Symbol Classification in Receiver of OFDM Carrier Signal with Deep Learning and Whale Optimization Algorithm

Orthogonal Frequency Division Multiplexing (OFDM) remains a cornerstone in modern wireless communication systems, owing to its resilience to multipath fading and spectral efficiency. In OFDM systems, accurate symbol classification is paramount for successful data demodulation. This paper proposes a novel methodology for symbol classification in the receiver of an OFDM carrier signal, using a synergic combination of deep learning and feature selection with the Whale Optimization Algorithm (WOA). The deep learning component, embodied in a convolutional neural network (CNN), is adept at extracting intricate features from the received OFDM symbols, while the WOA facilitates efficient feature selection by optimizing a subset of attributes that contribute most to categorization accuracy. This dual approach not only enhances the discriminative power of the classification model but also reduces the computational complexity by focusing on the most relevant features. Experimental findings confirm the effectiveness of the proposed framework, demonstrating superior symbol classification performance compared to conventional methods. Moreover, the integration of feature selection with the WOA ensures the identification of an optimal subset of features, further improving classification accuracy and generalization capability. This study combines DL with metaheuristic feature selection to improve symbol classification in OFDM receivers, thereby making wireless communication systems more reliable and efficient.