Shallow waters SLAM experiments on meredith AUV using forward looking sonar

This paper describes the sea trial experiments conducted to validate the physical implementation of online EKF-SLAM onboard the Meredith Autonomous Underwater Vehicle (AUV) in shallow coastal waters. This paper provides details of the hardware and software implementations of the EKF-SLAM algorithms running real-time on the AUV to perform feature-based navigation. The computation resources onboard Meredith and the hardware configuration that facilitated the real-time SLAM demonstration are also featured in this paper. The localization estimates from the SLAM algorithm were validated with ground truth provided by GPS and the deployed targets. Dead-reckoning navigation solution based on DVL and FOGs that were used for vehicle autonomy were also made available for results validation.