Mission planning optimization for fleets of underwater gliders to increase the accuracy of hydrodynamic forecasting systems

Autonomous underwater gliders offer an attractive option for gathering data in the upper ocean because of their relatively easy, flexible, and low cost deployment, operation and recovery. Compared to AUVs that use propellers, autonomy is much longer, at the expense of speed. For this reason, multiple gliders would be required to adequately cover large areas in a short time (e.g. a regional basin at the synoptic time scale). Such a fleet mission drastically reduces errors of hydrodynamic forecasting systems when the oceanographic glider measurements are assimilated. Given an initial map of errors in a parameter of interest, the glider fleet should be directed to sample in regions that would most reduce the error in the mission time available. To do this, while taking into the account ocean currents that cause deviations from planned sampling trajectories, a method has been developed which calculates optimum waypoints, i.e. the set that minimizes the variance of the parameter of interest after data are assimilated. The method has been applied for the optimization of temperature measurements south of Cyprus in the eastern Mediterranean Sea for 1, 2, and 3 glider fleets with initial waypoints in a number of user-defined configurations. Results show that optimized open trajectories depend heavily on initial conditions, but that pre-planning can avoid situations where gliders do not reach the area of interest because of strong currents. Future work is needed to generate an initial set of waypoints that can reliably lead to the globally optimum set of trajectories, as well as to enforce closed trajectories (pre-determined recovery points).