FOSS4G-Asia 2023 Seoul

As the maritime traffic increases, maritime accidents are on the rise. Drones are being used to detect distressed vessels quickly in the open ocean. The current system does not have a plan to operate drones according to the search target, which means it is inefficiently operated and the search time is delayed, reducing the chance of survival. The operation plan should be developed based on tests in various environments and conditions. However, real-world implementations of distress situations have limitations. In the real world, implementing human distress in the ocean is limited by safety issues.
In order to find the optimal way to operate the drone for marine search and rescue, this study aims to construct a simulation environment with various environments and conditions.
To realize the simulation environment, Unity Engine was used. The environmental aspects of the simulation include the sea (sea colour, wave height), objects (vessels, people, etc.), and the environment (day, night, sunny, rain, snow, fog, etc.). In terms of functionality, camera angles can be set, and adding a drone function allows the drone to move and shoot. In this paper, the environment factor was fixed to daytime clear weather, and the focus was on the object to be searched. The simulation was developed with five types of distress vessels (2-tonne fishing vessel, 5-tonne fishing vessel, cargo vessel, passenger vessel, and motorboat) and six types of distressed persons (baby, adult, and with or without life jackets). The distress vessels can be set to sink, capsize, or drift, and the distress persons are drifting on the surface of the water. The altitude of the drone was limited to a maximum of 150m to comply with Korean law, and the camera angle could be set to (-30︒, -45︒, -60︒, -90︒). To reflect real-world coordinates, Cesium was integrated into Unity to simulate a marine environment. To do this, Cesium's coordinate system was adapted to be compatible with Unity's coordinate system, and the positions of the distressed person and the marine police in the distress scenario were expressed in terms of actual Earth longitude, latitude, and height. This maintains the scale and realism of the simulation environment.
It is expected that future simulation results will be used to develop a drone operation plan to effectively search for various distress ships and distressed person to minimize human casualties.