Date of Award
MS in Electrical Engineering
Baseband digital communication in electro-magnetic measurement while drilling (EM-MWD) systems is often corrupted by non-white surface noise. The inability to reliably decode the transmitted signals in a noisy environment limits the depth at which EM-MWD systems can operate. Correlation receivers, which are optimal in the presence of additive white Gaussian noise, can be sub-optimal in the presence of various types of field noise at different drilling sites.
This thesis investigates the application of artificial neural networks (ANN) as communication receivers in EM-MWD baseband digital communication systems. The performances of various ANN architectures and training algorithms are studied and compared with conventional correlation receivers via computer simulations. Standard symbol error rate (SER) test results show that the NN receiver is able to adapt to site-specific noise and thus outperforms the traditional correlation receiver.