Data-Driven Model-Free Tracking Reinforcement Learning Control with VRFT-based Adaptive Actor-Critic

Rădac, Mircea-Bogdan; Precup, Radu‐Emil

doi:10.3390/app9091807

Cited by 42 publications

(18 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Compared to the methods in [32,33], the whole operation process is analyzed in this paper. Reinforcement learning based algorithms were employed to solve optimal control problems in [34]. An adaptive actor-critic learning framework is carefully designed with the virtual reference feedback tuning.…”

Section: Related Workmentioning

confidence: 99%

Long Short-Term Memory Neural Network Applied to Train Dynamic Model and Speed Prediction

Tang

Gao

2019

Algorithms

View full text Add to dashboard Cite

The automatic train operation system is a significant component of the intelligent railway transportation. As a fundamental problem, the construction of the train dynamic model has been extensively researched using parametric approaches. The parametric based models may have poor performances due to unrealistic assumptions and changeable environments. In this paper, a long short-term memory network is carefully developed to build the train dynamic model in a nonparametric way. By optimizing the hyperparameters of the proposed model, more accurate outputs can be obtained with the same inputs of the parametric approaches. The proposed model was compared with two parametric methods using actual data. Experimental results suggest that the model performance is better than those of traditional models due to the strong learning ability. By exploring a detailed feature engineering process, the proposed long short-term memory network based algorithm was extended to predict train speed for multiple steps ahead.

show abstract

Section: Related Workmentioning

confidence: 99%

Long Short-Term Memory Neural Network Applied to Train Dynamic Model and Speed Prediction

Tang

Gao

2019

Algorithms

View full text Add to dashboard Cite

show abstract

“…It is noted that most of the schemes mentioned above need to establish neural networks to design controllers, which makes preparing the external testing signals and training processes inescapable. Recently, some useful results have been reported for unknown multiagent systems, such as Model-Free Adaptive Control (MFAC) [23]- [24], Q-Learning [25]- [27], Iterative Feedback Tuning (IFT) [28]- [29], Simultaneous Perturbation Stochastic Approximation (SPSA) [30]- [31], Iterative Learning Control (ILC) [32]- [38], Virtual Reference Feedback Tuning (VRFT) [39].…”

Section: Introductionmentioning

confidence: 99%

Data Driven Distributed Bipartite Consensus Tracking for Nonlinear Multiagent Systems via Iterative Learning Control

Zhao

2020

IEEE Access

View full text Add to dashboard Cite

This paper explores a data-driven distributed bipartite consensus tracking (DBCT) problem for discrete-time multi-agent systems (MASs) with coopetition networks under repeatable operations. To solve this problem, a time-varying linearization model along the iteration axis is first established by using the measurement input and output (I/O) data of agents. Then a data-driven distributed bipartite consensus iterative learning control (DBCILC) algorithm is proposed considering both fixed and switching topologies. Compared with existing bipartite consensus, the main characteristic is to construct the proposed control protocol without requiring any explicit or implicit information of MASs' mathematical model. The difference from existing iterative learning control (ILC) approaches is that both the cooperative interactions and antagonistic interactions, and time-varying switching topologies are considered. Furthermore, through rigorous theoretical analysis, the proposed DBCILC approach can guarantee the bipartite consensus reducing tracking errors in the limited iteration steps. Moreover, although not all agents can receive information from the virtual leader directly, the proposed distributed scheme can maintain the performance and reduce the costs of communication. The results of three examples further illustrate the correctness, effectiveness, and applicability of the proposed algorithm. INDEX TERMS Iterative control (ILC), Bipartite consensus, Data-driven control (DDC), Multi-agent systems (MASs), Nonlinear discrete-time systems.

show abstract

“…The influence of uncertainty on matrix calculation was solved by means of set membership degree method and band position representation method. In addition to the above documents, adaptive actor-critic data-driven model-free tracking reinforcement learning control based on virtual reference feedback tuning, a programmable logic controller (PLC)-based fractional water level control method, and a Takagi-Sugeno fuzzy model based on linear matrix inequality fault-tolerant control were also proposed [42][43][44].…”

Section: Introductionmentioning

confidence: 99%

Disturbance Observer and L2-Gain-Based State Error Feedback Linearization Control for the Quadruple-Tank Liquid-Level System

Meng

et al. 2020

Energies

View full text Add to dashboard Cite

This paper proposes a fresh state error feedback linearization control method with disturbance observer (DOB) and L2 gain for a quadruple-tank liquid-level system. Firstly, in terms of the highly nonlinear and strong coupling characteristics of the quadruple-tank system, a state error feedback linearization technique is employed to design the controller to achieve liquid-level position control and tracking control. Secondly, DOB is purposed to estimate uncertain exogenous disturbances and applied to compensation control. Moreover, an L2-gain disturbance attenuation technology is designed to resolve one class of disturbance problem by uncertain parameter perturbation existing in the quadruple-tank liquid-level system. Finally, compared with the classical proportion integration differentiation (PID) and sliding mode control (SMC) methods, the extensive experimental results validate that the proposed strategy has good position control, tracking control, and disturbance rejection performances.

show abstract

Data-Driven Model-Free Tracking Reinforcement Learning Control with VRFT-based Adaptive Actor-Critic

Cited by 42 publications

References 57 publications

Long Short-Term Memory Neural Network Applied to Train Dynamic Model and Speed Prediction

Long Short-Term Memory Neural Network Applied to Train Dynamic Model and Speed Prediction

Data Driven Distributed Bipartite Consensus Tracking for Nonlinear Multiagent Systems via Iterative Learning Control

Disturbance Observer and L2-Gain-Based State Error Feedback Linearization Control for the Quadruple-Tank Liquid-Level System

Contact Info

Product

Resources

About