Local extrema refinement strategy for quadrotor's nonlinear attitude-systems with tensor product model transformation

The classical sampling method that is equi-interval sampling is often used in tensor product (TP) model transformation, however, the information of local extrema for the model can easily be omitted by this method. In this paper, a local extrema refinement strategy based TP model transformation method for quadrotor nonlinear attitude subsystem is proposed. Linear matrix inequalities based Lyapunovs theorems are applied to parallel distributed compensation (PDC) controller design, the stabilization and external disturbance results of the PDC controller designed under the classical sampling method and the local extreme refinement sampling method are compared to fully demonstrate the effectiveness of our proposed method.