AI-Augmented, Ripple-Diamond-Chain Shaped, Rateless Routing in Wireless Mesh Networks With Multibeam Directional Antennas

IEEE Access

卷:6

Sunil Kumar¹  Fei Hu²  Ke Bao² 

[1]Electrical and Computer Engineering Department, San Diego State University, San Diego, CA, USA

[2]|Electrical and Computer Engineering Department, The University of Alabama, Tuscaloosa, AL, USA

关键词: Multibeam directional antennas;    multihop routing;    wireless mesh network;    fuzzy logic;    reinforcement learning (RL);    Q learning;   

DOI  :  10.1109/ACCESS.2018.2820053

来源: DOAJ

【 摘 要 】

Use of multibeam directional antennas (MBDAs) allows a node to simultaneously communicate in multiple directions on its different beams, which can significantly improve the data throughput compared with omni-directional or single-beam directional antennas. However, conventional multihop wireless routing protocols cannot effectively exploit the benefits of MBDAs. In this paper, we present a novel routing scheme for a wireless mesh network (WMN) equipped with MBDAs, which has the following two features: 1) Ripple-Diamond-Chain (RDC) shaped routing: To exploit the multidirection transmission capability of MBDAs, we use rateless codes to obtain loss-resilient symbols for data packets, which can be simultaneously transmitted on multiple beams of a node. Then we use ripples to differentiate the nodes in each hop in the tree topology of WMN. While the symbols are dispatched on multiple beams of a main path node in one ripple, they converge into another main path node in the next ripple. The entire route thus looks like a diamond chain, which can exploit the MBDA benefits. 2) Artificial Intelligence (AI)-Augmented Path Selection: Our routing scheme is augmented by using two AI algorithms. The fuzzy logic is used to define a weighted link quality, in order to adapt to different quality-ofservice requirements. The reinforcement learning is used to select the best main path based on the cumulative reward in all the links. In the simulations, we use the video as well as other time-sensitive traffic to evaluate efficiency of RDC routing, as well as intelligent path determination in WMN.

【 授权许可】

Unknown