Research On Unmanned Surface Vessel Collision Avoidance System Based On Evolutionary Genetic Algorithm

In recent years, with the increasingly consciousness about maintaining maritime rights in the world, USV as the main strategic weapons on the sea, the study on USV autonomous control has attached more attention, and the USV development will be more intelligent, systematism and standardization. One of the key technologies about USV autonomous control is intelligent collision avoidance system. This article combines robot path planning method with Convention on the International Regulations for Preventing Collisions at Sea to construct USV collision avoidance system based on evolutionary genetic algorithm.The article analysises USV collision avoidance and collision avoidance mechanism, for the USV encouter situation is carried on the detailed divison. It researches on the caculate method about USV motion parameters and the risk degree of collision from the perspective of time and space is caculated carefully.The thesis constructs USV collision avoidance decision-making system based on evolutionary genetic algorithm. The algorithm adopts the heuristic initialization method of population, According to the actual requirements of USV autonomous navigation, The economy、smoothness and safety evaluation factors are introduced to the fitness function. And in the evolutionary operators, it adds delete operation, repair operation, smooth operation based on the traditional evolutionary operation, then the population can updated.The system establishes an optimization iterative process, when the iteration is completed, it can conclude a reasonable collision avoidance path.The USV collision avoidance decision-making system software is devised with MATALB software. The USV navigation parameters and dynamic target movement parameters are enterde the through human-computer interaction interface, and the USV navigationg environment is imitated by some circles in different radius. It does a large number of simulations in two cases:the first one is that there only static obstacles exist; the second one is that both static obstacles and dynamic targets are in the simulation environment. Finally, the output path that can avoid the static obstacles and dynamic target effectively, and it can ensure the path smoothness and economy. The simulation proves that the USV collision avoidance decision-making system based on evolution genetic algorithm can real-time decides the collision risk, avoiding static obstacles and dynamic target effectively, realizing the USV autonomous collision avoidance decision.