How to estimate channel parameters more effectively, intelligently and accurately is the key problem to realize the requirements of intelligent and adaptive short-wave communication system. Based on the detailed analysis of chirp signal and fractional Fourier transform, an intelligent channel parameter estimation system is constructed. By building and training the regression model of multilayer fully connected neural network, the estimation error of Doppler shift is reduced. The simulation results show that the hierarchical channel estimation algorithm improves the precision of channel parameter estimation and the anti-noise performance of the system.

中文翻译：

---

基于神经网络回归模型的智能信道参数估计系统

如何更有效，智能，准确地估计信道参数是实现智能自适应短波通信系统需求的关键问题。在详细分析线性调频信号和分数阶傅里叶变换的基础上，构建了智能信道参数估计系统。通过建立和训练多层全连接神经网络的回归模型，减少了多普勒频移的估计误差。仿真结果表明，分层信道估计算法提高了信道参数估计的精度和系统的抗噪声性能。