Goal-oriented safe active learning for predictive control using Bayesian recurrent neural networks
It addresses the challenge of safe online data collection for model adaptation in learning-based MPC, a key problem for control systems in uncertain environments.
The paper proposes a goal-oriented safe active learning algorithm for model predictive control using Bayesian recurrent neural networks, achieving economic performance comparable to an MPC with full system knowledge while ensuring safety and finite-time exploration termination.
A key challenge in learning-based model predictive control (MPC) is to collect informative data online for model adaptation while ensuring safety and without penalising control performance. In this paper, we propose an online model adaptation scheme embedded within an MPC framework in which the last-layer parameters of a recurrent neural network are recursively updated via Bayesian learning. This is achieved by means of a goal-oriented safe active learning algorithm that alternates between an exploration phase, where the MPC actively explores system dynamics to collect informative data for model adaptation while still pursuing the main control objective, and a goal-reaching phase, where it focuses exclusively on the main control objective. The algorithm is complemented with theoretical guarantees of (i) recursive feasibility, (ii) safety, (iii) termination of exploration in finite time, and (iv) close-to-optimal performance. Simulation results on a benchmark energy system demonstrate that the proposed framework achieves economic performance comparable to that of an MPC with full system knowledge, while progressively improving model accuracy and respecting operational safety constraints with high probability.