No abstract

In this paper, we provide an analysis of CDMA range extension due to soft handoff. A simulation of GSM hard handoff and CDMA soft handoff is presented. In these simulations, a collection of 90 second calls are tracked through the system and the margin required to meet a constant call reliability is determined. A comparison of these margins for both GSM and CDMA indicate that, under variety of conditions, the shadow fade margin required by the GSM system is about 2.t3.6 dB higher than that needed by the CDMA soft handoff. I. INTRODUC~IONIn order to reduce the number of handoffs per call ("ping-pong" effect), hysteresis is usually introduced in hard handoff algorithms. This hysteresis, typically in the range of 5-8 dB, causes an overlap between the cell coverage areas. For a two cell geometry, the coverage overlap may be computed as follows: Ignoring shadow fades, and assuming an 8 dB hysteresis with a path loss exponent of 4, the overlap can be determined by equating the signal strengths received from the two cell sites at the point of handoff,where r denotes the normalized distance from the serving site. This results in r = 1.23, i.e. 23% overlap in range. For the mobile to be able to communicate at this range, an extra margin of about 40 log(1.23)=3.5 dB is required. With the addition of shadow fading, one must take into account the statistical nature of handoff procedures.A recent analysis presented in [ 11 compares the shadow fade margin required for isolated cells with that of soft handoff cellular systems. In this paper, we extend the results of [ 11 to compute the margin required for the cellular systems with Mobile Assisted Handoff (MAHO). In [l], it is shown that the shadow fade margin required by an isolated cell to meet 90% coverage reliability at the range of r = 1.25 is 14.2 dB. A similar calculation for soft handoff, for which the best cell is chosen as the server, leads to a required margin of 6.2 dB at the cell edge ( r = 1). Hence, the range advantage of soft handoff is 8 dB over an isolated cell with 25% range overlap. Although, this analysis may be sufficient for comparing CDMA with A M P S , it is not representative of digital cellular systems with MAHO such as GSM.In this paper, first, we provide a simple two cell analysis which determines a rough estimate of the differences between the soft handoff and MAHO driven hard handoff schemes. Next, performance of IS-95 CDMA and GSM systems are evaluated via simulations. Results of a comparative study of the two handoff methods and the sensitivity of call reliability to parameters such as handoff hysteresis, shadow fading, and path loss are presented. II. ANALYSIS OF TWO-CELL GEOMETRYIn order to gain some insight into the handoff mechanisms, this section presents a simple analysis of a twe-cell geometry in a one-dimensional space. Following the notation in [l], we assume the following propagation parameters: path loss exponent p = 4, and lognormal shadow fading standard deviation G = 8 dB. Also, it is assumed that the lognormal variables are correlat...