Command filter-based event-triggered control for stochastic MEMS gyroscopes with finite-time prescribed performance.

Area of Science:

• Control Systems Engineering
• Neural Networks
• MEMS Technology

Background:

• Stochastic microelectromechanical system (MEMS) gyroscopes exhibit complex nonlinear dynamics and external disturbances.
• Existing control strategies often struggle with finite-time performance and computational complexity.
• Efficient resource utilization in control systems remains a significant challenge.

Purpose of the Study:

• To develop an adaptive neural control strategy for stochastic MEMS gyroscopes.
• To achieve prescribed performance within a finite time.
• To enhance control accuracy while reducing communication load.

Main Methods:

• Utilizing radial basis function neural networks to model unknown system dynamics and disturbances.
• Employing finite-time prescribed performance functions and command-filtered backstepping for robust control.
• Implementing a switching threshold event-triggered control law to optimize communication resources.

Main Results:

• Guaranteed convergence of output tracking error to a small residual set.
• Ensured semi-global ultimate uniform boundedness of closed-loop system signals in probability.
• Demonstrated effectiveness and superiority through numerical simulations.

Conclusions:

• The proposed adaptive neural control strategy effectively manages stochastic MEMS gyroscope dynamics.
• Finite-time prescribed performance and reduced communication load are achieved.
• The control method offers a superior solution for MEMS gyroscope control applications.