Vinay Naidu

◦ Working with Dr. Daniel Stilwell and Dr. Lynn Abbott on a robust 3D pipeline for identifying and characterizing methane eruption sites using advanced 3D Side Scan Sonar technology. Conducted data preprocessing through Voxel Grid Filtering, Depth Thresholding, and SOR Filtering. Implemented novel Multi-Ransac Line Segmentation Model for detecting the plume lines and used graph theory with DFS to cluster the lines with in the given threshold. ◦ My research tackles this highly nonlinear and ambiguous problem by leveraging multi-agent decentralized estimation with a team of autonomous underwater vehicles (AUVs). 1. Decentralized Multi-Agent Estimation – Each AUV independently estimates the target’s motion using only bearing measurements, ensuring robustness against communication constraints. 2. Resolving Port-Starboard Ambiguity – Bearing measurements from a planar hydrophone array introduce a mirrored ghost target, making direct localization difficult. My work integrates Bayesian filtering techniques to suppress incorrect hypotheses over time.

Spearheaded Autonomy and GNC for India's First Commercial Autonomous Underwater Vehicle- Neerakshi. Developed a Novel Hybrid Architecture with Moos-Ivp and ROS2.

Project completed during the internship are, Universal robot 5 robot description (Universal Robot Description Format) in ROS. Denavit-Hartenberg forward kinematics, and Inverse Kinematics to visualize the robot's motion in Gazebo, Delta robot forward, and inverse kinematics using python to visualize the robots pose for the given position of the end-effector.