Collision Risk Assessment and Costmap Construction Methods for AUVs Using Forward-Looking Sonar

Autonomous underwater vehicles (AUVs) have been widely employed in a variety of tasks, ranging from environmental exploration and resource surveying to complex tasks that achieve superhuman performance. In this paper, a collision risk assessment method is proposed to estimating the probability of collisions between AUV s and obstacles. In the proposed method, an assessment model composed of a distance risk function and an orientation risk function is defined to calculate the collision risk. A rule base is designed to classify the collision risk of different levels while sending warning signals. In addition, a costmap construction method using forward- looking sonar is proposed, which estimates the collision cost and sonar cost to ensure the safety of AUVs. And the aspect ratio of the costmap is changed so that the planned path can meet the minimum turning radius constraint of the AUVs. The effectiveness of the proposed methods is demonstrated in a series of simulations.