Progressively communicating rich telemetry from autonomous underwater vehicles via relays

Abstract

As analysis of imagery and environmental data plays a greater rolein mission construction and execution, there is an increasing needfor autonomous marine vehicles to transmit this data to the surface.Without access to the data acquired by a vehicle, surface operatorscannot fully understand the state of the mission. Communicating imageryand high-resolution sensor readings to surface observers remainsa significant challenge - as a result, current telemetry from free-roamingautonomous marine vehicles remains limited to 'heartbeat' statusmessages, with minimal scientific data available until after recovery.Increasing the challenge, long-distance communication may requirerelaying data across multiple acoustic hops between vehicles, yetfixed infrastructure is not always appropriate or possible. In thisthesis I present an analysis of the unique considerations facingtelemetry systems for free-roaming Autonomous Underwater Vehicles(AUVs) used in exploration. These considerations include high-costvehicle nodes with persistent storage and significant computationcapabilities, combined with human surface operators monitoring eachnode. I then propose mechanisms for interactive, progressive communicationof data across multiple acoustic hops. These mechanisms include wavelet-basedembedded coding methods, and a novel image compression scheme basedon texture classification and synthesis. The specific characteristicsof underwater communication channels, including high latency, intermittentcommunication, the lack of instantaneous end-to-end connectivity,and a broadcast medium, inform these proposals. Human feedback isincorporated by allowing operators to identify segments of data thatwarrant higher quality refinement, ensuring efficient use of limitedthroughput. I then analyze the performance of these mechanisms relativeto current practices. Finally, I present CAPTURE, a telemetry architecturethat builds on this analysis. CAPTURE draws on advances in compressionand delay tolerant networking to enable progressive transmissionof scientific data, including imagery, across multiple acoustic hops.In concert with a physical layer, CAPTURE provides an end-to- endnetworking solution for communicating science data from autonomousmarine vehicles. Automatically selected imagery, sonar, and time-seriessensor data are progressively transmitted across multiple hops tosurface operators. Human operators can request arbitrarily high-qualityrefinement of any resource, up to an error-free reconstruction. Thecomponents of this system are then demonstrated through three fieldtrials in diverse environments on SeaBED, OceanServer and BluefinAUVs, each in different software architectures.