This paper proposed an optimal control scheme based on the actor-critic neural network(NN) for the complex mechanical manipulator system with dynamic disturbance. The actor's goal is to optimize control behavior, ...
详细信息
This paper proposed an optimal control scheme based on the actor-critic neural network(NN) for the complex mechanical manipulator system with dynamic disturbance. The actor's goal is to optimize control behavior, while the critic's goal is to evaluate control performance. The optimal control update law in the scheme can guarantee the system error and the weight estimation error SGUUB, and its stability and convergence are proved based on the direct Lyapunov method. Finally, the connecting rods on two degrees of freedom are tested to verify the effectiveness of the proposed optimal control scheme.
This paper proposes a data-driven robust regulation method of input-affine nonlinear systems with mismatched disturbances. First, the relationship between the robust control problem and the optimal control problem is ...
详细信息
This paper proposes a data-driven robust regulation method of input-affine nonlinear systems with mismatched disturbances. First, the relationship between the robust control problem and the optimal control problem is built, which indicates that the robust control of original systems can be the solution of the optimal control problem of the auxiliary system. Then, within the framework of adaptive dynamic programming, we present a data-driven algorithm to solve the optimal control problem. To implement the data-driven algorithm, we use two kinds of neural networks (NNs): actor NNs are used to approximate the sub-control policies of the augmented control and a critic NN is applied to estimate the value function. To learn the unknown parameters of actor and critic network weight vectors, the Monte Carlo integration method is employed. Finally, we provide a third-order benchmark model of the armature-controlled DC motor to illustrate the applicability of the developed control strategy.
This paper concerns with a novel generalized policy iteration (GPI) algorithm with approximation errors. Approximation errors are explicitly considered in the GPI algorithm. The properties of the stable GPI algorithm ...
详细信息
This paper concerns with a novel generalized policy iteration (GPI) algorithm with approximation errors. Approximation errors are explicitly considered in the GPI algorithm. The properties of the stable GPI algorithm with approximation errors are analyzed. The convergence of the developed algorithm is established to show that the iterative value function is convergent to a finite neighborhood of the optimal performance index function. Finally, numerical examples and comparisons are presented.
In case of linear quadratic regulator (LQR), the updated terminal cost receding horizon control (UTCRHC) enhances the standard RHC in terms of the stability and the convergence to the optimal solution. Based on the ac...
详细信息
ISBN:
(纸本)9781467322478
In case of linear quadratic regulator (LQR), the updated terminal cost receding horizon control (UTCRHC) enhances the standard RHC in terms of the stability and the convergence to the optimal solution. Based on the action-dependent (AD) value functions, known as Q-function, this paper proposes two modified RHC methods, AD-RHC and AD-UTC-RHC and proves the stability and convergence by using the matrix equivalence of the established control schemes. Additionally, in case of the horizon size N = 1, as like UTC-RHC is connected with heuristic dynamicprogramming (HDP), we shows the connection between the AD-UTC-RHC and ADHDP which has the learning abilities of the optimal solution without using the explicit descriptions of the system dynamics.
暂无评论