|
For Full-Text PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
|
An Adaptive Congestion Control for Random Access Channels in Mobile Communication Systems
Hideaki YOSHINO
Publication
IEICE TRANSACTIONS on Communications
Vol.E86-B
No.2
pp.732-742 Publication Date: 2003/02/01
Online ISSN:
DOI:
Print ISSN: 0916-8516 Type of Manuscript: PAPER Category: Wireless Communication Technology Keyword: traffic, congestion, control, random access, mobile,
Full Text: PDF>>
Summary:
An adaptive congestion control scheme that can be applied to various random access protocols in mobile communication systems is proposed. Its main features are scalability for handling increasing numbers of mobile terminals and adaptability for coping with drastic changes in traffic load. These are achieved by controlling the traffic load adaptively to maintain maximum throughput even under overload conditions. Procedure for measuring and estimating offered traffic and a method of setting control thresholds that maximize the average throughput are analytically derived, and the algorithm for adaptively controlling the permission rate is described. This scheme was applied to both the slotted ALOHA and ICMA/CD protocols. For each protocol, three control parameters--the unsuccessful rate, optimal traffic, and control thresholds--were analytically derived. Then stationary throughput characteristics were numerically evaluated. We found that the scheme could achieve high throughput by regulating transmission adaptively depending on the offered traffic. The preferred range of the permission base rate that enables adaptive control and limits the amount of processing at terminals was also clarified. Since one of the main advantages of our scheme is its adaptability to drastic variations in traffic load, we simulated its transient characteristics with three types of time-variant input models. The results indicate that this control scheme achieved nearly theoretically optimal throughput even during an overload for each input model.
|
|