed. Both optimal solution and heuristic algorithm are evaluated on various sets of realistic data, and their efficiency is verified, with the main conclusion being that the optimal solution always ends up monitoring more ships than the heuristic, while the heuristic algorithm is significantly faster than the optimal and scales to larger problems.
