Random access protocol and its impacts on air interface in cellular networks

Abstract

Cellular technology has already evolved through its first and second generations. Currently it is on the verge of the third generation, characterized by the addition of different types of data transmission techniques. In the near future, optimization of radio resource allocation with different kinds of traffic sources having different quality of services will be one of the main challenging problems. This thesis intends to find few solutions to optimize the air interface of current cellular systems with both kinds of traffic, current voice and low speed data traffic and future different types of data traffics.

Nowadays, GSM, TDMA 136, IS-95 and PDC are the most successful second-generation cellular systems. The random access part and the traffic channel part are utilized, respectively, for call reaching in the base transceiver station and information transmission in the uplink direction of a wireless cellular network such as the Global System for Mobile communications (GSM). Optimization of the random access part and the traffic channel part are studied, respectively, in the first part and the second part of the thesis. Overall radio interface optimization considering these two parts together is studied in the third part of the thesis. Suitable cases are confirmed by simulations.

Some form of Slotted ALOHA is used in all current digital cellular networks. The multiple power level transmission system in the original Slotted ALOHA concept, has one inherent problem. It exhibits a throughput-collapse phenomenon whereby the throughput decreases exponentially to zero as the mean number of aggregate transmission attempts increases beyond optimum. Few throughput-collapse preventing algorithms are presented first in this thesis. The random access protocol with the retransmission cut-off scheme is used in the access channel for initial access into the TDMA based traffic channel. The analysis of the random access channel accompanied by retransmission cut-off and its optimization is devised. The number of slots occupied by a terminal can be changed dynamically in the newly developed General Packet Radio Services (GPRS) system. An analysis of the GPRS transmission system with and without RLC/MAC block (radio packet) dropping is carried out.

A call is either rejected or blocked depending on its inability to succeed either in the random access part or in the traffic channel part. It is shown that the number of random access slots provisioned in the wireless access system has a significant impact on the overall call set up probability and traffic channel efficiency. The number of random access slots at optimum performance is derived considering four parameters: average call arrival rate, number of retransmissions, number of traffic channels and average call holding time. The results show that the number of random access slots at optimum performance depends on the average call arrival rate and the number of retransmissions. Moreover, it is independent of the number of traffic channels and average call holding time.

The models presented in this thesis could be used as a guideline for the optimization of radio resource allocation in the future evolution of cellular networks.