Autonomous Underwater Vehicle Planning for Information Exploitation

Abstract

The ability of an Autonomous Underwater Vehicle (AUV) to dynamically plan safe routes and maneuvers in dangerous environments is directly relevant for the future of the use of AUVs in the exploration and exploitation of the underwater environment, specifically the littorals and inland waters. This thesis builds upon the existing body of knowledge of the REMUS AUV dynamics and kinematics and develops a control scheme for a real-time optimized vehicle trajectory that will permit continuous and autonomous collection and exploitation of external sensor data, which will facilitate full 360-degree, 2-dimensional mapping of the underwater environment surrounding the vehicle while preventing the vehicle from coming into contact with mapped objects in the water. The developed control schema will seek to generate a trajectory in real-time that optimizes a key parameter of interest, the Information Gain, while minimizing a specified cost function of constraints, such as kinematic limits and obstacle avoidance criteria.