Background: the research background refers to the optimization and control. The subject refers to the path planning of Autonomous Underwater Vehicles in real ocean environments. The field of study is a combination of mechanical and marine engineering. Aim: the aim of doing the research is improvement of safety for underwater vehicle path planning in addition to its improvements of optimization. Such an aim is done by using a modern kind of optimization that is particle swarm optimization with a modified type of penalty function in addition to the modification of population initialization.
Methodology: first, a review on the recent researches in the fields is done and a modern type of optimization in path planning (such as particle swarm optimization with penalty function), environment and problem modelling is done and then, the problem constraints are identified and modelled. Then, using evolutionary methods with some innovations, it is tried to solve the environment problems and improvement of them. The method used for the problems solution, is particle swarm optimization with a new type of penalty function for the exact constraint satisfaction. For verification, an objective function which is the total distance, is considered. In addition, it is tried to tune the objective function for correct calculation of it during optimization. The objective function, is a multiobjective one in real environment other than verification that is a combination of time and energy usage. Finally, by modifying obstacle avoidance constraint and coefficient decrement plus final angle-based calibration, it is tried to improve the optimization problem solution. Other case innovations such as velocity approximation strategy and triangular based population initialization is proposed for both increment in calculation speed and reduction of objective function.
Conclusions: the results show that, not only by modifying penalty function, the constraints are satisfied, but also by particle swarm