论文部分内容阅读
针对统计量算法盲检测多进制振幅键控(MPSK)信号的缺陷,提出了一种幅值相位型连续多值复数Hopfield神经网络算法,构造了适用于MPSK信号的幅相型离散多电平激活函数,并分别在异步和同步更新模式下证明了该神经网的稳定性.当该神经网的权矩阵借助接收数据补投影算子构成时,该幅相型离散Hopfield神经网络可有效地实现MPSK信号盲检测.仿真试验表明:该算法所需接收数据较短,可到达全局真解点,并且适用于含公零点信道.
In order to overcome the shortcomings of statistical algorithm blind detection of MPSK signal, an amplitude-phase continuous multivalued complex Hopfield neural network algorithm is proposed to construct amplitude-phase discrete multi-level Activate the function, and prove the stability of the neural network in asynchronous and synchronous update mode respectively.When the weight matrix of the neural network is composed of the received data complement projection operator, the amplitude and phase discrete Hopfield neural network can be effectively implemented MPSK signal blind detection.Experimental results show that the proposed algorithm needs less received data, can reach the global true solution point, and is suitable for the common zero channel.