Next generation mobile communication systems are expected to provide a wide range of multimedia services with variable QoS (Quality of Service). The major issues of the next generation systems are performance enhancement and efficient resource managem...
Next generation mobile communication systems are expected to provide a wide range of multimedia services with variable QoS (Quality of Service). The major issues of the next generation systems are performance enhancement and efficient resource management.
3GPP (3rd Generation Partnership Project) is standardizing LTE (Long Term Evolution), which is a framework for the evolution of the 3GPP radio access technology such as HSDPA/HSUPA (High Speed Downlink/Uplink Packet Access). The target of LTE is to provide significantly higher data rates and the radio transmission is based on OFDMA (Orthogonal Frequency Division Multiple Access) for high spectrum efficiency and robust performance. The OFDMA divides a broadband wireless channel into a number of subcarriers that are transmitted simultaneously. Hence, the OFDMA has a potential for efficient handling of broadband radio resources since subcarriers can be dynamically assigned to different users.
In order to employ the OFDMA to next generation systems, the wireless system requires the channel state information of entire system bandwidth to allocate downlink resources in time varying broadband channel. However, the OFDMA/FDD (Frequency Division Duplex) system needs much signaling overheads for mobile stations to feedback the information to the packet scheduler which is located in the base station. Therefore, an efficient CQI (Channel Quality Information) report mechanism should be considered to reduce the extensive signaling overhead in a fast moving environment. In addition, the performance of a mobile system heavily depends upon packet scheduling and the frequency domain packet scheduling is needed to increase system throughput by exploiting multi-user diversity and frequency-selectivity. The allocation of subcarriers should consider not only throughput performance but also fair allocation among mobile stations. Also, the complexity of scheduling algorithm should be lessened to manage a large number of mobile stations since the packet scheduler takes into account much information, such as wireless channel state, buffer status, and QoS requirement.
This paper proposes the Multi-Level feedback scheme for the efficient report of channel state information. Furthermore, the paper proposes the OPF (OFDMA Proportional Fair) packet scheduling for the sub-optimal resource allocation scheme which guarantees proportional fairness with reduced complexity.
Proposed Multi-Level feedback scheme continuously measures downlink channel state and averages over multiple time slots for reporting. Each channel state per subband is quantized into a discrete value LMCS which designates the possible MCS (Modulation and Coding Scheme) level. Then, the scheme merges the channel state information of adjacent subbands at each reporting time to reduce the size for reporting. The procedure is divided into the frequency domain merging and the Multi-Level mapping. The frequency domain merging compares the channel quality of adjacent subbands and merges it into one MCS level when the difference between adjacent subbands is lower than a certain threshold. The Multi-Level mapping transforms channel state information into indicators which is composed of 2bit per subband. Then, the channel information is reported to the base station for packet scheduling.
Proposed OPF packet scheduling algorithm is based on the conventional Proportional Fair scheduling and the paper extends the scheme for throughput performance and proportional fairness on OFDMA systems. Firstly, the paper redefines the instantaneous channel state into which is the sum of the channel state of selected subbands. Secondly, a weighted variable which represents the delay of incoming data is added for QoS. Then, the algorithm selects mobiles according to the priority function and resource allocation is done by using the ratio of the proportional fairness. In addition, the moving average per user is calculated once per scheduling period to update the average value Rk (t).
For comparison, a system-level simulation approach is used and it assumes an OFDMA/FDD system which has 10 MHz channel bandwidth. The channel model is based on SCME (Spatial Channel Model Extended). The simulation result shows that the proposed Multi-Level feedback scheme reduces signaling overheads up to 73% compared to the full CQI feedback scheme. In addition, the scheme gives the best throughput performance among partial CQI feedback schemes. Also, the proposed OPF packet scheduling follows proportional fair throughput with reduced complexity which is similar to the complexity of the single carrier packet scheduling algorithm. Furthermore, the algorithm increases system performance by efficiently utilizing the frequency diversity.