POWER PREDICTION IN MOBILE COMMUNICATION-SYSTEMS USING AN OPTIMAL NEURAL-NETWORK STRUCTURE

This paper presents a novel neural-network-based predictor for receive d power level prediction in direct sequence code division multiple acc ess (DS/CDMA) systems, The predictor consists of an adaptive linear el ement (Adaline) followed by a multilayer perceptron (MLP). An importan t but difficult problem in designing such a cascade predictor is to de termine the complexity of the networks. We solve this problem by using the predictive minimum description length (PMDL) principle to select the optimal numbers of input and hidden nodes, This approach results i n a predictor with both good noise attenuation and excellent generaliz ation capability. The optimized neural networks are used for predictiv e filtering of very noisy Rayleigh fading signals with 1.8-GHz carrier frequency, Our results show that the optimal neural predictor can pro vide smoothed in-phase and quadrature signals with signal-to-noise rat io (SNR) gains of about 12 and 7 dB at the urban mobile speeds of 5 an d 50 km/h, respectively, The corresponding power signal SNR gains are about 11 and 5 dB, Therefore, the neural predictor is well suitable fo r power control applications where ''delayless'' noise attenuation and efficient reduction of fast fading are required.