Kalman filtering and systolic processors
Hen‐Geul Yeh · 2005
A new parallel computing structure of the systolic array type for implementing the Kalman filter for state-space estimation is presented. This corresponds to the adjustment of Kalman filter algorithms to the type of the Faddeev algorithm. The Faddeev algorithm provides a wide range of matrix computational capabilities and it maps easily into a concurrent systolic array architecture. The special arrangement for the data flow mapping from Kalman filter to Faddeev algorithm, and from Faddeev algorithm to architecture are presented.