Abstract:
Inter symbol interference and rapid time variations makes doubly selective
channels difficult to equalize. The rapid time variations entail a receiver
that also adapts itself to the channel variation. This increases the
implementation complexity requirement for the receiver. Therefore faster
techniques are much needed to establish real time communication over
such severe channels. The focus of this research report is on faster
equalization techniques for doubly selective channels. The block
equalization technique was used as it emphases on low complexity
equalization. MMSE equalizer is reformulated as the problem solving a
system of linear equations. This allows the application of algorithms from
linear algebra predominantly iterative methods for solution of system of
linear equations. Jacobi, Gauss Seidel, Steepest Descent and Conjugate
Gradient are used. Jacobi and Gauss Seidel are good for square matrices of
lower order, but worse for higher order matrices. Steepest Descent takes
more time and more number of iterations to converge than Conjugate
Gradient. This report concludes by proposing the Conjugate Gradient
method is suitable for low complexity block equalization problem from the
perspective of run time and bit error rate.
