Memory-based attack-tolerant T-S fuzzy control of networked artificial pancreas system subject to false data injection attacks

Abstract

This paper studies the memory-based attack-tolerant control issue of networked artificial pancreas system subject to false data injection attacks. First, a T-S fuzzy modeling approach and a T-S fuzzy observer based on sampled outputs are utilized to deal with the system nonlinearity and estimate full system state, respectively. Second, to maintain the system security in the presence of malicious attacks in the communication network, a radial basis function neural network-based estimator is introduced to approximate the real attack signal. Then, with the utilization of estimated state, memory state related to sampled period and approximated attack, a memory-based attack-tolerant T-S fuzzy controller is constructed to restore the blood glucose levels within a safe range. Third, some novel sufficient conditions formed in a set of linear matrix inequalities are derived to solve the controller and observer gains. Finally, the advantages of the proposed blood glucose regulation scheme are confirmed through illustrative simulation outcomes.