I am working on problems at the intersection of Robotics x Computer Vision at Autonomous Field Robotics Lab. My research is focused on increasing the robustness of visual-inertial SLAM for underwater environments. I have worked on projects involving comparing various state estimation algorithms in the underwater domain, developing a deep pose estimation framework for underwater relative localization, and using GoPro as an inexpensive sensor for underwater structure mapping. I recently finished working on robust underwater state estimation by swithing between less-accurate odometry estimate and accurate VIO to get track AUVs when VIO fails.
š Iām currently working on tightly-coupled fusion of magnetometer with Visual-Inertial SLAM.