Path planning and obstacle avoidance for multiple autonomous unmanned surface vehicles

Today, the use of unmanned surface vehicles becoming increasingly widespread. There are many benefits of doing demanding tasks by unmanned surface vehicles. It is a necessity for the unmanned surface vehicles to be moved in coordination with one another like schools of fish. Moreover, the easy accomplishment of civilian and military missions on the sea will also be possible if unmanned surface vehicles act together. In this study; priority has been given to safety of life and property at sea and various simulations have been carried out with the aim of planning path and avoiding obstacles where mobile and fixed obstacles exist. Simulations of multi-autonomous unmanned surface vehicles were conducted around the Prince Islands in Istanbul and the Istanbul Bosphorus which are located in Turkish territorial waters. In addition to path planning and obstacle avoidance algorithms, we also propose a method for sensor-independent visual navigation of unmanned autonomous surface vehicles using artificial intelligence technologies. Even if the electronic sensors are malfunctioning or subjected to electronic jamming, synthetic images which are derived from European Ship Simulator and real ship images are used and so that unmanned sea vehicles can continue autonomously without being affected by this situation. It is proposed to design the unmanned surface vehicles to conform to the existing rules, as there is no article in the international maritime rules for unmanned vehicles yet.