Machine-Learning-Driven Synthetic Panoptic Segmentation Dataset for MASV
Currently leading research, under the supervision of Dr. Md Masud Rana, on the development of an AI enabled Maritime Autonomous Surface Vehicle (MASV) for real world canal monitoring and management within a canal system maintained by the Lower Neches Valley Authority (LNVA). The research aims to create an autonomous system capable of continuously monitoring surface vegetation, underwater conditions and water quality while reducing dependence on labor intensive field inspection and delayed laboratory based assessment.
The work integrates a modular robotic platform, AI based visual perception, autonomous navigation, environmental sensors, simulation and remote monitoring tools into a unified system. The canal environment is being recreated and tested in Gazebo Harmonic using ROS 2 Jazzy, allowing robotic, AI, sensing and navigation functions to be evaluated before physical deployment. The longer term goal is to develop the physical MASV and connect it with a web/API based monitoring system for live environmental data, remote control, mapping and AI assisted interpretation of canal conditions.
Conducted at Lamar University through the Texas Hazardous Waste Research Center (THWRC), the research is intended to support a practical, real world canal monitoring system rather than a simulation only prototype.