Abstract:Control algorithms such as model predictive control (MPC) and state estimators rely on a number of different parameters. The performance of the closed loop usually depends on the correct setting of these parameters. Tuning is often done manually by experts based on a simulation model of the system. Two problems arise with this procedure. Firstly, experts need to be skilled and still may not be able to find the optimal parametrization. Secondly, the performance of the simulation model might not be able to be carried over to the real world application due to model inaccuracies within the simulation. With this contribution, we demonstrate on an industrial milling process how Bayesian optimization can automate the tuning process and help to solve the mentioned problems. Robust parametrization is ensured by perturbing the simulation with arbitrarily distributed model plant mismatches. The objective is to minimize the expected integral reference tracking error, guaranteeing acceptable worst case behavior while maintaining real-time capability. These verbal requirements are translated into a constrained stochastic mixed-integer black-box optimization problem. A two stage min-max-type Bayesian optimization procedure is developed and compared to benchmark algorithms in a simulation study of a CNC machining center. It is showcased how the empirical performance model obtained through Bayesian optimization can be used to analyze and visualize the results. Results indicate superior performance over the case where only the nominal model is used for controller synthesis. The optimized parametrization improves the initial hand-tuned parametrization notably.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to automatically adjust parameters through Bayesian optimization in a CNC machining center to improve the performance of the model predictive controller (MPC) and state estimators (such as the Kalman filter KF), and ensure their robustness. Specifically, the paper focuses on: 1. **Limitations of Manual Parameter Tuning**: Traditional parameter - tuning methods rely on experts' manual adjustment based on system simulation models. This method is not only time - consuming but also likely to lead to sub - optimal solutions. Moreover, due to the inconsistency between the simulation model and the actual system, excellent performance in simulation may not be reproducible in practical applications. 2. **Robustness Problem**: To ensure that the control strategy can still maintain good performance in the face of differences between the model and the actual system, the paper introduces random model - plant mismatch and minimizes the expected integral reference tracking error through optimization, while ensuring acceptable behavior in the worst - case scenario and maintaining real - time capabilities. 3. **Optimization Challenges**: The main challenges in the optimization process include: - Black - box Optimization Problem: The impact of control parameters on performance cannot be directly analyzed and needs to be evaluated through simulation or experiments. - Time Efficiency: Simulations and experiments can be very time - consuming, so an efficient sampling algorithm is required. - Noise Influence: The objective function may be affected by noise, which requires the optimization algorithm to be able to handle noisy data. To solve the above problems, the paper proposes a two - stage Bayesian optimization framework combined with anomaly detection techniques to optimize the parameters of MPC and KF simultaneously and ensure robustness under different environmental conditions (such as different reference trajectories or operation modes). In this way, the paper aims to find a set of controller parameters that can maintain good performance under various uncertain conditions.

Robust Parametrization of a Model Predictive Controller for a CNC Machining Center Using Bayesian Optimization

A novel robust tuning strategy for model predictive control

Active Design of Dynamic GP Models for Model Predictive Control Using Expected Improvement

Model-Based Robot Learning Control with Uncertainty Directed Exploration

Closed-loop Performance Optimization of Model Predictive Control with Robustness Guarantees

Automatic Tuning of Robust Model Predictive Control in Iterative Tasks Using Efficient Bayesian Optimization

What is the Best Way to Optimally Parameterize the MPC Cost Function for Vehicle Guidance?

A Bayesian optimization framework for the automatic tuning of MPC-based shared controllers

MPC Controller Tuning using Bayesian Optimization Techniques

Robustly Stable Model Predictive Control Based on Parallel Support Vector Machines with Linear Kernel

From Parametric Model-based Optimization to robust PID Gain Scheduling

Differentiable Robust Model Predictive Control

Vehicle Cabin Climate MPC Parameter Tuning Using Constrained Contextual Bayesian Optimization (C-CMES)

Stability-informed Bayesian Optimization for MPC Cost Function Learning

A Robust Tube-Based Smooth-MPC for Robot Manipulator Planning

Parameter-Adaptive Approximate MPC: Tuning Neural-Network Controllers without Retraining

Fast Calibration of a Robust Model Predictive Controller for Diesel Engine Airpath

Safe Bayesian Optimization for Complex Control Systems via Additive Gaussian Processes

Model Predictive Control meets robust Kalman filtering

Robust Precision Position Tracking of Planar Motors Using Min-Max Model Predictive Control

Efficient Calibration of Embedded MPC