Building an automatic control system of maneuvering ship in collision situation with genetic algorithms

The automatic control system of maneuvering ship in collision situation has been studied using improved Genetic Algorithms (GA). The navigational traffic rule at sea is appended to the fitness of GA and a new collision avoiding system at sea was suggested. In the system, navigating obstacles and other moving ships are identified by ARPA system (Automatic Radar Plotting Aids) and then the degree of probable collision threat in future, which is generated by those obstacles, can be predicted with a stochastic predictor. The stochastic predictor is designed such that the probability density map of these obstacles existence is derived from Markov process model in enough future time. Evaluation of the collision threat degree and navigational traffic rules at sea and efficiency of the path is considered respectively in the GA. The chromosome having variable length and analogous crossover guarantees the optimum or semi-optimum path to be obtained and the specified mutate guarantees the original path to be kept and make self-ship reach the goal place safely. This control system is examined with the experimental system that is equipped on "Shioji Maru" which is the training ship of our university. The experimental result has been so far satisfactory