Secure Cloud-WSNs with Advanced Intrusion Detection System using Kepler Optimization and Bi-ConGRU Network Approaches
Sheela Pitta, S. Gopalakrishnan, S. Ravi Chand · 2024
As WSNs are popularly used in the contemporary world in various fields such as environmental monitoring, smart cities, monitoring and controlling of infrastructure and several essential activities, a detailed cloud security of the WSNs is vital. However, the establishment of these networks has come with new enormous security concerns such as peoples' unauthorized access, data leakage, and cyberattacks to these networks and all that leads to a compromise and poor performance of these networks. To solve these problems, it is necessary to apply high-quality and effective approaches to improve intrusion detection and meliorate the network's robustness. Based on the identified issues, this paper aims to discuss the critical problems and challenges in WSN cloud security with advances in intrusion detection systems that can address the emerging threats. Conventional approaches are not very effective in generalizing the threats or accurately detecting and preventing them as they cannot analyse the high dimensional data and learn from the new attack types. To eliminate these issues, the following approach has been developed including the Kepler Optimization Algorithm, KePO, and the Bi-Directional Convolutional Gated Recurrent Unit Network, BiConGRUNet. Thus, the KePO algorithm is used for feature selection resulting in the dataset improvement to boost up the intrusion detection system. The BiConGRUNet model takes advantage of state of the art deep learning methods, thus allowing for the two-dimensional analysis of network traffic. The current strategy focuses on considerable assessment with the help of benchmark datasets, including CICIDS-17 and UNSW-NB15, to prove the efficiency of the presented model.