导弹与制导技术

图像制导中的一种改进增量学习RBF神经网络

  • 梁涛 ,
  • 李庆震 ,
  • 赵久奋 ,
  • 刘宁
展开
  • 1 第二炮兵工程学院,西安 710025
    2 61683部队,北京 100090

梁涛(1983-),男,陕西人,硕士研究生,研究方向:飞行器设计。

收稿日期: 2010-06-02

  网络出版日期: 2025-05-28

A Modified Incremental Learning RBF Neural Network in Image Guidance

  • LIANG Tao ,
  • LI Qingzhen ,
  • ZHAO Jiufen ,
  • LIU Ning
Expand
  • 1 The Second Artillery Engineering College, Xi'an 710025,China
    2 No.61683 Unit, Beijing 100090,China

Received date: 2010-06-02

  Online published: 2025-05-28

摘要

为改进现有人工神经网络在图像制导系统的自动目标识别算法中的不足,加快收敛速度,通过对基于RBF神经网络自动目标识别技术的研究,提出以隐层神经元价值函数作为指标的生长修剪策略,构造出一种基于增量学习的网络资源分配网络(IL-RAN),对目标进行了在线实时识别。仿真结果表明,利用该改进算法可产生规模较小的网络,并使得网络参数的选择与识别误差建立了联系,减少了总体的计算量,计算时间也大大缩短。

本文引用格式

梁涛 , 李庆震 , 赵久奋 , 刘宁 . 图像制导中的一种改进增量学习RBF神经网络[J]. 弹箭与制导学报, 2010 , 30(5) : 63 -65,72 . DOI: 10.15892/j.cnki.djzdxb.2010.05.066

Abstract

To cover the shortage of existing artificial neural network system in automatic target identification algorithm of the image guidance system,the automatic target identification technology based on radial basis function neural network was studied to accelerate the convergence.The value function of hidden level neuron was created as an index of neuron growing and pruning strategy.An incremental learning resource allocating network(IL-RAN) was constructed to learn and online classify the target.The experimentalanalyses proved that with this algorithm,the smaller network within the error bound can be constructed,the connection between the network paramater selection and idintification error was established.Both the calculation and time were reduced.

参考文献

[1]
Pati Y E,Krishnaprasad P S.Analysis and synthesis of feedforward neural network using discrete affine wavelet[J].IEEE trans.on Neural Networks,1993,4(1):73-75.
[2]
Szu H,Telfer B A.Neural network adaptive wavelets for signal representation and classification[J].Optical Engineering,1992,31(9):1907-1916.
[3]
Szu H,Telfer B,Garcia J.Wavelet transforms and neural networks for compression and recognition[J].Neural Networks,1996,9(4):695-708.
[4]
Mallat S.A theory of multiresolution signal decomposition: the wavelet transform[J].IEEE Trans.PAMI,1989,11(7):674-693.
[5]
张学工, 边肇祺. 模式识别[M].2版. 北京: 清华大学出版社, 1999.
[6]
瓦普尼克. 统计学习理论的本质[M]. 张学工,译. 北京: 清华大学出版社, 1999.
[7]
Platt J, et al. Large margin DAGs for multiclass classifi-cation[C]// Advances in Neural Information Processing Systems, MIT Press, 2000, 12. 547- 553.
[8]
S Vijayakumar, S Wu. Sequential support vector classi-fiers and regression[C]// Proc. of SOCO, 1999, Genoa, Italy: 610- 619.
[9]
Platt J. A resource allocating network for function inter-polation[J]. Neural Computation, 1991, 3(2): 213- 225.
[10]
J Platt. Sequential minimal optimization: A fast algo-rithm for training support vector machines, Microsoft, Res. Tech Rep. MSR-TR-98-14[R]. 1998.
[11]
V Kadirkamanathan, M Niranjan. A function estimation approach to sequential learning with neural networks[J]. Neural Computation, 1993, 5(6): 954- 975.
[12]
N Sundararajan, P Saratchandran, L Yingwei. Radial basis function neural networks with sequential learn-ing: MRAN and its applications[M]. Singapore: World Scientific, 1999.
文章导航

/