3.0

Average Course Rating

This course will cover principles and techniques for designing, implementing, and analyzing linear and nonlinear state estimators for dynamical systems for which traditional least-squares and linear Kalman filtering approaches might not be sufficient. In particular, emphasis is placed on state space systems that are characterized by partial observability and/or non-Gaussian uncertainties that, generally, arise in applications governed by complex non-linear stochastic dynamics and measurement processes. First, a brief review of matrix theory, state-space models and realizations, probability theory, dynamic system motion models, least-squares estimation, Luenberger observers, and linear Kalman filters (continuous and discrete versions) is presented. Then, these concepts are extended to advanced state estimation concepts and applications, to include: extended Kalman filtering, unscented Kalman filters, Cubature and Cubature-Quadrature Kalman Filters, and particle filtering along with various application examples.