Modulation Recognition of OTFS Signal for UAV Communication System

With the continuous expansion of drone application scenarios, the communication requirements and modes of drones have become more diversified and complex. However, high-speed information transmission has always been a focus of attention. To provide a guarantee for highly reliable communication of drones, Orthogonal Time Frequency Space (OTFS) technology overcomes the influence of multipath and Doppler effects in traditional communication systems in high-speed moving environments. At the same time, in the multipath channel of unmanned aerial vehicle communication systems, traditional methods for identifying OTFS signal subcarrier modulation methods exist some problems such as low recognition accuracy and incomplete identification methods. Therefore, deep learning can be used to study the recognition of OTFS signal subcarrier modulation methods. The article uses Convolutional Neural Network (CNN), Long Short Term Memory (LSTM) Neural Network and LCDNN for modulation recognition. Meanwhile, Residual Network (Res Net) is used as the base model. The result of simulation experiment shows that the LCDNN has higher recognition accuracy, short training time for model and good network performance.