摘要
利用连续型Hopfield神经网络实现无线通信信号的盲检测是一种较为有效的方法,但是其抗干扰性能较差,在低信噪比等复杂环境下算法的误码率过高。为了解决连续型Hopfield盲检测算法的不足,文中对传统的激活函数进行了改进,并给出了一种新的激活函数,新激活函数有效地降低了算法对噪声的敏感度,极大地提高了算法的抗干扰能力。仿真表明,在低信噪比、大数据量等复杂环境下,改进后的算法表现出了较强的抗干扰能力和稳健性,性能得到了显著的提高。
It＇s more effective to use continuous Hopficid neural network to blindly detect wireless communication signal,but its anti jam- ming performance is poor and the algorithm bit error rate is a little high in complex environments, such as in low SNR environment. To conquer the above shortcoming ,a new kind of activation function is put forward in this thesis,which can effectively reduce the algorithm sensitivity to noise and greatly improve its anti-jamming capability. Simulation results demonstrate that the improved algorithm has better anti-jamming capability and robustness in complex environments, like low SNR or massive date environment. The improved algorithm shows strong anti-interference ability and robustness ,performance has been improved significantly.
出处
《计算机技术与发展》
2012年第12期207-210,共4页
Computer Technology and Development
基金
国家自然科学基金资助项目(60772060)
