Several approaches to fixed-point implementation of matrix inversion

Matrix inversion is a general problem in a wide variety of applications. The problem becomes even more challenging when an efficient hardware implementation is required. In this paper, three matrix inversion implementations in fixed-point are presented and compared. The fixed-point implementation of such methods is considered and the needed processing elements are introduced. The characteristics of each approach are described and their hardware implementations are compared.