Final Mtech.
Project
Topic Learning Approach to Link Adaptation in
WirelessMan scenario.
My final project
aimed at applying learning methods to link adaptation. Although the
method used for learning is generic enough to find many applications in
low memory footprint scenarios, with requirement for little or no
knowledge about the function to be optimized , we have
concentrated on handling a small problem of threshold
determination for switching from one coding profile to another. The
data used was generated with 802.16-OFDM
PHY and SUI model for
the medium (due to lack of time no MAC layer simulations were done,
although that shall be an important extension of the studies). The
coding for the entire PHY and SUI models was done in MATLAB. The BER and FER stats generated
were then used to create a network model in C/C++ with support for
obtaining errors at integral values of SNR.
Armed with this model and the C/C++ implementation of learning
automaton the convergence statistics were obtained for the learning
method.
The essential resources have been included below:
Thanks
-- Avinash