作者: Stéphane Dussy
DOI: 10.1007/978-1-4471-0265-6_5
关键词: Control theory 、 Stability (learning theory) 、 Linear matrix inequality 、 Filter (signal processing) 、 Convexity 、 Diagonal 、 Representation (mathematics) 、 Mathematics 、 Lyapunov equation 、 Digital filter
摘要: Implementation of a filter in fixed-point digital hardware results small errors during each step the algorithm, induced by inexact arithmetic structure real system (e.g., finite word-length effects). These can lead to unexpected behaviour since stability and performance are usually ensured for an idealised representation system, without taking into account erratic effects quantisation. This chapter addresses this finite-precision issue proposing convex formulation robust diagonal problem. approach guarantees wide class quantised systems, with direct application design filters.