Autonomous Underwater Data Muling Using Wireless Optical Communication and Agile AUV Control
Friday, February 21st, 2014
ABSTRACT: Underwater exploration and surveillance currently relies on subsea cables and tethers to relay data back to the user. The cause for this is that water heavily absorbs most electromagnetic signals, preventing effective radio communication over large distances, and that underwater communication with acoustic signals affords only bit rates on the order of Kilobits per second.
In this talk we present a novel design and implementation for an underwater data muling system. This system allows for automatic collection of underwater datasets without the need to physically connect to or move the sensors by using mobile robots to travel to the sensors and download the data using wireless optical communication to bring it back to the base station.
The system consists of two parts. The first part is a modular and adaptive robot for underwater locomotion in six degrees of freedom. We present a hardware design as well as control algorithms to allow for in-situ deployment without the need for manual configuration of the parameter space. To achieve this we designed a highly parameterizable controller and methods and algorithms for automatically estimating all parameters of this controller. The second part of the data mulling system is a novel high-bandwidth optical underwater communication device. This device allows for transfer of high-fidelity data, such as high-definition video and audio, images, and sensor logs. Finally we present algorithms to control the robots path in order to maximize data rates as it communicates with a sensor while using only the signal strength as a measurement.All components and algorithms of the system have been implemented and tested in the real world to demonstrate the validity of our claims.
BIO: Marek Doniec is a postdoctoral associate working with Nicholas Roy at the Computer Science and Artificial Intelligence Lab at MIT. His research interests include robotics, instrumentation, sensor networks, underwater wireless optical communication, and multi-robot systems. He holds a Ph.D. in Computer Science and Electrical Engineering from MIT, a M.S. in Computer Science from Yale University and a Vordiplom (B.S.) in Computer Science from the University of Karlsruhe. He has been an active member of research labs in the United States, Germany, France, and Singapore. Additionally, he has experience in industry, having worked at Rethink Robotics developing controls for a humanoid arm.
February 21, 2014