Innovative Methods for Optimally Mixing a Diverse Suite of Control Effectors for Marine Vehicles

The control of modern underwater vehicles presents a number of challenges, including significant nonlinearities in the vehicle dynamics and a potentially large suite of control effectors with widely differing properties. Current generation control architectures for submarines are often characterized by independent control loops for each actuator or for different vehicle axes, which are not able to explicitly address interactions among loops, particularly nonlinear interactions, or to combine the full suite of control effectors to produce a desired vehicle response. In Phase I, the team began developing inner-loop control approaches that address these deficiencies, and demonstrated these methods through their application to a fast-attack submarine model. The approaches combine model predictive control with a quadratic or piecewise linear programming based control allocator that explicitly accounts for actuator saturation nonlinearities. The Phase II effort will focus on refining the design methodology so that it can be used to rapidly implement a control law for any specific vehicle of interest. The proposed Phase II research effort will also add path planning and state estimation capabilities to provide a complete guidance and control system for submersibles. The broad applicability of the design methodology will be validated through its application to multiple diverse vehicle models.