This concept covers the exact recursive least squares (RLS) lattice filter, which uses order-recursive and time-recursive updates of forward and backward prediction errors to avoid explicit matrix inversion. Central to the theory are the angle parameter (gamma), which measures the component of a unit vector orthogonal to a data subspace and can be updated by order recursion, and the parameter delta, which requires time recursion and links forward/backward prediction error energies (sigma) across lattice stages via projection operators on Krylov-like shifted-data subspaces. The concept belongs to adaptive signal processing / adaptive filtering theory, extending least-squares projection theory into a stage-wise (order-recursive) and QR-decomposition-based (unitary rotation) implementation of adaptive prediction filters.
This concept covers the exact recursive least squares (RLS) lattice filter, which uses order-recursive and time-recursive updates of forward and backward prediction errors to avoid explicit matrix in…