The term "Wi-Fi" refers to the wireless local area network that affords internet access in numerous familiar places, including stores, restaurants, cafes, and college campuses. The Internet's unavailability, which allows for many attacks, has delayed and sometimes completely disrupted these services, although they were made possible by this technology. Therefore, due to their inherent weakness, wireless interfaces have spurred the study and suggestion of traffic investigation and anomaly detection systems. This study introduces a scalable and modular algorithm structure for setting up deep learning (DL) based on a Convolutional Deep Belief Network (CDBN) with a genetic algorithm (GA) named OCDBN to detect malicious frames reliably and To detect attacks, with features optimized via a GA. This model uses a dissimilarity measure that deals with non-numerical and numerical features. It also uses the Aegean Wi-Fi Intrusion Detection (AWID) dataset to test how well the suggested Algorithm works. This work found up to 12 of the 14 attack classes in the AWID dataset recognized with high confidence just by looking at a single frame, as long as the appropriate features are perceived