Linear quadratic Gaussian control under randomly varying distributed delays

A state estimation and state feedback control law has been formulated in a stochastic setting, based on the principles of minimum variance filtering and dynamic programming, for application to processes that are subjected to randomly varying distributed delays. The proposed estimation and control law for delay compensation is built upon the concept of the conventional Linear Quadratic Gaussian (LQC), hereafter called Delay Compensated Linear Quadratic Gaussian (DCLQG). Although the certainty equivalence property of LQG does not hold for DCLQG in general, the combined state estimation and state feedback approach of DCLQG offers a suboptimal solution to control of randomly delayed processes. Specifically, DCLQG is applicable to analysis and synthesis of control systems for vehicle management of future generation aircraft and autonomous vehicles where a computer network is employed for distributed processing and on-line information exchange between diverse control and decision-making functions. Results of simulation experiments are presented to demonstrate the efficacy of the proposed DCLQG algorithm for flight control of an advanced aircraft.