An autonomous air-sea rescue system using Particle Swarm Optimization

The focus of this paper is the application of Particle Swarm Optimization to an autonomous air-sea rescue system. This heuristic technique is used to search an area of the sea for survivors of a sinking ship, and is compared with the Random Search, which is used as a benchmark. The search is carried out by four small unmanned helicopters that are deployed from a stationary platform, which in this case, is an autonomous surface vessel. The helicopters are instructed to search a given region for survivors by measuring the temperature of each point visited in the search space using an infrared camera. It has been found that on average, the Particle Swarm Optimization algorithm covers less ground than the Random Search but detects more targets, indicating the effectiveness of the heuristics.