Cross-Layer Design Of Dynamic Resource Allocation For MIMO-OFDM System

1. InstructionThe fallibility of wireless communication is mainly made by time-varying and multi-path of wireless fading channel. For wireless channel has randomness, frequency selectivity and time-varying characters, we need to do real-time monitoring to channel state in order to effectively use wireless resources. Allocating information bits and power adaptively to users based on their channel gains is one of advantages for separating a real channel into several subchannels in Orthogonal Frequency Division Multiplexing (OFDM) system.Multiple Input Multiple Output (MIMO) technology uses multiple antenna (or antenna array) and multiple channel in both signal sending side and receiving side. It does not need to increase bandwidth because the signals in sending side share same frequency band. Since MIMO system can create multiple parallel space channels and transmit information independently through them, it could improve transmission data rate absolutely.MIMO-OFDM technology is the combination of MIMO and OFDM. It improves the quality of transmission signals through using antenna array. It greatly increases the tolerance of noise, interference and multi-path because of using time, frequency and space diversities.While using PHY layer technologies of MIMO-OFDM system, and combining PHY layer techniques with its up layers, cross-layer design has become a research hotspot. It is significant to study cross-layer resources allocation algorithm.2. Cross-layer design of dynamic resource allocation for MIMO-OFDM systemThe proposed algorithm adopts MAC-PHY layer resources allocating strategy. MAC layer implements access control and queue scheduling by inquiring user's queuing state information and traffic status in PHY layer. Queue scheduling obeys FIFO (first in first out) rules. According to control parameters provided by MAC layer, PHY layer completes allocation of subchannels and bits. In order to reduce complexity, we propose a suboptimal solution that separates the problem into two steps: subchannel allocation and bit allocation. The bit allocation algorithm is greedy method. The user with the minimum normalized transmit rate is allowed to choose the best subchannel, which maximizes the total throughput and provides a fair resource sharing among all users as well. On the other hand, the bit allocation algorithm is a proportional greedy approach. At first, the selection of the user with the minimum normalized transmit rate is conducted, then one more bit is allocated to the subchannel, which requires the least additional power to transmit one more bit from this user's predetermined subchannel set.3. Simulation result and conclusionThe proposed algorithm counts data queuing information of users in MAC layer. Under the control of MAC layer, according to channel state and proportion fairness restrict, PHY layer implements resource allocation for users. The simulation results indicate that through using the proposed algorithm the average spectral efficiency is greatly improved and the power consumption is reduced. The performance of proposed algorithm is obviously inferior to non-cross layer algorithm.