Research On Blind Detection Algorithm Based On Complex Continuous Hopfield Neural Network

In the development of artificial neural network,Hopfield neural network plays an important role,and it does not depend on the statistics,compared to other similar algorithms,the required amount of data is shorter,so more in line with modern reliable and high-speed communication requirements.The main factors affecting the quality of communication are as follows: the anti-jamming ability of the system and the rate of system processing signal.So they play an important role in HNN practical application.This paper started from these two factors,in the transmission signal for the premise of QPSK,mainly to make the following innovative works:(1)Based on the Hopfield Neural Network(HNN)model,a new type of complex-continuous Hopfield type positive feedback neural network(PFHNN)is proposed to improve the anti-jamming capability of the system.In this paper,a positive feedback mechanism is added to the structure of CSHNN system,and a new energy function is constructed and its stability is proved.Then,a VS-PFHNN blind detection algorithm with adaptive positive feedback coefficient is proposed.Finally,a double Sigmoid structure is added to improve the convergence rate of the algorithm.The experimental results show that the PFHNN blind detection algorithm can improve the anti-jamming performance of the HNN network,while the VS-PFHNN blind detection algorithm improves the anti-jamming performance of the HNN network compared with the PFHNN algorithm.(2)In order to improve the rate of HNN signal processing,the third chapter of this paper deeply studies the double Sigmoid complex continuous Hopfield neural network(CS-DSHNN)blind detection algorithm.In this paper,a new activation function is designed,and a new HNN complex system(CSHNN)is constructed.The new energy function is given and its stability is proved.Then,a new Double-Sigmoid Complex-Continuous Hopfield neural network CS-DSHNN is proposed,and a new energy function is given and its stability is proved.Experiments show that CSHNN algorithm improves the anti-jamming ability of the network,CS-DSHNN blind detection algorithm has obvious faster convergence rate.(3)In the fourth chapter,we study the concrete influence of the parameters in the Sigmoid function selected in Chapter 3 of this paper on the performance of the whole network.The function contains k and ? two parameters,respectively,when separately changes the two parameters this paper first studied the convergence rate of the entire HNN network.Then make the two parameters change at the same time,observe the effect of the cost function of the whole network three-dimensional map,select the optimal combination of parameters.The experimental results show that the change of the parameter k has a great influence on the convergence rate of the network,and the change of the parameter ? has a small influence on the convergence rate of the network.When the two parameters change at the same time,the selection of the zero point of the cost function is the optimal combination of parameters,so that the HNN network blind detection algorithm can obtain the optimal performance under the same initial conditions.