Hierarchical wideband multicarrier CDMA systems: Reverse link analysis

Abstract

The reverse link performance of hierarchical wideband multicarrier code-division multiple-access (MC-CDMA) systems is studied. An MC-CDMA system divides the system bandwidth into several equal narrow subbands that are used to transmit multiple signal waveforms in parallel. MC-CDMA systems are known to be robust to multipath fading and narrowband interference. We propose a hierarchical MC-CDMA system where the microcell(s) occupies a fraction of the available subbands. It is shown that such an architecture is a viable way of supporting microcell traffic while protecting the existing macrocell traffic. The effects of microcell Nakagami fading and power control error are also studied.Index Terms-Carrier-to-interference ratio (CIR), frequency-selective fading, hierarchical architectures, macrocell, microcell, multicarrier code-division multiple-access (MC-CDMA), wideband.

BER performance of MC-CDMA systems with MMSE in Nakagami-m fading

“…Both of the perfect and imperfect power control 400 quasisynchronous uplink MC-CDMA systems with MMSE receiver are considered. For perfect power control systems, the received power of all the users should be equal to the desired signal power P of the BS, i.e., a; = P, k = 1, 2,oo., K. For imperfect power control systems, since it has been experimentally verified that the imperfections in power control can be modeled as a log-normal random variable [17], [18], the received power of user k should be replaced by a; = P1Os,/ 1 O , where t; k is a Gaussian random variable with mean 0 and standard deviation 0' E for every MS k and represents the power control error, measured in dB [18]. Fig.2 shows the influence of power control imperfections on the uplink average BER for the quasisynchronous uplink MC-CDMA systems with MMSE receiver.…”

“…It's easy to show that xn is gamma distributed RV with parameters m and On' Then the characteristic function (CF) of xn and y can be respectively given by (fix" (s)=(1-s0)mr m (16) and q}y (s)= rr:)I-sO)mr m (17) By changing the variables in (15) and using the alternative representation of the Q-function [15], [16] Q(x)= � f "/ 2 exp ( -� 2 2 t e (x? :O) (18) 7r 0 2sm e r the average BER can be derived as Similarly to [8] and [13], when using (19) to evaluate the average BER, it is still necessary to take the "semi-analytic" approach whereby a sample average of (19) is taken where each evaluation is performed by generating a random selection of the fading amplitudes {fJ k,n } of all the K -1 interfering users and then numerically evaluates the integral in (19).…”