DiffTune-MPC: Closed-Loop Learning for Model Predictive Control

Keyphrases

• Model Predictive Control 100%
• Closed-loop Learning 100%
• Cost Function 27%
• Parameter Space 9%
• Learning Methods 9%
• Analytical Gradients 9%
• Performance Evaluation 9%
• High Dimension 9%
• Control Parameters 9%
• Closed-loop 9%
• Controller 9%
• Performance Metrics 9%
• Potential Difference 9%
• Autonomous Systems 9%
• Robotic System 9%
• Learning Model 9%
• Learnability 9%
• Planning Horizon 9%
• Closed-loop Performance 9%
• Future Behavior 9%
• Control Sets 9%
• Auxiliary Problem 9%
• Generalization Ability 9%
• Sequential Quadratic Programming 9%
• Metric Function 9%
• Nonlinear MPC 9%

Engineering

• Closed Loop 100%
• Predictive Control Model 100%
• Cost Function 27%
• Simulation Result 9%
• Parameter Space 9%
• Open Loop 9%
• Metrics 9%
• Potential Difference 9%
• Control Parameter 9%
• Autonomous System 9%
• Robotic System 9%
• Planning Horizon 9%
• Auxiliary Problem 9%

Chemical Engineering

• Predictive Control Model 100%
• Auxiliaries 9%