Probabilistic Deming Deep Recursive Network and Schmidt-Samoa Cryptosystem for Cyber Attack Detection and Secure Transmission in Wireless Networks

Abstract

A wireless network is an elastic data communications system that exploits wireless media to broadcast data over the network. In wireless network, cyber attack detection is vital for maintaining the security and integrity of communication systems since these networks are more susceptible to different threats due to their open, dynamic, and broadcast nature. Therefore, cyber-attacks are becoming more difficult, targeting systems that handle or store sensitive information. Due to the rapid increase in cyber-attacks, the design of a detection mechanism discovers the harmful effect and attacks. The developed mechanisms for cyber attack detection unable to achieve higher detection accuracy and security in the data transmission. Therefore, a novel method referred as Probabilistic Deming Regressive Deep Recursive Network based Kupyna Schmidt-Samoa Signcryption (PDRDRN-KSSS) technique is proposed for efficient cyber attack detection in wireless network with better accuracy and security. PDRDRN-KSSS performs two different processes such as classification and secure data transmission. The classification is performed using Probabilistic Deming Regressive Similarity Indexed Deep Recursive Network for cyber attack detection.  The designed network obtains number of data samples and their features as input.  Then, the feature selection is performed using Deming Regression analysis to choose the more relevant features from the dataset for attack detection. After that, the cyber attacks detection is performed through classifying the data samples with selected feature using the Tversky Similarity Index. Tversky Similarity Index is used to analyze the training data and mean of the particular classes. Based on similarity value, data samples are classified into normal and different types of attack nodes (i.e., Fuzzers node, Analysis node, Backdoors node, DoS node, Exploits node, Generic node and Reconnaissance node). Followed by this, the Cramer's correlated Kupyna Schmidt-Samoa Signcryption is used for securely transmit the normal data samples. The cryptographic technique performs key generation, signcryption, and unsigncryption. During the key generation process, pair of public key and private keys is generated. Then, the signcryption is carried out using encryption and signature generation.  The sender encrypts the data sample using the recipient’s public key to promise confidentiality and generate the signature their own private key to guarantee the integrity. The ciphertext and the signature are sent to the recipient. The recipient confirms the signature using the sender's public key and then decrypts the ciphertext using their private key to get the original data. With this, secure data transmission is achieved in wireless networks. Experimental evaluation is carried out on metrics such as data confidentiality rate, integrity rate, attack detection accuracy and attack detection time with respect to different number of data samples. The results confirm that the proposed PDRDRN-KSSS method attains better accuracy, integrity, confidentiality with lesser time.