International Journal of Wireless Information Networks, Vol. 4, No. 2, 1997 1068-9605 / 97 / 0400-0113$12.50 / 0 Ó 1997 Plenum Publishing Corporation 113 Effect of Propagation Modeling on LOS Microcellular System Design Dongsoo Har, 1 Henry L. Bertoni, 1,3 and Seongcheol Kim 1,2 Linear microcells employing low base station antennas have been proposed for emerging PCS systems. In such systems, the characteristics of the propagation on line-of-sight (LOS) radio links will fundamentally determine the system performance. In this context, it is meaningful to discuss the effect on system design of the two commonly used propagation models, the two-slope model and the single-slope model. In this paper we present several methods to compare the frequency reuse pattern and base station separation based on these two propagation models, and verify the importance of correct propagation modeling for ef®cient system design. We consider interfering signals coming from two cochannel base stations in the ®rst tier, and include slow fading loss when comparing system performances. Based on system design results and subsequent performance, the two-slope model is recommended for LOS PCS microcells. KEY WORDS: Linear microcells; reuse factor; slow fading; two-slope model; quality of service. 1. INTRODUCTION Because the buildings con®ne the radiation from base station antennas, cell coverage in urban environ- ments, as well as coverage along a highway, can be described in terms of linear cells. To assist the design of linear cellular systems, several measurements [1±4] and performance predictions [5, 6] have been reported. Two line-of-sight (LOS) models have been identi®ed in the literature, the simplest of which can be obtained by a simple single-regression ®t (single-slope model) to the measured signal in dB as a function of distance on a logarithmic scale [6, 7]. On the other hand, it has been shown that the use of a two-segment regression ®t can achieve a more accurate representation having smaller slow fading ¯uctuation [1, 8]. The two-slope model, 1 Center for Advanced Technology in Telecommunications, Polytechnic University, Brooklyn, New York 11201. 2 Present address: AT&T Bell Laboratories, Holmdel, New Jersey 07733. 3 Correspondence should be directed to Henry L. Bertoni, Center for Advanced Technology in Telecommunications, Polytechnic Univer- sity, 6 MetroTech Center, Brooklyn, New York, 11201. which is the two-segment ®t together with the standard deviation of the measurements to the ®t, can be char- acterized by the presence of the breakpoint of the slope index. We examine in detail the difference in the cell layouts obtained using these two models in order for the systems to meet required quality of service (QOS). The study reported here is limited to the downlink, so that power control need not be considered, and to sys- tems with a speci®ed 17 dB C / I protection ratio, such as AMPS and IS-54. We also examine the effect on the design of considering the two nearest interferers, as opposed to considering a single interfering signal. These approaches are considered to achieve a design that has the minimum frequency reuse factor NR, and hence the minimum density of base stations to handle the antici- pated traf®c with a given QOS. Depending on the reuse factor NR, the cell size can be made larger and it is assumed that power can be increased, so that the system is interference-limited rather than noise-limited. The ®rst method is an analytic approach to achieve a given bumping probability at the cell boundary. The second method uses numerical integration to limit the