Fuzzy/Neural Congestion Control for Third Generation DS-CDMA Cellular Systems

碩士 === 國立交通大學 === 電信工程系 === 87 === In this thesis, we study congestion control in DS-CDMA cellular systems. We propose a fuzzy technique for congestion control in DS-CDMA with PRMA (packet reservation multiple access) as its MAC (medium access control) protocol. We design a fuzzy congesti...

Full description

Bibliographic Details
Main Authors: Bo-way Chen, 陳伯偉
Other Authors: Chung-ju Chang
Format: Others
Language:en_US
Published: 1999
Online Access:http://ndltd.ncl.edu.tw/handle/61385666756630034405
Description
Summary:碩士 === 國立交通大學 === 電信工程系 === 87 === In this thesis, we study congestion control in DS-CDMA cellular systems. We propose a fuzzy technique for congestion control in DS-CDMA with PRMA (packet reservation multiple access) as its MAC (medium access control) protocol. We design a fuzzy congestion controller, named fuzzy channel access function (FCAF), based on knowledge from the channel access function (CAF) used in conventional DS-CDMA/PRMA cellular systems. We also design a fuzzy/neural congestion control for DS-CDMA/FRMA (frame reservation multiple access), where FRMA evolves from PRMA. FRMA separates the contention traffic from the reservation traffic so that the reservation traffic will not be disturbed by the contention traffic, unlike PRMA. The fuzzy/neural congestion controller is constituted by a pipeline recurrent neural network (PRNN) interference predictor, a fuzzy performance indicator, and an access probability controller (APC). This APC can be either fuzzy access probability controller (FAPC) or neural-net access probability controller (NAPC). By taking more system performance parameters into consideration, the access probabilities given by FAPC and NAPC can be more accurate. Simulation results show that the DS-CDMA/PRMA system with FCAF performs better than that with conventional CAF in overall performance. Furthermore, the DS-CDMA/FRMA system with fuzzy/neural congestion controller overrides the DS-CDMA/PRMA system with CAF in the voice packet dropping ratio, the corruption ratio, the utilization, and data packet delay. This approach also outperforms the DS-CDMA/PRMA system with FCAF in the voice packet dropping ratio under medium traffic load, the corruption ratio and the utilization for all traffic loads. Moreover, NAPC outperforms FAPC. If the requirement of the voice packet dropping ratio is set to be $10^{-2}$, the DS-CDMA/FRMA system with NAPC has an improvement of $10.76\%$ in capacity, the DS-CDMA/FRMA system with FAPC has $7.59\%$ improvement, and the DS-CDMA/PRMA system with FCAF has $5.06\%$ improvement, with comparison to the DS-CDMA/PRMA system with CAF.