Dynamic neural network-based robust observers for uncertain nonlinear systems

“…The update law in (19) is motivated by the fact that if the full state were available for feedback and if the approximation error, , were zero, then using P 1 · · · P n T = F 1 · · · F n T + G 1 · · · G n T θ, the parameters could be estimated via the least squares esti-…”

Section: Purgingmentioning

confidence: 99%

See 1 more Smart Citation

Simultaneous state and parameter estimation for second-order nonlinear systems

Kamalapurkar

2017

2017 IEEE 56th Annual Conference on Decision and Control (CDC)

View full text Add to dashboard Cite

In this paper, a concurrent learning based adaptive observer is developed for a class of second-order nonlinear time-invariant systems with uncertain dynamics. The developed technique results in uniformly ultimately bounded state and parameter estimation errors. As opposed to persistent excitation which is required for parameter convergence in traditional adaptive control methods, the developed technique only requires excitation over a finite time interval to achieve parameter convergence. Simulation results in both noise-free and noisy environments are presented to validate the design.

show abstract

“…However, the sliding mode observer is only appropriate for the pure feedback nonlinear systems. To handle the uncertain nonlinearities of system, the NN observer [10,11] has been widely used in the complicated nonlinear system. Inspired by the high gain observer and NN observer, this paper presents a variable gain NN observer to estimate the unknown states and nonlinearities.…”

Section: Introductionmentioning

confidence: 99%

Synchronization and Tracking Control for Dual‐motor Driving Servo Systems with Friction Compensation

Zhao

Ren

2018

Asian Journal of Control

View full text Add to dashboard Cite

A nonsingular fast terminal sliding mode (NFTSM) controller is designed by incorporating the variable gain neural network (NN) observer, which is utilized to guarantee motor speed synchronization and load position tracking of dual-motor driving servo systems. By designing the variable gain NN observer, the states and uncertain nonlinearities of servo systems are estimated with fast convergence rate and small steady-state error, where the effects from external disturbance are suppressed as well. Based on the estimated states, the cross-coupling synchronization strategy and NFTSM tracking scheme are designed to achieve the rapid speed synchronization and precise load tracking, where the NNs are introduced to approximate and compensate friction nonlinearities. In particular, a novel nonlinear synchronization factor characterizing the degree of speed synchronization is proposed to achieve switching between synchronization control and tracking control, which is proven to deal with the coupling problem of synchronization and tracking. Finally, the comparative simulations and experiments are included to verify the reliability and effectiveness.

show abstract

Dynamic neural network-based robust observers for uncertain nonlinear systems

Cited by 38 publications

References 27 publications

Model-based reinforcement learning for output-feedback optimal control of a class of nonlinear systems

Model-based reinforcement learning for output-feedback optimal control of a class of nonlinear systems

Simultaneous state and parameter estimation for second-order nonlinear systems

Synchronization and Tracking Control for Dual‐motor Driving Servo Systems with Friction Compensation

Contact Info

Product

Resources

About