Automatic Modulation Recognition Using Multi Channel Prioritized Input Deep Learning Hybrid Model

Automatic Modulation Classification (AMC) is crucial in cognitive radio systems for identifying modulation schemes in wireless communication. This paper introduces a novel approach using a Multichannel prioritized Convolutional Long Short-Term Deep Neural Network (MP-CLDNN) to enhance AMC performance. Here prioritization of features is done by employing LMS algorithm. The prioritized feature is multi-channelized to CLDNN model. CLDNN architecture leverages the combined strengths of Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and fully connected Deep Neural Networks (DNN), providing an efficient and accurate framework for modulation classification. The CNN layer extracts hierarchical features from input signals, capturing local patterns in time and frequency domains. The LSTM layer models the temporal dependencies of the signals, which are vital for distinguishing modulations with time-varying characteristics. Finally, the DNN layer integrates these features to make the final classification decision. Compared to traditional methods, the CLDNN-based approach shows significant improvements in classification accuracy, especially in low Signal-to Noise Ratio (SNR) conditions. The proposed framework is applicable to a wide range of wireless communication systems, including 5G, Internet of Things (IoT), and military applications, where real time modulation identification is critical for efficient spectrum management and signal processing.