Investigation into the sensitivity of the power predictions of a microcellular ray tracing propagation model

This paper investigates the sensitivity of the three-dimensional (3-D) rag tracing microcellular model presented in [1], [2]. The variation of the rec eived power is examined for different rag permutations, wall characteristic s, antenna position offsets and database inaccuracies, Predictions of the d ifferent configurations in line-of-sight (LOS), non-LOS (NLOS), and deep sh adow areas are compared with each other and also with narrowband measuremen ts, The analysis illustrates that although the model produces reliable resu lts with five orders of reflection and one order of diffraction, higher ord ers of reflection and double diffracted rays enhance the model's performanc e in deep shadow areas, It is also shown that good agreement with measured results can be obtained for wall conductivity in the order of 10(-3) S/m an d values of relative permitivity around five, The sensitivity analysis to t he antenna positioning and database inaccuracies indicates that the receive r positions which suffer the highest power deviations are those at the boun daries of the LOS areas, as wed as the positions in the deep shadow regions . In general, for antenna offsets up to 1 m, the predictions of the model a re not significantly affected. Finally, the building databases with Im maxi mum displacement do not have severe effects on the predictions, but databas es with less accuracy can seriously degrade the performance of the model.