Artificial intelligent adaptive control for DS-CDMA with open-loop power control operating over a low Earth orbiting satellite line

In this paper, we present the results of using a neural network based adaptive coprocessor to decrease the performance degrade caused by the variations of received signal power. According to the statistical characteristics of the receiver outputs, the neural network based coprocessor adaptively adjusts the soft-decision quantization thresholds used for forward error correction decoding. Based on the assumption of open-loop power control and Gaussian multiple access interference, simulation results indicate that, under the condition of having a coded bit error rate of 10^-3 with E_b/N₀ range from 10 dB to 6 dB, the neural network based coprocessor can provide about 9% to 16% increase in capacity (corresponding to 40% to 60% bit error rate improvement) of the DS-CDMA system over a low Earth orbit satellite (LEOS) channel in cellular commucations