Two approaches for autonomous underwater vehicle global path planning in large range ocean environment

In this paper, two approaches for autonomous underwater vehicle are presented and used to solve the problem of global path planning in large range oceain environment. The first planning approach is so called genetic algorithm based on domain knowledge, and the second planning approach is based on A^* algorithm. In the two approaches, the planning workspace is represented as a set of equal grid based os known and certain large-scale chart data, which referred to as the grid-based model. The grid-based workspace model and its data structures are introduced in this paper. Some problem about the first approach are discussed in the paper, such as the coding scheme based on decimal grid-coordinate and variable-length chromosome, the generating method of initial population, the fitness evaluation function, the evolve strategy and some superiority genetic operators are all designed and introduced in detail. And some measures are adopted to improve the searching capability and to speed up convergence of the CA planning algorithm. And this paper also proposed the realization method of A^* to search an optimal planning path for AUV. The simulation result shows that the CA which adopting a method of variable length codes makes the path described simply and clearly, has the character of high speed global convergence, and can more effectively solve the problem of path planning for AUV, the A^* algorithm can find a relative optimal path to grid-based model in a little time, and both the two approaches can satisfy the system requirement of real time.