Wiberg, D. M.Ljungquist, D.Powell, T. D.A recursive parameter estimator in discrete time is presented for multiple-input, multiple-output linear stochastic systems that are bilinear in the parameters and state. It is globally convergent to minima of the asymptotic negative log likelihood function, and approximates the fast transient response of the optimal nonlinear filter used as a parameter estimator.en-USparameter estimationnonlinear filteringrecursive estimationsystem identificationoptimal estimationestimation theoryglobal convergenceSystems IntegrationTechnical Report