In this paper, the basic principle of recursive least squares-back propagation neural network is discussed first.Then, the advantages of this kind of network compared with the common BP networks are proved through experiments.Last , according to the real demands of aeronautic commander system of fighting,the RLS-BP network and common BP network are applied to the multi-object identification system respectively so that the difference between them can be shown on restraint speed and identification effect.The result of experiments exhibits that RLS-BP network is applied easily in practice and it's stability is great better .
|