The Application Investigation of Neural Networks for Uplink Power Control in CDMA Mobile Communications

倪梁方,郑宝玉
DOI: https://doi.org/10.3969/j.issn.1673-5439.2005.01.001
2005-01-01
Abstract:In this paper,a kind of Neural Networks power control scheme is proposed,which focuses on the application of this networks to the problem of minimum transmitted power control.At first,the minimum transmitted power control model is derived.The principle diagram of this kind of Neural Networks to solve the model is then obtained in terms of steepest descent method,and the analysis is given to the theoretical basis and practical performance of minimum transmitting power control with this networks while the approach to calculate the parameters of the networks in terms of the requirements of power control is investigated.Finally,the performance of this controller is simulated.The simulation results show that the proposed neural networks power control scheme achieves a higher average signal-to-interference ratio(SIR) and a lower average deviation of SIR,of individual signal received by base stations,and larger fraction of users who converge to above 90% of target SIR than the minimun power assignment algorithm.This good controller performance would lead to a small outage probability and substantial capacity improvements.
What problem does this paper attempt to address?