A Key Management Protocol for Heterogeneous Sensor Networks Based on Zero Trust Security and Chaotic Neural Networks
Guogang Li, Tong Xie, Cheng Zou, Wenlong Fu · Research Square · 2022
Abstract Aiming at the node security risks and key management vulnerabilities in heterogeneous sensor networks, a key management protocol for heterogeneous sensor networks based on zero-trust security and chaotic neural networks (KMPHSN-ZTSCNN) was proposed. Based on the singular matrix decomposition of difficulty and Hopfield overload chaos neural network classification features, using blockchain and zero-knowledge proof to realize sensor network node registration and authentication, it relies on channel state information (CSI) and adjustable mathematical function to generate dynamically changing keys to complete continuous verification and achieve zero-trust security authentication to ensure data security. The protocol can dynamically allocate different keyspace sizes according to the security level of the group, node storage capacity and computing capacity, and can adapt to the asymmetric structure of heterogeneous sensor networks. Theoretical proof and experimental performance analysis show that the protocol is feasible and can meet the security requirements of heterogeneous sensor networks.