Abstract

As part of the research for development of a leader-follower formation between unmanned underwater vehicles (UUVs), this study presents an optical feedback system for UUV navigation via an optical detector array. Capabilities of pose detection (translational offsets in x, y, z-axes and also yaw rotation) and control in a static-dynamic system (e.g. UUV navigation into a docking station) and a dynamic-dynamic system (e.g. UUV to UUV leader-follower system) are investigated. In both systems, a single light source is utilized as a guiding beacon for a tracker/follower UUV. The UUV uses an optical array consisting of photodiodes to receive the light field emitted from the light source.

For UUV navigation applications, accurate pose estimation is essential. In order to evaluate the feasibility of underwater distance detection, the effective communication range between two platforms, i.e. light source and optical detector, is determined. Based on the light attenuation in underwater, the geometry and dimensions of an optical detector array are calculated, and the boundary conditions for the developed pose detection algorithms along with the error sources in the experiments are identified. As a test bed to determine optical array dimensions and size, a simulator, i.e. numerical software, is developed, where planar and curved array geometries of varying number of elements are analytically compared and evaluated. Results show that the curved optical detector array is able to distinguish 5 degree- of-freedom (DOF) motion (translation in x, y, z-axes and pitch and yaw rotations) with respect to a single light source. Analytical pose detection and control algorithms are developed for both static-dynamic and dynamic-dynamic systems. Results show that a 5 x 5 curved detector array with the implementation of SMC is reasonably sufficient for practical UUV positioning applications.

The capabilities of an optical detector array to determine the pose of a UUV in 3-DOF (x, y and z-axes) are experimentally tested. An experimental platform consisting of a 5 x 5 photodiode array mounted on a hemispherical surface is used to sample the light field emitted from a single light source. Pose detection algorithms are developed to detect pose for steady-state and dynamic cases. Monte Carlo analysis is conducted to assess the pose estimation uncertainty under varying environmental and hardware conditions such as water turbidity, temperature variations in water and electrically-based noise. Monte Carlo analysis results show that the pose uncertainties (within 95% confidence interval) associated with x, y and z-axes are 0.78 m, 0.67 m and 0.56 m, respectively. Experimental results demonstrate that x, y and z-axes pose estimations are accurate to within 0.5 m, 0.2 m and 0.2 m, respectively.

Presenter Bio

Firat Eren received the B.S. degree in Mechatronics Engineering from Sabanci University, Istanbul, Turkey, in 2008, and the M.S. degree in Mechanical Engineering from the University of New Hampshire, Durham, NH, USA, in 2011 He is currently a Ph.D. candidate in Mechanical Engineering at the University of New Hampshire. His research interests include system dynamics and control, navigation, detection and control of underwater vehicles, optical systems, and communication.

Publication Date

4-17-2015

Document Type

Presentation

Download

Share

COinS