Cross-layer Optimization For MIMO-aware Multi-hop Wireless Networks

Multiple Input Multiple Output(MIMO)technology,which can significantly improve the transmission capacity constraints and the reliability of communication,has brought a great breakthrough in the field of wireless communications.At the same time,there are still a lot of application scenarios waiting to be explored in depth.The wireless multi-hop network has been widely used in wireless network communication field due to its good robustness,flexible structure,high bandwidth,distributed deployment and so on.How to effectively combine the two technology above and give full play to their potential,will greatly improve the quality of existing communications.Based on the full understanding at home and abroad about the multi-hop wireless network resource optimization research,this paper have the in-depth study of the scheduling problem of MIMO technology in wireless multi-hop networks.Aiming at the scene of MIMO multi-hop wireless networks and utilizing the benefit of degrees of spatial freedom brought by multi antennas to provide MIMO channel mode for each transmission link,we introduce the predictive queue and two-tier queue conception to improve the network model and give a distributed implementation of the optimization problem,the main research work is as follows:(1)Propose the distributed cross-layer optimization based on MIMO-aware multi-hop wireless network based on channel mode.Aiming at a long time MIMO multi-hop wireless network average utility maximization problem,this paper propose a dynamic scheduling model with MIMO-aware channel mode which make each node of network can sense the current network state so as to select the appropriate MIMO channel mode to carry out data transmission to meet the communication needs.Compared with the traditional MIMO multi-hop wireless network,the dynamic MIMOaware channel model make the scheduling more intelligent and also can obtain better network utility.To ensure the stability of network operation,we use the Lyapunov optimization algorithm and then through the dual decomposition algorithm eventually realize the coupling separation to give a cross-layer resource allocation optimization of distributed implementation.(2)Propose the distributed cross-layer optimization based on predictive queue.On the basis of(1),this paper introduce the predictive service model which combine the predictive queue to the original data queue.Each node in the network based on a predictive window which can predict the future and send a range of slot packets.Through the prediction of the future data and decision making ahead of schedule,the whole network can effectively reduce the time delay of the data packet under the condition of ensuring the utility.In this paper,using the equivalent queue,the queue in the real data and predict are equivalent to a macroscopic sum queue and combined with Lyapunov drift theory and dual subgradient algorithm for distributed problem solving.Then according to the mapping and updating rule partition specific for each specific item optimization decision.(3)Propose the distributed cross-layer optimization based on two-tier queue.On the basis of(1),this paper introduce the two-tier queue to make up for the deficiency of the classical back pressure algorithm.By separating the whole network architecture,we construct the queues on the network layer and data link layer to make the network layer of each node is responsible for routing decisions and the data link layer of each node is responsible for link scheduling decisions so that the original joint optimization scheme can be separated.Based on Lyapunov optimization algorithm and dual decomposition algorithm to achieve optimal network utility and distributed implementation.