The future success of personal communication services depends on its ability to provide transmission services with high transmission rate and low transmit power. Serial interference cancellation (SIC) technique is among the more promising approaches to increase the system capacity in many interference limited wireless communication systems. For systems employing SIC, the traditional power control strategy is no longer optimal. Assuming that limited feedback information is available at the transmitter, we study, in this thesis, the joint optimization of power allocation, SIC decoding order and bit allocation for systems employing SIC. In particular, we study two important SIC-based wireless communication systems, the multirate CDMA system with SIC and the multiple antenna system using V-...[ Read more ]

The future success of personal communication services depends on its ability to provide transmission services with high transmission rate and low transmit power. Serial interference cancellation (SIC) technique is among the more promising approaches to increase the system capacity in many interference limited wireless communication systems. For systems employing SIC, the traditional power control strategy is no longer optimal. Assuming that limited feedback information is available at the transmitter, we study, in this thesis, the joint optimization of power allocation, SIC decoding order and bit allocation for systems employing SIC. In particular, we study two important SIC-based wireless communication systems, the multirate CDMA system with SIC and the multiple antenna system using V-BLAST algorithm.

The two main standards, W-CDMA and cdma2000, of the 3^rd Generation wireless communication systems are both multirate CDMA system and SIC is likely to be employed at the base station receiver to improve system capacity. Different services with different data rates and quality of service (QoS) requirements have to be supported. We formulate the joint power allocation and the SIC decoding order problem as a constrained optimization problem with the objective of minimizing the total transmission power while every user meets the data rate and QoS requirements. The optimal power control scheme with the optimal SIC decoding order is derived. We find that ranking users in the descending order according to their channel gains gives the optimal SIC decoding order, regardless of the users' data rate and QoS requirements. This finding shows that the commonly suggested SIC orders, based on either the received data rate or the received power, are not optimal in multirate CDMA systems.

Systems with multiple transmit and multiple receive antennas, also known as multiple input multiple output (MIMO) systems, can achieve very high spectral efficiency using a recently proposed V-BLAST architecture. V-BLAST is a SIC-based algorithm with the optimal SIC decoding order, optimal assuming equal transmit power and data rate in transmitting antennas. However, the performance of V-BLAST will be degraded when some of the decorrelated channels are in deep fade. Allowing limited feedback information to the transmitter can help to overcome this problem. We optimize the bit-only, power-only, as well as the bit-and-power allocation schemes for the MIMO systems with V-BLAST algorithm. Since the SIC ordering in V-BLAST is no longer the optimal decoding order when the data rate and the power are not the same for all transmitting antennas, we further optimize the SIC decoding order together with the bit-only and the bit-and-power allocation schemes. Because of the high complexity in solving for the optimal ordering, sub-optimal SIC decoding order are proposed to reduce the complexity significantly with the performances approaching the optimal ordering.