Article Preview
Top1. Introduction
The presence of a Line of Sight (LoS) component in the fading channel makes the study of Rician fading very important. LoS or near LoS links are possible in many small cell deployments of 3GPP-LTEA systems, as well as future millimeter wave links. As is well known, the k-factor in Rician fading generalizes the Rayleigh (K=0) and AWGN (K=infinity) channels. Considerable amount of work has been done on the effect of Rician fading in various scenarios. For example, Wu et al. (2010) analyses the performance of Optimum Combining Receivers under Rician fading conditions. Nandi et al. (2010) talk about optimal transmit powers in wireless sensor networks under Rician fading conditions, while Jayaweera et al. (2003) propose an optimal signaling scheme for improvement in capacity in multiple antenna scenarios undergoing Rician fading, where the knowledge of k-factor is available at the transmitter. Reducing transmit power is an important requirement for mobile terminals on the uplink. There are many works which talk about methods to reduce the transmit power. For example, Bavanianet al. (2009) propose a power control strategy based on the total received power in the uplink of cooperative base stations. In our work, we consider a multipoint to point link in a typical uplink scenario, where transmit power is the main constraint. Hence, we propose a power control scheme with limited feedback with the aim of minimizing the transmitted power, while maintaining the same rate. Since both the streams are of interest, we look at a joint detection receiver (Grant and Cavers, 2009) and calculate the sum rate for the transmitted streams.
The motivation for this work is as follows. Consider two transmitters transmitting to a single receiver at the same time and frequency as shown in Figure 1. Assume both the streams are QPSK modulated and are received at different powers P1and P2 at the receiver, based on the path loss encountered. To understand how the sum rate varies with the received power of the two streams, let us initially consider the case where the first stream is received at -90dBm and the noise floor is at -100dBm. The power of the second stream is varied from -100dBm to -80dBm and the variation of the sum rate is plotted. It can be seen from Figure 2 that in case of Rayleigh fading, the sum rate increases monotonically with increase in the power of the second stream. However, in the case of Rician fading, the sum rate drops as the power of the second stream approaches the power of the first stream and then rises again. This is because in the equal power region, the constellation points overlap and the 
approaches zero and the constellation points are no more uniquely decodable.