Path Planning For An Autonomous Underwater Vehicle

In the last decade, researchers and organizations have shown great interest in the investigation of underwater environment for many applications such as military, oceanographic research, and more recently offshore with the desire to exploit natural resources at over 1000 meters deep. Remotely operated vehicles (ROV) have long been developed and have been in use but more recently autonomous underwater vehicles (AUV) have been designed. The need for high capacity autonomous navigation in the cluttered underwater environment has necessitated the interest for more research into the problems faced by these vehicles. One of such problems faced by AUV concerns how the vehicle should have the ability to plan paths around obstacles in order to safely operate.This thesis is focused on the path planning mission. The objectives addressed here is to enable AUV to traverse its environment from any given current position to a desired goal position. The path to this goal location must enhance vehicle maneuver around any obstacle to reach the goal through the shortest, safest, minimal cost and optimal path in terms of time and energy consumption. Many approaches to solve the path planning problems have been proposed in recent years, but all have limitations of one form or the other.This thesis focuses on planning paths in a plane. Here, in order to easily take into account any modification of the environment; we simply represent obstacles as polygons. For generating an optimal shortest path we have transformed the working space to the configuration space, using visibility graph methods combined to A* search algorithm. The advantage of using this combination is that it offers a relatively fast and effective way to solve for safe paths around obstacles. However, the resulting shortest paths from the consecutive position taken by the vehicle are jagged. To solve the jagged paths, a trajectory smoothing component (using polynomial data fitting techniques) is developed; the parameters are set in this thesis. Then by combining all those process the path planner is created.The resulting path planner integrated onboard the AUV is able to achieve the proposed mission. In order to carry out an assessment of the path planner's performance, different variants of the simulation environment are created using MATLAB. In this simulation, the vehicle is represented as a polygon. The resulting paths generated demonstrate the efficiency of the path planner based on A* search algorithm for obstacle avoidance and optimizing the time and energy consumption.