Research of Acquiring Eigenvector of Real Symmetric Matrix
Luo Feng, Kun He, Wu Shunjun · 2006
A systolic array processing method that works out the eigenvectors based on the Jacobi algorithm for eigenvalue/eigenvector decomposition (EVD) is presented in this paper. According to systolic array processing architecture, it is of highly parallel and can get eigenvectors when eigenvalues are worked out. It is easy to perform, for it works with coordinate rotation to avoid complicated algorithm such as multiple, divide and evolution etc. At last some simulations show the precision and delay can be satisfied when the method is applied to real time domain