井下认知无线电EWA信道选择算法

Channel EWA selection strategy in coal mine cognitive radio

摘要:

为解决特殊的井下恶劣无线环境造成通信极易中断的问题,提出基于经验加权吸引策略的信道选择算法. 深入分析报酬反馈结构,建立更加细致和合理的多级的报酬反馈体制. 采用简单多次重复概率实验的方法验证了基于EWA学习的信道选择算法的有效性. 通过与传统单一固定信道传输和Q学习算法的深入对比研究表明,基于EWA学习的信道选择算法根据无线信道特性的在线学习经验对优选信道快速捕获的准确性和稳定性方面具有更好的性能.

Abstract:

In oder to solve extremely unstable link quality and communication failure caused by strong electromagnetic noise and interference in special coal mine environment, this paper proposes an Experience-Weighted Attraction(EWA) selection algorithm. Based on the system structure analysis in depth, it establishes more detailed and reasonable multi-layer reward feedback system. Research is conducted by the method of simple repeated probability experiments through software simulations. By comparing with traditional fixed single channel transmission and Q learning, channel selection strategy with EWA learning has more advantageous performance on sensitivity and stability.