A novel blind detection algorithm based on improved Compound Sine Chaotic Neural Networks
Qingxia Meng, Shujuan Yu, Huan Liu, Yun Zhang · 2015
In order to improve the performance of the blind detection algorithm and the phenomenon of premature convergence of Hopfield Neural Networks(HNN) as well as speed up convergence of Transient Chaotic Neural Networks(TCNN), a new blind detection algorithm based on Improved Compound Sine Chaotic Neural Networks(ICSCNN) is proposed in this paper, constructing a new energy function and proving the stability of ICSCHNN in asynchronous update mode and synchronous update mode separately. The algorithm uses sequence with chaos initialization as the transmitting signal and the proposed network has more flexible transient chaos dynamics characteristics and stronger global search ability owing to its adoption of non-monotonic activation function constituted by compound sine and sigmoid function, time-varying gain, piece-wise exponential annealing function. Simulation results show that compared with the second order statistics algorithm(SOS), blind detection algorithm based on HNN, blind detection algorithm based on TCNN, the novel algorithm not only reduces the error rate significantly but also requires shorter data size, thereby improves the performance of blind detection.