Path Coverage Optimization for USV with Side Scan Sonar for Victim Recovery

Abstract

Uncrewed marine surface vehicles (USV) with side scan sonar are increasingly being used to locate submerged victims who drowned in open water. This work demonstrates a novel algorithm that automates path planning by optimizing transect orientation of a Boustrophedon path through a convex polygon for sonar quality. The orientation maximizes the length of the transects while minimizing the variation in length. The algorithm uses a weighted sum to score possible paths. The weightings are explored by simulation with four convex polygons of different sizes representing locations in Texas and Washington where marine search and recovery exercises have been conducted or have been planned. The overall weighting (0.5, 0.5) was demonstrated using the Hydronalix AMY USV at Lake Sahuarita, Arizona, confirming that the best scored orientation does produce a more favorable path for sonar than the worst scored orientation. In addition, the path for the worst scored orientation was more difficult to execute.