Neural network state observer-based robust adaptive fault-tolerant quantized iterative learning control for the rigid-flexible coupled robotic systems with unknown time delays

Abstract

In this work, the neural network state observer-based robust adaptive quantized iterative learning output feedback control (RAQILOFC) is investigated for rigid-flexible coupled robotic systems (RFCRSs) with unknown time delays and actuator faults. To deal with hysteresis quantization and actuator defects, a novel fault-tolerant RAQILOFC is designed first, based only on accessible system output data. Then, using the fault-tolerant RAQILOFC laws in combination with the neural network state observer, the two given angular positions are tracked while concurrently suppressing the flexible vibration. Simultaneously, uncertainties associated with system dynamics and unknown time delays are taken into account while designing controllers for RFCRSs. It is demonstrated that the fault-tolerant RAQILOFC technique would converge to and remain inside a predefined small compact set after a finite number of cycles. Additionally, it is shown that the system signal sequences are bounded in presence of unknown time delays and hysteresis quantization. Finally, a numerical example is conducted to illustrate the proposed neural network state observer-based fault-tolerant RAQILOFC strategy’s efficacy.