Suboptimal Markovian smoothing estimates based on continuous curves of solutions of the algebraic Riccati inequality (Q1614333)

Consider the steady state smoothing problem for linear stochastic models of the form \[ dx= Fx dt+ B_1 du+ B_2dv,\quad dy= Hx dt+ R^{1/2} du, \] where \((du,dv)\) is a standard \(p\)-dimensional Wiener process and \(B_2\neq 0\). Suboptimal Markov smoothing estimates are derived using continuous curve solutions to an algebraic Riccati inequality, which in turn are constructed on the basis of a result on continuous dependence of solutions to an algebraic Riccati equation on the data matrices.