Analysis of dynamic characteristics and performance evaluation of a two-axis gimbal sensor system

Abstract

A two axis gimbal sensor system captive in a dynamic platform such as an autonomous aerial vehicle or a high precision homing missile must stabilize a line-of-sight toward a target against the external motion induced by aerodynamic forces and maneuvering of a dynamic platform. It is well known that target pointing and tracking performance of a dynamic platform is largely affected by ability to decouple a sensor system from the body motion of a dynamic platform. This thesis addresses on the analysis of dynamic characteristics and performance evaluation for the stabilization loop of a two-axis gimbal sensor system with multivariable inputs such as reference, disturbance and noise. First of all, space stabilization concept and mathematical modeling for a two-axis gimbal sensor system is formulated. Dynamic model for the stabilization loop of a two-axis gimbal sensor system is derived and system parameters have been verified by practical experiments. The validity of the derived dynamic model is illustrated by the comparison of the experimental and simulation results. To satisfy desired design and performance specifications, such as gain margin, phase margin and body-motion decoupling performance etc., for a two-axis gimbal sensor system with multivariable feedbacks such as reference, disturbance and noise, three kinds of controller are designed. Firstly, the Lead-PI controller design using a loop-shaping method based on the empirical frequency response for the stabilization loop of a two-axis gimbal sensor system is realized. Secondly, the LQG/LTR controller design using modified uniform singular value to overcome the simultaneous gain matching problem at low and high frequencies is implemented. Thirdly, the robust $H_\infin$ controller design approach combining model matching and mixed sensitivity problem to improve simultaneously command following, disturbance rejection and noise suppression performance in the presence of external disturbances and model uncer...