An encryption method for LAN feedback sharing information based on robust multi-view subspace clustering

With the purpose of capturing the complex structures and laws in the information, obtaining comprehensive and consistent information clustering results, improving the personalization of the generated key, preventing the attacker from cracking the encrypted information by analyzing the fixed key patterns, and increasing the difficulty of cracking for the attacker. In this paper, we study the encryption method of LAN feedback sharing information based on robust multi-view subspace clustering to improve the security of information encryption. In LAN environment, the sending node and receiving node measure the channel characteristics by sending ordinary pilots to each other in LAN to get the measurement of complex channel coefficients when transmitting feedback sharing information. Through robust multi-view subspace clustering algorithm, the values of the sending nodes measuring themselves are clustered to capture the complex structures and laws in the measured values, and to obtain the clustering results with comprehensive consistency of the measured values. Generate the initial key according to the clustering result and improve the personalization of the initial key. The dynamic composite chaotic system is used to process the LAN feedback sharing information, generate chaotic sequences, design S-boxes to combine the initial key, generate the wheel key, encrypt the LAN feedback sharing information, and provide an extra level of security for encryption. The experimental results show that the clustering normalization mutual information of this method reaches 0.92, which is 18% higher than single view clustering. It can effectively cluster the measurement values of the sending nodes themselves, effectively encrypting the shared information in the local area network and improving the security of shared information. Under different attack types, the avalanche effect value is 0.989–0.999, which means that the encryption effect is better and can effectively increase the difficulty for attackers to crack.