| 基于RBF神经网络的PID无线AQM算法 |
| |
| 引用本文： | 陈亮.基于RBF神经网络的PID无线AQM算法[J].南通纺织职业技术学院学报,2009,9(3):14-18. |
| |
| 作者姓名： | 陈亮 |
| |
| 作者单位： | 南通纺织职业技术学院,南通,226007;南京理工大学计算机科学与技术学院,南京,210094 |
| |
| 摘 要： | PID算法能较好地控制队列长度，但算法中比例积分微分系数较为敏感，基于试凑和经验的设定往往使控制效果难以保证，根据Ad Hoc网络环境参量时变的特点，推导了无线TCP／AQM模型，将递推计算修正功能引入PID算法，设计了一种基于RBF神经网络的PID的AQM，该算法可以在线调整PID控制器增益．仿真表明：在动态拓扑、无线分组丢失的AdH oc网络中，该算法取得了较好的队列控制效果．
|
| 关 键 词： | 无线自组网 拥塞控制 主动队列管理 PID RBF |
PID Congestion Control Algorithm Based on RBF Neuron Network in Ad Hoc Network |
| |
| CHEN Liang.PID Congestion Control Algorithm Based on RBF Neuron Network in Ad Hoc Network[J].Journal of Nantong Textile Vocational Technology college,2009,9(3):14-18. |
| |
| Authors: | CHEN Liang |
| |
| Affiliation: | CHEN Liang ( 1. Nantong Textile Vocational Technology College, Nantong 226007, China; 2. Nanjing University of Science and Technology, Nanjing 210094,China) |
| |
| Abstract: | PID can regulate queue size successfully,but the Proportional-Integral-Differential control gain is sensitive. So the control performance, which is dependent on the gain by experience and trial method, is hard to ensure. According to Ad Hoc network time-varying characteristic, the paper deduced TCP/AQM model in wireless network, and introduced recursion and modification gain to PID controller. It proposed a PID based on RBF neuron network AQM,which can modify PID gain dynamically. Simulations demonstrated that new algorithm performance is good under wireless packet loss and dynamical topology condition. |
| |
| Keywords: | PID RBF |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
