Direct versus iterative methods for fixed-point implementation of matrix inversion | IEEE Conference Publication | IEEE Xplore