Project Objectives

The research objectives of this project include 1) understanding how the mobile acoustic channel fluctuates in the micro-scale of tens of milliseconds and 2) investigating the optimal ocean sampling strategy for given spatial correlation properties. The educational objective is to foster STEM career interests among middle school students in Alabama through underwater robotics competitions.
As the defining feature of our planet, the oceans are vital to our nation as energy and food sources, recreational destinations, science exploration space, and national defense barriers. Autonomous underwater vehicles (AUVs) represent a critical innovation in the exploration of our oceans and lakes. This project envisions an underwater mobile multi-modal communication network to support ocean sampling and AUV missions. We plan to address three research thrusts: 1) mobile acoustic channel modeling, 2) mobile adaptive acoustic communications, 3) integrated sampling, communications, and navigation. The gained science insights will help us to develop innovative signal processing techniques at multiple layers, from the channel modeling to the system level.

Current Efforts

• 1) We are developing recurrent neural networks to perform channel estimation for the underwater acoustic channel. The recurrent neural network uses a stacked structure to improve the channel estimation for time variation.
• 2) We are examining the impact of sampling schemes on interpolation performance for underwater acoustic field reconstruction.
• We are developing machine learning algorithms for underwater data muling with mobile nodes to minimize vehicle travel distance and improve data transmission fairness.
• 3) We are developing a simulation tool in ROS Gazebo to simulate a fleet of aquatic platforms for underwater sampling.
• 4) We are performing two major outreach events: MATHCOUNTS-Tuscaloosa for regional middle schools and the underwater robotics competition for local high schools.