Neurodynamics for Robust Kinematic Control of Serial Manipulators Under Deception Attack

With the deployment of advanced communication and networking techniques, remote control of manipulators has become feasible and attractive. It is known that neurodynamics, as an important part of computational intelligence, has been extensively adopted for addressing the kinematic control problem of manipulators. However, in the context of remote control, traditional neurodynamic methods cannot be directly adopted due to the existence of cyberattacks. In this brief, we deal with the remote control problem of manipulators subject to a deception attack, where an attack can modify the control commands sent to the controlled manipulator. To resolve this issue, we presented a neurodynamics-based dynamic controller synthesized by nonlinear activation functions and aided by a dynamic attack estimator. Theoretical analysis is provided to guarantee the robust control performance of the proposed method. The simulation comparison with state-of-the-art methods validates the advantages of the proposed method.