摘要
研究一种前向型神经网络的改进学习算法并基于TI的TMS320C5402定点数字信号处理器开发系统实现该算法的训练学习。测试结果表明:网络学习速率提高,网络的输出动态响应具有超调小、响应快和鲁棒性强的优点,为DSP在实时性控制系统中的应用奠定了一定基础。
Study on a kind of ameliorative BP arithmetic and realize it on TI's TMS320C5402 fixed digital signal processor. Test results show: the study ratio improved, the output of Neural-network system own features as small overtop, quick response, strong robust. The research settles a foundation for DSP application in automation control system.
出处
《天津科技大学学报》
CAS
2004年第2期41-43,52,共4页
Journal of Tianjin University of Science & Technology
