Progressively Communicating Rich Telemetry from Autonomous Underwater Vehicles via Relays

Christopher Murphy, Ph.D., 2012
Hanumant Singh, Advisor

Without access to data acquired by autonomous marine vehicles, surface operators cannot fully understand the state of a mission.  Communicating imagery and high-resolution sensor readings to surface observers remains a challenge -- as a result, telemetry from free-roaming autonomous marine vehicles remains limited to 'heartbeat' status messages with minimal data available until recovery.  Long-distance communication may require relaying data across multiple acoustic hops between vehicles, yet fixed infrastructure is not always appropriate.

I propose mechanisms for interactive, progressive communication of data across multiple acoustic hops.  These mechanisms include wavelet-based embedded coding methods, and a novel image compression scheme based on texture classification and synthesis.  I then present CAPTURE, a telemetry architecture that employs advances in delay tolerant networking to enable progressive transmission of data, including imagery, across multiple acoustic hops. Specific design considerations include high-cost vehicle nodes with persistent storage and significant computation, and human surface operators.

CAPTURE provides an end-to-end software solution for communicating science data from AUVs.  Automatically selected imagery and time-series data are progressively transmitted across multiple hops.  Human feedback is incorporated by allowing operators to request arbitrarily higher resolution versions of data.  These components are demonstrated through field trials.