Energy-Efficient Power Allocation for HARQ Systems

Introduction

There is a rapid increase in the number of devices communicating using different wireless technologies. It is predicted that there will be about 50 billion connected devices by the year 2020. To support this rapid growth, ideally more radio resources are needed. However, the radio resources like frequency spectrum and transmit power available for wireless communication are limited like any other natural resource. Hence it is becoming increasingly important to efficiently utilize the available radio resources. Another important aspect of communication over wireless networks is the reliability of the data. Often, reliability over wireless networks is affected by fading, which causes the signal strength to vary at the receiver and this results in a loss of data packets.

One way to overcome the reliability problem in wireless systems is by employing adaptive transmission techniques. These techniques will also be a fundamental feature for future generation wireless systems. An important factor in facilitating the adaptive transmission schemes is the channel state information (CSI) feedback from the receiver to the transmitter. In the presence of CSI at the transmitter, a common adaptive transmission technique used at the physical layer is dynamic link adaptation. In link adaptation, the transmitter dynamically adapts the modulation order, coding rate and/or other signal transmission parameters (Goldsmith, 2007; Goldsmith, & Chua, 1998). Commonly used link adaptation strategy in practical wireless systems is adaptive modulation and coding (AMC). The basic idea of AMC is to exploit the good channel conditions by transmitting with higher order modulation and coding scheme (MCS), and if the channel quality degrades, the transmitter switches to a lower MCS level to maintain a fixed target error probability. By selecting the best MCS level for a given channel quality, the link adaptation results in higher average throughput performance compared to the non-adaptive systems (Catreux, Erceg, Gesbert, & Heath, 2002). However the errors in the channel estimation, quantization of channel state information, and non-ideal feedback channel lead to degradation in performance even in the presence of link adaptation. To further improve the robustness of the wireless systems against the link adaptation inaccuracies, hybrid automatic repeat request (HARQ) schemes are adopted at the medium access control (MAC) layer in systems like LTE and LTE-advanced (Dalhman, Parkvall, & Sköld, 2011). Moreover, for the emerging small-cell systems, HARQ schemes can be seen as a fast retransmission mechanism to improve reliability without causing significant delays.

In HARQ schemes, forward-error correction (FEC) schemes are combined with automatic repeat request (ARQ) schemes (Lin, & Costello, 2004). If a data packet is received in error, the transmitter re-sends the information about the erroneous packet to the receiver. Upon receiving additional information about the erroneous packet, the receiver makes another try to decode the erroneous packet, and it may or may not combine the information received across different transmissions of the same data packet. HARQ systems use acknowledgment (ACK) and negative acknowledgment (NACK) signal feedback to determine whether retransmission of a data packet is required or not. Although retransmission schemes consume additional radio resources in terms of channel uses and transmission power, they are necessary to improve the transmission reliability in fading channels. There are many different HARQ schemes proposed in the literature, and they are mainly classified into two types (Lin, Costello, & Miller, 1984). In type-I HARQ schemes, the receiver does not combine the information across different (re)transmissions. By way of contrast, in type-II HARQ schemes, the receiver combines the information from different (re)transmissions to decode the packet. These type-II HARQ schemes are further classified into: