Nonlinear control based on H-infinity theory for autonomous aerial vehicle

Abstract

Existence of modeling errors, external disturbances, and inaccurate design assumptions make robustness a desired property for any control system in real-life applications. Linear robust controllers are widely seen as acceptable solutions when systems are controlled close to known equilibrium points and a known trajectory. Unfortunately, linear robust control designs become partially ineffective when these conditions are not met. The nonlinear and unsteady aerodynamics of aircraft in the presence of external disturbances and adverse conditions make application of linear robust controllers challenging. This paper presents a nonlinear version of robust H-infinity controller based on L2 gain and dissipativity concepts. The nonlinear H-infinity approach allows larger perturbations from the trim condition and delays any control degradation and risk of instability compared with the linear versions. The nonlinear H-infinity controller requires the solution of a Hamilton-Jacobi-Isaacs equation which is a limiting factor due to its complexity. This paper applies a state-feedback Taylor series expansion of the value function to iteratively solve the partial differential equations in incremental steps for an increasing degree of accuracy. A large UAS in trajectory tracking is used for performance comparison and robustness analysis in one hand with a linear robust controller, and on the other, with a robust nonlinear model predictive controller.