Robust estimators for target tracking
Abstract
Several non-adaptive robust estimator design methods are discussed in this thesis. Constant estimator gains that guarantee the acceptable performance of the estimator in the presence of target model uncertainties are computed. Two approaches are considered for designing robust estimators: a theoretical approach and a trajectory-based approach. In the theoretical approach, cost functions are related to the estimation error variance obtainable from the Lyapunov equation under the assumption of Gaussian white noise system input. Several methods for designing estimators robust to parameter variations in the plant model are presented. A mechanism for trading off small estimator error variance and low sensitivity to unknown parameter variations is developed using the estimator gains computed from an iterative algorithm. A robust estimator design method that considers both real parameter variations and the unknown inputs is also presented. Trading off the robustness to real parameter variations and the robustness to unknown inputs is possible using a design parameter. In the trajectory-based approach, the cost functions are related to the estimation error with target trajectories generated by a realistic (deterministic) pilot input. An estimator is designed to have acceptable performance in tracking several different target trajectories with a capability to trade off the mean and maximum estimation error. These robust estimator design methods are applied to the aircraft tracking problem. A tracker with a mathematical model capable of using both position data and orientation data is designed to be robust to the uncertainties of the aircraft being tracked. This tracker's performance is compared with the performance of the conventional $\alpha - \beta$ tracker that uses only position data.
Degree
Ph.D.
Advisors
Andrisani, Purdue University.
Subject Area
Aerospace materials
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.