PAPR aware power allocation in OFDMA uplink

This thesis investigates the power allocation scheme and essential design constraints to be considered in multicarrier systems particularly in the case of orthogonal frequency division multiple access (OFDMA) system in multiuser (MU) scenario. The compatibility between multicarrier system and multiple input multiple output (MIMO) system is exploited in designing the power allocation algorithm for a cellular network with multiusers. The multicarrier MIMO system facilitates dynamic resource allocation due to the decomposition of physical resources into multiple domains. The energy efficiency and interference management are the crucial aspects especially in uplink (UL) transmission. Limiting the power consumption of mobile terminals (MT) in uplink (UL) is inevitable due to the limited amount of available energy. Furthermore, the traditional multicarrier system introduces a dynamic peak power variation with respect to average power causing erroneous circuit behavior. This phenomenon is usually quantified as peak to average power ratio (PAPR). High PAPR drives the high power amplifier (HPA) into non-linear region to result in significant degradation in the system performance in terms of power efficiency. In this thesis an iterative power allocation algorithm is proposed to minimize the sum power and PAPR. This thesis presents the power allocation strategy such that the PAPR is controlled during the power allocation (minimization) stage in frequency domain. The optimal power allocation is achieved by joint optimization of transmit power and receive beamformers (TX-RX) using convex optimization technique. The original problem is not jointly convex with respect to TX-RX. Therefore an iterative algorithm is proposed to optimize TX and RX alternatingly such that by calculating TX for given fixed set of RX and vice versa until convergence. The statistical approach is adopted to reduce the PAPR by actually minimizing the signal power variance (SPV) due to the fact that the large number of independent and identically distributed complex OFDMA symbols tends to follow Gaussian probability density function characterized by certain mean and variance. The non-convex constraints in the formulation are transformed into convex form using the successive convex approximation (SCA) with required change of variable (COV). The algorithm guarantees to maintain the user-specific quality of service (QoS) defined by the rate constraint. Hence, equipped with the potentials of future generation technologies and using convex optimization as a tool, this thesis offers a sum power and PAPR minimization scheme for MU SIMO-OFDMA UL transmission