Worcester Polytechnic Institute (WPI) researchers Nikolaos Gatsonis and Michael Demetriou have been awarded $347,843 by the U.S. Navy to develop a computational model that will use data from a network of autonomous underwater vehicles (AUVs) to estimate in real time the concentration and source of a contaminant liquid or gas plume as it moves and diffuses through ocean waters.
The estimator, once deployed on a system of AUVs, could help guide responses to chemical or biological threats, accidents, and environmental disasters, said Gatsonis, principal investigator (PI) of the project and professor and head of the Department of Aerospace Engineering. Demetriou, also a professor in the Department of Aerospace Engineering, is co-PI on the project.
“Locating the plume source and estimating in real-time the concentration of a liquid or gas plume in ocean waters is complex but critical,” Gatsonis said. “With better, faster detection and estimation of underwater plumes, the Navy and other agencies could respond to intrusions, submerged airplanes or vehicles, people who are trapped, and contaminants that should be contained or remediated.”
Gatsonis and Demetriou will develop the algorithms that can be rapidly executed on computer hardware on AUVs. The model will allow at least three AUVs to patrol an area until a plume is detected, collect sensor data, and then begin estimating the plume’s physical parameters. In addition, the model will allow AUVs to cooperate with one another, move to new locations to learn more about the plume, and continue working even if one or more AUVs fail.
WPI students, including one graduate student and multiple teams of undergraduate students, will contribute to the research. Part of the project will involve testing software on land with autonomous vehicles.
The three-year project builds on Gatsonis and Demetriou’s interdisciplinary research that couples advanced computational methods for real-time estimation with vehicle guidance and control. Gatsonis is an expert on development of computational methods and their application to the simulation of complex physical and aerospace systems. Demetriou focuses his research on control of systems and on optimal ways of moving or scheduling sensors that are distributed in space. The researchers previously developed a similar model, supported by Air Force Office of Scientific Research and DARPA grants, that would enable unmanned aerial vehicles to collect data and estimate the concentration of gas plumes in the atmosphere.
