计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (12): 71-76.

• 网络、通信、安全 • 上一篇    下一篇

基于极端学习机的MANET移动性预测模型

张青林   

  1. 安徽经济管理学院 信息工程系,合肥 230051
  • 出版日期:2014-06-15 发布日期:2015-05-08

Prediction model of mobility in MANET based on extreme learning machines

ZHANG Qinglin   

  1. Department of Information Engineering, Anhui Economic Management Institute, Hefei 230051, China
  • Online:2014-06-15 Published:2015-05-08

摘要: 针对移动自组织网络移动性在管理无线网络带宽资源可用性方面的重要性,为了更好地规划连续服务可用性和有效能源管理以提升网络的整体服务质量,提出了一种基于极端学习机的MANET移动性预测模型。利用ELM对MANET中的任意节点进行建模;假设已知每个移动节点当前的移动性信息(位置、速度和运动方向角度),以这种方式预测节点未来的位置和相邻节点之间未来的距离;基于几个标准移动性模型,产生更加真实、精确的移动性预测,从而更好地捕捉任意节点直角坐标系之间现有交互/相关性。使用标准移动性模型的仿真结果验证了所提模型的有效性,实验结果表明,提出的预测模型明显改进了传统基于多层感知器的模型,此外,当预测相邻节点之间未来距离时,避免了当前算法对预测精度的限制。

关键词: 移动自组织网络, 多层感知器, 移动性预测, 极端学习机, 服务质量

Abstract: For the importance of mobile ad hoc networks(MANET) in managing resource availability of wireless network, a prediction model of mobility in MANET based on Extreme Learning Machines(ELM) is proposed to program availability of continuous service and effective energy management so as to improve total quality of service of network. ELM is used to model and predict mobility of arbitrary nodes in MANET. Each mobile node is assumed to know its current mobility information(position, speed and movement direction angle), and future node positions are predicted along with future distances between neighboring nodes. More realistic and accurate mobility prediction is generated based on several standard mobility models so as to capture better the existing interaction/correlation between cartesian coordinates of arbitrary nodes. The effectiveness of proposed model has been verified by simulation using standard mobility models illustrate. Simulation results show that proposed prediction model has improved over conventional model based on multi-layer perception. It circumvents the prediction accuracy limitations in current algorithms when predicting future distances between neighboring nodes.

Key words: mobile ad hoc networks, multi-layer perception, mobility prediction, extreme learning machines, quality of service