Robust Reconfigurable Control for Recovery from Stern and Bow Plane Jams in Underwater Vehicles

Stern and bow plane jams have been shown to cause catastrophic results in underwater vehicles. In this paper, the design of a fully automated robust reconfigurable controller that prevents excessively large depth excursions that could cause broaching or dangerously deep depths is presented. This paper extends our previous research results by incorporating bow plane jams, and by using a more realistic use of ballast changes during the simulation. Using LQR, three sets of gain matrices (one for normal operation, and one for each jam type) are computed for the system model at different speeds. The appropriate gain is selected depending on the jam type and speed. Simulation results are presented to show the performance of the reconfigured control system. The controller is also tested for robustness using the Edge theorem, and it is shown to stabilize a sufficiently wide range of system models