Sensitivity of handoff algorithms to variations in the propagation environment

The authors have developed a model (1992) for analyzing the performance of handoff algorithms based on relative signal strength measurements in a lognormal fading environment. Handoff decisions are made by averaging the received signal level over an interval d_av and using a hysteresis margin h (that is the mobile is handed off to a new base station when the averaged signal level from that base exceeds that from the current base by h dB). When d_av and h are increased, the mean number of handoffs is reduced, but the delay in handing off increases causing a reduction in link quality. Hence, the parameters d_av and h need to be chosen to optimize the tradeoff between these two criteria. In this paper, the model is exercised to obtain certain criteria for designing practical handoffs algorithms, in particular, for designing algorithms that are robust with respect to variations in the radio propagation environment. The lognormal fading process is characterized by its standard deviation σ_s and its correlation distance d_o. The sensitivities to variations in σ_s and in d_o and also to jumps in received signal strength (as with the street corner effect in microcells) are analyzed. These sensitivities indicate how to choose (d_av, h) pairs for robust performance in the presence of the variations