Linear Filters for Linear Systems with Multiplicative Noise and Nonlinear Filters for Linear Systems with Non-Gaussian Additive Noise
Erik I. Verriest · American Control Conference · 1985
Exact optimal least squares linear filters with precomputable gains are derived for the class of discrete linear systems with state update and output corrupted by white noise multiplying a linear function in the state. The derived method is then applied to obtain suboptimal nonlinear filters for linear systems with non-Gaussian additive noise.