A Study Of Energy-Efficient Technologies In Wireless Communication Systems

The development of the information society suggests greater system capacity and higher data rate for wireless communication systems. On the one hand, in order to meet these requirements, wireless communication systems have to increase the number of wireless base stations and com-puter rooms, which leads to a significant increase in energy consumption and exacerbates the problem of spectrum scarcity. On the other hand, since the traffic load is fluctuating, base stations are not full load all the time and the system has a larger energy saving space. To study energy saving and spectrum utilization in wireless communication systems has a very practical significance. Cognitive radio (CR) technology is designed to enhance the spectrum efficiency, which means study energy-saving technologies in CR systems can better solve the high energy consumption and low spectral efficiency problems in wireless communication systemThis thesis summarizes the existing energy-saving technologies in wireless communication system and discusses the importance of energy saving in cognitive OFDM systems. This thesis proposes two energy-efficient bit allocation algorithms for orthogonal frequency division multiplexing (OFDM)-based CR systems. Both algorithms are designed to maximize energy-efficiency (EE) of CR systems, subject to the maximal-power and sum-interference constraints. Therein, Continuous-Rounding algorithm achieves optimal power allocation via sub-gradient iterations, and then round the real number of bits downwards to the nearest integer to meet the discrete-bit requirements. Increment Minimization per Bit algorithm allocates the bits to different subcarriers one by one to maximize EE, until the total power is exhausted or the interference constraint is violated. Simulation results show that both algorithms obtain the favorable performance, which is much close to optimal algorithm, with a lower complexity.This thesis further investigates the energy-efficient power allocation for OFDM based relay-aided cognitive radio networks (CRNs) with im-perfect spectrum sensing considered in the system model. The object is to maximize the energy efficiency measured by the "Throughput per Joule" metric subject to the total transmit power and interference constraints. To solve the problem efficiently, the primal problem is transformed into a convex optimization problem using the fractional programming method. Optimal and suboptimal algorithms are proposed, and simulation results reveal that the proposed algorithms can obtain a good tradeoff between energy and capacity. Meanwhile, the suboptimal algorithm can reduce the computational complexity hugely with a marginal performance loss compared with the optimal algorithm.