Resilience to deception attacks in consensus tracking control of incommensurate fractional-order power systems via adaptive RBF neural network

This paper addresses the consensus control problem for incommensurate fractional-order multi-agent power systems vulnerable to sensor deception attacks within their communication networks. To overcome this problem, an adaptive Radial Basis Function neural network (RBFNN) state observer is employed to estimate the system uncertainties in the presence of malicious attacks. Additionally, a compensation mechanism is implemented to mitigate the effects of unknown time-varying gains in the system output induced by deception attacks. Within the adaptive backstepping framework, the proposed control strategy ensures that all system signals remain bounded while enabling all followers to achieve consensus with the leader's trajectory. The effectiveness of the proposed control approach is further validated through MATLAB simulations, demonstrating its ability to enhance stability and reliability in power systems.