Toward improving unmanned underwater vehicle sensing operations through characterization of the impacts and limitations of in situ environmental conditions

Many uncrewed underwater vehicles (UUV) missions are currently planned, executed, and analyzed with minimal environmental awareness. One common UUV operation is to detect and classify targets of interest within the water column or on the seafloor utilizing sonar, video imagery, and LiDAR (Light Detection and Ranging). When UUVs operate independent of environmental conditions, and conditions are unfavorable, poor data quality and resulting possible missed detections are more likely. Often the only remedy is costly secondary deployments occurring hours to days later, if at all. If deleterious conditions can be determined heuristically (e.g. low-quality data can be predicted) based on environmental conditions, then the UUV or its operators can take corrective actions before the deployment has concluded.Here, we compare historic UUV high frequency sidescan sonar and video imagery data quality with UUV observed in situ environmental conditions. Analysis of the video imagery found that qualitative measures of observable water clarity (i.e. none, poor, moderate, good) were significantly correlated (95% Confidence Intervals) with onboard UUV turbidity value measurements. We also analyzed several instances of operator observed refraction in the high frequency sonar imagery and found potential linkages to variations in ambient temperature and sound speed as observed by the vehicle in situ in the searchable sonar area.