Learning to Optimize in Model Predictive Control
This work addresses computational bottlenecks in MPC for robotics by improving optimization efficiency, though it is incremental as it builds on existing MPC and imitation learning methods.
The paper tackles the problem of noisy updates in sampling-based Model Predictive Control (MPC) when using few samples for computational efficiency, and shows that learning to optimize the update rule via imitation learning outperforms standard MPC with the same sample count in simulated robotics tasks.
Sampling-based Model Predictive Control (MPC) is a flexible control framework that can reason about non-smooth dynamics and cost functions. Recently, significant work has focused on the use of machine learning to improve the performance of MPC, often through learning or fine-tuning the dynamics or cost function. In contrast, we focus on learning to optimize more effectively. In other words, to improve the update rule within MPC. We show that this can be particularly useful in sampling-based MPC, where we often wish to minimize the number of samples for computational reasons. Unfortunately, the cost of computational efficiency is a reduction in performance; fewer samples results in noisier updates. We show that we can contend with this noise by learning how to update the control distribution more effectively and make better use of the few samples that we have. Our learned controllers are trained via imitation learning to mimic an expert which has access to substantially more samples. We test the efficacy of our approach on multiple simulated robotics tasks in sample-constrained regimes and demonstrate that our approach can outperform a MPC controller with the same number of samples.