摘要
认知引擎是认知无线电关键技术之一,其核心是利用人工智能算法完成认知学习、推理与决策功能。自适应满足环境变化和用户需求。提出一种基于ART1和FAM神经网络的认知引擎。该引擎基于MATLAB 802.11a仿真平台模拟无线通信环境,通过对环境信息的学习训练,结合信道特征和用户需求推理决策出系统最优工作参数,实现认知无线电自适应配置。仿真结果表明,该认知引擎能有效实现认知无线电学习推理功能,且算法精度和稳定性均优于SVM及BP网络模型。
Cognitive engine （CR） is one of the key technologies of cognitive radio, and its core is utilizing ar- tificial intelligence to achieve the ability of learning, reasoning and decision. This paper presents a cognitive engine based on ART1 and FAM neural network to adapt varied environment and user＇s needs. In the wireless environment based on 802. 11 a simulation platform, the engine can learn the environment informant, reason the best working parameters and realize the CR configurations on the conditions of channel characters and user＇ s needs. The simulation results improve that the proposed cognitive engine can effectively realize the learning nnrl r^n~nnin^r fimctinn nnrt it.~ nr^r~urAr^v nnrt ~tnhilitv nre, both ha, Her thnn thn~ nf SVM nnd RP netwnrk.
出处
《西南科技大学学报》
CAS
2012年第4期75-79,共5页
Journal of Southwest University of Science and Technology
基金
国家自然科学基金资助项目(61072138)
四川省科技厅应用基础研究项目(自筹)(2010jy0173)
关键词
认知引擎
认知无线电
神经网络
Cognitive engine
Cognitive radio
Neural network
Learning
Inference
