Abstract
The crucial areas for future improvements by network operators are the Quality of Service (QoS) and network capacity. In order to meet the capacity demands, there is a tendency to shift the operating frequency towards higher frequency bands. There is also a drive to move from a macrocell type network to a microcell type network. As most neighbourhoods have some sort of vegetation that can expand over the years, the previously insignificant shadowing effects will become more pronounced. In addition to that, the movement of vegetation structure will introduce an additional adversity for high frequency radiowave propagation. Two series of measurement campaigns, controlled environment: for signal operating at 0.9, 2, 12 and 17GHz, and outdoor environment: for signal operating at 1.8GHz had been carried out as the main components of this research work to investigate this impact. In the controlled environment measurement campaign, the radiowaves propagation experiments were conducted in a laboratory environment of an anechoic chamber, where a controllable wind generator was implemented to simulate the wind influence. Two vegetation samples of different sizes and shapes were used and placed in between the transmitter and receiver antennas inside the chamber. On the other hand, the outdoor measurements were conducted in real environments with the influence of natural wind conditions. The transmitted signals from existing base stations were measured using a scanning receiver to capture signal variations and deep fades. Three experimental sites were identified which included single-tree, line-of-trees and group-of-trees configurations. The impact of vegetation movement under different wind influences was analysed in terms of the first- and second-order statistics of the received signals. The results indicated similar wind dependency behaviour of the signal for both measurement campaigns and the fading amplitude can be represented by a Rician distribution. It was also observed that the received signal variations tend to increase as wind speed increases from calm to windy conditions. Once in the windy state, it has been found that any further increment of the wind speed has less of an effect on the received signal. Apart from analysing the signal over different wind speeds, the work also compared the impact at different frequencies, sizes of vegetation and vegetation structures. Based on the k-factor values derived from the results of controlled environments, it was noted that higher frequency signals and larger vegetation size would result in lower k-factor, thus indicated higher random multipath contributions. This is particularly obvious during the transition period from calm to windy condition. In addition, the comparison over different vegetation structures of outdoor experiments indicated that the changing rate of fast-fading distribution was related to the flexibility and density of the vegetation. Using the results of the outdoor measurement campaign, an empirical model of this dynamic behaviour has been proposed for use in future planning of more robust wireless systems providing services in vegetated areas. The model was developed to predict k-factor values lognormally distributed around an estimated exponential curve. The constants incorporated in the model were determined from outdoor measurements data of Sitel and 2. Finally, the predicted k-factor values generated from the newly developed model were assessed against the measured data of Site 3, which showed a good agreement between them, thus confirms its validity and accuracy.