Adaptive Algorithms
Matrix analysis: eigenanalysis, SVD, QR, LU, Cholesky factorization. Wiener filters, linear prediction, lattice filters. SGD, LMS, NLMS, RLS, QRD-RLS. Kalman filters including square-root forms and extensions to nonlinear systems (EKF, UKF, particle filters). Performance analysis and robustness. Optimization problems, KKT conditions. Signal manifold estimation, adaptive subspace (GROUSE). Multiple discriminant analysis, FKT. Neural networks as adaptive nonlinear systems, representation theorem, backpropagation. A major focus of the course is configuring algorithms to fit specific applications.
Prerequisite: ECE 211
Credits: 3.00
Course Code: ECE 416