摘要
移动节点位置预测是机会认知网络进行有效数据采集和消息转发的基础,提出了一种基于社会关系的移动节点位置预测算法.该算法基于位置对应用场景进行建模,通过节点的移动规律挖掘节点之间的社会关系.该算法以1阶Markov模型为基础对节点的移动性进行初步预测,然后,利用与其社会关系较强的其他节点位置对该节点的预测结果进行修正.最后,基于UCSD WTD数据集对算法进行仿真实验.结果表明,基于社会关系的移动节点位置预测算法与1阶Markov预测模型相比获得了更好的预测精度,并且算法具有较好的可扩展性.
Mobile node trajectory prediction is a basis for effective data collection and data dissemination in opportunistic cognitive networks. A social-aware node mobile trajectory prediction algorithm was proposed, in which the location of other nodes were used to update the prediction results based on the 1-order Markov model, and the accuracy of the prediction model was improved by sparse of state transition matrix. Finally, the feasibility of prediction model was verified based on the UCSD WTD dataset. The results showed that the location prediction model based on social relationship could get better accuracy than that of the 1-order Markov prediction model. Therefore, the algorithm has high value in practical applications.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2014年第12期1701-1705,共5页
Journal of Northeastern University(Natural Science)
基金
国家杰出青年科学基金资助项目(61225012
71325002)
教育部高等学校博士学科点专项科研基金优先发展领域资助项目(20120042130003)
教育部高等学校博士学科点专项科研基金资助项目(20110042110024)
中央高校基本科研业务费专项资金资助项目(N110204003
N120104001)
辽宁省科技计划项目(2013217004
关键词
机会认知网络
位置预测
MARKOV模型
社会关系
opportunistic cognitive networks
trajectory prediction
Markov model
social-aware