AUTONOMOUS FREQUENCY ALLOCATION FOR ROVING BASE STATIONS

Abstract

Roving base stations are useful in areas that are devoid of communications infrastrncture (fixed line and cellular) or have no long-term need for such architecture. It is also useful when there is a sudden, temporary surge in traffic load at a particular location. However, mobility brings about an immediate problem -- fluctuating signalto-interference ratio (SIR) due to the unpredictable movements. As the cells are now mobile, cochannel cells may be moving towards each other and create an unacceptable level of interference that deteriorate the performance of the systems. In this thesis, 3 autonomous frequency allocation algorithms are proposed. These autonomous assignment algorithms must be rapid, effective and efficient. The first algorithm, the Basic Autonomous Frequency Allocation (BAFA), allows each base station (BS) to select an operating frequency that minimises interference from other BSs. The BS periodically measures the level of interference of all operating frequencies. It then chooses the frequency channel with the lowest interference as its operating frequency. In the second method, Weighted Reserve Autonomous Frequency Allocation (WRAFA), special frequency groups are reserved for use in regions congested with other BSs. The BSs select their operating frequency in the same manner as that of the first algorithm. The reserve frequency groups are restricted in their usage by weighing clown the interference with a factor ?. In the third technique, Dual Selection Autonomous Frequency Allocation (DSAFA), the non-uniform nature of the cell coverage region is exploited. Each BS is allocated with 2 sets of frequencies based on the interference level measured. In general, the region covered by each set of frequency is unique. Thus, using 2 sets of frequency enlarges the area covered by the cell. This algorithm further enhances the SIR and GoS performance of the roving BSs. In terms of SIR performance, BAFA is the worst performer. Despite that, it is still capable of achieving a 7-dB improvement over the situation without reallocation. The WRAFA can achieve higher SIR, albeit with compromises in the GoS. The WRAFA also outperforms the other 2 algorithms in terms of GoS when traffic demand is low. The WRAFA's poor performance at high traffic intensity is attributed to allocated resources. At higher traffic intensity, the DSAFA is the best performer since it is capable of covering a bigger area with the same amount of resources.