Strengthening security in IoT-based smart cities utilizing cycle-consistent generative adversarial networks for attack detection and secure data transmission

The main purpose of Smart Environments (SE) is to conveniently improve the human's daily life. Internet of Things (IoT) is a developing network for smart objects. Privacy-based security is a significant issue in any real-world smart environments centered on the IoT system. Security susceptibility in the IoT-centered systems provides a risk of security affecting smart environment applications. In this manuscript, Strengthening Security in IoT-Based Smart Cities utilizing Cycle-Consistent Generative Adversarial Networks for Attack Detection and Secure Data Transmission (IoT-SC-CCGAN-ADSDT) is proposed. Here, input information is gathered from NSL-KDD. The NSL-KDD input is pre-processed. Then, the important features of the pre-processed data are selected by using Wild horse optimizer (WHO). After feature selection, the chosen features are provided to cycle-consistent generative adversarial network classifier for classifying the attack and normal data. The selected features are sent to the use after the prediction of outcomes using Advanced Encryption Standard (AES). The AES is optimized using Chameleon Swarm Algorithm for transmitting the data in a safer way. After transmitting the data securely, the normal data outcomes obviously shown in LCD monitor. To show these results, major problems in the smart cities are simply detected. The proposed model is activated using java. The efficiency is examined with performance metrics, like precision, sensitivity, specificity, accuracy, computational time, encryption time, decryption time, security level. The proposed IoT-SC-CCGAN-ADSDT approach provides 96.68%, 7.142%, 94.65%, and 97.58% greater accuracy compared to the existing DL-IOT-SCA, IoT-SC-PCA, IoT-SCA-DL methods respectively.