New tighter upper bounds on the performance of convolutional code over exponentially correlated Rayleigh fading channel

This paper presents tighter bounds on the pairwise error-event probability and therefore very tighter upper bounds of bit error rate(BER) for communication systems employing convolutional code over Rayleigh fading channel with exponential correlation coefficient. To obtain our bound, we start by proposing a closed form upper bound for Gaussian Q-function and as well as deriving a new lower bound of each channel covariance matrix eigenvalue. The simulation results show that our proposed bound is tighter than those previously developed, especially when the covariance parameter q increases. A gain of over 6.5dB was achieved for q = 0.95 at BER=10
^-5
from the better previous bound, whereas for the lower values of q we reach the simulated curve.