Obstacle Avoidance Control for the REMUS Autonomous Underwater Vehicle

Abstract : As the Navy continues its development of unmanned underwater vehicles, the need for total autonomous missions grows. Autonomous Underwater Vehicles (AUV) allow for advances in mine warfare, harbor reconnaissance, undersea warfare and more. Information can be collected from AUVs and downloaded into a ship or battle group's network. As AUVs are developed it is clear forward-look sonar will be required to be able to detect obstacles in front of its search path. Common obstacles in the littoral environment include reefs and seawalls which an AUV will need to rise above to pass. This thesis examines the behavior and control system required for an AUV to maneuver over an obstacle in the vertical plane. Hydrodynamic modeling of a REMUS vehicle enables a series of equations of motion to be developed to be used in conjunction with a sliding mode controller to control the elevation of the AUV. A two-dimensional, 24 deg. vertical scan forward look sonar with a range of 100 m is modeled for obstacle detection. Sonar mappings from geographic range-bearing coordinates are developed for use in MATLAB simulations. The sonar 'image' of the vertical obstacle allows for an increasing altitude command that forces the AUV to pass safely over the obstacles at a reasonable rate of ascent and pitch angle. Once the AUV has passed over the obstacle, the vehicle returns to its regular search altitude. This controller is simulated over different types of obstacles.