Adaptive Quantized Control for Nonlinear Systems With Mismatched Disturbances and Malicious Attacks

ABSTRACT For a class of uncertain strict‐feedback nonlinear systems subject to malicious sensor/actuator attacks, global adaptive quantized control via output feedback is investigated. A nonbackstepping method is proposed based on a nonidentification adaptive mechanism, which develops a pair of novel single‐dynamic‐gain observer and controller. For strict‐feedback nonlinear systems with mismatched disturbances, the differentiability assumption on sensor attack signals is removed, and uncertain information about the upper bound of attack signals is tolerated. The unknown nondifferentiable input/output gains induced by input quantization and sensor/actuator attacks are simultaneously addressed in time‐varying coupled matrix inequalities, where their upper bounds may be unknown and compensated by a single‐dynamic gain. Finally, it is shown that global output regulation with any preset accuracy can be achieved under a hysteretic logarithmic quantizer and state‐dependent sensor attacks, and the control method can be extended to the tracking problem under sensor and actuator attacks.