Energy Efficient Adaptive Multi-Antenna Systems

As the continuous development of the mobile Internet, explosive growth of new services and mobile applications, the mobile users require higher and higher data rates. Through deploying multiple antennas at the transmitters and receivers, multi-antenna systems can highly increase the spectral efficiency (SE) and transmission reliability through employing the multiplexing/diversity gain. Therefore, multi-antenna technology has become one of the key technologies for the future mobile communication systems. The research on the conventional multi-antenna technology mainly focuss on the spectral efficiency and tries to derive the achievable capacity from the information theory point of view. However, the high energy consumption is becoming a serious problem for the mobile networks and green mobile networks have drawn increasing attentions recently. Thus, how to increase the energy efficiency (EE) of the multi-antenna systems is significant in practice. From the standpoint of SE, employing higher transmit power, more transmit antennas and higher bandwidth can increase the SE. However, it is not the case from the standpoint of EE. Due to the existence of power consumed by circuit, signal processing and feeder etc. except for the transmit power, increasing the transmit power, transmit antenna number and bandwidth cannot always increase the EE and would even decrease it. Therefore, a big challenge for the future green mobile networks is how to increase the EE with guaranteed SE. This thesis focuses on the EE of the downlink multi-antenna systems and studies the EE of single user (SU) multiple input multiple output (MIMO), multiuser (MU) MIMO and multi-cell MIMO from a system level. Through EE based adaptive transmission technologies, the EE can be improved significantly with some SE constraints.Firstly, we study the EE maximization problem for the MIMO broadcasting channels under dirty paper coding (DPC). We propose a new optimization framework, in which transmit covariance optimization and active transmit antenna selection (ATAS) are designed jointly. This framework can reach the EE upper bound and point out the directions of energy efficient multiuser MIMO transmission. To optimize the EE under a fixed transmit antenna set, we propose an energy efficient iterative waterfilling scheme, through transforming the problem into a concave fractional optimization via uplink-downlink duality. It is proved that the proposed scheme converges to the global optimality. After that, ATAS is employed to determine the active transmit antenna set and to turn off the rest inactive antennas to save the circuit power.After that, this thesis addresses EE optimization for the downlink MIMO systems with linear precoding and imperfect channel state information at the transmitter (CSIT), and proposes multimode transmission technology to solve this problem. Due to the existence of imperfect CSIT, there exists a tradeoff between the interuser interference and the multiplexing gain, so two transmission schemes including SU-MIMO with singular value decomposition (SVD) and MU-MIMO with block diagonalization (BD) are considered here. Considering the effect of transmit power, bandwidth, transmission schemes, transmit antennas and receive antennas in a comprehensive manner, we propose joint bandwidth/power adaptation and mode switching to improve the EE for the following two scenarios, which is valuable in practice.1) Over the scenario when the transmit antenna number is larger than or equal to the total receive antenna number, the mode is defined as the following parameters, i.e. transmission schemes (SVD or BD), transmitter antenna number, receiver antenna number and user number. Under a fixed mode, we develop a joint transmit power and bandwidth adaptation scheme for both SU-MIMO and MU-MIMO, and employ capacity prediction schemes to combat the effect of the imperfect CSIT. After that, we propose a mode selection scheme based on the ergodic capacity, which provides guidelines on the preferred mode under different scenarios.2) Over the scenario when the transmit antenna number is smaller than the total receive antenna number, the mode is defined as the following parameters, i.e. transmission schemes (SVD or BD), active transmitter antenna set, active receiver antenna set and active user set. Under a fixed mode, we extend the above joint transmit power and bandwidth adaptation scheme in scenario1) to the case with maximum transmit power constraints. After that, we propose a low complexity joint active transmit and receive antenna selection scheme, which can achieve performance close to the optimal exhaust search based on the simulation results. Finally, we apply the algorithms in non-real time sessions and we can see the energy is saved significantly through employing the delay tolerance based on the energy-delay tradeoff.Finally, we address the EE optimization problem for the multi-cell MISO systems, and introduce a novel EE metric of network EE (NEE) for the multi-cell systems. We propose three transmission schemes with different cooperative levels and further propose a cooperative idling (CI) scheme to improve the NEE with each user’s SE constraints. Based on the NEE metric, we develop energy efficient power control strategies for three schemes with linear precoders. After that, CI is proposed. Through employing the micro-sleep cooperatively among different base stations and switching off the power amplifiers (PA) cooperatively, the NEE can be further improved.