.png){kind=icon}
.png){kind=icon}
Return
Data-Driven Predictive Control for Continuous-Time Nonlinear Systems: A Nonzero-Sum Game Approach
.png){kind=icon}
PRE
AI
DOI：10.1109/JAS.2025.125660.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Dear Editor, This letter proposes a reinforcement learning-based predictive learning algorithm for unknown continuous-time nonlinear systems with observation loss. Firstly, we construct a temporal nonzero-sum game over predictive control input sequences, deriving multiple optimal predictive control input sequences from its solution. To obtain the Nash equilibrium solution of the temporal nonzero-sum game, we solve the problem through policy iteration of reinforcement learning. Then, we train the actor neural network and critic neural network to estimate the control policy and action value function, respectively, using the collected offline and online input-state data. Compared to traditional predictive control methods, the proposed framework does not rely on an explicit model and obtains a data-driven controller design. Finally, the effectiveness of the proposed algorithm is validated through a numerical example.
Journal
I
IF:
0
Papers:
116
Citations:
0