LQG based model predictive control for gust load alleviation

Abstract

A linear quadratic Gaussian (LQG) based model predictive control (MPC) method is proposed to alleviate the dynamic gust loads of flexible aircrafts flying through turbulence, utilizing look-ahead information of the turbulence via light detection and ranging (LIDAR) systems or on board alpha probe. The new method features both infinite prediction horizon and infinite control horizon. The forepart of the infinite control sequence consists of a few online optimized variables, and the rest are outputs of an LQG controller, designed offline using an improved LQG method. The advantages of the proposed method are twofold. Firstly, the stability property of the controlled system is improved due to application of the infinite prediction horizon and the LQG controller. Secondly, adoption of an infinite control horizon not only improves the control performance, but also greatly reduces the number of online optimized control variables whilst retaining control performance. Furthermore, a technique to tackle the effects of control delay is also designed. The effectiveness and advantages of the proposed approach are demonstrated through numerical results using a general transport aircraft model.