GR and BP neural network-based performance prediction of dual-antenna mobile communication networks

The performance of a dual-antenna mobile communication network in 2-Rayleigh fading is investigated in this paper. Exact average symbol error probability (SEP) expressions with selection combining (SC) are derived for q-ary phase-shift keying (PSK) and pulse-amplitude modulation (PAM). Exact expressions are also given for the channel capacity. It is important to predict the performance of mobile communication networks in complex wireless environments. Thus, we propose generalized regression (GR) and back-propagation (BP) neural network-based SEP prediction methods. The theoretical results are used to generate training data. The proposed prediction methods are compared to the extreme learning machine (ELM), locally weighted linear regression (LWLR), support vector machine (SVM), and radial basis function (RBF) neural network methods. The results obtained verify that the proposed methods provide better SEP predictions.