Fixed-time Event-Triggered Adaptive Control of Manipulator System with Input Deadzone and Model Uncertainty

Haoran Zhan,Chen Wang,Qing Guo,Xinglong Wu,Tieshan Li
DOI: https://doi.org/10.1016/j.neucom.2024.128265
IF: 6
2024-01-01
Neurocomputing
Abstract:There exist universal model uncertainty and input deadzone in mechatronic motion plant due to unknown model parameters and actuator physical feature, which will degrade the motion performance and stability. In this study, a fixed-time event-triggered adaptive control is proposed in n-DOF manipulator system to address these problems. Firstly, the dynamic model of n-DOF manipulator with joint friction and input deadzone is constructed by Lagrange technique. Then a radial basis function neural network (RBFNN) is designed to approximate uncertain dynamics. Meanwhile, an event-triggered mechanism (ETM) is used to reduce the control update frequency and to save the communication resource. Furthermore, an adaptive estimation law is adopted to compensate unknown input deadzone parameters. Based on backstepping iteration technique, a fixed-time convergent controller is presented to guarantee the system state errors convergence to zero neighborhood in finite time irrelevant to arbitrary initial state. Finally, the effectiveness of the proposed control scheme is verified by sufficient comparative simulation and experimental results.
What problem does this paper attempt to address?