A Method For Enforcing State Constraints in Adaptive Control

Few methods exist for enforcing state constraints on a system with an adaptive control law. This paper develops an adaptive control law that enforces the system state to evolve in an elliptical region without using any sliding mode or switching type behavior as previous existing methods. Elliptical regions are chosen due to the structure of most nonlinear control stability proofs. No known constraint enforcement method for direct adaptive control can guarantee the system evolves in a more general constraint set. The new adaptive control law offers the potential ability to use less control effort than existing constraint enforcement methods since all possible system uncertainty does not have to be dominated to ensure constraint enforcement (which can lead to control saturation). The control law also offers the benefit that it behaves as a common flight tested and verified direct adaptive control law when the system is not at the constraint set. The potential ability of this adaptive law to enforce constraints is illustrated on a model for aircraft wing rock.