Studies on QoS improvement for OFDM based wireless networks

Abstract

The success of 3G and next generation wireless technologies depend on the users? satisfaction and affordability to different types of services such as interactive communications, web surfing entertainments etc. Quality of Service (QoS) determines the success of wireless technology to deliver a variety of services like voice and video. The chief detractors for QoS are latency, jitter and packet loss. Implementation of QoS in the wireless networks has a great role in order to meet the delay and packet loss constraints for real time service in spite of varying nature of the wireless link. All the technologies beyond 3G (B3G) like WiMAX, WLAN HSDPA, LTE use OFDM as the modulation technique as they are designed to support multimedia services. These technologies are planned to support peak downlink (DL) data rates up to 128 Mbps and peak uplink (UL) data rates up to 56 Mbps. To experience high data rates with best Quality of Service (QoS), Fast Link Adaptation algorithms and Fair Resource Allocation algorithms are identified as key technologies. Adaptation in physical layer parameters like Modulation Order, Code Rate and Power according to the prevailing channel conditions has been identified as the key solution to reduce the packet loss and to increase the throughput of the system. This research work proposes a modification to the existing Exponential Effective SNR Mapping (EESM) Link Adaptation algorithm to support Multimedia Transmission. The algorithm adopts grouping of subcarriers within an OFDM symbol based on their channel quality. Packet Error Rate (PER) prediction is performed for the two groups separately and appropriate Modulation and Code Rate (MCS) are selected to achieve the desire target. The proposed algorithm reduces the packet loss upto 20% when compared to the conventional EESM algorithm without grouping. Up to 1Mbps throughput enhancement has been demonstrated in poor channel conditions and 50% feedback overhead reduction in the control channel has been proved by the algorithm.