A neurodynamic optimization approach to robust pole assignment for synthesizing piecewise linear control systems

This paper presents a neurodynamic optimization approach to robust pole assignment for synthesis of piecewise linear control systems via state feedback. The robust pole assignment is formulated as a pseudoconvex optimization problem with linear equality constraints where a robustness measure is considered as the objective function. The robustness is achieved by means of minimizing the spectral condition number of the closed-loop eigensystem. Two recurrent neural networks with guaranteed global convergence are applied for solving the optimization problem in real time. Simulation results are included to substantiate the effectiveness and demonstrate the characteristics of the proposed approach.