Head-On Object Size Determination Using a Forward Looking Sonar for Obstacle Avoidance

Abstract

Autonomous underwater vehicles (AUVs) are increasingly used for a wide variety of commercial, scientific, and military applications. However, shallow coastal environments present challenges to autonomous navigation that are not as prevalent when operating in deeper waters. Shallow water environments can contain a variety of obstacles, from naturally-occurring shoals and coral heads to man-made hazards like offshore structures, moorings, and even fishing gear. Therefore, AUVs must be able to detect and avoid these objects, a task that becomes more difficult in very shallow water. First, breaking waves generate water that is well-mixed with suspended sediment and entrained air bubbles, reducing visibility and making cameras unreliable for vision-based navigation. Second, hydrodynamic loads from waves and currents can produce unsteady vehicle motions which degrade the imagery produced by side-looking sonar systems. For these reasons, forward-looking sonar (FLS) is an effective sensor for detecting obstacles in these environments. Avoiding them, however, remains difficult because the seafloor and free-surface severely limit a vehicle’s ability to maneuver in the vertical plane. As a result, AUVs must be able to plan and execute obstacle avoidance maneuvers in the horizontal plane.