Autonomy system for USV/UUV coordinated sampling

Abstract

An autonomy system that allows a Boeing-Liquid Robotics Wave Glider, an unmanned surface vehicle (USV), to optimally track a Teledyne-Webb Research Slocum Glider, a buoyancy-driven, unmanned underwater vehicle (UUV), by exploiting environmental information is described, along with operational results from a 2-month experiment offshore Iceland. To conduct collocated oceanographic observations with this heterogeneous mix of unmanned assets, autonomy routines were developed to predict Slocum Glider surfacing locations and plan a Wave Glider path that minimized geographic separation between the two platforms. Prediction and planning engines are tuned to the unique performance characteristics of each platform. To ensure survivability of the Wave Glider in the extreme seas present at our study site, autonomy was developed to switch the Wave Glider control between mission objectives and platform survivability. The best estimate of a Slocum Glider's path during a flight segment is based on commanded and measured information from the Slocum Glider just before diving: average speed over ground (SOG), desired heading, depth averaged current (DAC), and the duration of the dive profile. For the Slocum, the DAC is computed from the offset between the expected and actual surfacing location during the previous dive, divided by the segment duration. This time-invariant treatment of the ocean current structure is found to be inadequate to accurately predict vehicle surfacing locations in regions where tidally-driven or other rapidly time-evolving flow patterns dominate. An improved model for predicting Slocum Glider trajectories in the Iceland Basin that incorporates real-time current measurements from a Wave Glider-mounted acoustic doppler current profiler (ADCP) and a short-term predictive tidal model is used to improve the estimate of depth average current. This capability is unique in that no additional information is explicitly exchanged between the vehicles (say from an acoustic modem). The improved path plan is uploaded to the USV in real-time via satcomms, allowing the surface asset to remain close to the UUV for coordinated scientific sampling.