Cross Layer Optimization in Wireless Downlink System

• Main Author: Tran, Sang Van
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01-01-2012
• Subjects: Electrical engineering
• ISBN: 9781267256591; 1267256591

To improve the robustness and reliability of wireless transmission, two complementary link adaptation techniques are employed: AMC at the PHY layer and HARQ at the MAC layer. OFDMA and MIMO are also introduced at the PHY layer to provide flexibility in allocating resources and to increase system throughput. These advanced signal processing concepts are now integral components of 3GPP LTE standard. MIMO-OFDMA system exploits the scheduling concept in frequency domain (FD), time domain (TD), and spatial domain and selects the most efficient modulation and coding with spatial multiplexing or spatial diversity depending on the time-varying wireless channel condition to increase system capacity or improve system reliability. In OFDMA systems, wireless resources are divided into resource blocks (RBs), and scheduling these RBs in an optimized manner to increase overall system spectral efficiency (SE) and satisfy all active users QoS requirements is a main thrust in the design of LTE system. This work provides a holistic approach of cross layer optimization with the intent of maximizing SE by combining AMC, HARQ, OFDMA, and MIMO concepts in scheduling resources for an LTE downlink (DL) system. It formulates close-formed equations for calculating average SE for different MIMO configurations characterized by Rayleigh fading channel. A new online algorithm is developed to optimize SE for both Rayleigh and non-Rayleigh fading channel. With the developed algorithm, it is possible to determine how many retransmissions are required in addition to the initial transmission in advance depending on the current wireless channel condition. Latency can be reduced up to 24 ms when sending the initial transmission and all of its retransmissions sooner than waiting for retransmission requests as performed previously. An optimized scheduling algorithm of the LTE DL system taking into account the combination of frequency-time packet scheduling (FTPS) with added spatial dimension or MIMO and QoS awareness is developed. Simulations using proven LTE models are performed to compare SE obtained from closed-form equations and the developed algorithm for different system configurations. Results demonstrate that the proposed scheduling algorithm including both FTPS and spatial domain outperforms the FTPS only scheduling algorithm in overall system capacity.