Factorized Kalman Filter Mapping on Linear Array Processor

Abstract

A mapping of factorized Kalman filter with n-states and m-measurements on (n + m + 1)-cell linear array is described. The approach is to combine the tone and measurement updates of the covariance matrix and use fast Givens rotations to triangularize an appropriate matrix. The fast Givens rotations do not require computations of arithmatic square-roots and are preferred over Householder transformations. The rotations of the rows in this matrix are done in parallel to distribute the computations. The cell-to-cell communication is minimized by arranging the matrix in a near triangular form. An extended Kalman filter, commonly used in target tracking applications, is implemented to demonstrate the mapping on Warp linear array. With the availability of high-performance multi-computer chips in the near future, the results open new solutions to on-board target tracking and other Kalman filtering applications.