Abstract

This thesis investigates indoor propagation by measurement and simulation using ray-tracing. An automated site survey measurement system is described. This site survey measurement system is used for measurements in two environments, a hallway and a microwave lab. Measurements were taken along a path at 1.5 cm intervals and at each point 30 time samples were taken, at 2 second intervals. Both the temporal and spatial variation of the signal strength are represented with Rician probability distribution functions. The Rician-K factor is computed by three methods found in the literature, called in this thesis the K G /K 24 method, the K 12 method and the K MLE method. The value of the Rician-K factor for the temporal data is investigated. Since the measured data revealed little time variation, high Rician-K factor values were found. The spatial variation of the received signal is investigated by taking the average of the temporal data at each distance point. The spatial variation of the Rician-K factor is investigated. The measured data showed rapid variation with distance, and thus small values of Rician-K factor were found. The two measurement environments are modeled for analysis by ray-tracing. The results obtained from the simulations are compared with the measured results. The comparison is based on four factors obtained from measured and simulated data: small-scale fading, path loss index, large-scale fading and space-varying Rician K factor. The measured and simulated small-scale fading data matched poorly in the hallway. This is attributed to the wall model in the ray-tracing simulation, which is a poor representation of the actual wall construction. For the microwave lab, a more detailed representation and the location of the metal file cabinet and equipment gave good agreement between measured and simulated small-scale-fading. The path loss index indicates good agreement in the hallway and in the microwave lab between the large scale fading of the measurements and the simulations. Of the three methods for calculating the Rician-K factor, K G /K 24 is found to be the best for the large Rician-K factor values determined for time variations. However, the K MLE method is the best to find the much smaller Rician-K factor values in space variations.